paperID
stringlengths 36
36
| pwc_id
stringlengths 8
47
| arxiv_id
stringlengths 6
16
⌀ | nips_id
float64 | url_abs
stringlengths 18
329
| url_pdf
stringlengths 18
742
| title
stringlengths 8
325
| abstract
stringlengths 1
7.27k
⌀ | authors
stringlengths 2
7.06k
| published
stringlengths 10
10
⌀ | conference
stringlengths 12
47
⌀ | conference_url_abs
stringlengths 16
198
⌀ | conference_url_pdf
stringlengths 27
199
⌀ | proceeding
stringlengths 6
47
⌀ | taskID
stringlengths 7
1.44k
| areaID
stringclasses 688
values | embedding
stringlengths 9.26k
12.5k
| umap_embedding
stringlengths 29
44
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d5cd9b77-27a2-4289-bd66-935cd8e78cfc | an-algorithm-unrolling-approach-to-deep-blind | 1902.03493 | null | https://arxiv.org/abs/1902.03493v3 | https://arxiv.org/pdf/1902.03493v3.pdf | Deep Algorithm Unrolling for Blind Image Deblurring | Blind image deblurring remains a topic of enduring interest. Learning based approaches, especially those that employ neural networks have emerged to complement traditional model based methods and in many cases achieve vastly enhanced performance. That said, neural network approaches are generally empirically designed and the underlying structures are difficult to interpret. In recent years, a promising technique called algorithm unrolling has been developed that has helped connect iterative algorithms such as those for sparse coding to neural network architectures. However, such connections have not been made yet for blind image deblurring. In this paper, we propose a neural network architecture based on this idea. We first present an iterative algorithm that may be considered as a generalization of the traditional total-variation regularization method in the gradient domain. We then unroll the algorithm to construct a neural network for image deblurring which we refer to as Deep Unrolling for Blind Deblurring (DUBLID). Key algorithm parameters are learned with the help of training images. Our proposed deep network DUBLID achieves significant practical performance gains while enjoying interpretability at the same time. Extensive experimental results show that DUBLID outperforms many state-of-the-art methods and in addition is computationally faster. | ['Vishal Monga', 'Yuelong Li', 'Junyi Geng', 'Yonina C. Eldar', 'Mohammad Tofighi'] | 2019-02-09 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 1.91842958e-01 -3.95810813e-01 -2.18161836e-01 -1.53429717e-01
-3.56097519e-01 -2.32165501e-01 4.79767323e-01 -4.76011962e-01
-1.59824222e-01 4.50712085e-01 4.10828680e-01 -4.69420642e-01
-1.50683448e-01 -1.90814152e-01 -6.61868930e-01 -8.26498866e-01
6.99310601e-02 -8.12557042e-02 -2.08741695e-01 3.81548703e-02
4.22016650e-01 2.47155577e-01 -9.59269285e-01 -1.73000157e-01
9.94850338e-01 8.80877197e-01 1.97916418e-01 5.46368778e-01
6.75756931e-02 1.08058810e+00 -3.06758851e-01 -1.77388459e-01
3.11318845e-01 -5.89506924e-01 -6.87537968e-01 3.17465812e-01
4.64690566e-01 -7.26699710e-01 -8.61387610e-01 1.30335021e+00
6.11660182e-01 1.19453631e-01 6.74242795e-01 -8.89016569e-01
-1.34562838e+00 4.04832959e-01 -9.54172432e-01 4.35805440e-01
-1.20200748e-02 -1.19283140e-01 7.52555311e-01 -1.19194651e+00
1.42007545e-01 1.12682986e+00 9.07060206e-01 5.53844988e-01
-9.35896635e-01 -5.88724375e-01 9.07640010e-02 6.99061751e-01
-1.35365188e+00 -5.11742115e-01 9.61112380e-01 -3.25232446e-01
6.90502822e-01 1.04481518e-01 3.45713675e-01 8.77380371e-01
2.33404152e-02 1.20589817e+00 1.06262958e+00 -5.97718120e-01
2.83579051e-01 -2.84523189e-01 3.50360483e-01 7.36460209e-01
3.93006295e-01 3.58075202e-01 -4.70835328e-01 7.71668181e-02
1.06025016e+00 2.83428282e-01 -8.37194502e-01 -4.18414980e-01
-1.20555472e+00 8.46130252e-01 8.78016114e-01 2.77402282e-01
-4.10985708e-01 2.95887858e-01 3.14333677e-01 3.20166945e-01
5.68224728e-01 3.22334439e-01 -2.07683802e-01 1.83911517e-01
-1.25809884e+00 4.36027721e-03 6.23175561e-01 6.22747838e-01
6.91894710e-01 4.53152478e-01 5.42589743e-03 1.10138655e+00
3.22589040e-01 2.49584824e-01 8.77264142e-01 -9.02500093e-01
2.68886954e-01 3.80872458e-01 8.39073062e-02 -1.35033834e+00
-2.18849286e-01 -6.52433276e-01 -1.65152597e+00 1.89077884e-01
5.62844314e-02 -1.75485596e-01 -1.03596544e+00 1.38854277e+00
-1.28713071e-01 5.68879783e-01 4.19668816e-02 1.28205979e+00
5.99924505e-01 7.71262825e-01 -4.34998959e-01 -8.19888636e-02
1.10027981e+00 -1.41659522e+00 -9.87108946e-01 -4.55077916e-01
2.08131909e-01 -5.89378655e-01 7.89755106e-01 5.43284833e-01
-8.64220977e-01 -4.99352932e-01 -1.28384829e+00 6.20157123e-02
-2.54981488e-01 2.91500121e-01 7.25602925e-01 6.85218453e-01
-1.31978512e+00 5.33749521e-01 -8.83724391e-01 -3.16177428e-01
5.98480225e-01 2.74368554e-01 -1.05686821e-01 -3.10584486e-01
-1.16115129e+00 8.88188362e-01 4.40477729e-01 6.13014638e-01
-1.02601838e+00 -3.58554006e-01 -7.39409626e-01 -2.73465067e-02
2.29576468e-01 -8.05284798e-01 1.29810917e+00 -1.20315599e+00
-1.62173748e+00 6.95242465e-01 -3.38095963e-01 -7.49974608e-01
4.02808487e-01 -5.62183917e-01 -4.40879136e-01 1.27690792e-01
-2.14063361e-01 4.61796075e-01 1.53485668e+00 -1.46427131e+00
-2.57885307e-01 -2.36370832e-01 -9.91360769e-02 1.60822481e-01
-7.82315910e-01 5.29017933e-02 -5.44971585e-01 -1.30821514e+00
3.91135812e-01 -8.30050647e-01 -2.81966984e-01 1.80559695e-01
-4.27692324e-01 6.38077408e-02 1.00787807e+00 -1.03974628e+00
1.56091774e+00 -1.93612683e+00 5.18380761e-01 -1.04066007e-01
5.48427284e-01 6.63602412e-01 -1.16047546e-01 1.67333797e-01
-4.28783178e-01 -1.41113862e-01 -7.17628002e-01 -3.74247998e-01
-2.52674788e-01 -3.11324354e-02 -4.01074648e-01 8.70887876e-01
-1.40001535e-01 9.20367897e-01 -7.54269779e-01 -6.47223219e-02
2.76700318e-01 5.28621793e-01 -6.18478537e-01 4.12230402e-01
8.84022415e-02 3.71967405e-01 -1.56220719e-01 5.89033067e-01
8.83246958e-01 -5.11598885e-01 -1.28215730e-01 -3.42591971e-01
-2.89669298e-02 -2.52121598e-01 -1.02578497e+00 1.82492566e+00
-4.66915101e-01 1.12832510e+00 2.10655868e-01 -1.38754857e+00
7.34227538e-01 3.58773053e-01 1.14562228e-01 -3.11829299e-01
2.83743441e-01 2.49021098e-01 -2.52890378e-01 -6.23138905e-01
7.33853996e-01 1.71658944e-03 5.44756114e-01 5.98884046e-01
-2.33029395e-01 1.21355399e-01 -2.86054820e-01 2.06163883e-01
7.96075225e-01 8.26582219e-03 4.21385258e-01 -2.02830896e-01
7.81526864e-01 -2.55259544e-01 3.59322071e-01 8.36604953e-01
-2.34265253e-01 8.06851447e-01 -9.60531309e-02 -6.07004881e-01
-8.94527614e-01 -6.41022086e-01 -1.70164946e-02 7.21926689e-01
5.24580300e-01 -7.17092901e-02 -1.06415772e+00 -4.54794675e-01
-3.96932721e-01 4.58828807e-01 -5.15967548e-01 -2.09469303e-01
-6.52692139e-01 -1.09315073e+00 4.94419813e-01 4.00487423e-01
1.07906008e+00 -8.15508068e-01 -1.08707659e-01 1.25720158e-01
-1.26054823e-01 -8.36332142e-01 -7.31849253e-01 -1.28520533e-01
-1.06601834e+00 -8.87450218e-01 -1.59086299e+00 -1.16161489e+00
9.87963736e-01 8.99191141e-01 9.06460822e-01 1.94630012e-01
1.04712509e-02 2.23110050e-01 -4.25577074e-01 -4.24416512e-02
-2.98335999e-01 6.30035429e-05 2.05016106e-01 3.16469789e-01
2.92772651e-01 -7.82034099e-01 -8.76888454e-01 3.03920388e-01
-1.31181479e+00 6.78739250e-02 9.25924122e-01 1.11575687e+00
2.48631254e-01 1.76654816e-01 7.08315551e-01 -5.94548225e-01
8.42087030e-01 -5.33606589e-01 -4.48191613e-01 1.86263159e-01
-7.82418728e-01 1.75329462e-01 5.79556108e-01 -4.72372949e-01
-9.91029680e-01 -2.79361933e-01 9.53125283e-02 -7.29338408e-01
-9.09069702e-02 8.00304711e-01 1.09894693e-01 -4.16515678e-01
6.90817654e-01 6.31087184e-01 4.39256430e-02 -8.15026879e-01
5.60641170e-01 1.09522653e+00 7.15203464e-01 -2.31588587e-01
1.08565056e+00 4.69238400e-01 -3.84077221e-01 -7.79181778e-01
-7.66231179e-01 -3.80406380e-01 -3.49938601e-01 -1.59233585e-01
7.42167652e-01 -1.07639623e+00 -3.53617728e-01 1.13546789e+00
-1.27167070e+00 -2.38998786e-01 2.71610469e-01 4.66030598e-01
-3.45312804e-01 8.43185246e-01 -7.37472296e-01 -5.43004215e-01
-5.52728593e-01 -1.21728194e+00 5.95476925e-01 3.79202724e-01
1.20671175e-01 -1.20819187e+00 -4.19035628e-02 3.59405220e-01
8.59492421e-01 -2.80211955e-01 7.21332788e-01 -3.03537339e-01
-4.81476247e-01 -3.27527761e-01 -7.56864727e-01 6.78368747e-01
4.54918027e-01 -7.85012186e-01 -8.85155499e-01 -6.84125006e-01
3.89734864e-01 -1.97466001e-01 1.17427301e+00 7.18628228e-01
1.51965868e+00 -5.26821971e-01 -2.44698286e-01 8.69585335e-01
1.37248123e+00 -1.23608328e-01 6.70183957e-01 5.31922221e-01
9.29550886e-01 1.73455626e-01 -9.50454697e-02 1.44542918e-01
2.68894225e-01 6.61591411e-01 4.57766354e-01 -3.13037813e-01
-3.57652962e-01 -4.43765707e-02 5.09897053e-01 1.19737089e+00
-9.73301977e-02 -1.24698803e-01 -7.42199361e-01 6.64915800e-01
-1.88449001e+00 -7.38340259e-01 2.35791784e-02 1.97754860e+00
8.55283082e-01 -1.98862731e-01 -1.93383604e-01 1.82887673e-01
1.13574123e+00 5.12438655e-01 -6.99111819e-01 -1.23553947e-01
-1.54275820e-01 -1.82768553e-01 6.37605309e-01 5.86572766e-01
-1.24124491e+00 9.76869762e-01 5.80961323e+00 8.51112127e-01
-1.00719213e+00 1.44179806e-01 5.56382716e-01 4.36540663e-01
5.45681678e-02 3.96675058e-02 -3.82336974e-01 3.82194638e-01
5.54971218e-01 1.78453755e-02 8.70701253e-01 6.81245327e-01
3.79735291e-01 2.17223108e-01 -8.67818356e-01 1.34253025e+00
4.85060006e-01 -1.40296459e+00 2.22885981e-01 -1.96717069e-01
1.15257311e+00 -5.56127690e-02 2.78011292e-01 -4.52975407e-02
2.22252324e-01 -1.05264068e+00 5.07673025e-01 4.19541985e-01
7.13405907e-01 -4.32231754e-01 7.61283815e-01 2.34490201e-01
-9.37605619e-01 -1.90077722e-01 -3.66669744e-01 -7.43955970e-02
2.66846269e-01 6.99400067e-01 -3.64184767e-01 6.09558225e-01
5.21788061e-01 1.34492064e+00 -4.84461635e-01 1.46260083e+00
-3.06652606e-01 9.29588974e-01 1.75044745e-01 3.13220531e-01
4.00042236e-01 -2.57152438e-01 8.67334068e-01 1.06142747e+00
4.01536435e-01 -6.31926209e-02 -1.40674934e-01 8.95182967e-01
-3.35175782e-01 -2.06431478e-01 -3.72491270e-01 9.41882469e-03
3.32051218e-01 8.05998087e-01 -4.47999716e-01 -3.95774543e-01
-5.01828492e-01 1.47455740e+00 5.53921349e-02 8.06987047e-01
-6.85213029e-01 -3.90023440e-01 5.70495605e-01 -3.42004448e-01
4.45356995e-01 -3.95168513e-01 -1.89555541e-01 -1.55827904e+00
-1.15715757e-01 -1.12709761e+00 -3.56332981e-03 -9.20922637e-01
-1.54833591e+00 6.49426401e-01 -3.66303831e-01 -1.58528805e+00
2.90830433e-02 -5.31340957e-01 -5.58308423e-01 9.29991066e-01
-2.02828336e+00 -9.27685916e-01 -5.24206340e-01 6.31823421e-01
9.29168463e-01 -3.65964234e-01 6.62421644e-01 3.48077267e-01
-9.10546005e-01 6.13743782e-01 5.96381307e-01 3.10087949e-01
7.20725834e-01 -1.05494475e+00 3.56742382e-01 1.33148754e+00
1.52876183e-01 8.01253796e-01 6.58253610e-01 -4.11122918e-01
-1.48980689e+00 -1.18078053e+00 4.25163776e-01 6.50277287e-02
6.57243550e-01 2.43815348e-01 -1.08736384e+00 6.65853620e-01
6.00189269e-01 -5.77983372e-02 2.86078155e-01 -2.11284965e-01
-4.81963575e-01 -2.65019722e-02 -8.37134302e-01 5.84982514e-01
7.86604106e-01 -5.76895714e-01 -8.09602559e-01 2.87454784e-01
5.26852906e-01 -4.26327288e-01 -2.58258373e-01 5.42551801e-02
3.58567625e-01 -1.01461911e+00 1.15872431e+00 -4.15831774e-01
5.54020643e-01 -4.51899916e-01 -6.45930171e-02 -1.47425556e+00
-4.51910883e-01 -9.07973766e-01 -5.91364622e-01 9.62906122e-01
1.97294424e-03 -7.06844449e-01 6.91097677e-01 2.28798524e-01
-2.00217858e-01 -5.92482567e-01 -5.89348018e-01 -8.84880364e-01
1.99337706e-01 -2.62399316e-01 2.88051397e-01 1.16780388e+00
-2.96507955e-01 1.70577064e-01 -1.00213981e+00 3.63248974e-01
1.00660968e+00 5.73818050e-02 3.95486951e-01 -9.42244887e-01
-2.87878782e-01 -6.50127709e-01 -3.39962393e-01 -1.83393240e+00
2.93561101e-01 -9.76508081e-01 1.80616140e-01 -1.65486765e+00
2.96418518e-01 -1.52060881e-01 -3.60911965e-01 3.51868421e-01
-6.10737562e-01 3.44880909e-01 -5.91327101e-02 8.57271016e-01
-3.11171442e-01 7.13979006e-01 1.05454469e+00 -4.52093363e-01
-2.45210603e-01 -5.35477325e-02 -8.60041261e-01 8.24927270e-01
8.33748937e-01 -2.69012481e-01 -2.20610693e-01 -1.07258892e+00
4.53356728e-02 -1.12302065e-01 5.40915012e-01 -1.09901953e+00
5.27209699e-01 2.81139433e-01 1.87418684e-01 -1.94644138e-01
-2.34891800e-03 -8.70385826e-01 -2.12036893e-01 3.40512812e-01
-4.07923877e-01 -1.68277100e-01 7.47769326e-02 9.28533256e-01
-5.49028873e-01 -2.22186893e-01 8.29009414e-01 1.17040649e-02
-9.13552284e-01 3.53082001e-01 -4.93254870e-01 -2.90131867e-01
5.14864743e-01 -2.91438311e-01 -1.43889338e-01 -6.30838335e-01
-3.76180470e-01 1.81059726e-02 3.55309486e-01 4.17267084e-01
8.21436465e-01 -1.35794723e+00 -8.27812731e-01 1.52072474e-01
-1.74092680e-01 -1.76453859e-01 3.29183608e-01 9.38152671e-01
-7.14485705e-01 5.56759298e-01 -3.99586186e-03 -3.81127357e-01
-9.20784652e-01 6.13663733e-01 3.86416197e-01 -1.19890712e-01
-7.79716253e-01 9.94765759e-01 3.54272634e-01 -1.64612114e-01
4.36528325e-01 -4.41031940e-02 -2.00734109e-01 -1.73620358e-01
8.94362807e-01 3.72274846e-01 -8.51112828e-02 -6.10170901e-01
-1.03571266e-01 6.99385524e-01 -2.54110426e-01 1.20837651e-01
1.46631026e+00 -4.60178226e-01 -4.62001294e-01 -5.00969142e-02
1.29055560e+00 -2.20543757e-01 -1.28688943e+00 -6.81303740e-01
-5.07168621e-02 -4.44183886e-01 7.31977403e-01 -6.88318133e-01
-1.51966774e+00 9.45778549e-01 9.38642263e-01 2.48008296e-01
1.43972123e+00 -3.49140823e-01 1.09632277e+00 4.52764452e-01
1.31777555e-01 -6.38331056e-01 1.06502682e-01 3.72097671e-01
1.03253567e+00 -1.39613903e+00 8.87912735e-02 1.11473398e-02
-2.29042575e-01 1.15573871e+00 3.42151940e-01 -2.83178091e-01
5.97839952e-01 -4.04921137e-02 8.24293643e-02 1.11595402e-02
-4.14219080e-03 1.46444798e-01 2.56055772e-01 4.91033912e-01
2.44825825e-01 -2.36239806e-01 -1.88374355e-01 5.15849948e-01
2.90252566e-01 3.82580310e-01 4.04980123e-01 6.70908272e-01
-4.84463453e-01 -9.37146127e-01 -8.05434406e-01 3.82195920e-01
-4.15821075e-01 -5.18811524e-01 -1.81951508e-01 2.11853474e-01
-2.52875447e-01 8.87363911e-01 -2.91488290e-01 -3.57746452e-01
-1.54601112e-01 -2.89315492e-01 2.62374282e-01 -9.04155374e-02
-1.19392321e-01 -2.29011774e-01 -3.45044553e-01 -2.62504578e-01
-8.52840066e-01 -2.53929377e-01 -7.61634707e-01 -3.01409006e-01
-5.34959853e-01 6.85911179e-02 4.59457755e-01 1.02689505e+00
3.87073249e-01 3.66029650e-01 6.89958513e-01 -1.40179372e+00
-5.40085316e-01 -1.11516142e+00 -4.73525047e-01 -8.60784277e-02
8.62173140e-01 -5.44198990e-01 -5.98363280e-01 5.08979738e-01] | [11.579588890075684, -2.6643214225769043] |
970c69d3-f926-4ca8-a913-7ed57d2429ec | computationally-enhanced-approach-for-chance | 2306.14527 | null | https://arxiv.org/abs/2306.14527v2 | https://arxiv.org/pdf/2306.14527v2.pdf | Computationally Enhanced Approach for Chance-Constrained OPF Considering Voltage Stability | The effective management of stochastic characteristics of renewable power generations is vital for ensuring the stable and secure operation of power systems. This paper addresses the formidable task of optimizing the chance-constrained voltage-stability-constrained optimal power flow (CC-VSC-OPF) problem, which is hindered by the implicit voltage stability index and intractable chance constraints. Leveraging a neural network (NN)-based surrogate model, the stability constraint is explicitly formulated and seamlessly integrated into the model. To perform uncertainty propagation without relying on presumptions or complicated transformations, an advanced data-driven method known as adaptive polynomial chaos expansion (APCE) is developed. To extend the scalability of the proposed algorithm, a partial least squares (PLS)-NN framework is designed, which enables the establishment of a parsimonious surrogate model and efficient computation of large-scale Hessian matrices. In addition, a dimensionally decomposed APCE (DD-APCE) is proposed to alleviate the "curse of dimensionality" by restricting the interaction order among random variables. Finally, the above techniques are merged into an iterative scheme to update the operation point. Simulation results reveal the cost-effective performances of the proposed method in several test systems. | ['Jingtao Zhao', 'Shu Zheng', 'Wei Gu', 'Huan Long', 'Yijun Xu', 'Zhi Wu', 'Yuanxi Wu'] | 2023-06-26 | null | null | null | null | ['management'] | ['miscellaneous'] | [-1.72151580e-01 -4.63793159e-01 -2.64089614e-01 4.64547910e-02
-2.93529361e-01 -7.48744488e-01 2.08455592e-01 -1.52745157e-01
2.90635020e-01 1.12294447e+00 -1.78033382e-01 -5.58171690e-01
-8.82265389e-01 -5.89525163e-01 -1.25783771e-01 -1.25129199e+00
-2.22929329e-01 -2.44463757e-01 -3.59636813e-01 -2.51018792e-01
4.59409475e-01 6.27063394e-01 -1.12827790e+00 -7.28595972e-01
1.59847176e+00 1.13018048e+00 1.31227911e-01 1.82004031e-02
3.18864256e-01 2.66346306e-01 -4.54625458e-01 -4.70785499e-02
2.72870094e-01 -3.58910531e-01 -1.04348376e-01 -8.26393589e-02
-5.82065284e-01 -3.22060198e-01 -1.39886305e-01 1.26428699e+00
5.24901986e-01 2.82026410e-01 6.74635947e-01 -1.55581844e+00
-6.13867402e-01 4.15794104e-01 -6.67305052e-01 2.37082779e-01
4.80094738e-02 9.75376666e-02 8.52123857e-01 -9.35362518e-01
1.00612976e-01 7.91931033e-01 4.03868288e-01 -7.56264031e-02
-1.35032070e+00 -6.02241457e-01 1.65862873e-01 2.76675433e-01
-1.79860795e+00 -6.94475994e-02 1.41661572e+00 -3.08525681e-01
6.85068369e-01 5.13015568e-01 8.22573423e-01 4.01419312e-01
3.60700428e-01 3.40151280e-01 9.63182509e-01 -2.13991106e-01
4.28374797e-01 1.16634436e-01 -1.35449573e-01 2.45807186e-01
7.19704270e-01 1.78293794e-01 -1.41306460e-01 -2.78990448e-01
6.72994018e-01 -6.61979616e-02 -5.15716314e-01 -5.30156791e-01
-6.88956916e-01 7.42616355e-01 4.00820792e-01 3.10407400e-01
-4.42379832e-01 -2.48336673e-01 1.56363193e-02 -1.13770351e-01
2.37844199e-01 4.09981251e-01 -3.50600868e-01 -1.07194379e-01
-1.06097162e+00 -7.81242847e-02 6.14953399e-01 9.08199310e-01
2.48672292e-01 1.01088285e+00 -5.77866985e-03 3.91272128e-01
5.33508778e-01 6.61176741e-01 3.29639405e-01 -8.21719050e-01
4.41190362e-01 6.73324764e-01 4.33446944e-01 -1.39476812e+00
-3.80908102e-01 -1.11759579e+00 -1.05917501e+00 1.89620093e-01
-7.36883879e-02 -4.16418761e-01 -2.22515196e-01 1.56269109e+00
3.32949162e-01 -1.87511854e-02 -4.53291945e-02 8.85139227e-01
-1.12313792e-01 9.21143413e-01 -1.81679666e-01 -9.83682334e-01
9.51357126e-01 -2.75808454e-01 -8.70743275e-01 3.11039150e-01
7.97173753e-02 -4.50586706e-01 6.94739342e-01 3.62528294e-01
-9.16901827e-01 -1.34503007e-01 -1.50210714e+00 5.81669807e-01
-2.01880053e-01 3.86691988e-01 5.34364402e-01 7.46298730e-01
-6.77210689e-01 5.46931863e-01 -7.08602726e-01 2.86370337e-01
2.06977054e-01 2.74499267e-01 1.15321137e-01 6.14068389e-01
-1.20558476e+00 9.49927986e-01 6.70332193e-01 9.97866273e-01
-4.74400401e-01 -7.11015344e-01 -6.07482791e-01 2.89222062e-01
5.44899464e-01 -4.84504104e-01 4.69401211e-01 -5.45206964e-01
-1.73861980e+00 -4.00377303e-01 -7.56007954e-02 -1.74339131e-01
3.86892080e-01 1.57145590e-01 -4.40678835e-01 2.50452369e-01
-1.55700922e-01 -3.64864081e-01 8.30543399e-01 -1.29711854e+00
-2.53918499e-01 -2.30380967e-01 -3.19745839e-01 3.23801309e-01
-6.22173905e-01 -2.61579722e-01 1.06834836e-01 -8.07692528e-01
3.50668222e-01 -6.10353053e-01 -4.54196304e-01 -3.83938909e-01
-5.31744182e-01 -1.45703346e-01 6.82359338e-01 -1.09784508e+00
1.75332177e+00 -1.94224584e+00 4.05231744e-01 7.82232761e-01
-1.19134717e-01 2.17549741e-01 3.52032095e-01 6.82072103e-01
-1.86124161e-01 -6.81891153e-03 -3.85367036e-01 2.62882918e-01
2.51366526e-01 1.19385205e-01 -2.57256806e-01 5.51766455e-01
3.20699066e-01 6.54862523e-01 -7.31631935e-01 -3.03756237e-01
5.32017529e-01 2.73369491e-01 -2.72190928e-01 -1.21236868e-01
8.24270993e-02 3.06851119e-01 -8.74153912e-01 7.02761054e-01
9.13984001e-01 -1.85096741e-01 2.69523561e-01 -4.34394836e-01
-4.99257207e-01 -3.44989300e-01 -1.65176415e+00 1.23579085e+00
-4.32067841e-01 -4.93867807e-02 3.80208284e-01 -1.21367180e+00
1.03659368e+00 2.74633586e-01 7.40994871e-01 -3.56062800e-01
1.71332374e-01 3.16708416e-01 -3.98181304e-02 -2.93617278e-01
3.16201270e-01 8.42638835e-02 -1.36414822e-02 1.58438161e-01
-1.74944401e-01 -1.29304275e-01 2.65038162e-01 7.17137232e-02
2.83759564e-01 1.17315911e-01 5.98231316e-01 -8.57716203e-01
1.05591393e+00 -1.88576564e-01 1.12636685e+00 -7.51405582e-03
-1.37863740e-01 -2.97896005e-02 5.05148530e-01 -4.44791466e-02
-8.22351396e-01 -9.51307535e-01 -4.50980484e-01 2.24807058e-02
2.36212596e-01 -1.33433953e-01 -2.89866179e-01 -2.74362683e-01
1.49789184e-01 1.17416561e+00 -7.68595561e-02 -2.26404294e-01
-2.96769291e-01 -1.05264103e+00 -7.63175711e-02 2.95195490e-01
2.85492450e-01 -2.23922357e-01 -3.28238159e-01 3.41085255e-01
1.11467175e-01 -5.08179009e-01 -1.93047956e-01 1.60333648e-01
-7.00537562e-01 -7.45983720e-01 -6.21141374e-01 -3.88578355e-01
8.63518536e-01 -1.16008753e-02 4.02993083e-01 -4.00852077e-02
-1.27311498e-01 8.58467910e-03 2.45803930e-02 5.80508634e-02
5.47225866e-03 -1.48271814e-01 4.94674742e-01 5.64853996e-02
-2.59015471e-01 -9.86324966e-01 -5.94323874e-01 3.26601893e-01
-5.82076728e-01 -1.15676679e-01 6.47334874e-01 1.09427738e+00
5.39814889e-01 9.43259656e-01 1.06905103e+00 5.41833304e-02
7.62535810e-01 -5.80899596e-01 -1.45394242e+00 5.88658035e-01
-1.06273437e+00 -4.67067845e-02 1.18981171e+00 -3.45494092e-01
-1.19351757e+00 1.97612345e-02 2.75589138e-01 -5.35554588e-01
5.45075834e-01 7.64858782e-01 -6.06642962e-01 -3.96890700e-01
-1.50261030e-01 6.80399895e-01 -7.52003491e-02 -2.87214994e-01
3.44473183e-01 6.65328562e-01 5.02974033e-01 -6.00658000e-01
1.39897442e+00 -3.21729034e-02 5.47015786e-01 -6.51526034e-01
-1.48936957e-01 -1.91812720e-02 -3.25534195e-01 -3.52650255e-01
3.62893939e-01 -6.87938035e-01 -1.21116018e+00 4.12446797e-01
-7.66058028e-01 5.46355844e-01 -8.05037394e-02 5.32460988e-01
-2.02644974e-01 6.76888049e-01 -2.69779056e-01 -1.39211321e+00
-4.21311051e-01 -1.05440760e+00 2.71477163e-01 6.42274439e-01
2.50637800e-01 -8.33873749e-01 -2.29329899e-01 9.92729142e-02
5.09706676e-01 4.68830228e-01 9.66166079e-01 -3.00835580e-01
-7.40007997e-01 -1.12389065e-01 1.73003152e-02 4.60547090e-01
1.99062899e-01 3.21156114e-01 -2.22532317e-01 -5.38316905e-01
4.09468442e-01 2.55875379e-01 -1.43529728e-01 3.87083352e-01
8.78279209e-01 -6.28435194e-01 -2.09152415e-01 7.23892689e-01
1.93431377e+00 8.12388420e-01 2.80877441e-01 2.05208793e-01
4.74149793e-01 3.89780819e-01 3.99912059e-01 9.15323913e-01
4.35191035e-01 3.60745668e-01 4.81396258e-01 2.75599927e-01
7.56586134e-01 -2.33882949e-01 8.09512809e-02 1.22888672e+00
-4.95055541e-02 -1.51511103e-01 -5.14591515e-01 2.81357437e-01
-1.57264280e+00 -7.17519760e-01 -1.10560030e-01 2.19802117e+00
5.81047595e-01 7.59130195e-02 -4.67269838e-01 4.27074879e-01
7.55066097e-01 5.11064008e-02 -8.24262917e-01 -4.02181566e-01
-2.49410406e-01 -3.15347731e-01 4.59380925e-01 4.02008414e-01
-8.69912088e-01 9.96961519e-02 5.36972857e+00 1.31024837e+00
-1.14759231e+00 -4.05908585e-01 5.95126569e-01 1.89128771e-01
-5.42456448e-01 1.92339271e-01 -5.99830031e-01 9.17920709e-01
6.14331365e-01 -8.87104452e-01 7.95444310e-01 8.30256701e-01
8.59431028e-01 -2.66124874e-01 -3.04522693e-01 7.46779561e-01
-2.04164475e-01 -1.06828153e+00 -2.28274956e-01 6.69910088e-02
1.02854609e+00 -6.67346954e-01 8.63537267e-02 3.06371171e-02
-1.35879323e-01 -5.63184679e-01 6.43773735e-01 7.21512377e-01
5.07167697e-01 -1.04630208e+00 6.51710212e-01 3.28974426e-01
-1.28834414e+00 -6.31246448e-01 -9.98627022e-02 4.62995730e-02
5.44415534e-01 7.78361619e-01 -2.14053169e-01 1.33192849e+00
5.28331995e-01 6.60667956e-01 -3.08663487e-01 9.77380931e-01
-3.13071281e-01 4.23796624e-01 -5.84095418e-01 -1.60513014e-01
-5.00298999e-02 -9.44765687e-01 9.08762217e-01 2.17951268e-01
5.64122319e-01 2.68600583e-01 4.23022285e-02 1.17742717e+00
4.68469709e-01 1.23882733e-01 -1.82350740e-01 -2.18645170e-01
1.06883013e+00 1.49086130e+00 -5.62398195e-01 1.07727259e-01
-2.50513971e-01 3.28466535e-01 -2.98762530e-01 6.71996355e-01
-8.29274118e-01 -5.68326950e-01 2.80624062e-01 -4.16547984e-01
3.02751482e-01 -3.76497716e-01 -6.70873523e-01 -1.13979185e+00
4.63665128e-01 -6.41319752e-01 2.10016102e-01 -6.84309244e-01
-1.35673428e+00 3.97667319e-01 4.66262102e-02 -1.66545975e+00
-5.57238281e-01 -2.25085840e-01 -1.05993187e+00 1.27813876e+00
-1.59292591e+00 -8.80343437e-01 2.03938171e-01 4.79778588e-01
1.52885333e-01 -3.22625756e-01 6.44058168e-01 3.17305148e-01
-1.38292742e+00 3.57693166e-01 7.63944089e-01 -4.48991090e-01
-1.68109685e-01 -9.88801479e-01 -5.32196164e-01 1.48324823e+00
-6.56878114e-01 7.66432762e-01 8.98081362e-01 -7.35144734e-01
-1.85457110e+00 -5.83006918e-01 6.34099007e-01 3.14189583e-01
1.12314832e+00 -1.26939073e-01 -7.39268661e-01 1.05285756e-01
2.01428637e-01 -1.58723712e-01 3.35281610e-01 -4.73425269e-01
3.74997169e-01 -4.67478573e-01 -1.20216060e+00 6.42465591e-01
2.94039369e-01 -3.40488166e-01 -4.73323256e-01 -3.77213433e-02
5.13701737e-01 -3.59693542e-02 -1.30850434e+00 6.42535448e-01
3.78210485e-01 -4.09917414e-01 1.06830108e+00 -1.81188330e-01
-7.50179812e-02 -8.12683880e-01 -8.72102603e-02 -1.36651123e+00
-2.40881383e-01 -1.15431011e+00 -4.97253239e-01 1.64491439e+00
2.49119475e-01 -1.03323746e+00 4.83221889e-01 8.76191735e-01
-1.13843583e-01 -1.05569720e+00 -1.40755069e+00 -9.76198792e-01
-8.47465843e-02 -1.60169206e-03 7.47612417e-01 9.46912706e-01
4.00319576e-01 -1.92973912e-01 -2.59810567e-01 7.15684831e-01
7.92576432e-01 2.66170293e-01 1.18602477e-01 -8.87019634e-01
-2.12836578e-01 -5.36241591e-01 -4.55688350e-02 -5.57337284e-01
1.15926648e-02 -5.58441639e-01 -1.14210017e-01 -1.25777960e+00
-3.11553776e-01 -4.49696034e-01 -6.07480347e-01 -6.55485783e-03
-3.00028682e-01 -4.57252860e-01 1.56563520e-01 2.50783920e-01
-4.00324352e-02 1.31528449e+00 1.03881896e+00 9.63954404e-02
-2.25749165e-01 2.41500035e-01 -4.71099794e-01 4.55876976e-01
8.52664113e-01 -1.37176007e-01 -7.82049596e-01 1.15549227e-03
1.27799720e-01 6.96378052e-01 9.77121964e-02 -8.30658197e-01
3.49212527e-01 -3.05036873e-01 3.12530041e-01 -9.77272987e-01
1.99596107e-01 -1.22886622e+00 5.24331450e-01 6.42920852e-01
2.48762757e-01 3.67837250e-01 7.54627511e-02 5.45700252e-01
-1.43772230e-01 -2.38870561e-01 2.98026055e-01 4.84554470e-01
-4.40653831e-01 5.42911291e-02 -2.48986751e-01 -4.29802865e-01
1.22617900e+00 -2.29279816e-01 -2.78900057e-01 -1.02473721e-01
-5.28208017e-01 8.12252343e-01 1.25433072e-01 1.26950711e-01
4.01078790e-01 -1.45370924e+00 -4.33853030e-01 2.61829108e-01
-5.27089179e-01 -2.71381944e-01 6.08160555e-01 9.53190386e-01
-3.74679238e-01 6.54774606e-01 1.90296452e-02 -4.71563756e-01
-5.09317100e-01 6.08137608e-01 2.94057012e-01 -3.02930027e-01
-1.94583461e-01 2.35725969e-01 -5.55372894e-01 -8.22134838e-02
-9.01271030e-02 -7.58134052e-02 -2.36117542e-01 2.03179032e-01
1.53144032e-01 7.62948036e-01 -1.90809086e-01 -5.62095225e-01
-4.59765911e-01 5.11371434e-01 5.03591537e-01 1.43408656e-01
1.41327608e+00 -6.20123923e-01 -3.67875844e-01 1.16340116e-01
8.84396315e-01 -4.57041301e-02 -1.37527668e+00 2.62180537e-01
-9.58050042e-02 -3.40033144e-01 3.43948811e-01 -8.10551286e-01
-1.26919401e+00 4.22219664e-01 4.10844624e-01 4.05721247e-01
1.40733278e+00 -1.02053547e+00 3.97634059e-01 2.61009663e-01
5.09598553e-01 -1.18635869e+00 -4.95202392e-01 1.18226081e-01
8.27281237e-01 -7.18859017e-01 2.69932300e-01 -4.41237032e-01
-3.92131984e-01 1.16221392e+00 4.15676177e-01 -3.98617238e-03
7.29983866e-01 1.93126380e-01 -4.94836807e-01 4.01237696e-01
-4.68364388e-01 4.14498210e-01 3.40236723e-01 1.92213088e-01
-2.80172944e-01 8.50530118e-02 -9.11615133e-01 1.16664767e+00
1.88259827e-03 -1.49062723e-02 4.42328185e-01 9.08235550e-01
-6.55111074e-02 -6.77644551e-01 -4.70648885e-01 2.81235963e-01
-2.42575899e-01 -3.08368821e-02 4.23325390e-01 7.78128922e-01
-4.50421162e-02 1.06447601e+00 -3.72742385e-01 -9.69573334e-02
1.28040060e-01 -2.35234380e-01 -9.70458314e-02 1.61934868e-01
-1.71337649e-01 2.43048862e-01 -2.18716681e-01 -3.31945539e-01
-4.01307307e-02 -7.36753881e-01 -1.06395555e+00 -1.04682423e-01
-8.30369055e-01 6.98455691e-01 8.84180844e-01 1.00600672e+00
1.47232652e-01 4.12730724e-01 1.29723978e+00 -7.13753879e-01
-1.16359210e+00 -5.51425755e-01 -8.65327060e-01 -2.85484135e-01
-5.08546736e-03 -7.62192488e-01 -8.18399668e-01 -5.08692503e-01] | [5.712527751922607, 2.6081104278564453] |
b24e284f-15f3-4360-9336-64eee0a466e5 | action-quality-assessment-with-temporal | 2207.0927 | null | https://arxiv.org/abs/2207.09270v1 | https://arxiv.org/pdf/2207.09270v1.pdf | Action Quality Assessment with Temporal Parsing Transformer | Action Quality Assessment(AQA) is important for action understanding and resolving the task poses unique challenges due to subtle visual differences. Existing state-of-the-art methods typically rely on the holistic video representations for score regression or ranking, which limits the generalization to capture fine-grained intra-class variation. To overcome the above limitation, we propose a temporal parsing transformer to decompose the holistic feature into temporal part-level representations. Specifically, we utilize a set of learnable queries to represent the atomic temporal patterns for a specific action. Our decoding process converts the frame representations to a fixed number of temporally ordered part representations. To obtain the quality score, we adopt the state-of-the-art contrastive regression based on the part representations. Since existing AQA datasets do not provide temporal part-level labels or partitions, we propose two novel loss functions on the cross attention responses of the decoder: a ranking loss to ensure the learnable queries to satisfy the temporal order in cross attention and a sparsity loss to encourage the part representations to be more discriminative. Extensive experiments show that our proposed method outperforms prior work on three public AQA benchmarks by a considerable margin. | ['Jingdong Wang', 'Yang Long', 'Yu Guan', 'Errui Ding', 'Jian Wang', 'Songyang Zhang', 'Desen Zhou', 'Yang Bai'] | 2022-07-19 | null | null | null | null | ['action-quality-assessment', 'action-understanding'] | ['computer-vision', 'computer-vision'] | [ 4.22983736e-01 -4.58551437e-01 -6.41106963e-01 -6.36337817e-01
-1.27721000e+00 -2.74330795e-01 3.10634762e-01 -2.54033238e-01
-1.75788328e-01 4.10222709e-01 6.42019033e-01 1.55047327e-01
-2.17582345e-01 -5.39470792e-01 -6.45274639e-01 -6.10729396e-01
-3.40772793e-02 -1.62390113e-01 6.26470566e-01 -5.83348423e-02
1.87607363e-01 8.21465552e-02 -1.65350330e+00 9.12237167e-01
9.38805580e-01 1.43519723e+00 -6.38662279e-02 4.49139118e-01
7.91600626e-03 1.10489893e+00 -5.77293873e-01 -3.02737266e-01
3.23391706e-01 -7.74420738e-01 -8.20973337e-01 2.54899144e-01
7.74100304e-01 -7.72332430e-01 -6.40299797e-01 9.02857959e-01
4.15628403e-01 3.35263044e-01 4.51399446e-01 -1.41851628e+00
-6.10124826e-01 2.50769138e-01 -8.09855938e-01 5.72703540e-01
4.71961141e-01 5.14131844e-01 1.44308436e+00 -6.67280793e-01
2.11678758e-01 1.54971147e+00 2.56520629e-01 5.61242104e-01
-9.68300283e-01 -7.23641694e-01 7.12544084e-01 8.54499757e-01
-1.10848415e+00 -5.22290111e-01 8.04328680e-01 -4.80960488e-01
8.05537999e-01 1.96723416e-01 6.90745652e-01 1.23594630e+00
1.48137510e-01 1.03588510e+00 1.04852450e+00 1.14767365e-01
2.19072685e-01 -6.56594038e-01 1.38173297e-01 8.56049240e-01
-2.16802552e-01 -7.75796100e-02 -9.66978014e-01 -1.14944177e-02
8.49290133e-01 1.67204946e-01 -2.42091417e-01 -3.58768940e-01
-1.17097843e+00 7.02439427e-01 4.48732048e-01 6.84818020e-03
-3.85062009e-01 5.10580063e-01 6.18542671e-01 1.45417988e-01
2.79860139e-01 7.54640857e-03 -4.56005007e-01 -4.59621817e-01
-8.16087663e-01 1.53239742e-01 -1.25795621e-02 6.24280512e-01
5.95869899e-01 -1.33320794e-01 -1.04768527e+00 8.35496962e-01
3.47925097e-01 3.47599059e-01 4.98226613e-01 -1.43804491e+00
8.01104844e-01 8.01243365e-01 9.30998549e-02 -1.01321852e+00
-1.66264623e-02 -1.99184239e-01 -4.52115625e-01 2.12215364e-01
5.29960573e-01 4.78399932e-01 -1.08044481e+00 1.95133698e+00
1.80136561e-01 3.31244111e-01 -2.89065123e-01 1.11196482e+00
5.45185030e-01 6.58586860e-01 2.99963802e-01 -3.58674675e-01
1.40748739e+00 -1.24988651e+00 -7.57057965e-01 -2.26427272e-01
4.66065377e-01 -4.18912381e-01 1.28868282e+00 2.89966553e-01
-1.17264175e+00 -8.12721372e-01 -1.01569378e+00 -3.16736460e-01
1.49526805e-01 1.76294729e-01 3.72127920e-01 2.20484868e-01
-7.60878444e-01 5.19629598e-01 -1.04709291e+00 -3.42117064e-02
7.57563233e-01 1.53842226e-01 -2.80443370e-01 -3.98380190e-01
-1.15401804e+00 4.64371443e-01 7.67412931e-02 1.08420864e-01
-1.27503300e+00 -4.88963246e-01 -9.84471381e-01 -6.02245284e-03
5.25466681e-01 -6.15269840e-01 1.22349286e+00 -1.12416637e+00
-1.56608343e+00 6.63452804e-01 -4.97129202e-01 -2.67669976e-01
4.95873630e-01 -5.02465725e-01 -3.02906841e-01 4.84259158e-01
4.46805239e-01 7.51175404e-01 8.87056589e-01 -7.57909894e-01
-9.05561209e-01 -4.14740741e-01 4.77512926e-01 3.12388301e-01
-4.40429151e-01 4.41606529e-02 -8.85838330e-01 -8.42650592e-01
9.69330370e-02 -7.02770114e-01 -1.13258168e-01 4.45491076e-01
-1.36851847e-01 -4.10557061e-01 6.06686294e-01 -7.18091786e-01
1.60641944e+00 -2.13458514e+00 3.30225706e-01 -1.35713771e-01
1.32087409e-01 2.86897384e-02 -4.06235069e-01 1.44273480e-02
2.51103807e-02 -6.53845593e-02 -2.29328319e-01 -2.46572196e-01
-1.04263097e-01 3.03907931e-01 -1.68492600e-01 3.52221698e-01
4.87578779e-01 7.48300433e-01 -1.13165534e+00 -6.07433856e-01
7.65274018e-02 3.00274849e-01 -7.67095327e-01 4.00554627e-01
-2.46080443e-01 4.02136236e-01 -9.26210582e-01 8.32299709e-01
3.04355323e-01 -3.58966589e-01 -1.53790534e-01 -4.40651655e-01
1.12958349e-01 5.25017679e-01 -1.01152790e+00 2.11736894e+00
-8.26326907e-02 3.20384502e-01 -3.54537189e-01 -9.59384263e-01
5.93600035e-01 1.63662940e-01 8.80883098e-01 -1.09936476e+00
-1.42150551e-01 6.02926500e-02 -8.88934657e-02 -6.43277407e-01
2.30780810e-01 1.94924757e-01 -1.70995668e-01 2.07003757e-01
-7.49362707e-02 5.49386680e-01 2.58765787e-01 1.05338730e-01
1.41264796e+00 5.87390244e-01 2.08065614e-01 2.26663396e-01
5.99078298e-01 -1.26836285e-01 1.18196833e+00 5.21837831e-01
-8.03606808e-01 8.95565152e-01 6.39205158e-01 -5.70413709e-01
-7.20822453e-01 -1.01294065e+00 1.75800294e-01 1.38042474e+00
3.01212251e-01 -5.91453791e-01 -6.01250947e-01 -1.05491126e+00
-1.87136412e-01 2.26551920e-01 -9.01014745e-01 -4.64057654e-01
-8.11472893e-01 -4.67401534e-01 2.11674437e-01 8.63003910e-01
7.07445979e-01 -1.12599504e+00 -6.51249826e-01 1.69082090e-01
-5.65800846e-01 -1.17874432e+00 -9.17922735e-01 -2.73640126e-01
-8.60122561e-01 -1.33867431e+00 -8.25088739e-01 -4.07367170e-01
3.08738142e-01 5.16222596e-01 1.11895931e+00 -7.81984925e-02
-8.78073052e-02 4.55495358e-01 -5.86265564e-01 2.58999407e-01
5.20834774e-02 -2.85119355e-01 -1.33444920e-01 3.71915936e-01
3.84227544e-01 -5.19912541e-01 -1.15632582e+00 4.96630937e-01
-8.69737744e-01 9.22722742e-02 5.76128006e-01 6.51022732e-01
9.06657100e-01 -1.48924053e-01 3.88597339e-01 -2.73989856e-01
2.75169641e-01 -2.99994588e-01 -1.99475184e-01 2.87253022e-01
-3.54510903e-01 1.07241489e-01 3.48410279e-01 -4.59994316e-01
-8.78773034e-01 9.33095664e-02 1.86592907e-01 -8.31396639e-01
7.71009829e-03 2.42268547e-01 -4.03068125e-01 3.06272835e-01
3.19341391e-01 2.84117550e-01 -3.48683774e-01 -3.38204235e-01
3.74436766e-01 1.86173007e-01 4.17949408e-01 -7.33199775e-01
4.96500462e-01 4.56061780e-01 -9.01569128e-02 -1.18187197e-01
-1.30020666e+00 -6.68567896e-01 -5.60411096e-01 -4.58320946e-01
1.32637799e+00 -1.09649003e+00 -4.95962471e-01 4.08907115e-01
-1.05423415e+00 -2.97679096e-01 -3.27715099e-01 3.97188455e-01
-8.06252658e-01 4.96050864e-01 -6.25592351e-01 -6.62368655e-01
-1.14281379e-01 -1.45763421e+00 1.40562701e+00 2.16815144e-01
9.19233635e-03 -1.49036869e-01 1.74369544e-01 7.58937418e-01
-1.52543392e-02 3.26525241e-01 7.58246124e-01 -6.12590015e-02
-7.70262897e-01 2.13965893e-01 -3.32287371e-01 4.32549059e-01
2.89274305e-01 -6.65343106e-02 -7.68137395e-01 -2.04200029e-01
-2.36515105e-01 -5.94102144e-01 1.31492579e+00 4.27576989e-01
1.57934308e+00 -3.11778039e-01 -6.52767159e-03 4.87908512e-01
1.05958855e+00 3.39507997e-01 8.28053057e-01 3.79951030e-01
8.93784881e-01 4.51057762e-01 1.02611196e+00 5.97282708e-01
5.38713396e-01 9.95079517e-01 5.22982121e-01 2.52356321e-01
-2.04368070e-01 -2.88725227e-01 7.97950029e-01 5.89244068e-01
-2.59601772e-01 -9.50765684e-02 -4.89213020e-01 5.91306388e-01
-2.17395401e+00 -1.24166167e+00 1.36351973e-01 2.13748789e+00
9.46125507e-01 1.54032916e-01 4.47274953e-01 1.19910233e-01
5.68188846e-01 7.16599047e-01 -6.97067499e-01 2.85141207e-02
1.82774469e-01 7.84524251e-03 1.52459711e-01 1.79438755e-01
-1.28195894e+00 7.88881600e-01 5.66146755e+00 9.38804030e-01
-7.88955569e-01 1.30286768e-01 7.48679936e-01 -5.47628045e-01
-2.18554437e-01 -1.39679745e-01 -5.28531015e-01 6.40230596e-01
7.89505780e-01 2.36796126e-01 2.33556405e-01 6.52928948e-01
5.94254017e-01 5.67558035e-02 -1.18036997e+00 1.10046530e+00
3.75411846e-02 -9.21904087e-01 3.73776019e-01 3.19730006e-02
6.80867374e-01 -2.84225941e-01 1.55766606e-01 5.29150426e-01
5.59619851e-02 -9.70321774e-01 9.41477537e-01 8.36972237e-01
6.86410904e-01 -6.67173862e-01 2.64047414e-01 -7.93685839e-02
-1.59467149e+00 -3.79970253e-01 -3.45819801e-01 -2.83883531e-02
2.08198011e-01 1.97978348e-01 1.18957922e-01 4.33334827e-01
9.43181634e-01 1.09880912e+00 -6.17008150e-01 1.02595365e+00
-9.05811116e-02 6.40541136e-01 4.71578985e-02 4.16520417e-01
4.66443151e-01 -8.83223712e-02 3.67992967e-01 8.41831863e-01
3.51759166e-01 3.93944561e-01 4.58231628e-01 6.14909112e-01
-6.16638474e-02 8.62944871e-02 -3.91562209e-02 -9.85815153e-02
2.54484266e-01 8.78574610e-01 -2.74341553e-01 -1.28675506e-01
-7.32793808e-01 1.30743742e+00 4.05733079e-01 3.97349179e-01
-1.08157170e+00 1.77858751e-02 9.88028526e-01 2.14243576e-01
5.47883272e-01 -1.19408756e-01 1.17815379e-02 -1.24891603e+00
4.21243638e-01 -1.10142076e+00 8.16891134e-01 -7.77186215e-01
-1.27573347e+00 3.46551299e-01 3.53219137e-02 -1.71408522e+00
-1.31138429e-01 -3.71466249e-01 -6.52811944e-01 5.40630043e-01
-1.57064843e+00 -1.25178587e+00 -5.35914063e-01 8.01082432e-01
9.88500595e-01 3.09268590e-02 4.83036846e-01 4.72103059e-01
-8.10861826e-01 6.71788156e-01 -4.01160628e-01 2.36491472e-01
8.78168344e-01 -1.02522266e+00 1.61787793e-02 9.04037595e-01
6.30907109e-03 2.65500635e-01 3.54179949e-01 -4.62506413e-01
-1.14819837e+00 -1.11838841e+00 5.36158025e-01 -4.97309744e-01
4.45563436e-01 -7.11356699e-02 -1.01097095e+00 5.23441494e-01
-9.78554636e-02 5.01640081e-01 5.95868230e-01 -2.23800298e-02
-7.58593738e-01 -4.36695218e-01 -7.91079879e-01 4.52612400e-01
1.30398798e+00 -6.81055069e-01 -4.80200261e-01 7.00194314e-02
8.95773053e-01 -1.74256966e-01 -8.11277986e-01 5.94520450e-01
8.31223965e-01 -1.20017910e+00 9.43969727e-01 -8.36736917e-01
8.79318774e-01 -5.70571840e-01 -3.40778440e-01 -7.87616193e-01
-5.18558860e-01 -3.65616292e-01 -5.08041084e-01 1.19391990e+00
1.28794581e-01 -1.02949202e-01 7.58987904e-01 5.85863113e-01
-2.27782741e-01 -1.11623335e+00 -1.18722844e+00 -6.14568949e-01
-2.72778869e-01 -5.04641414e-01 4.26012635e-01 6.59234464e-01
-2.83645302e-01 2.44207889e-01 -7.02934802e-01 1.18596338e-01
4.79452461e-01 1.97862923e-01 5.35835147e-01 -9.87702727e-01
-5.27130187e-01 -6.15849316e-01 -6.32387757e-01 -1.37872016e+00
2.00137366e-02 -4.33595270e-01 1.31109804e-01 -1.59098971e+00
6.43886924e-01 -1.21393844e-01 -9.44125593e-01 6.07737243e-01
-5.51110387e-01 2.57793128e-01 1.55119002e-01 3.95540297e-01
-1.26993632e+00 1.00596845e+00 1.59636021e+00 -4.13214326e-01
-9.51006040e-02 -1.13753729e-01 -6.28345191e-01 6.50745630e-01
4.23670381e-01 -5.62555432e-01 -6.59637630e-01 -5.49246788e-01
3.87488231e-02 2.12350488e-01 4.11744922e-01 -1.09042323e+00
-1.38107404e-01 -5.93604743e-01 4.78835523e-01 -6.45976007e-01
5.31842291e-01 -6.01060510e-01 -2.21202463e-01 2.77904958e-01
-5.82126677e-01 2.80761160e-02 -1.04353756e-01 9.76610780e-01
-3.77039552e-01 1.62246421e-01 8.07870209e-01 -2.81434935e-02
-9.35460567e-01 8.59301209e-01 7.39010721e-02 3.20733249e-01
9.57382441e-01 -2.96986043e-01 -2.46845454e-01 -3.70093644e-01
-5.62835693e-01 3.71369064e-01 3.23785841e-01 6.34816229e-01
7.15736330e-01 -1.88933408e+00 -6.86943650e-01 -1.15532964e-01
3.60618591e-01 -1.45852998e-01 6.45515442e-01 8.99072766e-01
-7.65931979e-02 2.42813289e-01 -4.88965809e-01 -6.66945577e-01
-1.23469424e+00 4.46069270e-01 4.73050088e-01 -4.75168765e-01
-6.22420013e-01 7.39371777e-01 6.59440935e-01 2.95049608e-01
4.00741577e-01 -4.12575930e-01 -3.61968070e-01 1.68070644e-02
8.47110987e-01 4.44300652e-01 -3.27207983e-01 -8.48443866e-01
-4.50643420e-01 6.82088971e-01 -1.34448797e-01 -8.00362155e-02
1.17614639e+00 -1.98604882e-01 2.62403399e-01 4.54500675e-01
1.24277532e+00 -2.44833484e-01 -1.85374331e+00 -3.73375177e-01
-2.96820551e-01 -7.96032190e-01 6.74851984e-02 -6.13808990e-01
-1.24557066e+00 9.63354766e-01 8.07898164e-01 -2.70425856e-01
1.52396333e+00 3.39261442e-02 1.00700700e+00 1.05437860e-01
1.72560304e-01 -1.03119779e+00 8.15476418e-01 3.10842335e-01
9.84589279e-01 -1.35168862e+00 -5.61255924e-02 -2.26990759e-01
-7.70688891e-01 9.50916052e-01 9.53541279e-01 -4.24636565e-02
2.50492156e-01 -4.17966694e-01 -3.89171392e-02 -1.50621533e-01
-7.90593386e-01 -3.52825433e-01 7.23909318e-01 4.31417495e-01
4.55001622e-01 -1.31083027e-01 -3.32435369e-01 8.26589167e-01
4.38510269e-01 2.48769484e-02 1.19494125e-01 8.04181874e-01
-5.47242820e-01 -1.01531756e+00 -1.21070910e-03 3.74955475e-01
-6.82616591e-01 1.29516810e-01 -1.08116597e-01 4.69383299e-01
1.41062662e-01 9.48264778e-01 -4.48405147e-02 -4.67185438e-01
6.36125565e-01 -4.50190343e-02 5.32790840e-01 -4.80238438e-01
-2.71201521e-01 1.29198238e-01 3.37144211e-02 -1.43752170e+00
-7.28517592e-01 -8.20761561e-01 -1.23185909e+00 1.01550922e-01
1.22048959e-01 -3.16660345e-01 -1.28096342e-01 1.00844419e+00
2.81469941e-01 7.88296819e-01 7.04550862e-01 -6.58786476e-01
-5.13019741e-01 -8.19960833e-01 -4.74072784e-01 7.03342140e-01
4.69815969e-01 -9.56931770e-01 -5.95047772e-02 1.08012177e-01] | [8.431574821472168, 0.7042214274406433] |
f88b4aae-923b-4da8-b9f4-a0f374797813 | grounded-keys-to-text-generation-towards | 2212.01956 | null | https://arxiv.org/abs/2212.01956v1 | https://arxiv.org/pdf/2212.01956v1.pdf | Grounded Keys-to-Text Generation: Towards Factual Open-Ended Generation | Large pre-trained language models have recently enabled open-ended generation frameworks (e.g., prompt-to-text NLG) to tackle a variety of tasks going beyond the traditional data-to-text generation. While this framework is more general, it is under-specified and often leads to a lack of controllability restricting their real-world usage. We propose a new grounded keys-to-text generation task: the task is to generate a factual description about an entity given a set of guiding keys, and grounding passages. To address this task, we introduce a new dataset, called EntDeGen. Inspired by recent QA-based evaluation measures, we propose an automatic metric, MAFE, for factual correctness of generated descriptions. Our EntDescriptor model is equipped with strong rankers to fetch helpful passages and generate entity descriptions. Experimental result shows a good correlation (60.14) between our proposed metric and human judgments of factuality. Our rankers significantly improved the factual correctness of generated descriptions (15.95% and 34.51% relative gains in recall and precision). Finally, our ablation study highlights the benefit of combining keys and groundings. | ['Jianfeng Gao', 'Snigdha Chaturvedi', 'Bill Dolan', 'Sudha Rao', 'Michel Galley', 'Baolin Peng', 'Faeze Brahman'] | 2022-12-04 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [ 2.88827419e-02 6.26428902e-01 -6.83244541e-02 -3.83751035e-01
-1.39132488e+00 -8.02803993e-01 1.16033053e+00 1.99685663e-01
-3.72979343e-01 1.13180876e+00 5.61617076e-01 -1.16651915e-01
-4.00380380e-02 -9.78935540e-01 -4.34468061e-01 -1.86070710e-01
2.81295955e-01 6.30620718e-01 -2.44111288e-02 -6.05827451e-01
4.96567011e-01 -2.31564730e-01 -1.48466241e+00 6.10986173e-01
1.47313631e+00 6.85851038e-01 3.89688641e-01 5.63245654e-01
-3.53651434e-01 7.37344980e-01 -9.40096855e-01 -9.51254249e-01
-2.53237765e-02 -5.99256575e-01 -1.09992218e+00 -3.14025372e-01
4.71423715e-01 -3.27349365e-01 7.08592832e-02 7.86824703e-01
5.49727142e-01 1.86316475e-01 8.53779137e-01 -1.36260808e+00
-1.07176733e+00 1.09589744e+00 1.09743044e-01 -1.27696604e-01
8.68045926e-01 2.08071515e-01 1.38693607e+00 -1.08658862e+00
8.65512490e-01 1.09773815e+00 4.35056627e-01 7.71093845e-01
-8.10901999e-01 -5.59926748e-01 6.41048560e-03 9.47916973e-03
-1.30719686e+00 -4.43521321e-01 3.12386751e-01 -5.27518928e-01
9.55567181e-01 5.43429434e-01 1.95489287e-01 1.32525682e+00
-1.59633383e-01 6.54155314e-01 9.30885136e-01 -4.45583969e-01
8.13567340e-02 4.06030297e-01 -1.56388059e-01 4.73616302e-01
4.95023370e-01 -6.47146404e-02 -6.98925436e-01 -1.85851008e-01
4.45054919e-01 -4.29202586e-01 -4.26242113e-01 1.87561333e-01
-1.65441859e+00 9.48778033e-01 2.40300253e-01 3.86557221e-01
-4.22611475e-01 -2.44067624e-01 3.39655817e-01 1.43230155e-01
3.45167190e-01 1.27284157e+00 -2.45172948e-01 -3.67701679e-01
-9.70911503e-01 8.35144937e-01 1.01689434e+00 1.12456548e+00
5.95270991e-01 -2.61695623e-01 -9.01338518e-01 8.06685865e-01
4.22587153e-03 6.26821816e-01 5.93682945e-01 -7.68059790e-01
8.58816803e-01 7.08642125e-01 6.80901587e-01 -1.14517558e+00
-2.12864250e-01 -4.21203494e-01 -7.74229586e-01 -3.63361746e-01
2.48918906e-01 -2.98303068e-01 -4.35623169e-01 1.75550842e+00
8.65900889e-02 -3.37348074e-01 2.57046998e-01 9.34410274e-01
1.04524446e+00 8.42178285e-01 1.94945887e-01 -1.66397810e-01
1.18324685e+00 -9.15476441e-01 -7.84779966e-01 -7.68061802e-02
7.46065021e-01 -8.73710871e-01 1.47388601e+00 3.00099909e-01
-1.02352130e+00 -4.59109962e-01 -8.07573557e-01 -1.13463901e-01
-5.29186964e-01 4.16850924e-01 5.29132783e-01 3.05038512e-01
-1.04483831e+00 5.31495631e-01 -2.16318816e-01 -4.47921932e-01
-5.66481873e-02 -6.80367202e-02 -2.66692489e-01 1.29512891e-01
-1.67751133e+00 8.58005226e-01 7.27046669e-01 -3.21859449e-01
-8.07941556e-01 -7.29996681e-01 -7.97818363e-01 -5.65346219e-02
4.55476284e-01 -1.17870414e+00 1.49183500e+00 -4.22237188e-01
-1.25278139e+00 7.86569536e-01 -4.26591970e-02 -4.85495210e-01
6.40240431e-01 -5.56786060e-01 -5.97153544e-01 9.61767510e-02
7.69337118e-01 8.88491869e-01 4.80452061e-01 -1.23970687e+00
-7.97359288e-01 2.87286490e-01 4.40504283e-01 3.63674074e-01
-4.86987293e-01 -1.33544818e-01 -1.61869183e-01 -7.80620396e-01
-1.77132905e-01 -6.05388641e-01 -1.72045380e-01 -3.93881083e-01
-9.25855517e-01 -5.15217423e-01 1.48593873e-01 -6.83758199e-01
1.67568696e+00 -1.57567608e+00 -3.37628186e-01 -9.35551375e-02
2.29854822e-01 2.80545652e-01 -2.76141733e-01 1.03043473e+00
2.68888235e-01 5.37950099e-01 -1.43915489e-01 -1.07445695e-01
4.11254913e-01 -2.17996553e-01 -8.17086458e-01 -4.16892260e-01
2.89859891e-01 9.91807282e-01 -1.30030894e+00 -5.85356653e-01
-2.89771799e-02 1.71655834e-01 -6.20006382e-01 4.47485089e-01
-5.51629364e-01 2.24323303e-01 -6.05609953e-01 3.79537642e-01
2.20692649e-01 -4.79836047e-01 -2.02395961e-01 -1.27481520e-01
-1.42532317e-02 8.52981985e-01 -9.95161712e-01 1.77100408e+00
-7.35628724e-01 3.49408448e-01 -7.47825980e-01 -2.16865033e-01
1.01637554e+00 5.13766408e-01 -2.99477633e-02 -7.29183197e-01
-1.05664767e-01 3.78928959e-01 -2.94355094e-01 -6.14381254e-01
1.24018455e+00 -8.38415474e-02 -4.91184533e-01 7.47213900e-01
5.61284646e-02 -3.37603748e-01 6.62586093e-01 7.89644480e-01
1.05971873e+00 1.80363983e-01 4.55351144e-01 -1.10522568e-01
4.67528492e-01 2.76540279e-01 1.67982012e-01 9.81980622e-01
2.21840248e-01 6.34408414e-01 3.23058695e-01 -2.05676004e-01
-9.50378895e-01 -9.68560636e-01 1.60936818e-01 8.66731763e-01
1.57211035e-01 -9.85821605e-01 -8.47882032e-01 -9.08103824e-01
-2.33904392e-01 1.29460478e+00 -4.59112048e-01 -1.03692085e-01
-3.04261476e-01 -3.42488468e-01 7.29436696e-01 1.44745275e-01
5.03339410e-01 -1.19119680e+00 -5.32267094e-01 3.50096852e-01
-9.99497235e-01 -1.29033804e+00 -4.82268631e-01 -6.23692036e-01
-5.32158792e-01 -9.18638527e-01 -7.21453190e-01 -5.02969861e-01
5.17760098e-01 9.34736058e-02 1.59786558e+00 -1.27914205e-01
1.99175581e-01 2.22787827e-01 -7.50795186e-01 -2.55223483e-01
-7.27333307e-01 3.69943351e-01 -2.26246454e-02 -2.14890748e-01
2.53608853e-01 -3.03895950e-01 -6.01579249e-01 3.37444365e-01
-9.28721607e-01 5.80303013e-01 6.45808101e-01 7.63641059e-01
4.67737198e-01 -2.40286782e-01 9.42183852e-01 -8.77183616e-01
1.35276234e+00 -5.28173208e-01 -1.43737718e-01 4.90874231e-01
-7.54968464e-01 2.72294462e-01 8.78687978e-01 -2.36140341e-01
-1.16509902e+00 -5.55287123e-01 -1.14916131e-01 1.57563776e-01
-1.03944264e-01 6.23739779e-01 -1.79803431e-01 5.20495951e-01
1.03186262e+00 2.73724109e-01 -6.48603678e-01 -2.54341841e-01
7.84509361e-01 7.07331061e-01 6.85967982e-01 -7.48825133e-01
1.02993751e+00 2.66233459e-02 -5.46145499e-01 -3.59453440e-01
-1.14713144e+00 -2.88398743e-01 -2.19167203e-01 -1.41971827e-01
6.68709040e-01 -1.04527807e+00 -4.48144466e-01 3.62974331e-02
-1.52005661e+00 -2.24340513e-01 -3.51615012e-01 2.91356832e-01
-5.97902119e-01 3.09706390e-01 -2.42853075e-01 -7.29040563e-01
-8.68174434e-01 -5.42599678e-01 1.18544102e+00 1.65937856e-01
-6.86147213e-01 -8.10123742e-01 1.66713789e-01 4.37324554e-01
4.70007122e-01 4.30870742e-01 7.38204777e-01 -8.97457004e-01
-4.96919543e-01 -2.47335076e-01 -1.66220218e-01 7.79532418e-02
1.38261855e-01 8.63114092e-03 -7.63071418e-01 8.78369361e-02
-3.52929711e-01 -5.58729827e-01 5.63982487e-01 -2.02313244e-01
8.46749187e-01 -9.61581230e-01 -2.78206468e-01 1.91683531e-01
1.22947729e+00 2.97255274e-02 5.99153459e-01 4.36419427e-01
5.31178892e-01 7.39819407e-01 1.19097745e+00 7.20560908e-01
7.14549839e-01 7.75249481e-01 1.15329333e-01 8.15768912e-02
-1.83022574e-01 -8.71615112e-01 3.14460695e-01 8.60094666e-01
5.81115037e-02 -7.08531976e-01 -8.61334860e-01 8.25773776e-01
-1.67814672e+00 -1.15629923e+00 -1.73569053e-01 2.00835657e+00
1.09567380e+00 1.14565469e-01 -1.09992869e-01 -1.08299464e-01
6.66915476e-01 1.65567547e-02 -1.16036430e-01 -2.76808530e-01
-1.59290105e-01 1.22812286e-01 -2.13707879e-01 4.95410293e-01
-8.19248438e-01 1.11596656e+00 5.72799015e+00 1.03064823e+00
-8.12775612e-01 1.84616875e-02 4.95656699e-01 8.20946023e-02
-8.13364923e-01 1.13571738e-03 -1.04102540e+00 5.43695331e-01
8.88085186e-01 -7.72952497e-01 1.15138844e-01 7.73781002e-01
3.90895277e-01 4.64734547e-02 -1.24562645e+00 9.08201098e-01
2.08507031e-01 -1.38503778e+00 7.47583687e-01 -2.24931002e-01
9.50166523e-01 -5.01510739e-01 -1.08816989e-01 5.68643034e-01
5.10186136e-01 -1.00226295e+00 9.38213766e-01 5.22347569e-01
1.10854352e+00 -5.48893988e-01 7.15918422e-01 5.54675341e-01
-9.32815135e-01 2.21558943e-01 -1.94432214e-01 -1.57541931e-01
4.99747485e-01 8.24366093e-01 -1.43415177e+00 7.68616676e-01
2.43283883e-01 3.64940584e-01 -6.10072374e-01 8.59447360e-01
-7.19879389e-01 3.39191169e-01 1.60087198e-02 -4.40620601e-01
3.26716334e-01 1.47420511e-01 6.18119061e-01 1.39827371e+00
7.31247306e-01 8.50181729e-02 -1.08065736e-02 1.36366856e+00
-2.04085454e-01 3.50936264e-01 -6.53331041e-01 -2.23793581e-01
7.99507260e-01 1.44701183e+00 -3.94098878e-01 -6.02758050e-01
-6.01833779e-03 1.05129921e+00 3.48552138e-01 3.04004550e-01
-8.04947197e-01 -7.43865073e-01 3.23214918e-01 1.61751986e-01
-9.64903645e-03 -9.01876204e-03 -1.91290408e-01 -1.29836285e+00
2.94317275e-01 -1.03150344e+00 3.44798297e-01 -1.11099517e+00
-1.32030320e+00 9.48230982e-01 1.18006513e-01 -1.40866911e+00
-8.04611742e-01 -4.81220558e-02 -7.37511158e-01 9.30606902e-01
-1.45756769e+00 -1.01269329e+00 -5.00863910e-01 4.71377492e-01
6.04390204e-01 -3.77594270e-02 9.29132700e-01 2.45665848e-01
-1.63869664e-01 7.62925744e-01 -4.26387042e-01 1.34984821e-01
9.61413324e-01 -1.40586686e+00 7.93989480e-01 1.06577909e+00
3.25620770e-01 9.90827382e-01 9.58811998e-01 -7.65201807e-01
-8.49073112e-01 -1.29413247e+00 1.70344007e+00 -6.34343147e-01
5.91781139e-01 -3.73494267e-01 -6.80147350e-01 2.76723206e-01
2.52789706e-01 -6.41610384e-01 6.52125478e-01 -6.54974282e-02
-2.54305661e-01 1.59885287e-01 -1.00887823e+00 8.95757377e-01
1.24142945e+00 -5.41232526e-01 -8.80286455e-01 4.49783713e-01
1.03625941e+00 -3.44580799e-01 -6.60319805e-01 2.69992232e-01
2.73304850e-01 -9.25361633e-01 6.88997149e-01 -4.86631691e-01
8.58366311e-01 -4.25725728e-01 2.28889883e-02 -1.38679206e+00
-1.14797361e-01 -1.00261617e+00 -1.12431087e-01 1.70379019e+00
7.38830447e-01 -3.62066537e-01 3.75540435e-01 7.48337269e-01
-3.61284465e-01 -6.36476278e-01 -5.72566152e-01 -8.88257504e-01
-9.35643837e-02 -2.94207424e-01 8.85648906e-01 1.01380098e+00
2.50824273e-01 5.73301256e-01 -2.85580188e-01 -1.21233249e-02
2.13374034e-01 2.21020177e-01 9.23197150e-01 -1.00141144e+00
-2.04295829e-01 -3.18475306e-01 4.63069789e-02 -1.16564667e+00
-8.79778638e-02 -1.01285887e+00 1.65114537e-01 -1.99364233e+00
2.24510476e-01 -4.82527763e-01 1.07651941e-01 5.06776154e-01
-5.31332910e-01 -7.64157809e-03 1.33067191e-01 1.71675071e-01
-9.17519987e-01 6.94154978e-01 1.37535453e+00 7.32377023e-02
-3.11287165e-01 -2.00962514e-01 -1.20258546e+00 3.74599278e-01
7.48288035e-01 -4.75231349e-01 -5.08020103e-01 -3.70172530e-01
6.08946681e-01 9.61859524e-02 4.38519865e-01 -1.06956458e+00
1.43958449e-01 -2.52828568e-01 -7.18126893e-02 -4.60997462e-01
-4.91323695e-02 -1.99128896e-01 1.12054303e-01 1.25640675e-01
-7.37066984e-01 9.81606022e-02 -9.29479003e-02 1.63652733e-01
-4.08877641e-01 -3.13308537e-01 9.56028029e-02 -1.19717501e-01
-7.02932537e-01 1.86752051e-01 -6.23392388e-02 6.08822823e-01
7.48236597e-01 -6.12008832e-02 -7.11889565e-01 -6.55094743e-01
-2.16990665e-01 1.10207103e-01 3.05435926e-01 6.46281362e-01
6.12608492e-01 -1.47875154e+00 -1.08569992e+00 -2.24379852e-01
5.46290398e-01 -8.29718336e-02 1.25337884e-01 4.66778159e-01
-2.25926101e-01 8.22162151e-01 7.28937937e-03 -2.18594477e-01
-7.88585901e-01 2.43077666e-01 2.29179394e-02 -7.17652798e-01
-3.54160994e-01 7.02616870e-01 2.52849221e-01 -4.20199543e-01
-9.41730440e-02 -3.92146766e-01 -2.74890661e-01 1.28759190e-01
7.39597023e-01 2.96385199e-01 1.49174988e-01 -4.02811110e-01
2.64247810e-03 1.54249698e-01 -1.45838067e-01 -4.43866700e-01
9.20497537e-01 -1.23401135e-01 4.87308428e-02 2.08436087e-01
7.69550085e-01 1.93532422e-01 -8.17340791e-01 -6.46047071e-02
1.68981329e-01 -3.79580677e-01 -4.23567981e-01 -1.41684079e+00
-3.64356846e-01 7.38337159e-01 -6.89676702e-02 4.20462132e-01
9.09278989e-01 -5.37510216e-02 1.12319148e+00 6.46924138e-01
6.22662485e-01 -9.29397106e-01 2.88560510e-01 6.92004144e-01
1.26950490e+00 -1.24060535e+00 -2.80651599e-01 -4.37224418e-01
-1.00296247e+00 9.61869597e-01 8.14937890e-01 3.01579922e-01
-1.84478790e-01 -1.87011108e-01 1.48899958e-01 -1.94221273e-01
-9.55312312e-01 -2.31726587e-01 5.09532571e-01 7.11977959e-01
6.29842818e-01 9.82345864e-02 -6.11417413e-01 1.03274930e+00
-9.65734363e-01 3.29194441e-02 5.30242741e-01 4.11990762e-01
-4.87732440e-01 -1.01102173e+00 -1.05707400e-01 3.92236739e-01
-4.10231799e-01 -3.81692976e-01 -5.72916746e-01 7.41606891e-01
8.00316781e-02 1.34643245e+00 -2.97950268e-01 -4.76791948e-01
3.75181019e-01 -3.81019562e-02 5.78153580e-02 -9.03030157e-01
-7.82986999e-01 -5.05047798e-01 5.36167979e-01 -4.24193323e-01
-2.38486484e-01 -2.84779340e-01 -1.19146609e+00 -2.67659664e-01
-3.20093274e-01 5.93142033e-01 5.00458896e-01 8.16498578e-01
5.96941829e-01 2.57745534e-01 5.46578825e-01 -3.89509022e-01
-5.98800600e-01 -1.10503149e+00 -2.89582133e-01 7.99329400e-01
4.74393219e-02 -4.50855821e-01 -2.85645962e-01 1.34152994e-01] | [11.815940856933594, 8.918071746826172] |
c74599f4-f1de-44e7-aa12-5998389a042f | group-shift-pointwise-convolution-for | 2109.12629 | null | https://arxiv.org/abs/2109.12629v1 | https://arxiv.org/pdf/2109.12629v1.pdf | Group Shift Pointwise Convolution for Volumetric Medical Image Segmentation | Recent studies have witnessed the effectiveness of 3D convolutions on segmenting volumetric medical images. Compared with the 2D counterparts, 3D convolutions can capture the spatial context in three dimensions. Nevertheless, models employing 3D convolutions introduce more trainable parameters and are more computationally complex, which may lead easily to model overfitting especially for medical applications with limited available training data. This paper aims to improve the effectiveness and efficiency of 3D convolutions by introducing a novel Group Shift Pointwise Convolution (GSP-Conv). GSP-Conv simplifies 3D convolutions into pointwise ones with 1x1x1 kernels, which dramatically reduces the number of model parameters and FLOPs (e.g. 27x fewer than 3D convolutions with 3x3x3 kernels). Na\"ive pointwise convolutions with limited receptive fields cannot make full use of the spatial image context. To address this problem, we propose a parameter-free operation, Group Shift (GS), which shifts the feature maps along with different spatial directions in an elegant way. With GS, pointwise convolutions can access features from different spatial locations, and the limited receptive fields of pointwise convolutions can be compensated. We evaluate the proposed methods on two datasets, PROMISE12 and BraTS18. Results show that our method, with substantially decreased model complexity, achieves comparable or even better performance than models employing 3D convolutions. | ['Yu Qiao', 'Lixu Gu', 'Shanshan Wang', 'Wanli Chen', 'Diping Song', 'Cheng Li', 'Jin Ye', 'Junjun He'] | 2021-09-26 | null | null | null | null | ['volumetric-medical-image-segmentation'] | ['medical'] | [ 1.52103305e-01 -7.46610686e-02 2.71774493e-02 -5.55537462e-01
-1.50498256e-01 -3.28095257e-01 4.21168774e-01 -3.09122000e-02
-8.40904117e-01 5.85443735e-01 -1.93061735e-02 -7.56555021e-01
-1.91318497e-01 -7.83752203e-01 -6.55398369e-01 -6.40851080e-01
-1.94748715e-01 -1.14599504e-01 3.91521335e-01 -2.28398219e-02
2.36431703e-01 1.02046347e+00 -1.05758989e+00 3.95531923e-01
8.22701037e-01 1.08463097e+00 5.31515598e-01 4.59513128e-01
-3.77364367e-01 3.57619584e-01 -3.20402265e-01 9.89293531e-02
2.80155867e-01 6.83981851e-02 -8.80315244e-01 1.12915546e-01
4.20768321e-01 -4.30671066e-01 -4.29886043e-01 7.81275094e-01
6.57466054e-01 9.91882980e-02 5.84798396e-01 -7.52874672e-01
-6.63605928e-01 1.87048361e-01 -6.47992074e-01 6.18676305e-01
-2.48186648e-01 1.48719281e-01 3.79894495e-01 -9.71964300e-01
2.54615843e-01 1.13239777e+00 9.59695399e-01 6.59229994e-01
-1.23103619e+00 -5.41300774e-01 9.99424979e-02 -2.43737966e-01
-1.51879644e+00 -1.68274611e-01 6.37127399e-01 -3.39739949e-01
1.20016372e+00 4.88215774e-01 5.33765554e-01 6.05502963e-01
2.28415430e-01 7.23343134e-01 1.16533589e+00 -1.20517902e-01
-1.02700936e-02 2.86609735e-02 1.43359914e-01 7.68278778e-01
1.39408395e-01 6.15022592e-02 -9.13227350e-02 -1.28502235e-01
1.41437531e+00 5.40382683e-01 -4.58010882e-01 -2.04760522e-01
-1.42603552e+00 8.26921046e-01 9.81197774e-01 6.17160320e-01
-5.13204336e-01 1.68206319e-01 2.29340240e-01 9.34708044e-02
5.37949204e-01 4.54323232e-01 -4.76802617e-01 2.05858484e-01
-9.59613562e-01 1.44762382e-01 9.91469920e-02 1.02958453e+00
7.20594168e-01 -1.82205126e-01 -3.75262707e-01 9.12175715e-01
-9.07909051e-02 4.79917049e-01 5.78655899e-01 -5.92245758e-01
5.29969752e-01 7.99257696e-01 -1.20541416e-01 -9.25327837e-01
-9.48510289e-01 -5.62547565e-01 -1.18054068e+00 1.54603541e-01
3.63382399e-01 -8.16393346e-02 -1.32839143e+00 1.38961649e+00
1.74744070e-01 5.30939773e-02 -2.01513320e-01 8.31465483e-01
1.00576746e+00 3.44261438e-01 4.66029271e-02 1.69560149e-01
1.39854836e+00 -8.10870290e-01 -5.13838053e-01 -1.32547706e-01
9.28820729e-01 -6.97775066e-01 1.08004165e+00 -1.58741862e-01
-1.26226640e+00 -6.01132631e-01 -8.88817072e-01 -7.25615919e-02
-5.37782609e-01 3.27471793e-02 9.14619982e-01 6.10738695e-01
-1.20578027e+00 7.11770535e-01 -1.02981901e+00 -2.02913955e-01
9.24586415e-01 7.02484548e-01 -3.90042841e-01 -1.84929907e-01
-1.07085943e+00 8.75277460e-01 3.91520768e-01 3.06098521e-01
-3.11542153e-01 -1.16685784e+00 -8.52265775e-01 1.76109180e-01
6.82227910e-02 -7.20542073e-01 1.00119436e+00 -3.05947900e-01
-1.09193349e+00 6.86525285e-01 -1.86911494e-01 -4.06520367e-01
4.49734688e-01 4.28479835e-02 -1.88294590e-01 2.07599834e-01
1.18672065e-02 9.08671081e-01 5.20857871e-01 -8.49246562e-01
-3.65588814e-01 -3.80734175e-01 1.83755666e-01 1.70494497e-01
-4.14550215e-01 -7.07537457e-02 -5.95318556e-01 -7.73113251e-01
5.79374313e-01 -8.57405961e-01 -5.38887024e-01 3.64636540e-01
-4.28792894e-01 1.53264314e-01 7.84946322e-01 -3.08010697e-01
1.13526320e+00 -2.26406813e+00 -3.25538844e-01 1.86935917e-01
6.01762295e-01 5.91598094e-01 5.86824305e-03 -1.07180052e-01
-1.92406520e-01 2.36811161e-01 -5.25836766e-01 -2.57765502e-01
-3.57460290e-01 4.60105062e-01 1.61613718e-01 3.11947674e-01
4.95750427e-01 1.21774733e+00 -7.59770393e-01 -5.90891540e-01
3.40663165e-01 7.22669363e-01 -8.05900335e-01 -2.71433264e-01
4.12912011e-01 5.22571325e-01 -6.14811599e-01 4.46953446e-01
1.18475997e+00 -5.55364490e-01 -3.00351113e-01 -2.98685282e-01
-2.10835412e-01 1.40208006e-01 -8.66627574e-01 1.67025316e+00
-5.49312472e-01 4.49237019e-01 -1.93855897e-01 -1.15493953e+00
9.08409119e-01 2.56922692e-01 5.84716201e-01 -8.08711231e-01
1.38535321e-01 2.32263327e-01 4.73317951e-02 -4.28306788e-01
1.70510560e-01 -3.26883912e-01 2.95710303e-02 1.79441690e-01
2.85861059e-03 1.72901843e-02 -1.71274543e-01 -1.18085556e-01
1.09008288e+00 -2.65274882e-01 1.41349345e-01 -4.91407275e-01
4.62234378e-01 -2.14455903e-01 4.10000652e-01 8.63142371e-01
-1.87000751e-01 7.50482738e-01 2.63186872e-01 -8.09981465e-01
-7.97405779e-01 -9.51598465e-01 -4.52481955e-01 5.52783847e-01
1.80188954e-01 -3.88129689e-02 -5.13058424e-01 -8.31618547e-01
1.44822866e-01 2.31636122e-01 -9.20716584e-01 -8.94818604e-02
-8.93809438e-01 -9.20818448e-01 6.47966623e-01 9.23344195e-01
8.24690998e-01 -8.76791418e-01 -9.35353994e-01 2.18539774e-01
4.80517112e-02 -1.07403183e+00 -6.81987882e-01 4.19076473e-01
-1.44769442e+00 -7.98032105e-01 -1.17514837e+00 -8.11311722e-01
1.16283798e+00 5.43103755e-01 8.87037277e-01 1.96839631e-01
-3.51491421e-01 -1.42763242e-01 -1.96324274e-01 -4.36479568e-01
1.33819640e-01 1.83836982e-01 -1.50320813e-01 -2.83915669e-01
3.80112916e-01 -4.95560169e-01 -1.02754319e+00 4.66569483e-01
-1.07207382e+00 3.50914776e-01 8.37472320e-01 1.08811271e+00
6.89751148e-01 -1.24511957e-01 4.40128207e-01 -1.00983763e+00
5.15641093e-01 -1.87814444e-01 -1.73292935e-01 2.43010540e-02
-3.54904890e-01 2.34286383e-01 5.82038581e-01 -4.63159025e-01
-1.01642227e+00 9.45353210e-02 -1.82541847e-01 -6.09855771e-01
-1.22877985e-01 4.10731137e-01 2.91446298e-01 -4.56041873e-01
7.81946063e-01 3.62542391e-01 1.93351477e-01 -7.05388069e-01
1.21812910e-01 4.54436511e-01 3.56304586e-01 -3.85244310e-01
5.86107850e-01 7.94412911e-01 1.99957192e-01 -6.97310984e-01
-4.61289942e-01 -4.21607792e-01 -8.61930132e-01 2.24723026e-01
8.69903386e-01 -5.34638166e-01 -5.07831395e-01 4.92445827e-01
-1.15760362e+00 -2.37268865e-01 -3.43183368e-01 6.72295034e-01
-3.94695550e-01 1.64109781e-01 -7.05077291e-01 -2.96150535e-01
-3.87239337e-01 -1.31821060e+00 9.12627637e-01 2.13635951e-01
-7.09359124e-02 -1.12372983e+00 -5.49083054e-01 -1.39297321e-01
7.67313778e-01 2.96602041e-01 1.09372616e+00 -5.26947379e-01
-3.69464189e-01 -2.16585442e-01 -5.72804868e-01 3.24824542e-01
3.73107880e-01 -7.04807580e-01 -9.09688652e-01 -2.48943329e-01
7.00949058e-02 2.36888349e-01 8.56662869e-01 6.88610375e-01
1.66658604e+00 -9.65725258e-02 -5.48983872e-01 9.23497081e-01
1.22793293e+00 3.14088076e-01 6.58086658e-01 1.62915587e-01
6.37001276e-01 2.54729807e-01 3.44612867e-01 3.42552066e-01
1.06417485e-01 5.08401453e-01 3.15118045e-01 -8.29616964e-01
-3.08217108e-01 8.10880959e-02 -5.13596714e-01 5.84371865e-01
-3.43432963e-01 2.40309134e-01 -8.30104470e-01 4.69521105e-01
-1.54382873e+00 -6.20121539e-01 -7.28468597e-02 2.00793934e+00
6.73152208e-01 1.74188033e-01 -2.91223437e-01 9.00973976e-02
5.86847544e-01 4.80885729e-02 -5.52139461e-01 -3.12439561e-01
1.01296669e-02 5.35140276e-01 8.43666732e-01 3.75838041e-01
-1.13453126e+00 5.92651308e-01 6.35538578e+00 8.43813658e-01
-1.47057664e+00 7.39889741e-02 7.19942927e-01 -3.55602682e-01
-1.84101582e-01 -3.27646613e-01 -6.87084138e-01 5.33142626e-01
3.42298865e-01 2.81323284e-01 3.88126671e-02 5.33273399e-01
2.78732449e-01 -1.04473568e-02 -1.06296647e+00 1.15852404e+00
-2.20256299e-01 -1.72916460e+00 1.18016563e-01 3.01906526e-01
5.87063074e-01 8.49433020e-02 2.36558840e-01 1.80588380e-01
-2.97339082e-01 -1.28727210e+00 3.01373750e-01 4.09227669e-01
8.82130444e-01 -6.07692599e-01 7.70281553e-01 3.67400795e-01
-1.08988440e+00 1.99313257e-02 -5.33332109e-01 -1.93713084e-02
1.04058474e-01 5.83013535e-01 -9.88157272e-01 4.17596459e-01
9.21755672e-01 3.48387837e-01 -4.71608013e-01 1.18493438e+00
2.21353725e-01 1.29713982e-01 -5.02483487e-01 -6.95117116e-02
6.30240738e-01 1.50013894e-01 2.58797109e-01 1.29233563e+00
4.17770177e-01 4.38793302e-01 -1.27289608e-01 9.15202439e-01
3.68295647e-02 -1.62039809e-02 -5.96889257e-01 4.56852168e-01
3.81314933e-01 9.23847735e-01 -9.14638281e-01 -3.50063682e-01
-4.61969376e-01 9.84221458e-01 2.03182757e-01 4.88608420e-01
-7.86856592e-01 -6.33079767e-01 6.38133109e-01 3.93584311e-01
5.56262732e-01 -3.27300340e-01 -5.71539760e-01 -7.29398429e-01
1.04461452e-02 -3.66017133e-01 2.68982053e-01 -5.25330782e-01
-1.05994916e+00 7.87818193e-01 7.45960400e-02 -1.17248416e+00
1.38656959e-01 -8.01837206e-01 -4.70960468e-01 1.22136557e+00
-1.80774713e+00 -9.52566981e-01 -3.65832925e-01 9.10581172e-01
2.57232308e-01 1.50033817e-01 8.04629982e-01 5.48714519e-01
-2.06812367e-01 7.83014059e-01 -3.56295519e-02 2.01271117e-01
4.33915406e-01 -1.01607370e+00 5.17935872e-01 4.02789652e-01
-3.66418719e-01 9.03776169e-01 1.16863012e-01 -3.13774914e-01
-1.02916598e+00 -1.13274050e+00 9.67345476e-01 -2.98082948e-01
1.31933719e-01 -1.94269761e-01 -1.04203761e+00 4.52413142e-01
-2.62853205e-01 7.34711230e-01 7.55885959e-01 -5.50274104e-02
-1.16677135e-01 1.74749549e-02 -1.47385263e+00 5.75485945e-01
1.37277365e+00 -3.91404569e-01 -4.00392950e-01 1.86654761e-01
6.29076302e-01 -6.57923937e-01 -1.10799062e+00 6.95789576e-01
4.12527412e-01 -8.19261074e-01 1.26900768e+00 -4.16533113e-01
1.52463734e-01 -3.47722203e-01 -4.23034653e-02 -1.15232706e+00
-5.67767680e-01 -2.80843586e-01 8.87050480e-02 4.67295796e-01
3.64525914e-01 -9.00122762e-01 7.51431465e-01 6.08985901e-01
-4.56771225e-01 -1.27959740e+00 -1.13390207e+00 -7.87400544e-01
3.97463232e-01 -3.43430400e-01 9.45090055e-01 9.16071475e-01
-1.17038898e-01 -1.34474576e-01 2.16378197e-02 1.45243987e-01
2.26765573e-01 5.00269011e-02 2.40810335e-01 -1.02089667e+00
-1.53469384e-01 -6.37625694e-01 -6.28601789e-01 -1.44119918e+00
-3.73556823e-01 -1.02126360e+00 -1.79869741e-01 -1.43804622e+00
5.50416149e-02 -9.63419795e-01 -5.48066437e-01 7.52697289e-01
-2.35580653e-01 5.49908996e-01 1.12772465e-01 1.61400437e-01
-5.00246249e-02 3.26469719e-01 2.01591420e+00 -1.96394652e-01
-2.95979083e-01 2.24458203e-02 -6.43717289e-01 7.59647071e-01
8.71639311e-01 -1.89438239e-01 -5.52313447e-01 -7.72656679e-01
-3.63590300e-01 -8.25237855e-02 4.07715738e-01 -1.00085962e+00
1.91601276e-01 1.24888401e-02 8.70540380e-01 -7.35801280e-01
3.59423131e-01 -8.19397211e-01 -9.04223546e-02 5.35945475e-01
-9.85633358e-02 4.58283015e-02 5.94311714e-01 6.69569671e-02
-1.46915376e-01 -1.86818376e-01 6.48402512e-01 -4.30317253e-01
-5.51336050e-01 6.69288099e-01 -1.70286730e-01 -4.07534719e-01
9.40290987e-01 -7.07760632e-01 -9.44118351e-02 1.79031238e-01
-1.00919211e+00 2.28117686e-02 1.91515371e-01 1.85143664e-01
8.80797684e-01 -1.47533405e+00 -4.04602945e-01 6.22555315e-01
-2.32376549e-02 5.31641841e-01 7.38480210e-01 1.10167480e+00
-6.28928363e-01 7.34805405e-01 -3.10229897e-01 -8.78023863e-01
-1.05792928e+00 5.88585019e-01 6.14615262e-01 -3.35369140e-01
-9.77321625e-01 9.17431593e-01 7.30910420e-01 -5.82330406e-01
3.35561559e-02 -8.59452367e-01 -2.00307667e-01 -3.20344031e-01
6.15771651e-01 9.10299122e-02 3.78396392e-01 -4.12161171e-01
-4.97686982e-01 5.74198067e-01 -3.75322878e-01 3.44717503e-01
1.36185455e+00 5.52086830e-02 1.88890427e-01 -3.65995914e-02
1.38189614e+00 -4.57940042e-01 -1.41294694e+00 -4.50732023e-01
-3.68556976e-01 -6.10756993e-01 2.89705157e-01 -6.82918072e-01
-1.26743865e+00 1.12406039e+00 7.39823759e-01 3.07888780e-02
1.20084786e+00 1.17411256e-01 8.76833618e-01 1.14745535e-01
2.81493068e-01 -7.45984793e-01 -1.21208191e-01 4.66851234e-01
7.95955420e-01 -1.03712714e+00 -9.98804942e-02 -6.12822294e-01
-2.92906135e-01 1.07021105e+00 6.01642907e-01 -6.56025931e-02
1.01400185e+00 3.09766233e-01 -4.11526486e-02 -4.84762549e-01
-1.09730497e-01 -1.11647896e-01 3.43005300e-01 6.92055941e-01
4.08943504e-01 1.10098548e-01 -2.75494933e-01 5.62955618e-01
2.98203602e-02 1.27247646e-01 9.69102755e-02 1.20183015e+00
-2.51648754e-01 -8.40879142e-01 -3.31808984e-01 6.49750590e-01
-2.97517717e-01 -3.16871911e-01 3.40663880e-01 8.70172083e-01
3.77342641e-01 4.92422372e-01 3.91283363e-01 -2.57064223e-01
6.10386789e-01 -1.19466990e-01 6.42192543e-01 -4.60850447e-01
-4.24019456e-01 5.69808893e-02 -3.77855927e-01 -4.99932557e-01
-4.62204039e-01 -3.97515446e-01 -1.44953096e+00 -4.13751125e-01
-3.07486176e-01 -2.06958294e-01 5.81259489e-01 8.17715704e-01
5.62996387e-01 8.45885932e-01 5.24146795e-01 -9.70393836e-01
-5.45464933e-01 -8.87223125e-01 -5.63447714e-01 2.65443325e-01
4.18368697e-01 -7.56053030e-01 2.63329912e-02 -2.19607070e-01] | [14.512409210205078, -2.5908234119415283] |
0ea92955-23d2-4626-87ab-c4175828abbb | towards-a-structured-representation-of | null | null | https://aclanthology.org/R13-1009 | https://aclanthology.org/R13-1009.pdf | Towards a Structured Representation of Generic Concepts and Relations in Large Text Corpora | null | ['Archana Bhattarai', 'Vasile Rus'] | 2013-09-01 | towards-a-structured-representation-of-1 | https://aclanthology.org/R13-1009 | https://aclanthology.org/R13-1009.pdf | ranlp-2013-9 | ['open-information-extraction'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.279925346374512, 3.847116470336914] |
9cd6bc87-84c8-490f-8a02-e9cacd60d69a | detecting-and-mitigating-hallucinations-in-1 | 2305.13632 | null | https://arxiv.org/abs/2305.13632v1 | https://arxiv.org/pdf/2305.13632v1.pdf | Detecting and Mitigating Hallucinations in Multilingual Summarisation | Hallucinations pose a significant challenge to the reliability of neural models for abstractive summarisation. While automatically generated summaries may be fluent, they often lack faithfulness to the original document. This issue becomes even more pronounced in low-resource settings, such as cross-lingual transfer. With the existing faithful metrics focusing on English, even measuring the extent of this phenomenon in cross-lingual settings is hard. To address this, we first develop a novel metric, mFACT, evaluating the faithfulness of non-English summaries, leveraging translation-based transfer from multiple English faithfulness metrics. We then propose a simple but effective method to reduce hallucinations with a cross-lingual transfer, which weighs the loss of each training example by its faithfulness score. Through extensive experiments in multiple languages, we demonstrate that mFACT is the metric that is most suited to detect hallucinations. Moreover, we find that our proposed loss weighting method drastically increases both performance and faithfulness according to both automatic and human evaluation when compared to strong baselines for cross-lingual transfer such as MAD-X. Our code and dataset are available at https://github.com/yfqiu-nlp/mfact-summ. | ['Shay B. Cohen', 'Edoardo M. Ponti', 'Anna Korhonen', 'Yftah Ziser', 'Yifu Qiu'] | 2023-05-23 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-4.58165491e-03 5.11379093e-02 -1.37652770e-01 -3.36472601e-01
-1.39507210e+00 -5.85968375e-01 7.68658519e-01 3.23862433e-01
-4.03470963e-01 9.56155241e-01 9.20945585e-01 9.89009589e-02
2.11060718e-01 -4.78682339e-01 -6.46602690e-01 -3.17457587e-01
3.49486381e-01 2.92579353e-01 -2.89657056e-01 -3.20611507e-01
2.66644806e-01 -3.22966352e-02 -9.57109869e-01 3.45067620e-01
1.31152081e+00 6.54903352e-01 1.26151815e-02 3.57485890e-01
2.94927299e-01 8.00753891e-01 -7.36336052e-01 -8.20714295e-01
-7.33052045e-02 -8.36165309e-01 -8.85339677e-01 -1.31686315e-01
7.15085864e-01 -3.16929936e-01 -3.65700841e-01 1.13273215e+00
8.10687900e-01 3.01121920e-01 7.64728904e-01 -9.67161655e-01
-9.66601372e-01 7.57285237e-01 -2.47998223e-01 2.38151222e-01
4.70751405e-01 2.70592362e-01 1.18323183e+00 -1.16822159e+00
5.05302131e-01 1.25848973e+00 7.97500551e-01 6.26013935e-01
-1.19438446e+00 -5.22001386e-01 -1.61244944e-01 2.49159932e-01
-1.14392662e+00 -8.16986442e-01 6.30664170e-01 -1.85099006e-01
9.76845741e-01 2.60165483e-01 3.27683806e-01 1.41298139e+00
8.71055499e-02 9.09875393e-01 1.10928988e+00 -3.22001427e-01
2.97107697e-02 2.08590686e-01 1.73691716e-02 5.97877800e-01
1.44990966e-01 -1.46653384e-01 -8.74271452e-01 4.04983526e-03
9.47827920e-02 -3.79446208e-01 -6.64616168e-01 9.73483250e-02
-1.28308058e+00 9.22320545e-01 4.63706642e-01 3.06937754e-01
-3.90345365e-01 8.80678464e-03 7.17257917e-01 3.88505429e-01
6.97891772e-01 7.86476374e-01 -2.63725907e-01 -4.06525224e-01
-1.28588510e+00 1.02549739e-01 7.62525499e-01 7.21800089e-01
5.20229042e-01 8.80925208e-02 -5.25376141e-01 1.21637845e+00
-1.03864513e-01 3.95922750e-01 7.94282734e-01 -9.81721282e-01
6.01075768e-01 2.36245558e-01 -2.05189615e-01 -9.72362220e-01
-1.37197196e-01 -6.12234592e-01 -9.93698418e-01 -2.23835304e-01
1.25520632e-01 -5.81930801e-02 -4.07385379e-01 2.17876267e+00
-1.31846517e-01 -1.93932638e-01 2.80120045e-01 7.65145540e-01
8.85248125e-01 7.87126124e-01 -3.86343002e-02 -3.01265866e-01
1.21055555e+00 -1.17495894e+00 -7.52993524e-01 -4.79131699e-01
4.93725210e-01 -8.64769220e-01 1.64392722e+00 2.64316678e-01
-1.41161406e+00 -2.72126377e-01 -1.20611668e+00 -4.48928356e-01
-7.35858977e-02 1.50579527e-01 1.45466819e-01 2.72965670e-01
-1.07829809e+00 8.71939719e-01 -6.83225036e-01 -4.61318851e-01
3.77785414e-01 -1.89606592e-01 -4.23319995e-01 -2.23652273e-01
-1.41262019e+00 1.53079522e+00 3.81101280e-01 -1.46619141e-01
-8.45249116e-01 -7.38326073e-01 -8.71355772e-01 2.00582206e-01
8.41159001e-02 -8.62772405e-01 1.35374355e+00 -7.66428709e-01
-1.34597480e+00 8.70841205e-01 -2.08398774e-01 -5.80499470e-01
6.66094720e-01 -4.71532822e-01 -3.19151253e-01 1.71861961e-01
3.60870630e-01 6.37224436e-01 6.13219321e-01 -1.05235910e+00
-1.73598573e-01 -6.85237870e-02 -1.38777703e-01 4.22767133e-01
-7.15366900e-01 -2.88395546e-02 -2.94084698e-01 -7.70391881e-01
-3.64579648e-01 -6.66400075e-01 2.27632046e-01 -1.74360111e-01
-6.11450136e-01 -3.95482779e-01 4.15671557e-01 -9.55033481e-01
1.27779520e+00 -1.90912521e+00 1.87645659e-01 -3.70756924e-01
1.47459611e-01 4.14381087e-01 -3.17354947e-01 7.70948231e-01
2.28323683e-01 5.04648052e-02 -6.70139253e-01 -5.57092488e-01
1.98561326e-01 -9.76287872e-02 -4.32521373e-01 3.30261230e-01
4.71178502e-01 9.83332813e-01 -1.00168955e+00 -4.08004463e-01
-9.22241062e-02 5.89549065e-01 -6.29971683e-01 2.13735342e-01
1.06840581e-01 2.58450001e-01 -7.88295548e-03 2.74208874e-01
3.08576554e-01 -2.27923140e-01 -1.63437933e-01 -1.84192002e-01
6.35363907e-02 8.46002638e-01 -4.25509900e-01 1.87454164e+00
-6.94951475e-01 6.98325396e-01 -3.30749780e-01 -8.66524220e-01
8.51651311e-01 4.12366509e-01 -1.60800125e-02 -7.74140894e-01
1.55972305e-03 3.91081929e-01 -7.40848631e-02 -3.08273047e-01
9.49381232e-01 -5.29484332e-01 -3.34562153e-01 6.57332957e-01
2.99614131e-01 -3.89707357e-01 4.34830844e-01 5.01616180e-01
1.16073823e+00 -9.31540430e-02 5.05019605e-01 -3.03887606e-01
2.95154065e-01 -1.11929648e-01 4.71066892e-01 6.23203874e-01
-3.07359904e-01 8.74712408e-01 5.05401015e-01 2.22156122e-01
-9.94162917e-01 -1.01191890e+00 -1.59727335e-01 8.70619714e-01
-1.92292571e-01 -5.74575365e-01 -9.05722260e-01 -5.08996785e-01
-1.40373260e-01 1.19979417e+00 -5.20618558e-01 -6.85572982e-01
-2.88504034e-01 -6.53706610e-01 8.81743789e-01 2.64730692e-01
5.27392209e-01 -1.14249003e+00 -4.29278076e-01 1.92483738e-01
-7.65454948e-01 -1.08426058e+00 -9.18298662e-01 -2.98181307e-02
-7.24986196e-01 -6.37824833e-01 -8.76995623e-01 -5.06456077e-01
2.26899043e-01 1.10556409e-01 1.40699077e+00 -4.07422297e-02
1.22664317e-01 1.73352629e-01 -3.25446874e-01 -2.17399597e-01
-7.44029164e-01 3.66665930e-01 1.47722900e-01 -3.06280315e-01
3.09335411e-01 -5.26418567e-01 -5.92178285e-01 -1.21391371e-01
-8.46685946e-01 6.02136031e-02 6.65233433e-01 8.85931671e-01
3.51313859e-01 -3.87143761e-01 9.61106122e-01 -8.58931124e-01
1.33984315e+00 -5.82365394e-01 1.77031323e-01 1.83270276e-01
-6.87279880e-01 6.14558980e-02 8.21702063e-01 -3.08745474e-01
-8.81080270e-01 -5.98736644e-01 -2.44102597e-01 -2.42567390e-01
9.94544327e-02 8.36432517e-01 -7.58073628e-02 4.67766970e-01
9.64519143e-01 3.04166019e-01 -6.81253374e-02 -3.63062352e-01
5.06958961e-01 7.51630843e-01 7.47365355e-01 -5.02166808e-01
5.80787003e-01 7.79112801e-02 -5.47581375e-01 -7.91992903e-01
-1.21837580e+00 -2.66615748e-01 -3.32293123e-01 -6.25808313e-02
6.08464777e-01 -1.02335572e+00 -2.99068600e-01 3.01544785e-01
-1.35601103e+00 -2.71680593e-01 -3.70470226e-01 5.48362255e-01
-5.73568881e-01 4.72300410e-01 -8.76345277e-01 -3.53599936e-01
-7.72706211e-01 -9.23591197e-01 8.52215946e-01 -1.23277940e-01
-7.36237049e-01 -1.17036486e+00 4.48359966e-01 4.39370513e-01
6.14148378e-01 1.94333404e-01 9.02913570e-01 -6.58356190e-01
4.27851304e-02 -2.44532853e-01 -1.44458726e-01 7.52073586e-01
1.70335025e-01 -2.45722353e-01 -9.79206085e-01 -3.73003513e-01
1.80058718e-01 -9.12460268e-01 1.12348568e+00 1.52591541e-01
7.00723529e-01 -7.28912950e-01 3.42111886e-01 2.55604893e-01
1.31892657e+00 -4.66829687e-01 7.12917805e-01 1.52029559e-01
6.36809170e-01 5.83237112e-01 4.62950528e-01 3.43138009e-01
6.28111005e-01 7.75476158e-01 -3.69484983e-02 -8.01658481e-02
-5.22757173e-01 -5.94022632e-01 9.11788642e-01 1.61879086e+00
1.35846987e-01 -4.09484148e-01 -7.47940123e-01 6.33986115e-01
-1.67618227e+00 -1.12042952e+00 1.11886278e-01 2.19550371e+00
1.42286026e+00 5.08563109e-02 4.89421599e-02 4.16157693e-02
6.39075279e-01 2.08643928e-01 -4.86740053e-01 -5.79645276e-01
-3.14356983e-01 6.21174723e-02 2.62492187e-02 7.10634112e-01
-7.40576446e-01 9.97991502e-01 5.84926271e+00 9.95327175e-01
-1.12272382e+00 3.21083426e-01 4.10455525e-01 -2.84086496e-01
-5.57358444e-01 -1.54632121e-01 -3.95310163e-01 5.66703320e-01
1.15649569e+00 -6.03154957e-01 3.03933024e-01 5.60037673e-01
3.71885329e-01 6.21267818e-02 -1.19009686e+00 6.14258409e-01
4.75860804e-01 -1.03438044e+00 2.40378127e-01 -2.00543165e-01
7.55801201e-01 1.97546288e-01 1.34569883e-01 4.54230458e-01
1.47293091e-01 -1.03642058e+00 9.10920024e-01 3.51752877e-01
8.89861166e-01 -7.47299790e-01 5.55034876e-01 3.72934461e-01
-5.66970885e-01 4.61548209e-01 -5.01783550e-01 1.05867006e-01
2.56774843e-01 7.87696064e-01 -8.27655911e-01 3.89660895e-01
3.50968421e-01 8.98143291e-01 -6.86904371e-01 8.38337421e-01
-5.21457016e-01 6.68573439e-01 8.48926511e-03 -2.69555170e-02
2.31695905e-01 -2.75650751e-02 6.37812555e-01 1.55732942e+00
5.19350827e-01 -2.52964109e-01 -2.25372329e-01 1.01440883e+00
-5.34899712e-01 2.71707386e-01 -6.94471836e-01 -1.28777385e-01
5.26892543e-01 1.14720011e+00 -3.76656801e-02 -4.92342800e-01
-3.05134177e-01 1.28583813e+00 7.20804930e-01 1.52213424e-01
-6.75849855e-01 -4.48019743e-01 3.52045596e-01 -7.70016387e-02
-5.79579622e-02 -6.32640794e-02 -2.50409067e-01 -1.46121347e+00
1.70921117e-01 -1.19325078e+00 2.22095087e-01 -8.53801668e-01
-1.49711001e+00 8.78250420e-01 -9.42064524e-02 -1.32464731e+00
-4.54636216e-01 -1.46160647e-01 -6.26079738e-01 8.21890712e-01
-1.57469571e+00 -1.07829571e+00 -9.40992162e-02 5.35645127e-01
7.14917600e-01 -1.17610380e-01 9.29024816e-01 2.85173386e-01
-4.80148137e-01 8.57543409e-01 -2.11597495e-02 -1.46719441e-01
1.15167606e+00 -1.19460428e+00 3.37965250e-01 1.16410923e+00
1.66378930e-01 6.65854335e-01 9.69012141e-01 -4.35075849e-01
-8.86381865e-01 -1.13517821e+00 1.28178096e+00 -4.45082545e-01
6.80680990e-01 -1.20689832e-01 -1.31846309e+00 4.48417395e-01
7.24346995e-01 -4.06552672e-01 7.36572385e-01 2.18575180e-01
-6.05077803e-01 1.40391782e-01 -9.68151808e-01 7.36442685e-01
8.04096103e-01 -8.27779412e-01 -8.52353930e-01 4.33551580e-01
6.41541421e-01 -1.50460795e-01 -8.67554069e-01 2.38299668e-01
2.88952500e-01 -1.00880611e+00 6.25655949e-01 -4.78445500e-01
1.06305385e+00 -6.39961334e-03 -1.50997221e-01 -1.92729366e+00
-2.90541589e-01 -6.04664445e-01 1.51167577e-02 1.41733968e+00
6.10099435e-01 -4.23422366e-01 2.68600702e-01 4.09668386e-01
-4.69365656e-01 -5.39320230e-01 -8.82548988e-01 -8.45225811e-01
6.35067999e-01 -2.35283688e-01 2.09460124e-01 1.14778674e+00
4.28933859e-01 8.99246216e-01 -6.16094410e-01 -3.03437322e-01
5.85145712e-01 -1.46848932e-01 5.23815513e-01 -7.73997843e-01
-3.14435691e-01 -6.32863224e-01 -6.90910667e-02 -7.76618779e-01
3.99987876e-01 -1.39883983e+00 2.86253721e-01 -1.68037534e+00
5.59507012e-01 2.97689587e-02 -2.08884165e-01 6.05837345e-01
-4.65981781e-01 4.74273533e-01 2.99176246e-01 3.85460883e-01
-6.29403234e-01 9.86426115e-01 1.18752801e+00 -9.38550010e-02
7.31454343e-02 -3.06926996e-01 -1.01207912e+00 5.19207537e-01
1.01751804e+00 -5.74892640e-01 -2.56826013e-01 -6.23127401e-01
7.51156881e-02 1.47227673e-02 3.04989666e-01 -9.78974938e-01
1.38021693e-01 2.10799649e-01 6.87436089e-02 -1.87891603e-01
3.95330608e-01 -9.89041030e-02 -1.96613088e-01 4.15519834e-01
-8.01281333e-01 1.71127111e-01 2.22052991e-01 2.87358940e-01
-4.74052697e-01 -3.16972345e-01 8.32553506e-01 -1.07744768e-01
-1.85695812e-01 3.51246223e-02 -2.78221577e-01 6.35165632e-01
3.48216355e-01 1.79669619e-01 -6.46180511e-01 -7.77060747e-01
-2.71780878e-01 9.24122408e-02 5.88154495e-01 4.37071055e-01
7.02566147e-01 -1.53041768e+00 -1.37069023e+00 -2.38703504e-01
3.61645073e-01 -4.83090878e-01 2.40410417e-01 1.04231358e+00
-8.33859220e-02 4.52467620e-01 -3.19925666e-01 -1.77052587e-01
-9.67577577e-01 1.68136477e-01 2.73867041e-01 -4.18683231e-01
-4.74647760e-01 6.42236650e-01 1.64345264e-01 -4.63510811e-01
5.79457730e-02 -4.07265462e-02 -6.40494749e-02 2.50702113e-01
5.65265894e-01 4.46700573e-01 4.07738477e-01 -7.57222831e-01
-3.04744780e-01 2.38572836e-01 -2.27417171e-01 -4.56362873e-01
1.27864742e+00 -1.19199902e-01 -1.04972877e-01 5.49634099e-01
1.42377460e+00 2.15537623e-01 -1.13919294e+00 -3.72389436e-01
-9.75100845e-02 -2.10264623e-01 1.85026247e-02 -9.79362428e-01
-7.92389095e-01 1.08339179e+00 8.78620446e-02 1.26594350e-01
9.89786506e-01 -7.95030743e-02 9.98585582e-01 3.47190559e-01
5.72430016e-03 -1.09294605e+00 3.32222134e-01 6.48650885e-01
1.49206531e+00 -1.18130934e+00 -2.58784126e-02 -1.44420257e-02
-1.07309830e+00 6.88767850e-01 3.16882610e-01 -3.68781574e-02
9.26865712e-02 -7.50554800e-02 1.21038340e-01 -5.36927804e-02
-1.01019001e+00 4.58268300e-02 4.45714921e-01 4.01513070e-01
8.03794146e-01 6.81168810e-02 -4.83270824e-01 6.42240882e-01
-7.38523066e-01 -1.04887031e-01 8.29040468e-01 4.77836102e-01
-3.49579155e-01 -8.97229850e-01 -7.96731785e-02 3.80737782e-01
-6.03544474e-01 -4.78400201e-01 -5.08544922e-01 4.87216085e-01
-4.20773476e-01 1.00498927e+00 -1.56914070e-01 -3.00039381e-01
4.69953209e-01 3.51798162e-02 5.02451599e-01 -7.19957113e-01
-6.08447790e-01 -1.65659994e-01 2.60115176e-01 -3.79121482e-01
-4.26731646e-01 -6.50439382e-01 -9.89712179e-01 -5.43160200e-01
-1.28979251e-01 1.65948212e-01 3.93026501e-01 8.25919151e-01
3.71385545e-01 4.32250977e-01 5.02268016e-01 -6.44642770e-01
-8.63776147e-01 -1.37545288e+00 -3.51751000e-01 6.78895533e-01
3.86951864e-01 -3.39640379e-01 -6.02657437e-01 -5.99609800e-02] | [12.130143165588379, 9.319857597351074] |
3db6a1be-de6d-4ad4-92c4-d39ca53b23f9 | phase-retrieval-exact-solution-based-on | 2005.05542 | null | http://arxiv.org/abs/2005.05542v1 | http://arxiv.org/pdf/2005.05542v1.pdf | Phase-retrieval exact solution based on window modulation | Quantitative phase imaging (QPI) is a promising tool for imaging complex
objects. By Combining self-reference interferometry and phase retrieval, this
paper proposes a general exact QPI for arbitrary complex-objects as well as a
one-shot exact QPI for transparent objects with small phase range (i.e., weak
scattering object). The optical configuration is similar to the Zernike phase
contrast, while the phase-shifting area is a little smaller. With three frames
of phase-shift, the algebraic relationship between the phase and the measured
intensity is built, providing an excellent approximate phase recovery. Then, an
efficient iterative optimization strategy is reported to turn that approximate
solution into an exact one. The iteration exhibits linear convergence property.
Thus exact phase map is achieved with a sufficient number of iterations. When
the object is a pure phase sample with a small phase ranges of [0,{\pi}], a
single intensity measurement will be enough for exact phase recovery by
selecting a tiny phase shifting value, based on a similar linear-convergence
iterative algorithm. The feasibility and accuracy of our method are verified by
both numerical simulations and experiments on diverse phase objects. | [] | 2020-05-12 | null | null | null | null | ['transparent-objects'] | ['computer-vision'] | [ 4.54231322e-01 -2.71900296e-01 1.12978205e-01 -2.54507452e-01
-3.52886647e-01 -1.38892472e-01 6.11028560e-02 -4.31260586e-01
-6.21741533e-01 8.17899048e-01 -5.25080979e-01 1.57237664e-01
-4.35325205e-01 -8.61206472e-01 -3.63281518e-01 -1.37051809e+00
9.19513628e-02 7.05666840e-01 5.10988176e-01 -2.37861928e-02
4.31652337e-01 1.13154888e-01 -1.07603908e+00 -3.93860787e-01
1.15420234e+00 9.59395230e-01 2.90341347e-01 1.38195843e-01
-9.71738920e-02 3.77056152e-01 -6.51881337e-01 1.02438159e-01
2.91303188e-01 -4.82527137e-01 -3.97429645e-01 -1.16366856e-01
-1.07655749e-02 -3.72686893e-01 -4.22377944e-01 1.37162101e+00
3.65418673e-01 -1.34044051e-01 4.89696711e-01 -1.00457251e+00
-3.57728124e-01 4.74445313e-01 -1.01082444e+00 -7.70184547e-02
5.35978615e-01 4.01538253e-01 5.26999176e-01 -5.83152652e-01
7.38353729e-01 8.26736569e-01 5.72859347e-01 2.53606081e-01
-1.13251913e+00 -8.33452940e-01 -4.27102089e-01 3.65066051e-01
-1.65488470e+00 -1.05713613e-01 1.07749987e+00 -1.47441179e-01
3.82655323e-01 1.75822109e-01 9.40316916e-01 -5.07917702e-02
4.20205295e-01 1.09214127e-01 1.73846018e+00 -4.85003710e-01
-1.47486869e-02 1.33615941e-01 3.30184937e-01 6.08058035e-01
6.76330566e-01 1.70033708e-01 -3.46381426e-01 6.98208064e-02
5.65474391e-01 1.25683501e-01 -9.57500696e-01 -4.76358950e-01
-1.59912026e+00 9.05648768e-02 6.15385234e-01 4.56221670e-01
-3.20372760e-01 -9.86846313e-02 -3.50604236e-01 3.02255809e-01
5.15136719e-02 5.66201985e-01 9.97124240e-02 4.18857932e-02
-9.03672397e-01 2.67252713e-01 9.46707606e-01 7.47174680e-01
1.23774970e+00 -2.30986737e-02 -2.08866253e-01 4.67017978e-01
5.28860271e-01 1.33589149e+00 1.38890311e-01 -8.17060590e-01
2.34099194e-01 4.00311440e-01 7.02725410e-01 -8.39810371e-01
-2.84979254e-01 -4.43828017e-01 -8.75546396e-01 4.37015176e-01
7.14701653e-01 -6.95982901e-03 -6.71798885e-01 1.58347070e+00
4.86797601e-01 1.83438718e-01 1.71262652e-01 1.13183463e+00
6.74480855e-01 8.78493011e-01 -5.63184798e-01 -1.20770049e+00
1.37491143e+00 -4.24373090e-01 -1.02275372e+00 -1.50743172e-01
-2.98619699e-02 -9.40036118e-01 7.63402700e-01 4.59194064e-01
-1.11967874e+00 -2.90547252e-01 -1.30601859e+00 4.56952751e-01
4.20499861e-01 -2.04522356e-01 1.82923079e-01 3.05904031e-01
-5.01630127e-01 6.36542559e-01 -5.37361622e-01 1.70551151e-01
5.95850451e-03 3.62890303e-01 -2.66643047e-01 -1.47083431e-01
-8.42363238e-01 5.00247300e-01 4.30202365e-01 4.38360065e-01
1.85727850e-02 -8.42858553e-01 -2.07377657e-01 -3.02974194e-01
2.16865435e-01 -6.01838410e-01 8.53905559e-01 -5.78526974e-01
-1.73914468e+00 8.09389174e-01 -5.24317026e-01 -6.19810633e-02
2.36579865e-01 1.36190269e-03 -3.39568466e-01 5.60102105e-01
2.11443201e-01 -6.38931841e-02 9.70508635e-01 -1.33090115e+00
-3.15887123e-01 -4.34118927e-01 -2.55792707e-01 3.50652993e-01
1.05625920e-01 2.84529962e-02 -1.13951899e-01 3.09461862e-01
1.01700306e+00 -9.93827760e-01 -1.78899869e-01 1.28777817e-01
-3.17586243e-01 1.41163215e-01 1.10724139e+00 -2.27651864e-01
1.18990064e+00 -1.94641995e+00 -2.17436358e-01 2.38642126e-01
4.49404597e-01 3.36598873e-01 1.79983035e-01 3.32260013e-01
-9.64749381e-02 -3.20518851e-01 -5.46656132e-01 5.07431328e-01
-5.37593901e-01 -1.77332610e-01 1.31180972e-01 7.59008050e-01
1.33401025e-02 9.03686047e-01 -9.64022696e-01 -5.32846332e-01
4.11969796e-02 1.22627392e-01 -5.77601552e-01 2.99039990e-01
3.46730798e-02 8.64423215e-01 -5.39561927e-01 5.69173396e-01
1.48611569e+00 -3.79996032e-01 1.27251685e-01 -8.94981265e-01
-9.45368290e-01 -8.75880048e-02 -1.64588630e+00 1.07773578e+00
-7.81165287e-02 4.88581449e-01 4.84957457e-01 -1.05785823e+00
9.83612597e-01 2.85875797e-01 8.40218425e-01 -9.66186941e-01
1.32979229e-02 4.87910390e-01 4.43973631e-01 -6.43515944e-01
1.71142630e-02 -4.90302801e-01 2.52000540e-01 6.33462548e-01
-2.93785870e-01 -5.04357696e-01 4.28012401e-01 -1.82533398e-01
8.64351094e-01 4.06715050e-02 4.04154837e-01 -5.69069386e-01
8.43086720e-01 6.41289800e-02 8.06876659e-01 5.94110310e-01
-2.52495259e-01 5.30291557e-01 1.21092923e-01 -3.02293509e-01
-8.88278961e-01 -1.06287146e+00 -4.73043919e-01 -3.30720305e-01
1.16136873e+00 -4.79237139e-02 -4.71930504e-01 1.87213972e-01
-3.67033809e-01 4.38001379e-02 1.27388844e-02 -8.41705408e-03
-1.04054284e+00 -1.25891626e+00 -3.37417930e-01 -1.82637125e-01
1.25204217e+00 -1.02772880e+00 -5.38948953e-01 3.44782770e-01
-5.52922070e-01 -9.99862969e-01 9.72737186e-03 -2.67656922e-01
-7.13113546e-01 -1.27819145e+00 -7.06391513e-01 -8.29514146e-01
6.60330176e-01 8.37509274e-01 1.03799665e+00 -7.03844987e-03
-2.93163449e-01 1.93996444e-01 2.72175055e-02 -1.99167281e-01
-1.58234596e-01 -5.79436600e-01 5.15387543e-02 3.50969955e-02
1.17059991e-01 -8.20001125e-01 -1.09648502e+00 4.86905754e-01
-6.38872743e-01 2.02843398e-01 6.13848686e-01 9.17130232e-01
7.62579501e-01 2.62213945e-01 1.03095271e-01 -4.64355052e-01
1.46625161e-01 5.62165454e-02 -1.14167619e+00 1.41945213e-01
-6.50863409e-01 7.13633047e-03 4.27176744e-01 -6.18848026e-01
-1.31305230e+00 -3.98178220e-01 -5.30225001e-02 1.53073147e-02
1.61013588e-01 3.94058019e-01 -2.03883752e-01 -6.57123148e-01
4.54396725e-01 4.55898583e-01 2.85533726e-01 -2.32956987e-02
-6.10276461e-02 5.16431630e-01 3.68399888e-01 -3.65096003e-01
1.42610085e+00 7.02508271e-01 5.63919008e-01 -1.14764535e+00
-7.10146725e-01 -4.90360856e-01 -3.42914134e-01 -3.14674288e-01
5.88846624e-01 -6.48343980e-01 -1.34769475e+00 8.60551953e-01
-9.23486710e-01 -1.04826950e-01 -3.22924376e-01 9.73532915e-01
-3.74618262e-01 6.03129268e-01 -5.98103344e-01 -7.07353055e-01
-4.91109878e-01 -1.06858075e+00 7.88490593e-01 7.27591574e-01
2.82728195e-01 -5.63630998e-01 1.62139580e-01 2.33372375e-01
4.35235411e-01 1.65223740e-02 6.18668437e-01 2.81178355e-01
-1.42962885e+00 4.17421460e-02 -4.68453228e-01 -1.33470997e-01
1.76569507e-01 7.83575624e-02 -6.86334550e-01 -2.53028601e-01
7.54395723e-01 5.88575937e-02 4.19425815e-01 7.09852815e-01
5.46508133e-01 -3.04993927e-01 -5.36875010e-01 8.78581464e-01
2.01885271e+00 6.16463125e-01 8.86549652e-01 2.83924192e-01
6.19041979e-01 1.83224127e-01 1.24095500e+00 1.82634085e-01
-8.04661289e-02 6.82227671e-01 1.61668763e-01 1.84589922e-02
-4.36123041e-03 2.40036249e-01 -2.64146458e-02 1.01467502e+00
-2.06720456e-01 -9.22937095e-02 -6.51873112e-01 1.27438381e-01
-1.40825844e+00 -1.14975297e+00 -6.96897984e-01 2.55233598e+00
1.00785005e+00 6.18692562e-02 -4.82648790e-01 2.66772270e-01
5.60141623e-01 1.85753688e-01 -3.41088235e-01 3.97199243e-01
-1.70951560e-01 3.66169333e-01 4.12977219e-01 8.24701786e-01
-7.33923495e-01 5.47155380e-01 6.29127645e+00 7.30469286e-01
-1.38839209e+00 -1.06082164e-01 1.00442395e-02 2.78603017e-01
-5.84693730e-01 3.97687942e-01 -8.48185778e-01 7.35887825e-01
2.69460976e-01 -3.34729970e-01 3.63364279e-01 2.60195807e-02
3.31535101e-01 -6.31516337e-01 -3.74460220e-01 1.26090252e+00
-3.89074326e-01 -1.11261666e+00 -1.85631812e-01 -6.09434806e-02
5.69474220e-01 -2.73446798e-01 -1.22215487e-01 -2.18074575e-01
-4.96828496e-01 -2.78932750e-01 2.40785077e-01 8.27993214e-01
9.62342203e-01 -3.55647594e-01 6.18392050e-01 4.65454519e-01
-1.24812365e+00 6.89003393e-02 -4.39743459e-01 -2.35608388e-02
3.57134491e-01 1.08016825e+00 -5.25791168e-01 6.87364280e-01
6.81908190e-01 6.86013162e-01 1.87177937e-02 1.26512599e+00
-1.92597002e-01 6.49476826e-01 -6.10449791e-01 -1.19580165e-01
-1.53391749e-01 -1.41278315e+00 1.03026259e+00 6.40181899e-01
5.80365598e-01 7.28345156e-01 9.34833959e-02 1.13471842e+00
3.94925207e-01 8.62017870e-02 -5.41710079e-01 1.72246680e-01
5.16678154e-01 1.40947485e+00 -6.72290683e-01 -1.24684162e-01
-5.05684316e-01 5.93736827e-01 -2.58998126e-01 8.14398527e-01
-5.80093205e-01 -7.42189229e-01 4.15894210e-01 5.09791076e-01
7.11115636e-03 -2.90677369e-01 6.61132857e-03 -1.35688138e+00
3.23079675e-01 -3.77027035e-01 -2.23919049e-01 -1.01153195e+00
-6.92255139e-01 2.46701837e-01 1.27073422e-01 -1.82715893e+00
6.27300469e-03 -5.16841233e-01 -8.21797013e-01 1.00042307e+00
-1.71767700e+00 -7.58280456e-01 -7.00928152e-01 3.47512662e-01
-3.80381495e-01 2.99700767e-01 4.77430850e-01 1.69447601e-01
-2.45069250e-01 -5.06802984e-02 2.97575474e-01 -8.76386166e-02
8.23423147e-01 -9.51772392e-01 -5.52218318e-01 8.72900903e-01
-3.93953741e-01 6.65041268e-01 9.35349822e-01 -4.72185254e-01
-1.69422543e+00 -4.08915907e-01 5.93525887e-01 2.20189601e-01
5.18013716e-01 9.67353135e-02 -1.04720688e+00 2.15873063e-01
4.18232083e-01 2.50379033e-02 2.76222199e-01 -5.41408122e-01
2.31493086e-01 -7.19573975e-01 -1.32112432e+00 5.18369794e-01
9.50084984e-01 -7.73348510e-02 -3.76702905e-01 3.46862912e-01
6.46724045e-01 -6.05949759e-01 -9.30590689e-01 8.74176919e-01
6.34860873e-01 -1.22113180e+00 9.67018604e-01 4.90266919e-01
6.87614605e-02 -9.12517488e-01 2.72163481e-01 -1.04867971e+00
-5.83703637e-01 -9.28294301e-01 1.18023738e-01 9.85070944e-01
-3.38443518e-02 -1.28766239e+00 7.46465743e-01 3.49879488e-02
3.21319140e-02 -5.47639668e-01 -6.69676840e-01 -6.47292852e-01
-5.07086873e-01 2.24197820e-01 2.73169667e-01 5.69648862e-01
-2.34063834e-01 4.73819703e-01 -1.40211418e-01 4.67242986e-01
1.35703993e+00 8.33430886e-01 6.55814767e-01 -1.50968814e+00
-2.93122977e-01 -1.34020135e-01 -3.18112224e-01 -1.41664338e+00
-1.82886511e-01 -4.33540106e-01 1.37907013e-01 -1.37064540e+00
4.61129636e-01 -8.11375499e-01 -1.27746090e-01 -1.55964077e-01
-3.32713693e-01 2.73557127e-01 -1.09336101e-01 7.27394700e-01
-2.53231317e-01 5.60178578e-01 1.92914927e+00 -1.79153249e-01
-4.13190007e-01 4.38797891e-01 -3.95519316e-01 5.95667899e-01
5.59214830e-01 -4.92437303e-01 -4.18658197e-01 -2.46744696e-02
3.09645623e-01 3.11325192e-01 1.78950816e-01 -1.20461679e+00
3.41237307e-01 -3.03606987e-01 5.05329743e-02 -8.85736704e-01
4.85845029e-01 -8.07594776e-01 4.69772577e-01 7.68076360e-01
4.94203180e-01 -6.33808732e-01 -2.93685615e-01 3.62536877e-01
-5.29844344e-01 -5.87147593e-01 1.10634661e+00 -2.31346130e-01
-7.19906449e-01 5.41012585e-01 -2.45237537e-02 1.11100748e-01
9.16566074e-01 -6.30668521e-01 -4.02214438e-01 -2.07203060e-01
-3.51703286e-01 -9.22534149e-03 6.40006602e-01 -5.51809788e-01
6.32722020e-01 -1.36505890e+00 -6.22171044e-01 4.94316965e-01
-4.26074909e-03 9.90985259e-02 2.83578515e-01 1.25459862e+00
-1.00584507e+00 2.31266722e-01 -6.21283129e-02 -1.14584970e+00
-1.26602292e+00 1.90530896e-01 5.94933093e-01 -3.71198773e-01
-6.09145999e-01 5.25139272e-01 2.87464261e-01 -1.67021811e-01
-5.44921398e-01 -2.46702269e-01 -8.13656747e-02 -1.39270976e-01
7.42418885e-01 2.22963393e-01 -2.54043967e-01 -4.52633172e-01
-2.36547321e-01 1.34392393e+00 3.45288873e-01 -2.52239317e-01
1.13734257e+00 -2.27833077e-01 -5.60847700e-01 3.78602922e-01
1.09160233e+00 3.77506971e-01 -1.29368997e+00 -5.04307747e-01
-4.68207896e-01 -7.24972427e-01 -2.88337499e-01 -1.52280182e-01
-8.98094893e-01 7.70970464e-01 6.73583090e-01 2.11788625e-01
1.28181612e+00 -2.29912311e-01 6.94772482e-01 5.22553384e-01
6.75254881e-01 -6.58562899e-01 1.29993618e-01 2.89695442e-01
4.42143977e-01 -1.20123208e+00 3.76986623e-01 -8.10654581e-01
-2.81392056e-02 1.01947999e+00 4.99296039e-01 -1.29273787e-01
8.32308710e-01 3.40935171e-01 -2.20202617e-02 -1.39698789e-01
-2.66387671e-01 2.55577061e-02 1.38919443e-01 6.21617019e-01
3.46408784e-01 1.69643387e-02 -7.37578154e-01 -1.03194647e-01
7.54122287e-02 1.76082656e-01 5.69092333e-01 8.91683638e-01
-7.93325782e-01 -1.01741815e+00 -5.46701193e-01 2.41469994e-01
-2.26891860e-01 9.45509672e-02 4.55106407e-01 9.28653181e-01
-1.13581665e-01 7.71872342e-01 2.49564186e-01 1.75748676e-01
4.04929549e-01 -2.70195425e-01 9.15674806e-01 -3.10764164e-01
1.97601348e-01 2.47721270e-01 -3.15213889e-01 -7.28871286e-01
-1.01789820e+00 -5.33271074e-01 -1.48201931e+00 3.80376470e-03
-4.62314367e-01 3.38033646e-01 2.66681343e-01 9.33325827e-01
-1.45137891e-01 2.94650178e-02 8.49079013e-01 -5.62470019e-01
-2.51599580e-01 -7.80939341e-01 -9.02474046e-01 3.24347883e-01
1.83005318e-01 -4.61224854e-01 -5.54025650e-01 -3.56811970e-01] | [11.064508438110352, -2.595252513885498] |
699326b0-ad16-4c51-8714-5ca423b43c90 | clinicalgpt-large-language-models-finetuned | 2306.09968 | null | https://arxiv.org/abs/2306.09968v1 | https://arxiv.org/pdf/2306.09968v1.pdf | ClinicalGPT: Large Language Models Finetuned with Diverse Medical Data and Comprehensive Evaluation | Large language models have exhibited exceptional performance on various Natural Language Processing (NLP) tasks, leveraging techniques such as the pre-training, and instruction fine-tuning. Despite these advances, their effectiveness in medical applications is limited, due to challenges such as factual inaccuracies, reasoning abilities, and lack grounding in real-world experience. In this study, we present ClinicalGPT, a language model explicitly designed and optimized for clinical scenarios. By incorporating extensive and diverse real-world data, such as medical records, domain-specific knowledge, and multi-round dialogue consultations in the training process, ClinicalGPT is better prepared to handle multiple clinical task. Furthermore, we introduce a comprehensive evaluation framework that includes medical knowledge question-answering, medical exams, patient consultations, and diagnostic analysis of medical records. Our results demonstrate that ClinicalGPT significantly outperforms other models in these tasks, highlighting the effectiveness of our approach in adapting large language models to the critical domain of healthcare. | ['Xiaohu Li', 'Longjun Fan', 'Zongxin Du', 'Guoxing Yang', 'Guangyu Wang'] | 2023-06-16 | null | null | null | null | ['question-answering'] | ['natural-language-processing'] | [ 5.33592068e-02 6.09786630e-01 -3.87625337e-01 -5.18168747e-01
-1.08633697e+00 -3.81801456e-01 1.56202450e-01 7.85676301e-01
-5.30009627e-01 8.31699729e-01 5.89706481e-01 -6.84100628e-01
-3.64205569e-01 -6.17974818e-01 -2.51501709e-01 -1.29193529e-01
-2.88723456e-03 9.88956630e-01 -2.35616770e-02 -3.68620604e-01
-7.47380182e-02 2.37965405e-01 -6.09126925e-01 8.55740726e-01
1.17659855e+00 5.59158921e-01 6.15115613e-02 7.68125594e-01
-4.33976054e-01 1.25380373e+00 -5.64087152e-01 -5.90479612e-01
-3.42354923e-02 -2.98280954e-01 -1.10240769e+00 -2.01833665e-01
-3.86361927e-02 -3.26569915e-01 -2.32657809e-02 6.01409376e-01
8.49214792e-01 -5.11343777e-03 2.57698625e-01 -7.36287773e-01
-7.71529436e-01 8.15546691e-01 -1.39929488e-01 1.74267456e-01
8.04706633e-01 2.97791749e-01 7.61201560e-01 -4.33946818e-01
7.64876485e-01 1.27607286e+00 8.08006048e-01 7.73252249e-01
-6.68522716e-01 -3.37681830e-01 2.66676471e-02 -1.23236310e-02
-1.08509564e+00 -2.49495611e-01 2.05263853e-01 -4.79318649e-01
1.34281409e+00 2.06999466e-01 3.86571378e-01 1.06115735e+00
5.36066353e-01 9.12583113e-01 8.36237669e-01 -5.67321181e-01
1.17852837e-01 4.66999799e-01 2.96771765e-01 8.11687827e-01
4.17578146e-02 -1.89794868e-01 -3.25157791e-01 -5.80463767e-01
6.11254156e-01 1.05514005e-01 -2.99820870e-01 2.13386610e-01
-1.36361170e+00 6.15861356e-01 1.22859374e-01 4.60026175e-01
-6.48736835e-01 -5.23795307e-01 6.63189411e-01 3.07249635e-01
2.66348898e-01 8.51596117e-01 -9.49817359e-01 -2.62197852e-01
-7.04471111e-01 1.61569983e-01 1.24867582e+00 8.91792357e-01
1.69461101e-01 -4.65641916e-01 -6.79395020e-01 8.42217863e-01
1.83210045e-01 3.74661982e-01 8.88634443e-01 -8.25047791e-01
6.99691236e-01 9.58745241e-01 -2.04016659e-02 -7.95948625e-01
-6.40614390e-01 -3.53556693e-01 -9.41842258e-01 -7.42714226e-01
2.03124672e-01 -4.34785873e-01 -8.57717872e-01 1.55463576e+00
3.88419718e-01 1.09573357e-01 7.55957723e-01 5.54174304e-01
1.23742127e+00 4.69514698e-01 7.33265281e-01 -3.56090516e-01
1.73322237e+00 -1.09445834e+00 -9.42238569e-01 -4.20026094e-01
1.11309087e+00 -8.07483733e-01 1.06755412e+00 3.00689369e-01
-1.22031569e+00 -2.85423905e-01 -2.78949708e-01 -2.18471259e-01
-3.60187113e-01 3.45499255e-02 7.13578045e-01 3.48951787e-01
-9.35868561e-01 1.36132345e-01 -8.15165460e-01 -5.16608357e-01
2.52711177e-01 1.44532740e-01 -2.97390938e-01 -3.37536752e-01
-1.46281540e+00 9.66703117e-01 6.00301564e-01 2.67334152e-02
-4.77169663e-01 -1.07977736e+00 -9.40577507e-01 4.90852259e-02
4.97671723e-01 -1.18982542e+00 1.59546864e+00 -3.32208782e-01
-1.43934190e+00 1.03466105e+00 -4.35785827e-04 -4.39802378e-01
5.58342755e-01 -2.70662814e-01 -6.12170696e-01 3.12178552e-01
-4.14387062e-02 2.75248408e-01 1.22628873e-02 -7.52406955e-01
-4.51548725e-01 -1.31212309e-01 2.74010420e-01 2.25056887e-01
-1.72168493e-01 1.03048824e-01 -5.65476418e-01 -4.26318616e-01
-2.93778628e-01 -6.32819593e-01 -8.59980166e-01 -3.36973369e-01
-5.00334978e-01 -1.60816103e-01 3.60888056e-02 -7.74263024e-01
1.58325970e+00 -2.08402753e+00 -3.06829154e-01 1.37037545e-01
3.32200050e-01 6.38889611e-01 -8.20805728e-02 7.41137803e-01
8.23805481e-02 2.81830192e-01 -7.70336986e-02 2.65452191e-02
-1.51195750e-01 5.33972561e-01 -1.02909833e-01 -3.13960731e-01
3.42281371e-01 1.24528909e+00 -1.20077860e+00 -1.08207369e+00
4.65077087e-02 2.39558175e-01 -7.80584693e-01 4.76464629e-01
-4.64196742e-01 5.35990059e-01 -1.03447521e+00 7.27877319e-01
2.25484669e-01 -8.86265159e-01 3.96478236e-01 1.03921488e-01
2.89380461e-01 4.02513444e-01 -7.28868365e-01 1.76521993e+00
-5.46933115e-01 -1.29568055e-01 1.05323605e-01 -6.63351178e-01
4.11833555e-01 8.12958896e-01 5.68316936e-01 -6.08553648e-01
3.22097726e-03 7.77884200e-02 6.65069818e-02 -1.46706200e+00
2.68447280e-01 -2.65948296e-01 -6.34821504e-02 5.21076918e-01
-1.96357265e-01 -1.34929404e-01 2.83759385e-01 4.09916461e-01
1.21830046e+00 -3.56331110e-01 8.18994164e-01 -5.42168245e-02
7.92430997e-01 4.00211096e-01 5.44620693e-01 8.62214148e-01
-2.70909667e-01 2.66975433e-01 4.58128422e-01 -4.06741619e-01
-3.66357505e-01 -7.11236954e-01 -2.43029505e-01 1.02107835e+00
-2.99518496e-01 -5.86807728e-01 -5.79144239e-01 -7.40386009e-01
-1.74762569e-02 7.30365634e-01 -4.09983039e-01 -1.02764726e-01
-6.74698055e-01 -9.36119795e-01 5.92475355e-01 5.95889390e-01
3.27233195e-01 -1.37426078e+00 -6.49072051e-01 6.56414628e-01
-4.89171267e-01 -1.60343516e+00 -3.63578707e-01 -2.26682082e-01
-9.53740418e-01 -1.34620249e+00 -4.54342067e-01 -9.42091107e-01
6.18309498e-01 -2.56272167e-01 1.46897042e+00 9.35643017e-02
-5.71457684e-01 6.90245807e-01 -3.22920263e-01 -6.07542098e-01
-9.78469610e-01 1.88098907e-01 -3.80867153e-01 -5.76942742e-01
6.38583541e-01 -1.25369743e-01 -6.43802047e-01 6.28132746e-02
-1.11028147e+00 1.74985215e-01 1.02423608e+00 9.19258058e-01
5.62298179e-01 -2.65970767e-01 7.73265183e-01 -1.56412363e+00
1.30637968e+00 -5.57676196e-01 5.82085252e-02 8.55689466e-01
-6.32071018e-01 -9.62366983e-02 5.05462885e-01 -4.03254002e-01
-1.22418654e+00 -3.92689288e-01 -5.31288743e-01 1.42860785e-01
-4.60838884e-01 1.13049614e+00 2.25439191e-01 2.26133019e-01
9.14079010e-01 7.09876642e-02 5.39731868e-02 -3.66354674e-01
2.00150728e-01 6.97717965e-01 4.46536362e-01 -7.31914818e-01
3.22932690e-01 -4.31389511e-02 -4.59697843e-01 -4.62027878e-01
-1.20700479e+00 -6.23862863e-01 -4.34616476e-01 2.92442560e-01
8.73600662e-01 -9.00314152e-01 -8.52830231e-01 3.02908979e-02
-1.06726921e+00 -2.92993575e-01 -4.71240729e-01 6.95819020e-01
-1.15859061e-01 5.14841616e-01 -1.11557317e+00 -6.33794665e-01
-8.35779846e-01 -1.11267912e+00 1.08020914e+00 1.22293957e-01
-4.19931084e-01 -1.43158209e+00 1.33536056e-01 7.87610471e-01
3.91302586e-01 3.54603082e-01 1.47513759e+00 -1.08926380e+00
-1.94211438e-01 -3.98373455e-01 -1.00662030e-01 1.89492121e-01
2.76077002e-01 -3.15194398e-01 -5.27795255e-01 -1.21093772e-01
1.34895533e-01 -6.30193949e-01 2.23507240e-01 1.76469356e-01
1.26514530e+00 -4.03860599e-01 -4.39307660e-01 3.03401977e-01
1.13934946e+00 2.52525419e-01 1.99722350e-01 1.18139014e-01
3.82000983e-01 7.02371061e-01 6.56291246e-01 4.34293270e-01
8.65993798e-01 4.96943407e-02 -1.72623381e-01 -3.56572628e-01
1.88251093e-01 -1.12414852e-01 -9.97152627e-02 1.30613351e+00
6.06729910e-02 -1.66010812e-01 -1.53057325e+00 5.75012982e-01
-1.93744147e+00 -5.57866812e-01 1.63029045e-01 1.63305891e+00
1.51609969e+00 1.96309835e-02 -4.48153257e-01 -4.98599201e-01
1.26837656e-01 -3.86689305e-01 -5.78282535e-01 -4.18709844e-01
7.26927444e-02 2.01103151e-01 1.31026739e-02 4.98511165e-01
-6.98975623e-01 8.59713018e-01 7.38128233e+00 5.20672441e-01
-8.84126306e-01 1.40829682e-01 6.77488327e-01 1.74909130e-01
-2.56687492e-01 -5.14940977e-01 -5.99623263e-01 1.45189032e-01
9.65055048e-01 -3.21331471e-01 -1.26698136e-01 5.83410144e-01
3.23168427e-01 2.50021424e-02 -1.30855620e+00 7.22604930e-01
6.27914071e-02 -1.37757826e+00 2.61512369e-01 -1.37742803e-01
5.79298913e-01 -7.54011869e-02 -2.13284805e-01 7.58790135e-01
5.89449465e-01 -1.15599418e+00 -1.54929996e-01 4.35599983e-01
6.25204146e-01 -4.47298549e-02 1.23007667e+00 7.83642769e-01
-7.21575320e-01 -1.48808882e-02 -2.62411386e-02 1.53376237e-01
3.45575273e-01 5.61900973e-01 -1.52249408e+00 9.35256422e-01
4.09217983e-01 4.44546014e-01 -4.62037891e-01 7.56088853e-01
-1.68312266e-01 5.66140294e-01 -1.37734964e-01 1.72824845e-01
2.97362924e-01 -3.93505543e-02 -5.46335690e-02 1.61429143e+00
-5.18083535e-02 7.07485557e-01 6.57999992e-01 4.02522683e-01
-4.17557806e-02 6.08604729e-01 -4.49357837e-01 -2.73972094e-01
2.11267725e-01 1.18598342e+00 -2.42785901e-01 -7.50600934e-01
-5.36902428e-01 3.54696780e-01 3.44942063e-01 4.42000061e-01
-6.27885044e-01 -2.11475119e-02 3.71837795e-01 5.22423089e-02
-3.69142562e-01 1.86956987e-01 -1.95394263e-01 -1.17653120e+00
-3.25471312e-02 -1.31374753e+00 9.49419200e-01 -6.09609604e-01
-1.53298509e+00 9.22802448e-01 9.81421694e-02 -8.68625224e-01
-5.63868225e-01 -6.95683718e-01 -2.64505774e-01 7.67186165e-01
-1.90384245e+00 -1.13427138e+00 -2.12739766e-01 7.96059251e-01
4.74959582e-01 -2.29760125e-01 1.24907196e+00 5.96602440e-01
-5.46012580e-01 6.67331278e-01 -3.07621926e-01 2.42944434e-01
1.02421045e+00 -1.11862862e+00 6.43888339e-02 3.09382617e-01
-5.92419565e-01 1.02291644e+00 3.79013330e-01 -6.60543561e-01
-1.26297557e+00 -1.10976553e+00 1.38423884e+00 -6.75655007e-01
6.86197519e-01 7.80323744e-02 -1.19766164e+00 8.19478095e-01
1.39329910e-01 -2.59493232e-01 1.45197415e+00 1.56789556e-01
-8.55175313e-03 1.77225754e-01 -1.40928698e+00 4.79339451e-01
8.07656467e-01 -5.34741879e-01 -1.05892205e+00 8.88265729e-01
8.61856163e-01 -9.04835999e-01 -1.38826561e+00 5.83873570e-01
2.51804590e-01 -3.46937686e-01 1.08130968e+00 -1.32614326e+00
6.68990374e-01 2.27891549e-01 2.79148370e-01 -1.01256943e+00
-1.48815498e-01 -6.67862117e-01 -2.48006806e-02 8.19936872e-01
6.49972379e-01 -8.89034688e-01 4.92743671e-01 1.05830550e+00
-1.01354159e-01 -1.04631102e+00 -4.98132944e-01 -1.27526954e-01
6.53191730e-02 -1.94146991e-01 5.58794856e-01 1.38033998e+00
4.47593689e-01 3.84308159e-01 -5.36747612e-02 3.56680751e-01
2.15227753e-02 2.39097923e-01 4.66010571e-01 -1.06459248e+00
-6.15527809e-01 -1.93511546e-01 1.21524230e-01 -9.21305001e-01
6.84859678e-02 -8.03708911e-01 -2.32837312e-02 -2.01738358e+00
2.52621204e-01 -4.57458109e-01 -3.42079699e-01 8.99591148e-01
-6.46435142e-01 -3.86651933e-01 -1.29486561e-01 1.13211527e-01
-7.28653431e-01 3.76417115e-02 1.59306765e+00 -1.39953569e-01
-3.57991070e-01 8.96505490e-02 -9.81625199e-01 8.36125314e-01
7.20752001e-01 -3.46762031e-01 -4.04090971e-01 -7.13224113e-01
3.17485631e-01 6.51083529e-01 -1.69087380e-01 -4.93950129e-01
5.72090089e-01 -4.35183853e-01 6.54547941e-03 -2.19140828e-01
8.45003799e-02 -7.57766247e-01 -1.90785423e-01 6.83146298e-01
-6.83560431e-01 2.76643515e-01 4.89113867e-01 3.00622195e-01
-5.46788454e-01 -1.63078919e-01 3.97076398e-01 -5.81546843e-01
-5.28429568e-01 2.55342901e-01 -4.49677050e-01 5.53877234e-01
9.86451864e-01 -2.96980678e-03 -4.11327004e-01 -1.13079503e-01
-8.88957143e-01 1.01330161e+00 -1.52552128e-01 3.12096775e-01
5.68075001e-01 -6.99060977e-01 -9.48159099e-01 1.14993371e-01
3.92344087e-01 4.20131177e-01 6.31838620e-01 9.71356332e-01
-8.67140472e-01 7.43707001e-01 6.45462573e-02 -4.85364407e-01
-1.28589571e+00 6.70127034e-01 2.81379402e-01 -1.12552166e+00
-7.34154344e-01 6.02500319e-01 2.50110239e-01 -8.44166160e-01
4.35163200e-01 -8.48021746e-01 -2.79854923e-01 -2.37018839e-01
6.08259976e-01 -2.58292764e-01 1.59438446e-01 5.88750839e-03
-4.05563235e-01 3.42089146e-01 -4.77099597e-01 2.74101228e-01
1.17543185e+00 -1.01125659e-02 -2.71961391e-01 2.44120121e-01
5.95366955e-01 5.86980535e-03 -2.79381394e-01 -7.33629107e-01
2.31585845e-01 4.15698849e-02 -3.00758511e-01 -1.48771620e+00
-5.56396127e-01 7.19141722e-01 1.68573663e-01 4.94821602e-03
1.06698167e+00 3.76971811e-02 9.43417966e-01 9.54449415e-01
2.36397833e-01 -9.86986697e-01 -4.10137512e-02 5.78838646e-01
6.22995555e-01 -1.43988228e+00 -3.97214293e-02 -6.21695399e-01
-9.36685860e-01 9.30035532e-01 7.68597484e-01 6.28603101e-01
7.51893342e-01 3.86519313e-01 8.24535668e-01 -3.89250547e-01
-1.12498271e+00 2.18429893e-01 1.46900520e-01 2.41669953e-01
8.08238506e-01 1.06082477e-01 -2.84703195e-01 7.15221524e-01
-2.58585602e-01 4.57920820e-01 2.62203842e-01 1.18889177e+00
-2.15297751e-03 -1.18428707e+00 -1.83657110e-01 4.72258449e-01
-8.86812449e-01 -3.89881372e-01 -2.67757833e-01 7.89964020e-01
-1.66660659e-02 8.48075628e-01 -4.31541622e-01 1.99789494e-01
7.15176642e-01 3.24504673e-01 1.49140611e-01 -1.12747836e+00
-1.30052090e+00 -1.42271444e-01 2.72905856e-01 -4.40632552e-01
-3.79805058e-01 -6.45989403e-02 -1.44286060e+00 -3.64487544e-02
4.96283434e-02 6.04728639e-01 2.88812786e-01 1.00361335e+00
7.46415496e-01 9.19985712e-01 -1.51633307e-01 2.52627403e-01
-8.13211143e-01 -9.21155274e-01 -8.90484527e-02 5.44834197e-01
3.43548745e-01 2.49029994e-02 1.81665629e-01 1.48926660e-01] | [8.725122451782227, 8.598854064941406] |
660bc6a4-1298-4a91-8756-7394c2d3fdde | zero-shot-motor-health-monitoring-by-blind | 2212.06154 | null | https://arxiv.org/abs/2212.06154v1 | https://arxiv.org/pdf/2212.06154v1.pdf | Zero-Shot Motor Health Monitoring by Blind Domain Transition | Continuous long-term monitoring of motor health is crucial for the early detection of abnormalities such as bearing faults (up to 51% of motor failures are attributed to bearing faults). Despite numerous methodologies proposed for bearing fault detection, most of them require normal (healthy) and abnormal (faulty) data for training. Even with the recent deep learning (DL) methodologies trained on the labeled data from the same machine, the classification accuracy significantly deteriorates when one or few conditions are altered. Furthermore, their performance suffers significantly or may entirely fail when they are tested on another machine with entirely different healthy and faulty signal patterns. To address this need, in this pilot study, we propose a zero-shot bearing fault detection method that can detect any fault on a new (target) machine regardless of the working conditions, sensor parameters, or fault characteristics. To accomplish this objective, a 1D Operational Generative Adversarial Network (Op-GAN) first characterizes the transition between normal and fault vibration signals of (a) source machine(s) under various conditions, sensor parameters, and fault types. Then for a target machine, the potential faulty signals can be generated, and over its actual healthy and synthesized faulty signals, a compact, and lightweight 1D Self-ONN fault detector can then be trained to detect the real faulty condition in real time whenever it occurs. To validate the proposed approach, a new benchmark dataset is created using two different motors working under different conditions and sensor locations. Experimental results demonstrate that this novel approach can accurately detect any bearing fault achieving an average recall rate of around 89% and 95% on two target machines regardless of its type, severity, and location. | ['Moncef Gabbouj', 'Onur Avci', 'Osama Abdeljaber', 'Turker Ince', 'Sadok Sassi', 'Amir Alhams', 'Ozer Can Devecioglu', 'Serkan Kiranyaz'] | 2022-12-12 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 6.63472831e-01 -3.27729317e-03 -7.89449140e-02 2.47929484e-01
-7.63845503e-01 -2.47582808e-01 3.09577525e-01 -9.51081067e-02
2.79415458e-01 5.00633061e-01 -4.51528549e-01 -1.43606588e-01
-6.91989576e-03 -6.75251186e-01 -9.29755867e-01 -9.23973083e-01
-1.67840555e-01 3.45685154e-01 3.28627497e-01 -8.73946100e-02
3.89004201e-02 4.85592097e-01 -2.04549408e+00 1.80100933e-01
7.81202912e-01 1.07812905e+00 1.50719255e-01 7.49314189e-01
6.15780592e-01 5.61487079e-01 -1.34229028e+00 3.54973286e-01
2.11096898e-01 -4.56102878e-01 -6.27293229e-01 2.74349123e-01
-6.45439606e-03 -3.85646135e-01 -5.23370206e-01 9.71629262e-01
5.24688363e-01 -2.28924863e-02 6.65315151e-01 -1.59224391e+00
-4.97126997e-01 1.27525896e-01 -1.93838954e-01 7.15807825e-02
6.05706811e-01 5.16639411e-01 4.44544166e-01 -8.33636284e-01
1.50419489e-01 8.02545965e-01 6.23198807e-01 6.22221053e-01
-8.67850244e-01 -5.24308443e-01 -1.55729622e-01 2.56392270e-01
-9.62034643e-01 -2.29067326e-01 1.09452224e+00 -3.79368067e-01
8.08130562e-01 1.42760217e-01 4.06785667e-01 1.44568956e+00
9.18320656e-01 5.89273453e-01 8.78186584e-01 -3.64889294e-01
3.76426786e-01 -3.69264752e-01 -1.23135850e-01 3.61716866e-01
2.62636960e-01 3.83756876e-01 -3.93652558e-01 1.11035183e-01
4.43283767e-01 1.85841918e-01 -4.70072895e-01 -3.74482460e-02
-1.06650960e+00 4.02723253e-01 1.20720439e-01 5.39671004e-01
-5.16494870e-01 1.40380813e-02 5.02300501e-01 7.87287414e-01
3.82356852e-01 4.82608289e-01 -4.60318685e-01 -2.60976195e-01
-4.73430097e-01 1.60307512e-01 6.09217703e-01 5.01996279e-01
3.79699439e-01 6.93827569e-01 -7.08994791e-02 6.16962075e-01
4.44426723e-02 7.11658955e-01 9.50637221e-01 -2.52489746e-01
2.24113941e-01 5.36585629e-01 1.85995638e-01 -1.02965724e+00
-3.55771124e-01 -5.52198231e-01 -7.43962944e-01 5.01654565e-01
5.33984341e-02 -2.13598579e-01 -1.12613487e+00 1.42471611e+00
3.17405313e-01 5.21280766e-01 2.35331491e-01 8.33038807e-01
3.31088185e-01 4.32419270e-01 -5.04318357e-01 -1.60671413e-01
9.23238695e-01 -4.96158361e-01 -8.88422906e-01 -5.96619070e-01
7.07820296e-01 -5.39942205e-01 1.00672543e+00 5.51258087e-01
-7.05557466e-01 -8.16364348e-01 -1.59641075e+00 7.71631360e-01
-2.96121418e-01 1.74973711e-01 4.50169481e-03 5.68183780e-01
-6.37551546e-01 6.91356778e-01 -9.33309197e-01 -7.92813823e-02
-3.14105824e-02 2.09019840e-01 -3.36955935e-01 -4.12811100e-01
-1.56302714e+00 1.03690708e+00 6.54876679e-02 4.58096117e-01
-1.44779885e+00 -4.59004641e-01 -7.70703733e-01 -3.89286906e-01
2.57421672e-01 -1.30276769e-01 1.13838255e+00 -9.08884227e-01
-1.42646730e+00 3.77238810e-01 1.93735465e-01 -5.03125787e-01
2.97091514e-01 -4.83834893e-01 -1.04137409e+00 1.31072521e-01
1.42484218e-01 -3.12910050e-01 1.37303567e+00 -1.29493511e+00
-5.36517978e-01 -3.47213000e-01 -9.34764668e-02 -1.30269527e-01
-1.44666478e-01 -2.89606005e-01 3.21228117e-01 -6.98899508e-01
8.17702338e-02 -8.58383954e-01 3.38503867e-01 -3.42033178e-01
-6.78637326e-01 -7.63112679e-02 1.44019055e+00 -7.04918087e-01
9.27619338e-01 -2.39422250e+00 -4.97290529e-02 1.72303692e-01
-2.36795843e-02 2.96550572e-01 -1.14517417e-02 4.70058054e-01
-3.17440331e-01 -4.36190546e-01 -4.11071330e-01 1.13682561e-01
-1.08312219e-01 3.07552159e-01 -1.10726528e-01 7.42469549e-01
7.62986600e-01 6.74517751e-01 -9.23789978e-01 3.36369038e-01
4.47896779e-01 2.87584335e-01 4.36752476e-02 5.36414862e-01
-4.00272459e-02 4.85508084e-01 -3.04680288e-01 8.40017676e-01
1.18325412e-01 2.00478315e-01 -2.09333226e-01 -2.10217908e-01
4.40350473e-01 -9.72684249e-02 -1.20402527e+00 1.16689682e+00
-4.97014493e-01 6.00929260e-01 -1.26100421e-01 -1.52422750e+00
1.18664169e+00 7.97210217e-01 5.34956932e-01 -6.85439646e-01
1.90625593e-01 4.74523455e-01 7.96906054e-02 -8.32981408e-01
-2.15835385e-02 -9.36766118e-02 -4.07593787e-01 1.94901332e-01
1.40581608e-01 -2.02411786e-02 -1.89834982e-01 -4.95917976e-01
1.77793920e+00 -9.41891596e-02 -3.59540522e-01 1.75630748e-01
4.07776386e-01 -1.47060826e-01 5.23689091e-01 3.57674956e-01
-1.81443647e-01 4.80318815e-01 3.12839299e-01 -1.16640404e-01
-6.25019252e-01 -1.02132785e+00 -1.72095206e-02 4.13058639e-01
2.38266721e-01 4.01773870e-01 -6.85523152e-01 -6.89197779e-01
2.40611911e-01 7.59252727e-01 -6.25828326e-01 -1.21793747e+00
-4.12521183e-01 -5.99967718e-01 7.37464488e-01 5.92395842e-01
3.27723533e-01 -1.11741078e+00 -8.82059634e-01 3.76469851e-01
1.58969924e-01 -9.00038362e-01 1.09195814e-01 5.69789469e-01
-4.96245682e-01 -1.53777361e+00 -5.81290424e-01 -9.24794257e-01
7.93766618e-01 -6.68516904e-02 5.97638488e-01 1.23822883e-01
-4.07119989e-01 3.48624289e-01 -5.85472882e-01 -3.51540059e-01
-9.17779684e-01 -3.49756688e-01 4.31882232e-01 1.94928423e-01
-1.46056470e-02 -3.09545159e-01 -3.24842930e-01 4.30515677e-01
-9.64986622e-01 -5.74485540e-01 7.74981916e-01 1.12780988e+00
3.14775407e-01 9.54524219e-01 1.22537565e+00 -6.30309939e-01
6.56130552e-01 -7.74676561e-01 5.00734849e-03 -3.96785103e-02
-8.22669685e-01 -1.97850108e-01 8.38954151e-01 -6.37722850e-01
-7.46461689e-01 -2.26839706e-01 -2.09897801e-01 -7.84285545e-01
-4.38081384e-01 6.36765838e-01 -4.64853197e-01 2.17604786e-01
6.90639198e-01 1.19262226e-01 3.69732797e-01 -4.75393444e-01
-5.19811213e-02 7.71303117e-01 9.99343276e-01 -3.01727444e-01
9.91173446e-01 -4.73572947e-02 1.31046074e-02 -4.96723711e-01
-4.56751674e-01 -1.38704330e-01 -4.21455860e-01 -6.20437026e-01
5.39946973e-01 -6.98021948e-01 -2.76897579e-01 1.31491041e+00
-7.90898383e-01 -3.62654954e-01 -3.86507690e-01 4.48694557e-01
-3.27880025e-01 8.72165486e-02 -5.86627781e-01 -7.65053570e-01
-1.88962936e-01 -1.15986979e+00 1.03782511e+00 -7.11159334e-02
-1.70652166e-01 -8.78027558e-01 -2.32192025e-01 7.74223953e-02
3.20521176e-01 1.00428843e+00 9.83360291e-01 -6.86533272e-01
1.34552673e-01 -1.13675594e+00 5.36124825e-01 9.76798952e-01
7.66720593e-01 -7.50652179e-02 -8.07928801e-01 -5.82500279e-01
3.83270115e-01 -3.11198086e-01 3.52760524e-01 3.18303984e-03
6.36168361e-01 -9.64953080e-02 -3.21315706e-01 -1.16045609e-01
1.26376629e+00 5.01840651e-01 4.28872138e-01 2.03137100e-01
7.56482542e-01 3.76888871e-01 1.03184485e+00 1.68341443e-01
-2.44593024e-01 4.08797383e-01 8.11318636e-01 -1.14589885e-01
-9.18200761e-02 -2.86414195e-02 9.18364465e-01 7.37016380e-01
3.38501602e-01 -4.48335856e-01 -7.44664490e-01 6.22113585e-01
-1.34578502e+00 -7.35304654e-01 -7.24368915e-02 2.19837379e+00
6.27637088e-01 6.99602842e-01 -1.57501906e-01 1.21269572e+00
9.38565850e-01 -8.69407281e-02 -9.60712612e-01 -2.84982890e-01
-1.51078682e-02 3.46063644e-01 1.62646845e-01 -3.77763622e-02
-8.64595234e-01 1.74526393e-01 5.59365606e+00 6.14626527e-01
-1.53286326e+00 -5.18063381e-02 3.39032024e-01 3.66103202e-02
5.24549298e-02 -3.94818097e-01 -2.21815616e-01 7.03413904e-01
1.20525348e+00 7.82217309e-02 1.47153005e-01 7.51235247e-01
9.95547399e-02 -1.59803301e-01 -1.23655534e+00 4.55016077e-01
1.79308727e-01 -6.54705405e-01 -4.40120101e-01 -3.13331693e-01
5.66701710e-01 -3.32105100e-01 7.40734264e-02 3.12498391e-01
-3.09983969e-01 -9.86425996e-01 7.61382163e-01 7.19235063e-01
1.03118861e+00 -9.78527069e-01 1.19242215e+00 4.69915390e-01
-9.73171830e-01 -3.74403805e-01 1.82746798e-01 -2.22443938e-01
-1.91705860e-02 8.16799700e-01 -7.87609935e-01 8.09399128e-01
5.93869507e-01 6.43216729e-01 -4.09746379e-01 4.56398547e-01
-2.41880015e-01 9.29332554e-01 -9.92679968e-02 2.66142815e-01
-2.68270195e-01 5.01745760e-01 6.52751923e-01 4.01681811e-01
5.50851464e-01 -5.03501415e-01 8.98232535e-02 5.49584329e-01
3.28772068e-01 -6.40614986e-01 -8.60155821e-01 3.73205799e-03
6.46866679e-01 7.55396187e-01 -4.85868126e-01 4.70906384e-02
-2.88821965e-01 1.10358405e+00 -2.86039919e-01 4.23073471e-01
-8.97783518e-01 -5.09452224e-01 6.19474769e-01 1.98498741e-01
1.00792721e-01 7.23082498e-02 -4.93263453e-02 -5.42977750e-01
4.26563293e-01 -9.49379861e-01 -4.56280448e-03 -8.00671458e-01
-1.32026756e+00 3.42804939e-01 -6.98248148e-02 -1.64049232e+00
-6.12708509e-01 -5.57464659e-01 -8.56942415e-01 7.17560530e-01
-1.03221858e+00 -8.23321164e-01 -1.62490189e-01 4.79691595e-01
5.76267600e-01 -2.98368424e-01 8.34857941e-01 2.87362695e-01
-7.84598887e-01 6.24002755e-01 9.44462121e-02 6.52341545e-02
5.06710410e-01 -1.01563549e+00 2.72029400e-01 1.09983540e+00
-2.73535520e-01 -2.17002966e-02 9.27350760e-01 -7.34428585e-01
-1.75931585e+00 -1.48783290e+00 2.78458744e-01 -2.18495414e-01
6.09621108e-01 -6.44647330e-02 -1.11956644e+00 4.01382625e-01
-1.93369180e-01 3.65226924e-01 1.18679479e-01 -5.61584890e-01
2.48665527e-01 -7.63458610e-02 -1.42601371e+00 1.96225330e-01
7.14208007e-01 -5.80807865e-01 -6.26705945e-01 3.40823293e-01
7.05532253e-01 -6.08303905e-01 -1.12883675e+00 9.05854821e-01
2.97101051e-01 -7.80107558e-01 5.98056376e-01 -1.14299878e-01
3.93826544e-01 -4.54528153e-01 -9.56927091e-02 -1.67432761e+00
2.35034246e-03 -3.41462553e-01 -6.06594205e-01 1.13928354e+00
1.10142067e-01 -9.68254387e-01 4.45117861e-01 3.39968875e-02
-7.64988661e-01 -1.33095479e+00 -1.10934353e+00 -9.76175308e-01
-1.88153401e-01 -7.00491488e-01 7.31826305e-01 7.56151021e-01
-1.28697157e-01 -2.27562953e-02 -1.37585178e-01 6.56174242e-01
1.26220480e-01 -1.54806182e-01 5.83813071e-01 -1.14008009e+00
-2.13000000e-01 6.62723780e-02 -9.11368608e-01 -2.94901401e-01
2.32212514e-01 -5.68929076e-01 5.56125522e-01 -1.22083259e+00
-6.52921319e-01 -2.57914245e-01 -5.20741940e-01 5.88206053e-01
-3.59400779e-01 2.39344284e-01 -5.21113634e-01 1.22870825e-01
1.33187398e-01 5.11356294e-01 1.10856760e+00 -3.93132120e-01
1.95735261e-01 2.26555333e-01 -1.76825188e-02 3.87495846e-01
8.48881721e-01 -3.55793297e-01 -5.31777918e-01 -6.43591583e-02
-1.41189203e-01 3.96885872e-01 7.90553391e-01 -1.57322788e+00
-1.69901490e-01 2.54137844e-01 3.71590734e-01 -4.64140832e-01
-3.93914506e-02 -8.66000891e-01 3.85036200e-01 8.15740168e-01
1.35162771e-01 1.45477712e-01 1.87585771e-01 7.88351893e-01
-4.85349625e-01 -3.69283222e-02 4.71512496e-01 4.39780146e-01
-6.83028698e-01 5.23871928e-02 -6.06018066e-01 -2.71311402e-01
1.32861781e+00 -4.18594956e-01 -1.55102566e-01 -4.36403528e-02
-6.23845398e-01 -6.59586024e-03 4.67108846e-01 8.43868196e-01
9.35655177e-01 -1.43404675e+00 -4.86061633e-01 7.74802983e-01
3.00408483e-01 2.43744031e-01 3.23426396e-01 7.57507145e-01
-4.79049645e-02 1.54503554e-01 -1.10408656e-01 -7.32374012e-01
-9.12326038e-01 4.53045487e-01 4.51496303e-01 -3.05021238e-02
-6.43137991e-01 6.20625198e-01 -4.43297148e-01 -2.65116012e-03
-1.00659383e-02 -4.49996531e-01 -1.49626713e-02 -2.25638673e-01
2.99050473e-02 4.21810120e-01 6.67515159e-01 -5.95832407e-01
-2.22851038e-01 2.62434125e-01 3.48254651e-01 3.35502982e-01
9.45781171e-01 2.76887089e-01 3.45567763e-01 1.05529642e+00
1.04783213e+00 -4.02952135e-01 -1.36654329e+00 1.08637780e-01
-1.31917208e-01 -1.27290234e-01 -2.82435287e-02 -7.58019090e-01
-1.32562244e+00 5.44657052e-01 9.85512495e-01 5.54305255e-01
1.29177058e+00 -1.23899192e-01 1.11998630e+00 -6.65836483e-02
6.36503279e-01 -9.02710140e-01 3.09416980e-01 9.37626213e-02
9.20124769e-01 -8.74291420e-01 -4.68748540e-01 6.22211322e-02
-2.80522287e-01 1.03144264e+00 5.95286548e-01 -5.56457043e-01
6.47980750e-01 4.54847097e-01 1.08662210e-01 -4.09014761e-01
-7.04571068e-01 2.27975950e-01 2.88391352e-01 9.46504414e-01
3.55654061e-02 1.00989193e-01 1.53142482e-01 5.58876634e-01
-6.23365529e-02 -2.64752328e-01 2.86329985e-01 1.37680686e+00
-2.91782469e-01 -7.44618654e-01 -6.10215485e-01 7.86496639e-01
-3.45388591e-01 5.54952562e-01 -6.63136318e-02 7.61512339e-01
3.31195861e-01 1.32580197e+00 1.27686858e-01 -9.23229337e-01
5.60454607e-01 2.88023293e-01 3.24426919e-01 -7.10572660e-01
-2.69039392e-01 -3.27769727e-01 -2.90323999e-02 -6.07083917e-01
-2.36143485e-01 -6.41286731e-01 -1.61779964e+00 1.81569502e-01
-6.99792147e-01 -7.97916129e-02 4.83240157e-01 1.28857827e+00
1.76896945e-01 1.26235747e+00 1.24322712e+00 -8.33078384e-01
-8.39430511e-01 -1.28242266e+00 -8.24992061e-01 6.43537879e-01
6.37830138e-01 -1.30660844e+00 -9.19894397e-01 4.02597934e-02] | [6.784180641174316, 2.3818764686584473] |
6f6b3371-5466-499a-ac7c-b6684c8a328e | energy-based-residual-latent-transport-for | 2211.0682 | null | https://arxiv.org/abs/2211.06820v1 | https://arxiv.org/pdf/2211.06820v1.pdf | Energy-Based Residual Latent Transport for Unsupervised Point Cloud Completion | Unsupervised point cloud completion aims to infer the whole geometry of a partial object observation without requiring partial-complete correspondence. Differing from existing deterministic approaches, we advocate generative modeling based unsupervised point cloud completion to explore the missing correspondence. Specifically, we propose a novel framework that performs completion by transforming a partial shape encoding into a complete one using a latent transport module, and it is designed as a latent-space energy-based model (EBM) in an encoder-decoder architecture, aiming to learn a probability distribution conditioned on the partial shape encoding. To train the latent code transport module and the encoder-decoder network jointly, we introduce a residual sampling strategy, where the residual captures the domain gap between partial and complete shape latent spaces. As a generative model-based framework, our method can produce uncertainty maps consistent with human perception, leading to explainable unsupervised point cloud completion. We experimentally show that the proposed method produces high-fidelity completion results, outperforming state-of-the-art models by a significant margin. | ['Nick Barnes', 'Jing Zhang', 'Saeed Anwar', 'Shi Qiu', 'Ruikai Cui'] | 2022-11-13 | null | null | null | null | ['point-cloud-completion'] | ['computer-vision'] | [ 4.24263835e-01 5.38210809e-01 1.64340675e-01 -3.40440512e-01
-9.95751202e-01 -4.90862489e-01 8.70274484e-01 -1.01135045e-01
1.05472013e-01 2.51535237e-01 1.45004824e-01 -1.72835472e-03
2.29871389e-03 -1.04791915e+00 -1.18247294e+00 -5.58302760e-01
3.12513530e-01 9.99027908e-01 -1.64103974e-02 2.26723701e-01
2.47426361e-01 5.90931356e-01 -1.48442864e+00 1.64468333e-01
1.16641295e+00 7.91359186e-01 6.59592986e-01 4.28292215e-01
-1.55046046e-01 4.64663446e-01 6.94167987e-02 -3.29201430e-01
8.41652676e-02 -2.43506446e-01 -6.91144824e-01 5.88334262e-01
1.28377020e-01 -3.10802191e-01 -3.68810058e-01 9.36704516e-01
-1.36812076e-01 -9.37078893e-02 1.02195108e+00 -9.59527671e-01
-6.40715957e-01 2.18350142e-01 -1.95430979e-01 -8.00450385e-01
3.09447348e-01 6.93658814e-02 1.00622594e+00 -1.18684733e+00
7.68056571e-01 1.16954970e+00 5.00757515e-01 2.69759357e-01
-1.60801291e+00 -2.13138521e-01 2.18028426e-02 -2.46634722e-01
-1.58239079e+00 -3.45271260e-01 1.16001618e+00 -6.89860106e-01
7.06207752e-01 9.34828743e-02 6.49912894e-01 8.47488225e-01
2.71205187e-01 7.73104608e-01 7.14801431e-01 -2.52041191e-01
7.03385770e-01 -1.05871841e-01 -5.46050847e-01 6.29741192e-01
1.02378555e-01 1.07731737e-01 -4.31635439e-01 -2.75419742e-01
1.11207414e+00 3.38162422e-01 -1.74952194e-01 -1.00058997e+00
-1.15469360e+00 7.35013783e-01 5.49065471e-01 -9.91730615e-02
-5.89248180e-01 4.07261282e-01 -3.36697638e-01 -2.07829282e-01
7.72264063e-01 1.60054594e-01 -1.57095328e-01 -1.67295244e-03
-1.30931139e+00 3.70520294e-01 6.60220206e-01 1.44016027e+00
1.22686446e+00 -3.06205247e-02 -1.06790371e-01 4.72727686e-01
9.27372336e-01 6.86161935e-01 -3.69054049e-01 -1.17643917e+00
3.88255954e-01 5.98548234e-01 1.44833446e-01 -8.26635659e-01
1.37468800e-01 -6.18815422e-01 -8.48298192e-01 3.55008036e-01
-4.07242924e-02 3.32262516e-01 -1.01980746e+00 1.58397436e+00
2.30867252e-01 4.05439973e-01 -1.77673777e-04 6.44430995e-01
3.23102981e-01 6.30550563e-01 -2.36129344e-01 -8.66521895e-02
1.09587026e+00 -7.02349961e-01 -4.25830513e-01 -1.54131204e-01
2.20298722e-01 -6.56445146e-01 7.92818725e-01 3.06736797e-01
-1.26515472e+00 -4.92534518e-01 -1.03700674e+00 -1.46502629e-01
1.76670209e-01 1.15638264e-01 4.06326830e-01 1.57447055e-01
-1.02085698e+00 8.09289753e-01 -1.36895180e+00 -5.95241264e-02
4.82385308e-01 1.58645689e-01 -2.29556128e-01 -3.17989379e-01
-4.50693995e-01 4.82222557e-01 1.66171834e-01 7.15173781e-02
-1.27609766e+00 -7.49637544e-01 -1.02881277e+00 2.09900141e-01
1.69173762e-01 -1.24556434e+00 1.04194176e+00 -3.33351403e-01
-1.57038474e+00 5.72574854e-01 -5.47348678e-01 -2.64523268e-01
5.68437934e-01 8.67482461e-03 1.54449031e-01 3.59833658e-01
1.10497840e-01 9.60295200e-01 1.15402400e+00 -1.74065769e+00
-1.28438145e-01 -2.41623163e-01 -2.20421404e-01 9.20357481e-02
2.55124599e-01 -6.17491066e-01 -7.51875937e-01 -6.09240472e-01
6.75673604e-01 -9.36876416e-01 -3.10286343e-01 2.48126641e-01
-4.22161907e-01 8.87788609e-02 7.33519614e-01 -5.14238894e-01
6.24454498e-01 -2.05442834e+00 6.07625008e-01 2.73426712e-01
3.61014038e-01 -4.05255169e-01 9.44526959e-03 5.78267097e-01
5.66765331e-02 -5.56544028e-02 -1.01233470e+00 -1.28506136e+00
2.44247958e-01 5.34488320e-01 -4.75229740e-01 5.85452616e-01
6.13988280e-01 1.03986824e+00 -9.18647528e-01 -2.80927569e-01
5.60530305e-01 8.65980983e-01 -9.63423073e-01 3.53061557e-01
-6.72440648e-01 8.74596477e-01 -4.94245917e-01 6.59816504e-01
1.12200952e+00 -4.39318359e-01 -4.58460040e-02 -6.57899603e-02
-2.23823920e-01 1.59679756e-01 -9.61617351e-01 2.63864350e+00
-4.03096497e-01 3.79955620e-01 1.75481603e-01 -7.21659660e-01
1.06538165e+00 2.52619326e-01 5.31479478e-01 -4.62394617e-02
-1.65334001e-01 2.97713280e-01 -4.93625611e-01 -1.64939433e-01
5.82782984e-01 -2.56426215e-01 1.05907299e-01 1.60924897e-01
1.63405448e-01 -6.54962659e-01 -4.25812602e-01 2.90336937e-01
1.03638101e+00 6.62469983e-01 -1.85363695e-01 -1.78012401e-01
2.93878019e-01 -2.29343131e-01 4.03069019e-01 4.81854290e-01
4.87707943e-01 1.25971735e+00 3.01530063e-01 -8.42668712e-02
-1.46946609e+00 -1.49691129e+00 -1.60553411e-01 8.70221928e-02
2.99202651e-01 -5.39160490e-01 -7.24303722e-01 -4.21628594e-01
-1.28065830e-03 9.74201977e-01 -4.42248732e-01 -1.59604967e-01
-3.42879951e-01 -2.20748976e-01 4.24906649e-02 3.79033357e-01
3.86448234e-01 -8.20093751e-01 -3.83130252e-01 2.42061168e-01
-3.46024811e-01 -1.19781375e+00 -3.77006143e-01 1.09446354e-01
-1.19326282e+00 -6.96241498e-01 -6.59745097e-01 -5.10591269e-01
1.03237450e+00 9.64255184e-02 1.20566678e+00 9.30805504e-03
1.43246744e-02 5.44894457e-01 -3.02978784e-01 4.74693179e-02
-6.09555781e-01 -6.03585914e-02 -2.25594848e-01 2.24781722e-01
-4.33107512e-03 -1.12989140e+00 -6.21873617e-01 4.00711857e-02
-1.15512812e+00 5.80198884e-01 7.78122127e-01 5.75907886e-01
9.89986062e-01 -1.45372823e-01 7.40582719e-02 -4.50978935e-01
2.28370968e-02 -5.76073527e-01 -6.64256573e-01 6.71287009e-04
-6.95032835e-01 5.36760390e-01 4.56798822e-01 8.52421075e-02
-9.75408018e-01 5.58817625e-01 -3.43251884e-01 -9.39791441e-01
-3.16147745e-01 2.42187589e-01 -5.02103209e-01 -1.12940278e-02
2.79373974e-01 6.15332067e-01 -2.90325377e-02 -8.52616429e-01
5.83661616e-01 3.31888258e-01 7.43510783e-01 -8.54509234e-01
1.03051662e+00 9.38626707e-01 2.36607373e-01 -5.92874825e-01
-2.86259085e-01 -4.45403636e-01 -1.03842139e+00 -1.82584405e-01
1.05189693e+00 -9.94091213e-01 -4.91714895e-01 1.24612845e-01
-1.61649621e+00 -1.37948468e-01 -5.59152484e-01 5.14265358e-01
-1.09774053e+00 5.85813403e-01 -2.52712935e-01 -8.38890910e-01
2.39441041e-02 -1.33383358e+00 1.82086313e+00 -2.41073489e-01
2.22361386e-01 -8.98775697e-01 1.70994818e-01 8.54177773e-02
4.12376001e-02 3.82232577e-01 7.78526366e-01 -1.02974184e-01
-1.39339185e+00 -4.31886362e-03 -1.63128391e-01 2.32375801e-01
-3.08694933e-02 -9.82230082e-02 -1.11412370e+00 -1.87465176e-01
2.72047937e-01 4.66384999e-02 1.00788164e+00 4.06082898e-01
1.10922658e+00 -1.78505421e-01 -4.57726121e-01 9.62923706e-01
1.67243433e+00 -2.57756740e-01 9.52813625e-01 -1.98063269e-01
8.67012918e-01 3.67154330e-01 1.95702121e-01 6.14213467e-01
6.13033056e-01 6.32941127e-01 9.38043714e-01 1.41221166e-01
-1.58229306e-01 -8.15912068e-01 1.75940990e-01 9.33700264e-01
-1.20601065e-01 -2.36945733e-01 -8.94171059e-01 5.78785658e-01
-1.92557204e+00 -6.46596909e-01 -7.47895837e-02 2.27925205e+00
4.75850821e-01 2.00062264e-02 -4.08356160e-01 -1.09929815e-01
5.55213988e-01 1.81621477e-01 -5.26488364e-01 3.98618951e-02
8.26493800e-02 1.06264815e-01 2.91653007e-01 8.62706304e-01
-7.26552606e-01 8.86398017e-01 6.04444504e+00 8.90365005e-01
-6.66360259e-01 1.61797702e-01 1.78958818e-01 3.59237909e-01
-1.14648032e+00 5.78220963e-01 -4.12186474e-01 3.83840501e-01
5.28183818e-01 2.15145260e-01 5.16671002e-01 8.59467745e-01
1.73508972e-01 1.16601393e-01 -1.33199859e+00 1.04884970e+00
2.21016686e-02 -1.55694187e+00 2.34827667e-01 6.25898421e-01
7.95872986e-01 1.19688235e-01 -1.32502183e-01 -5.26631102e-02
1.76245466e-01 -8.17382514e-01 1.15505028e+00 1.05035055e+00
1.00332487e+00 -6.08541667e-01 2.10074544e-01 7.50883281e-01
-1.34962535e+00 4.17709261e-01 -4.94702250e-01 1.31605849e-01
5.15163243e-01 8.44555259e-01 -5.56778133e-01 9.94372845e-01
2.46961266e-01 7.93887973e-01 -1.34048298e-01 9.45106268e-01
-3.78493696e-01 4.23989028e-01 -4.01180506e-01 5.52172244e-01
2.15095934e-03 -5.50350785e-01 1.01468968e+00 8.64463449e-01
5.66380799e-01 4.62394767e-02 1.94847971e-01 1.79745543e+00
-1.10326581e-01 -3.34403694e-01 -6.63732529e-01 9.21345949e-02
3.59890372e-01 1.16132569e+00 -7.97440410e-01 -2.16229811e-01
-1.68931380e-01 1.34126830e+00 3.65020066e-01 4.15536374e-01
-7.44337320e-01 8.30832720e-02 4.50411946e-01 1.29651055e-01
5.12941360e-01 -6.26032531e-01 -5.50086379e-01 -1.35535467e+00
2.31888652e-01 -8.63387585e-02 -3.81561011e-01 -1.03810251e+00
-1.07933760e+00 5.11954367e-01 1.06176294e-01 -1.64469135e+00
-4.86200124e-01 -3.17573875e-01 -6.16796494e-01 9.55856144e-01
-1.61681354e+00 -1.52295375e+00 -3.47843081e-01 3.04732114e-01
5.72985411e-01 1.55154169e-01 8.69055569e-01 -3.09288036e-02
8.52025300e-02 1.17343113e-01 2.62134373e-01 -2.16737702e-01
2.19244704e-01 -1.28415799e+00 6.30475342e-01 9.86977816e-01
2.41981596e-01 5.79857528e-01 7.04680502e-01 -9.48771715e-01
-1.53941119e+00 -1.22226191e+00 6.94239855e-01 -5.74048579e-01
4.32123244e-02 -7.18194485e-01 -9.10212874e-01 5.87534249e-01
2.72260141e-02 4.92112078e-02 1.70272559e-01 -3.65282238e-01
-2.05635145e-01 2.90691882e-01 -1.00878286e+00 3.59047353e-01
1.11936712e+00 -7.39858210e-01 -6.20062053e-01 1.55804336e-01
1.01997530e+00 -4.68384445e-01 -8.55560005e-01 4.78730083e-01
3.43402624e-01 -8.47607136e-01 9.99273777e-01 2.51757745e-02
7.79054403e-01 -5.30148804e-01 -4.49993551e-01 -1.22260094e+00
-4.35780585e-01 -7.19134510e-01 -4.64159071e-01 1.07836449e+00
1.73840746e-01 -2.71209717e-01 9.28881645e-01 5.32986760e-01
-7.74295688e-01 -7.20253646e-01 -9.63428915e-01 -7.51201808e-01
1.03056766e-01 -9.03594375e-01 8.86088908e-01 6.06099427e-01
-3.63906264e-01 -6.73348375e-04 -2.96814084e-01 5.20975173e-01
1.07321346e+00 2.67123818e-01 6.88343227e-01 -1.28035319e+00
-5.20476341e-01 -8.74720439e-02 -3.70654941e-01 -1.69660783e+00
2.06860632e-01 -1.01612937e+00 2.45626971e-01 -1.78580201e+00
1.90695584e-01 -3.85517925e-01 2.52093047e-01 2.00881347e-01
1.50813937e-01 1.44576743e-01 1.81430623e-01 5.96588612e-01
-4.10457730e-01 1.26521873e+00 1.27862835e+00 -1.47220537e-01
-1.19719729e-02 -4.09064665e-02 -4.74690646e-01 7.61630356e-01
1.84427425e-01 -6.18552208e-01 -4.45782512e-01 -6.46598577e-01
3.61367255e-01 1.82409376e-01 7.86914945e-01 -1.18126237e+00
4.37701017e-01 -7.68538266e-02 3.10345441e-01 -1.09385574e+00
6.49204493e-01 -1.00649762e+00 5.25449455e-01 1.18764095e-01
-1.35255694e-01 -4.35923427e-01 -7.07872435e-02 9.27651942e-01
-1.53901294e-01 -1.40707120e-01 4.98922259e-01 -1.61523502e-02
-2.55908966e-01 7.10399985e-01 -2.89452802e-02 -2.62814939e-01
6.31116986e-01 -2.12503567e-01 1.61446452e-01 -3.64252687e-01
-8.74888659e-01 4.42221798e-02 1.01212287e+00 2.91910112e-01
9.89392102e-01 -1.63610959e+00 -8.07106555e-01 5.83186030e-01
3.25654596e-01 7.39324450e-01 3.19720894e-01 5.89284897e-01
-6.26146138e-01 4.13781643e-01 2.50640571e-01 -1.19372690e+00
-4.79156733e-01 4.70170736e-01 1.53707162e-01 -1.24592558e-01
-1.01564407e+00 5.47564566e-01 5.26025534e-01 -6.64651275e-01
-1.61314175e-01 -6.52213216e-01 3.89170587e-01 -6.05202675e-01
6.24875128e-02 9.29864347e-02 3.30502167e-02 -7.38623977e-01
-1.54090062e-01 7.56895185e-01 4.41918105e-01 -4.71833944e-01
1.39637291e+00 -1.91922143e-01 -1.23481423e-01 3.57158959e-01
1.29243243e+00 9.95682254e-02 -1.84807539e+00 -2.57674873e-01
-3.71929049e-01 -5.73319137e-01 8.38804543e-02 -3.19425762e-01
-7.03376591e-01 1.01842403e+00 2.28568166e-01 -9.29770917e-02
8.59278619e-01 3.33001703e-01 5.79625010e-01 6.33517727e-02
6.62690938e-01 -7.28427708e-01 -1.45266727e-01 3.73185992e-01
1.08308506e+00 -1.10314894e+00 3.11760753e-02 -7.08275139e-01
-2.55766839e-01 1.04838550e+00 1.31229162e-01 -3.39987576e-01
7.90784180e-01 1.10618271e-01 -7.82381415e-01 -5.01145184e-01
-7.83505082e-01 -1.40113741e-01 5.87544084e-01 6.26124024e-01
3.47999483e-02 1.36454180e-01 1.51356786e-01 4.37320590e-01
-1.29267260e-01 6.20406121e-02 1.89645782e-01 7.13007569e-01
-3.67800415e-01 -1.22491229e+00 -3.89933586e-01 5.63054979e-02
1.88251212e-01 -7.57313147e-02 -1.55766279e-01 3.78114611e-01
2.17375860e-01 6.50974393e-01 3.43661249e-01 -4.51908857e-01
1.16158701e-01 8.02136585e-02 4.90774781e-01 -7.90211678e-01
2.39838779e-01 4.77569699e-01 -3.85346562e-01 -6.56689882e-01
-2.60832191e-01 -7.28661656e-01 -1.38469422e+00 -8.42254013e-02
-2.08508790e-01 1.52971253e-01 8.23579252e-01 1.10822761e+00
5.95853984e-01 5.34891188e-01 7.47853041e-01 -1.31240332e+00
-3.18036854e-01 -8.27203751e-01 -5.89476824e-01 2.63682157e-01
3.14909846e-01 -7.19943047e-01 -3.68238091e-01 1.85119972e-01] | [8.733274459838867, -3.439751148223877] |
f98767d8-a845-418d-b364-424c772018dd | csmcnet-scalable-video-compressive-sensing | 2108.01522 | null | https://arxiv.org/abs/2108.01522v1 | https://arxiv.org/pdf/2108.01522v1.pdf | CSMCNet: Scalable Video Compressive Sensing Reconstruction with Interpretable Motion Estimation | Most deep network methods for compressive sensing reconstruction suffer from the black-box characteristic of DNN. In this paper, a deep neural network with interpretable motion estimation named CSMCNet is proposed. The network is able to realize high-quality reconstruction of video compressive sensing by unfolding the iterative steps of optimization based algorithms. A DNN based, multi-hypothesis motion estimation module is designed to improve the reconstruction quality, and a residual module is employed to further narrow down the gap between re-construction results and original signal in our proposed method. Besides, we propose an interpolation module with corresponding training strategy to realize scalable CS reconstruction, which is capable of using the same model to decode various compression ratios. Experiments show that a PSNR of 29.34dB can be achieved at 2% CS ratio (compressed by 98%), which is superior than other state-of-the-art methods. Moreover, the interpolation module is proved to be effective, with significant cost saving and acceptable performance losses. | ['Yibo Fan', 'Jinjia Zhou', 'Xiao Yan', 'Bowen Huang'] | 2021-08-03 | null | null | null | null | ['video-compressive-sensing'] | ['computer-vision'] | [ 3.38834435e-01 -2.35604063e-01 -1.85987920e-01 2.56940466e-03
-2.85888255e-01 1.81313902e-01 1.98380858e-01 -4.71468747e-01
-3.52560610e-01 5.91961741e-01 4.31998104e-01 -2.78804421e-01
6.75464272e-02 -7.14044929e-01 -7.41559029e-01 -7.52637565e-01
-6.80258945e-02 -2.32115611e-01 1.01455897e-01 -1.59362257e-01
3.28900665e-01 2.73668289e-01 -1.25943387e+00 3.44847411e-01
7.99099267e-01 1.37060773e+00 8.58314216e-01 3.56361121e-01
1.33091748e-01 1.22750473e+00 -3.50343823e-01 -7.76158646e-02
2.97025800e-01 -6.84132159e-01 -2.97230631e-01 7.67397061e-02
5.45532480e-02 -1.03430128e+00 -1.00260615e+00 1.28001630e+00
6.27907813e-01 -7.93152303e-02 2.20383957e-01 -5.76815188e-01
-5.60335934e-01 9.01923239e-01 -5.02243161e-01 1.38037264e-01
2.36437127e-01 9.20976773e-02 6.56188071e-01 -9.71028447e-01
5.51811159e-01 1.12486649e+00 5.22454321e-01 5.78326941e-01
-7.26129115e-01 -7.01993942e-01 -3.40318531e-01 7.51144052e-01
-1.35986602e+00 -7.44228482e-01 1.01049793e+00 7.34488815e-02
6.66579485e-01 2.80962259e-01 7.95755208e-01 9.91827548e-01
1.56342268e-01 8.63666773e-01 5.10907352e-01 -3.92062783e-01
1.56602204e-01 -4.93511051e-01 -6.37533307e-01 5.78858495e-01
2.86973208e-01 7.68083557e-02 -4.46833938e-01 5.60450673e-01
1.19377851e+00 4.03636456e-01 -7.02633500e-01 -1.29197091e-02
-1.37577260e+00 5.55831075e-01 7.48521507e-01 7.12151408e-01
-4.83508080e-01 4.89913017e-01 2.21832812e-01 2.73505211e-01
2.56100036e-02 6.33836389e-02 -2.10488364e-02 1.33581400e-01
-1.43498135e+00 -1.05207833e-02 5.13755739e-01 8.55345249e-01
3.92987728e-01 8.51381838e-01 -1.27288952e-01 7.15157211e-01
5.61153889e-01 4.10717219e-01 8.19319367e-01 -1.49091876e+00
5.51327229e-01 1.63763359e-01 2.15469860e-02 -1.48191607e+00
-1.08197838e-01 -9.26892161e-01 -1.58375001e+00 -2.12657303e-02
-1.37264252e-01 5.16298134e-03 -6.49000168e-01 1.54335856e+00
-1.80055261e-01 5.56084692e-01 1.84874028e-01 1.15976834e+00
8.13514888e-01 8.46495330e-01 -3.31956565e-01 -5.51176488e-01
7.91498065e-01 -1.23014939e+00 -9.66786146e-01 -2.38552451e-01
2.78716803e-01 -6.20049894e-01 5.35588562e-01 5.57372689e-01
-1.40264368e+00 -8.74438643e-01 -1.50645506e+00 -7.91603234e-03
6.25406325e-01 3.52288276e-01 3.01924676e-01 3.37068945e-01
-1.11252689e+00 9.82939065e-01 -8.63763452e-01 2.25569263e-01
5.19203126e-01 1.45431876e-01 -7.63686895e-02 -3.99562418e-01
-1.20723295e+00 6.45028353e-01 6.01152956e-01 5.51655412e-01
-1.30230069e+00 -3.53694707e-01 -7.03487992e-01 4.66343284e-01
1.83647811e-01 -7.16004729e-01 8.22807431e-01 -1.17682886e+00
-1.54001999e+00 2.42003635e-01 -1.19875796e-01 -5.57086408e-01
6.54442430e-01 -1.67561337e-01 -6.54947937e-01 6.11637115e-01
-1.22497231e-01 5.18940389e-01 1.11061430e+00 -1.16978753e+00
-2.96538711e-01 -4.01857793e-02 -2.64269561e-01 1.62831545e-01
-4.96526301e-01 -2.14547858e-01 -5.76017380e-01 -9.56596017e-01
6.41253352e-01 -5.02826273e-01 -5.48793137e-01 1.61232308e-01
-2.12444216e-01 4.13088083e-01 8.48344803e-01 -1.18198073e+00
1.67318666e+00 -2.25547647e+00 2.78181791e-01 -6.70665354e-02
4.07804817e-01 5.82830906e-01 -3.20865810e-01 2.73078233e-01
6.48189411e-02 -6.32957965e-02 -5.13095021e-01 -4.55078840e-01
-3.72610360e-01 9.01910216e-02 -1.47560582e-01 5.00653744e-01
-1.47266299e-01 7.70076811e-01 -6.87799513e-01 -2.66104013e-01
3.64708781e-01 5.59358537e-01 -7.56703675e-01 1.74280673e-01
-5.94518073e-02 6.88645422e-01 -3.06493551e-01 6.18292332e-01
1.08042896e+00 -3.65870446e-01 5.61300755e-01 -5.08382201e-01
-2.49304790e-02 8.07789639e-02 -1.26795447e+00 2.23970795e+00
-2.65109390e-01 6.94754720e-01 4.17186916e-01 -1.38257086e+00
8.80768895e-01 3.76400352e-01 4.47843760e-01 -8.31783295e-01
3.13150823e-01 3.86215746e-01 1.37432329e-02 -7.19843268e-01
5.24638057e-01 -1.32648125e-01 5.12038469e-01 -3.59999314e-02
-2.31252849e-01 2.03722507e-01 -4.56918813e-02 7.96398073e-02
1.08283198e+00 1.98303387e-02 3.93242612e-02 -1.06562167e-01
8.16817164e-01 -4.26381439e-01 7.43857980e-01 5.38421154e-01
-1.26366794e-01 7.28867114e-01 5.58553822e-02 -3.98432136e-01
-1.39157486e+00 -8.06193829e-01 1.00810744e-01 3.16808581e-01
6.15222096e-01 -2.05424398e-01 -5.83774209e-01 8.36028457e-02
-4.93314415e-01 6.46696746e-01 7.38456920e-02 -2.24993035e-01
-8.77054811e-01 -4.25784022e-01 4.33432579e-01 3.98132712e-01
1.22455657e+00 -1.02862418e+00 -5.71322322e-01 5.81791520e-01
-7.02173591e-01 -1.30614126e+00 -1.95970356e-01 -3.80487919e-01
-1.29511189e+00 -7.16818035e-01 -9.50716436e-01 -1.04946589e+00
4.51512605e-01 5.50085068e-01 7.50997722e-01 4.82447475e-01
1.76090688e-01 -4.04571146e-01 -5.15582621e-01 2.71980464e-01
-5.31732380e-01 -3.29024971e-01 -8.68405476e-02 8.07644278e-02
-1.30998701e-01 -9.50845242e-01 -1.15978694e+00 1.38097018e-01
-1.26690888e+00 3.15200508e-01 8.32351089e-01 9.55149770e-01
3.79007131e-01 1.63097233e-01 5.63373685e-01 -1.75863758e-01
2.18750194e-01 -4.94529843e-01 -3.66173148e-01 -9.22041833e-02
-6.36504233e-01 -1.39340118e-01 9.37394798e-01 -2.75858939e-01
-9.41979051e-01 -7.84906894e-02 -4.32393640e-01 -8.39035034e-01
1.39663771e-01 6.29909515e-01 -1.54015467e-01 -1.03843227e-01
2.76323706e-01 7.57116795e-01 1.19226731e-01 -5.64673305e-01
5.62069193e-02 5.88804364e-01 6.62990868e-01 2.83016656e-02
5.69457710e-01 4.88972336e-01 1.37293071e-01 -6.79550529e-01
-5.18557191e-01 -3.39252986e-02 -1.53186783e-01 -2.61578262e-01
7.92451620e-01 -1.32951379e+00 -7.24967122e-01 6.18436813e-01
-1.23802245e+00 -1.80922270e-01 7.26294518e-02 8.23285103e-01
-4.17217970e-01 9.74123836e-01 -1.12652898e+00 -5.65857530e-01
-4.65766460e-01 -1.23790979e+00 4.94952261e-01 3.54954451e-02
4.77403194e-01 -6.76798761e-01 -4.84551579e-01 5.43624699e-01
8.13129365e-01 7.56294951e-02 5.35434961e-01 -2.94181593e-02
-1.08913672e+00 1.10081621e-01 -3.09806705e-01 6.97733343e-01
-2.43572637e-01 -5.13005257e-01 -7.30503857e-01 -5.75529754e-01
5.88736653e-01 -8.44763741e-02 1.15341735e+00 6.85121953e-01
1.54714739e+00 -5.30182242e-01 -8.32197294e-02 1.06456196e+00
1.81804383e+00 2.95071274e-01 1.32322013e+00 3.81132513e-01
6.59633636e-01 1.28280837e-02 3.07392687e-01 6.31836712e-01
1.71736181e-01 3.09799701e-01 9.02812123e-01 -1.96489114e-02
-3.00782293e-01 -3.37666571e-01 4.35218155e-01 1.35456717e+00
-1.72890261e-01 -4.26509380e-01 -3.52269977e-01 3.38408887e-01
-1.74080026e+00 -1.21200573e+00 -2.47304007e-01 1.94346309e+00
5.64944446e-01 5.34717217e-02 -4.87845331e-01 4.05248672e-01
7.47457087e-01 5.76738060e-01 -5.30214012e-01 2.01173678e-01
-3.57264996e-01 7.10878447e-02 3.61772090e-01 4.16049451e-01
-8.02450657e-01 7.74534047e-01 6.02641678e+00 1.11753166e+00
-1.23485363e+00 1.60888210e-01 6.02150977e-01 8.49029347e-02
-4.92414534e-01 -2.49725804e-02 -1.79403141e-01 7.32867658e-01
8.33688557e-01 2.37651095e-01 6.41560733e-01 6.17260933e-01
5.58767676e-01 1.37502477e-02 -7.52403855e-01 1.25744700e+00
1.07537016e-01 -1.73328960e+00 2.50546575e-01 -1.45887002e-01
7.33589470e-01 -9.55761299e-02 -8.28080773e-02 -1.09421737e-01
-3.10533464e-01 -1.08658230e+00 9.05900419e-01 5.74173272e-01
1.03343582e+00 -6.52040899e-01 8.20549190e-01 5.85658312e-01
-1.15042305e+00 -3.26298535e-01 -6.81551695e-01 -2.53736526e-01
4.54684377e-01 8.55766594e-01 -2.27344796e-01 9.00409818e-01
5.94771981e-01 1.05600178e+00 4.09879796e-02 1.00902736e+00
-2.10223362e-01 5.90078294e-01 -1.60098299e-02 2.50010222e-01
2.02424437e-01 -4.38405782e-01 6.91728592e-01 8.42870831e-01
9.84917402e-01 1.04028538e-01 -1.28953576e-01 8.76578331e-01
-2.73018330e-01 -1.61071360e-01 -4.71261114e-01 1.66193232e-01
5.18142700e-01 9.39693570e-01 -3.23118925e-01 -5.57649851e-01
-1.92826211e-01 1.26443613e+00 -2.05856323e-01 3.78034323e-01
-7.88527846e-01 -2.29528576e-01 1.69807434e-01 4.98411898e-03
7.41729736e-01 -4.32474345e-01 -1.17805205e-01 -1.50241625e+00
8.64583850e-02 -9.61475015e-01 -1.95192426e-01 -9.31878507e-01
-6.67495012e-01 4.80481237e-01 -5.01652181e-01 -1.67964959e+00
-5.92582375e-02 -3.55558544e-01 -5.68369865e-01 6.08678639e-01
-1.75839591e+00 -7.52149582e-01 -5.11221111e-01 6.15515530e-01
7.31358826e-01 -3.27123135e-01 4.29377824e-01 8.65748882e-01
-5.45188487e-01 5.75157166e-01 4.57861602e-01 9.97375026e-02
2.57055223e-01 -3.54467183e-01 -1.48495007e-02 1.37387669e+00
-1.93727866e-01 5.18783152e-01 6.42443299e-01 -5.74427843e-01
-1.74607563e+00 -1.10064733e+00 6.23319685e-01 6.03463173e-01
4.02893990e-01 2.63438970e-01 -8.62207472e-01 4.33133841e-01
2.85073817e-01 -4.41806875e-02 2.76070029e-01 -8.99047077e-01
-7.78673887e-02 -3.27141792e-01 -1.18422699e+00 5.37226677e-01
1.09467328e+00 -2.08294973e-01 -2.57458180e-01 2.07206812e-02
8.78902256e-01 -4.31876242e-01 -7.33279407e-01 5.42228878e-01
5.09799719e-01 -1.43034744e+00 1.14657843e+00 4.66183834e-02
1.17748737e+00 -4.52300370e-01 -5.87972045e-01 -8.96492064e-01
-4.68458623e-01 -5.51658094e-01 -4.97886568e-01 8.05793822e-01
4.22236174e-02 -3.01520467e-01 8.40232491e-01 -1.83629006e-01
-5.45837402e-01 -6.46026790e-01 -9.58959401e-01 -6.97253227e-01
-3.53132397e-01 -4.73294437e-01 4.20661122e-01 9.35325027e-01
-2.87055522e-01 1.91508189e-01 -9.37683284e-01 1.66616514e-01
8.51482272e-01 -8.17910116e-03 3.26509863e-01 -6.57427251e-01
-5.58975995e-01 -3.91821355e-01 -3.26319635e-01 -1.90552175e+00
-3.08123231e-02 -8.23309064e-01 -6.79612309e-02 -1.62714863e+00
1.37988195e-01 -2.51673013e-01 -4.09706384e-01 -1.11328885e-01
1.19260460e-01 3.26327920e-01 4.32214826e-01 5.65414131e-01
-4.42648888e-01 9.10034359e-01 1.75745881e+00 -2.34709829e-01
1.64001152e-01 -1.21049702e-01 -5.92347503e-01 6.02708161e-01
7.31782854e-01 -2.42854387e-01 -3.38451445e-01 -9.27395642e-01
-5.08935004e-02 8.13624918e-01 3.06933522e-01 -1.38305414e+00
3.43873918e-01 -6.06007921e-03 6.10333622e-01 -7.26682067e-01
4.60881770e-01 -9.65247989e-01 4.11481917e-01 1.05034864e+00
-2.21350476e-01 -1.34527594e-01 -2.59061813e-01 7.28599966e-01
-5.43924689e-01 -4.08359945e-01 9.16714489e-01 -3.14376980e-01
-8.08696985e-01 3.73504668e-01 -2.37936050e-01 -4.30116594e-01
6.49833620e-01 -2.48368308e-01 4.74598072e-02 -6.37400627e-01
-5.31139612e-01 -3.10364868e-02 3.68269354e-01 -1.14576161e-01
1.12088776e+00 -1.50449395e+00 -7.89194405e-01 2.35939234e-01
-5.93925416e-01 -1.15003847e-02 5.97047091e-01 8.20413649e-01
-1.23664677e+00 1.85502976e-01 -3.05316687e-01 -4.94710684e-01
-7.25260377e-01 5.15264094e-01 1.70679525e-01 -2.18280032e-01
-7.07939327e-01 6.90120757e-01 -2.20892981e-01 1.28958568e-01
1.92053184e-01 -1.01106972e-01 -2.08220586e-01 -3.56535494e-01
6.83209300e-01 5.99795341e-01 -2.47000352e-01 -4.17332858e-01
-9.52521861e-02 4.25396800e-01 4.18890387e-01 -2.61606388e-02
1.53226626e+00 -4.13730383e-01 -3.57222438e-01 -9.36179981e-02
1.24066305e+00 -2.83740200e-02 -1.34526837e+00 -2.02492073e-01
-4.68297631e-01 -5.82896411e-01 3.12774688e-01 -3.19866151e-01
-1.59598112e+00 9.60731149e-01 6.89233780e-01 2.83474270e-02
1.54343224e+00 -4.79071230e-01 1.39566863e+00 2.15965956e-01
2.93705761e-01 -7.92863190e-01 2.34673887e-01 3.25168550e-01
1.01948035e+00 -1.22252500e+00 2.33211860e-01 -1.37857154e-01
-2.44179726e-01 1.36048651e+00 4.02319551e-01 -3.49199831e-01
4.49478865e-01 2.84133181e-02 -3.34843367e-01 1.68328390e-01
-3.61707717e-01 2.38488406e-01 4.01665606e-02 1.87070414e-01
1.49300307e-01 -1.28929585e-01 -6.47174954e-01 2.64753938e-01
7.00246394e-02 2.08661020e-01 6.40436828e-01 6.30031705e-01
-7.51386106e-01 -7.54533947e-01 -2.05891594e-01 1.48134217e-01
-5.86835802e-01 -3.80940765e-01 5.43516338e-01 3.92306685e-01
3.53292413e-02 1.10698330e+00 4.17609587e-02 -7.96377361e-01
-2.02926815e-01 -5.69775462e-01 4.18229401e-01 9.17317718e-03
-1.04258597e-01 4.50459927e-01 -6.02293015e-02 -6.83942795e-01
-7.51801789e-01 -1.65242940e-01 -1.14530289e+00 -6.27752006e-01
-1.65641591e-01 -1.61210716e-01 6.01927757e-01 8.56904745e-01
3.20812941e-01 5.92533052e-01 9.62896764e-01 -1.01044548e+00
-5.27977824e-01 -1.00380886e+00 -6.18512034e-01 2.35612392e-01
5.54510057e-01 -8.48876238e-02 -5.25184810e-01 1.54741630e-01] | [11.157299995422363, -2.0167059898376465] |
1b25a0df-8f26-4fe3-9f3c-f45a146bde30 | learning-correspondency-in-frequency-domain | 2207.08602 | null | https://arxiv.org/abs/2207.08602v3 | https://arxiv.org/pdf/2207.08602v3.pdf | Pansharpening via Frequency-Aware Fusion Network with Explicit Similarity Constraints | The process of fusing a high spatial resolution (HR) panchromatic (PAN) image and a low spatial resolution (LR) multispectral (MS) image to obtain an HRMS image is known as pansharpening. With the development of convolutional neural networks, the performance of pansharpening methods has been improved, however, the blurry effects and the spectral distortion still exist in their fusion results due to the insufficiency in details learning and the frequency mismatch between MSand PAN. Therefore, the improvement of spatial details at the premise of reducing spectral distortion is still a challenge. In this paper, we propose a frequency-aware fusion network (FAFNet) together with a novel high-frequency feature similarity loss to address above mentioned problems. FAFNet is mainly composed of two kinds of blocks, where the frequency aware blocks aim to extract features in the frequency domain with the help of discrete wavelet transform (DWT) layers, and the frequency fusion blocks reconstruct and transform the features from frequency domain to spatial domain with the assistance of inverse DWT (IDWT) layers. Finally, the fusion results are obtained through a convolutional block. In order to learn the correspondence, we also propose a high-frequency feature similarity loss to constrain the HF features derived from PAN and MS branches, so that HF features of PAN can reasonably be used to supplement that of MS. Experimental results on three datasets at both reduced- and full-resolution demonstrate the superiority of the proposed method compared with several state-of-the-art pansharpening models. | ['Yanning Zhang', 'Xiuwei Zhang', 'Houjun He', 'Yan Zhang', 'Yinghui Xing'] | 2022-07-18 | null | null | null | null | ['pansharpening'] | ['computer-vision'] | [ 7.15928257e-01 -5.62716126e-01 1.17107388e-02 -1.69118032e-01
-6.03305340e-01 -1.82994902e-01 4.47567463e-01 -3.49196225e-01
-3.35417271e-01 6.12459421e-01 1.72103986e-01 2.07613140e-01
-5.89937925e-01 -1.22056246e+00 -5.56116223e-01 -1.06911588e+00
1.35434076e-01 -6.15387797e-01 1.82691723e-01 -4.94543701e-01
-2.06599589e-02 6.00837290e-01 -1.66099155e+00 3.10284406e-01
1.42128515e+00 1.10072637e+00 5.45625925e-01 3.99363667e-01
1.00612484e-01 5.48904240e-01 -3.02749664e-01 1.14430174e-01
5.57916403e-01 -3.84455711e-01 -3.08694094e-01 9.16297361e-02
3.80736083e-01 -5.20089865e-01 -4.63804483e-01 1.41445124e+00
3.79505485e-01 2.82165080e-01 1.22086465e-01 -6.99729323e-01
-9.40366864e-01 2.91257173e-01 -1.02671325e+00 3.31296653e-01
-1.33389696e-01 -4.56858948e-02 7.11515009e-01 -7.40139782e-01
2.00345576e-01 9.89391804e-01 8.27793002e-01 -1.50233030e-01
-1.00692904e+00 -6.37999177e-01 -1.49865314e-01 5.10696113e-01
-1.19818962e+00 -1.51532227e-02 1.22756684e+00 -7.02768192e-02
4.96054441e-01 3.71175498e-01 7.81266689e-01 4.54415679e-01
2.60467917e-01 3.71040910e-01 1.28054237e+00 -3.62801045e-01
-4.67701435e-01 -2.95125395e-01 4.66135480e-02 3.08771044e-01
1.37145326e-01 5.27031243e-01 -2.08702818e-01 2.25470647e-01
9.40267444e-01 2.14718848e-01 -8.68734896e-01 4.30849707e-03
-9.76881981e-01 7.08083987e-01 1.17881250e+00 6.65500164e-01
-6.73760951e-01 -3.06432486e-01 -5.70644950e-03 2.00427219e-01
5.50948203e-01 2.22562820e-01 -4.61409599e-01 5.24671674e-01
-1.09602129e+00 2.25966722e-01 1.43551782e-01 3.93359810e-01
1.08722138e+00 1.45409226e-01 -4.30913642e-02 1.11052346e+00
9.33355466e-02 5.87067366e-01 5.65039456e-01 -7.18655527e-01
3.69158357e-01 4.92131263e-01 -1.22659001e-02 -1.16488981e+00
-4.52926040e-01 -5.70362926e-01 -1.30079520e+00 1.46689922e-01
-4.39031646e-02 -8.96819457e-02 -9.08917725e-01 1.57532156e+00
2.63826042e-01 2.82205880e-01 6.87631592e-02 1.28189003e+00
8.30813408e-01 1.16775966e+00 -2.01788798e-01 -5.61357617e-01
1.29815185e+00 -9.14598346e-01 -9.02428865e-01 -1.53082401e-01
-9.32514668e-02 -9.93669927e-01 7.22553194e-01 3.12687248e-01
-8.06320846e-01 -1.14141607e+00 -1.22264636e+00 -1.86671749e-01
-4.59216863e-01 3.60315025e-01 6.95526898e-01 3.79661083e-01
-6.75021112e-01 7.55315423e-01 -5.60521960e-01 -9.37681720e-02
3.57287049e-01 -1.19223766e-01 -2.60001898e-01 -3.20244372e-01
-1.45869410e+00 8.66881251e-01 6.45033181e-01 5.79280436e-01
-2.82568723e-01 -9.59799826e-01 -6.90151513e-01 2.02117071e-01
1.50941774e-01 -3.62906814e-01 5.71024179e-01 -1.16257632e+00
-1.30435205e+00 2.61516809e-01 2.83345938e-01 -1.69880494e-01
2.48143762e-01 -4.26472187e-01 -1.09542727e+00 4.17283922e-01
7.85788223e-02 3.55343491e-01 1.14255440e+00 -1.12516320e+00
-9.98849928e-01 -2.34862179e-01 -6.89263046e-02 3.72331858e-01
-2.76965857e-01 -1.50335044e-01 -1.26558572e-01 -9.30111051e-01
3.59486938e-01 -3.32274497e-01 8.00008774e-02 -1.19258882e-02
-3.93474028e-02 2.35941917e-01 1.22351801e+00 -1.32676721e+00
1.00247335e+00 -2.30067801e+00 1.19345665e-01 -3.03909592e-02
-9.26349536e-02 5.97620130e-01 -3.35019410e-01 1.43363789e-01
-5.02124310e-01 -1.88113496e-01 -7.37453699e-01 4.57287848e-01
-5.41174650e-01 1.46642989e-02 -4.07891631e-01 5.30817688e-01
3.54483336e-01 5.45503497e-01 -5.78892052e-01 -2.96088457e-01
6.06615722e-01 8.50689113e-01 -1.55674085e-01 1.26367405e-01
7.75869237e-03 4.36458886e-01 -4.43643630e-01 5.68080485e-01
1.43468332e+00 3.07065815e-01 -3.84157121e-01 -8.63264024e-01
-6.07660055e-01 -2.29117706e-01 -1.14306581e+00 1.50656366e+00
-4.84640062e-01 3.15650254e-01 3.21440756e-01 -1.04176819e+00
1.09893548e+00 1.15473367e-01 5.71389198e-01 -9.59330320e-01
-1.09104767e-01 2.41265535e-01 -1.09593242e-01 -4.90437537e-01
4.05243993e-01 -3.40866923e-01 5.20632446e-01 -7.70687452e-03
-5.40864542e-02 -4.09429938e-01 -4.65987846e-02 -5.10667861e-01
2.98568130e-01 2.26537019e-01 1.34439632e-01 -1.29171476e-01
9.92733896e-01 -1.76163856e-02 6.76209271e-01 3.55428636e-01
-3.49089168e-02 6.58224225e-01 -1.52541324e-01 -5.19058704e-01
-1.11600399e+00 -9.43259597e-01 -4.60379779e-01 6.35527611e-01
4.97473657e-01 7.04348609e-02 -4.87873465e-01 -1.88148633e-01
-2.37061739e-01 5.83187521e-01 -4.12473589e-01 -3.54791254e-01
-7.72204459e-01 -1.18937862e+00 1.91132784e-01 1.90251663e-01
1.55760348e+00 -9.93530095e-01 -6.41757548e-01 2.36558989e-01
-3.50731850e-01 -1.00925636e+00 -3.79555970e-01 -7.01079741e-02
-7.79385269e-01 -1.12098718e+00 -8.49319637e-01 -6.21208310e-01
1.14479825e-01 7.63636947e-01 4.02676314e-01 -1.64582461e-01
-4.10933971e-01 -1.71336621e-01 -5.63679993e-01 -1.24547496e-01
-1.21643394e-01 -1.68023944e-01 -5.46386719e-01 4.23164666e-01
1.72133744e-01 -8.54836702e-01 -6.61936641e-01 9.32462066e-02
-1.45413613e+00 3.34110677e-01 1.02400887e+00 9.64412391e-01
6.34179890e-01 9.73987401e-01 4.35968697e-01 -3.93905640e-01
3.18104565e-01 -2.07105801e-01 -7.47216046e-01 2.11771160e-01
-5.00601232e-01 -4.46887672e-01 7.57638276e-01 -5.10851145e-01
-1.61025488e+00 -1.39255494e-01 -1.70310289e-01 -3.25361401e-01
-1.66493788e-01 6.39797866e-01 -2.18272597e-01 -3.95024925e-01
5.73210120e-01 5.99563181e-01 -3.82639244e-02 -6.86427414e-01
3.77834946e-01 7.58818626e-01 7.89095581e-01 -4.55782302e-02
1.07645094e+00 5.72335839e-01 3.84629332e-02 -1.10993040e+00
-9.54933763e-01 -3.77097577e-01 -3.64578694e-01 -1.51541218e-01
9.21905696e-01 -1.13368952e+00 -1.59201786e-01 7.86297917e-01
-9.10263479e-01 3.89037430e-02 -3.42432648e-01 6.34666741e-01
-3.63631666e-01 7.34160423e-01 -6.21182442e-01 -4.87882406e-01
-4.81997460e-01 -8.59216928e-01 8.46610069e-01 7.99695313e-01
5.88619173e-01 -4.87979680e-01 -6.81565180e-02 3.45138699e-01
7.22026944e-01 2.99706936e-01 9.20309424e-01 2.29429215e-01
-5.49126625e-01 -1.64875779e-02 -6.08552217e-01 7.14025497e-01
4.51835036e-01 -3.58463451e-02 -8.97966266e-01 -3.05129588e-01
6.19456947e-01 -3.30064036e-02 1.19408655e+00 7.37191021e-01
1.19510615e+00 -2.75314003e-01 6.30128523e-03 1.17073560e+00
1.88765848e+00 1.90201104e-01 9.76933777e-01 5.34619987e-01
5.82439363e-01 4.58736956e-01 7.97509730e-01 1.88320577e-01
-8.36690143e-02 4.91715163e-01 4.65895236e-01 -5.68288684e-01
-4.50697690e-01 -8.21707100e-02 1.31829426e-01 3.45106035e-01
-3.80340904e-01 1.73863918e-01 -2.95147032e-01 5.38171411e-01
-1.57546842e+00 -1.08091068e+00 -2.34659344e-01 1.88784635e+00
9.50695336e-01 -2.58618176e-01 -2.51199931e-01 4.12222236e-01
1.04904199e+00 6.67672276e-01 -4.40734088e-01 2.54990738e-02
-6.57668173e-01 3.97954464e-01 6.33755028e-01 4.18195188e-01
-1.43735254e+00 5.45117676e-01 4.65343380e+00 1.12551820e+00
-1.36913431e+00 9.46229231e-03 3.53135169e-01 3.66566598e-01
-6.94556609e-02 1.02599617e-02 -2.63568103e-01 3.95938575e-01
4.47932214e-01 5.18817315e-03 8.01341653e-01 4.49701458e-01
1.27497986e-01 -2.25054666e-01 -1.72164574e-01 1.06077170e+00
-1.41173983e-02 -9.83227253e-01 2.45320015e-02 -2.96426326e-01
7.67683685e-01 2.78868917e-02 8.72737169e-02 -2.51148939e-02
-1.36668712e-01 -8.54836702e-01 4.92882848e-01 7.82775044e-01
8.36167157e-01 -9.06888425e-01 7.92558134e-01 3.72475535e-01
-1.51046169e+00 -3.86027396e-01 -5.93030870e-01 1.92010656e-01
1.11287534e-01 9.65566695e-01 8.40286687e-02 1.42725921e+00
9.77272451e-01 1.04928052e+00 -2.99725056e-01 1.06750405e+00
-3.26620698e-01 2.95524836e-01 -3.62684846e-01 7.53791273e-01
1.31934255e-01 -6.65834844e-01 5.86879730e-01 9.96651888e-01
5.78144372e-01 5.55697978e-01 -2.27381545e-03 1.05240273e+00
2.97066182e-01 -1.19858421e-01 -1.93668649e-01 1.07588237e-02
1.60577595e-01 1.53335917e+00 -2.40081072e-01 -1.49809286e-01
-4.75740880e-01 8.20652127e-01 -2.33553246e-01 5.04899025e-01
-7.67779648e-01 -5.82243264e-01 4.18153822e-01 -2.24443123e-01
5.37508130e-01 1.04349162e-02 -1.68505624e-01 -1.14280736e+00
6.89115003e-02 -8.40876579e-01 4.17772889e-01 -1.02071047e+00
-1.20538259e+00 4.79109049e-01 -8.63837972e-02 -1.38482630e+00
4.43050295e-01 -3.18006575e-01 -5.93899429e-01 1.34598076e+00
-2.27792478e+00 -1.46553349e+00 -7.83898950e-01 6.91955149e-01
4.54677969e-01 2.06969157e-01 5.32960653e-01 3.41325194e-01
-3.49757373e-01 -1.49199516e-01 1.18749566e-01 -2.18118712e-01
5.16420007e-01 -7.46275663e-01 -1.38423815e-01 1.10874939e+00
-3.67023498e-01 -8.38482976e-02 6.61782801e-01 -6.25207067e-01
-1.02929497e+00 -1.25613558e+00 6.06927991e-01 5.17203331e-01
3.54053885e-01 3.37034017e-01 -1.21001875e+00 3.29580098e-01
5.62966280e-02 2.29321256e-01 1.69275060e-01 -5.54375589e-01
-2.65475273e-01 -5.65866351e-01 -1.24449134e+00 2.04303086e-01
6.11970127e-01 -5.33205509e-01 -8.25071871e-01 1.45563588e-01
7.58033454e-01 -2.67998159e-01 -9.20659065e-01 1.01585352e+00
3.85079563e-01 -1.28096128e+00 1.20475245e+00 1.95525274e-01
6.82551444e-01 -6.18297458e-01 -3.06152999e-01 -1.39523196e+00
-7.68134475e-01 -2.01388597e-01 1.48070604e-01 1.07106960e+00
-2.11575478e-01 -6.28915370e-01 1.22444443e-01 -2.64869124e-01
-1.90261275e-01 -2.95703590e-01 -7.19083726e-01 -4.71465677e-01
-4.55260351e-02 3.01804971e-02 7.59214282e-01 1.13073754e+00
-4.82586354e-01 4.34188396e-02 -4.38388675e-01 7.41129339e-01
6.50657475e-01 6.05427206e-01 2.32690856e-01 -1.14969361e+00
-1.34730607e-01 -4.30734873e-01 8.12180713e-02 -7.74609983e-01
-2.05876544e-01 -5.21866500e-01 -7.28042237e-03 -1.52667975e+00
1.51117155e-02 -1.97401017e-01 -3.69112492e-01 2.52649695e-01
-3.72292280e-01 1.88056126e-01 1.51841626e-01 2.90841043e-01
2.65902489e-01 9.14954662e-01 1.56891084e+00 -2.73365349e-01
-1.04658209e-01 -5.48729710e-02 -4.25791949e-01 6.87954545e-01
7.17073679e-01 -8.98351520e-02 -1.40937150e-01 -3.89164001e-01
-1.45455360e-01 3.37843567e-01 6.05673552e-01 -1.21843803e+00
1.29371837e-01 -2.04443917e-01 8.35414231e-01 -8.41417253e-01
3.58071804e-01 -1.04825830e+00 4.20722127e-01 4.28055584e-01
6.99111372e-02 -5.61998487e-01 3.21676046e-01 3.47507268e-01
-6.47473514e-01 8.56670737e-02 1.26535141e+00 -7.65261799e-02
-8.67111206e-01 3.14515233e-01 7.26676509e-02 -5.78279734e-01
8.21171761e-01 -2.32807800e-01 -5.58721602e-01 -1.14911370e-01
-3.98133934e-01 5.55175217e-03 1.94344014e-01 2.83451706e-01
6.76629245e-01 -1.22332573e+00 -8.70457768e-01 5.75148463e-01
-1.69222638e-01 7.76589438e-02 1.04589546e+00 8.00420523e-01
-5.98609865e-01 2.32252151e-01 -6.80868745e-01 -2.76176214e-01
-8.74433756e-01 5.58409333e-01 6.02865458e-01 -2.08438843e-01
-7.74261653e-01 6.01433694e-01 4.06146258e-01 -3.22187215e-01
-3.11225712e-01 -2.09133983e-01 -6.22681081e-01 4.28806171e-02
7.32367575e-01 4.81553555e-01 9.36730057e-02 -8.17454100e-01
5.82732782e-02 9.48335886e-01 2.13158220e-01 2.96630770e-01
1.75123382e+00 -2.38736123e-01 -4.57844675e-01 -6.28091469e-02
1.00981057e+00 -4.68911529e-02 -1.39425886e+00 -4.68776792e-01
-5.21830738e-01 -7.12520301e-01 5.26649296e-01 -8.35940659e-01
-1.41818941e+00 8.18433881e-01 9.27549243e-01 3.09564322e-01
1.87807560e+00 -4.08622265e-01 9.29350019e-01 -9.36073214e-02
-1.08403645e-01 -1.05655968e+00 -3.00245225e-01 1.26465186e-01
1.01030159e+00 -9.91779029e-01 1.51243284e-01 -6.85497165e-01
-1.62276447e-01 1.47436070e+00 4.75074083e-01 -3.13165724e-01
6.63731396e-01 -1.37511358e-01 -4.98971455e-02 -6.41633570e-02
-4.77138907e-02 -2.09549189e-01 4.26343858e-01 5.67638218e-01
4.75913621e-02 -5.38303182e-02 -4.32743400e-01 5.51115513e-01
-1.93018124e-01 1.96515664e-01 4.19186115e-01 4.76581037e-01
-6.57857060e-01 -6.49662018e-01 -7.31279731e-01 3.53002965e-01
-3.47049832e-01 -1.58552125e-01 2.89382398e-01 7.74713397e-01
7.35087931e-01 1.03551185e+00 -5.72040770e-03 -1.80857778e-01
3.75600278e-01 -3.11448008e-01 2.40782961e-01 -8.15741122e-02
-3.29228759e-01 4.09374148e-01 -2.60750473e-01 -4.57831651e-01
-8.03198516e-01 -3.53215694e-01 -8.07156563e-01 -2.09569350e-01
-5.18325388e-01 -5.36423214e-02 4.76431519e-01 7.75616646e-01
-1.19601443e-01 6.74885094e-01 8.42829764e-01 -9.36143756e-01
-4.53957140e-01 -1.09885311e+00 -1.06149948e+00 3.77590388e-01
7.00671554e-01 -4.71188605e-01 -4.37653184e-01 -2.53018606e-02] | [10.193730354309082, -1.9067124128341675] |
b5ae286c-1ff7-46d5-b286-1baa76bf00ed | structure-preserving-guided-retinal-image | 1805.06625 | null | http://arxiv.org/abs/1805.06625v2 | http://arxiv.org/pdf/1805.06625v2.pdf | Structure-preserving Guided Retinal Image Filtering and Its Application for Optic Disc Analysis | Retinal fundus photographs have been used in the diagnosis of many ocular
diseases such as glaucoma, pathological myopia, age-related macular
degeneration and diabetic retinopathy. With the development of computer
science, computer aided diagnosis has been developed to process and analyse the
retinal images automatically. One of the challenges in the analysis is that the
quality of the retinal image is often degraded. For example, a cataract in
human lens will attenuate the retinal image, just as a cloudy camera lens which
reduces the quality of a photograph. It often obscures the details in the
retinal images and posts challenges in retinal image processing and analysing
tasks. In this paper, we approximate the degradation of the retinal images as a
combination of human-lens attenuation and scattering. A novel
structure-preserving guided retinal image filtering (SGRIF) is then proposed to
restore images based on the attenuation and scattering model. The proposed
SGRIF consists of a step of global structure transferring and a step of global
edge-preserving smoothing. Our results show that the proposed SGRIF method is
able to improve the contrast of retinal images, measured by histogram flatness
measure, histogram spread and variability of local luminosity. In addition, we
further explored the benefits of SGRIF for subsequent retinal image processing
and analysing tasks. In the two applications of deep learning based optic cup
segmentation and sparse learning based cup-to-disc ratio (CDR) computation, our
results show that we are able to achieve more accurate optic cup segmentation
and CDR measurements from images processed by SGRIF. | ['Jiang Liu', 'Huazhu Fu', 'Zhengguo Li', 'Zaiwang Gu', 'Damon Wing Kee Wong', 'Jun Cheng'] | 2018-05-17 | null | null | null | null | ['optic-cup-segmentation'] | ['medical'] | [ 4.37645346e-01 -1.83017269e-01 5.60060978e-01 -2.15484217e-01
-1.18743181e-01 -7.67637640e-02 3.22432876e-01 -1.87638979e-02
-4.43574101e-01 6.76478744e-01 2.80433357e-01 -2.45605141e-01
-3.25088412e-01 -6.73738360e-01 -4.33734596e-01 -8.55920792e-01
6.56457022e-02 1.21987481e-02 4.75741625e-01 1.52861521e-01
5.08609533e-01 7.79063880e-01 -2.11259866e+00 3.26695859e-01
1.25155246e+00 1.04451895e+00 2.71682024e-01 7.54559338e-01
1.23907574e-01 5.09854853e-01 -2.84079790e-01 -1.29917100e-01
4.03023988e-01 -6.40099585e-01 -5.65520942e-01 4.67897713e-01
6.84389114e-01 -4.02483672e-01 1.60866638e-03 1.29851770e+00
6.60667539e-01 6.34996779e-03 6.78568244e-01 -3.36488932e-01
-2.04368070e-01 -3.33444566e-01 -6.41825974e-01 3.29482436e-01
-3.79299857e-02 2.27527425e-01 1.84417162e-02 -6.04307294e-01
5.66485882e-01 1.06382942e+00 4.05918151e-01 7.46502802e-02
-1.19308162e+00 -1.37819186e-01 -5.70598662e-01 4.22893882e-01
-1.01622927e+00 -4.65686440e-01 2.98252583e-01 -6.30367994e-01
7.20260382e-01 2.69785136e-01 1.06344879e+00 -3.17255050e-01
5.90735495e-01 8.32448080e-02 1.79326451e+00 -7.49626815e-01
3.77500802e-02 -2.27615144e-02 1.44624248e-01 6.31945372e-01
5.89949906e-01 4.44492489e-01 2.32873764e-02 1.77905887e-01
7.57624805e-01 -1.18591294e-01 -4.78455484e-01 8.51077363e-02
-6.99355304e-01 4.01098669e-01 4.64265823e-01 1.51403427e-01
-3.74261886e-01 -1.87016115e-01 1.26029611e-01 9.85650942e-02
5.72951436e-01 5.09008527e-01 4.81973849e-02 2.44763732e-01
-8.85082424e-01 1.34147838e-01 2.64006585e-01 1.74487740e-01
6.75772250e-01 -3.36036563e-01 -2.39300489e-01 9.13056731e-01
4.53097999e-01 4.16139901e-01 5.13523579e-01 -1.18020558e+00
-3.27649534e-01 8.73781323e-01 1.35308832e-01 -5.49016654e-01
-3.63013297e-01 -4.68140662e-01 -8.52764189e-01 1.08326781e+00
6.62847281e-01 -4.83293989e-04 -1.18430579e+00 6.96177185e-01
4.22628731e-01 1.67227432e-01 -3.03837925e-01 1.30946624e+00
5.42085290e-01 4.20975775e-01 -2.14657694e-01 -5.28564036e-01
1.38501942e+00 -6.05570853e-01 -7.46059000e-01 -6.59193322e-02
3.78641546e-01 -1.25493765e+00 7.01289415e-01 5.22686958e-01
-1.40314186e+00 -5.92359900e-01 -8.02708387e-01 -4.20541644e-01
1.33019939e-01 5.72827756e-01 3.99474710e-01 4.76208359e-01
-1.22860062e+00 5.58083057e-01 -7.36194134e-01 -3.31714064e-01
5.38227916e-01 2.37220287e-01 -3.39670837e-01 -5.65441310e-01
-3.18090916e-01 1.08749676e+00 1.03082448e-01 1.68118864e-01
-1.20881267e-01 -7.09475338e-01 -6.53661311e-01 -4.38465290e-02
-1.46990165e-01 -9.70879197e-01 6.95498407e-01 -8.67459953e-01
-1.48332667e+00 1.17133164e+00 -5.88979065e-01 -6.72697306e-01
2.92408913e-01 -2.33409017e-01 -1.73099726e-01 3.33464265e-01
-2.80497581e-01 1.06639951e-01 1.05022764e+00 -9.86076474e-01
-8.18372011e-01 -7.46117592e-01 -4.03254390e-01 1.01030491e-01
4.72279519e-01 4.14162219e-01 -3.25034261e-01 -8.43867660e-02
1.73608884e-01 -6.33306563e-01 -1.86060086e-01 2.12471589e-01
-1.64782345e-01 1.47709385e-01 5.45067847e-01 -8.86370122e-01
1.04211187e+00 -2.15354013e+00 -2.83551991e-01 7.41180032e-02
4.58963364e-01 9.45038140e-01 -5.72742261e-02 -8.80091414e-02
-6.98746694e-03 -2.54324645e-01 -2.04913169e-01 -2.93294247e-03
-7.40227282e-01 -6.24089465e-02 1.11451007e-01 7.25060642e-01
2.62330353e-01 6.08492196e-01 -5.69011569e-01 -2.91078180e-01
3.75677526e-01 9.11670923e-01 -3.12344700e-01 -4.44810651e-02
-9.00200382e-03 4.14469272e-01 1.82866037e-01 4.57876176e-01
9.10066605e-01 3.78490947e-02 -9.74924490e-02 -4.93982971e-01
-6.23957515e-01 -3.41779999e-02 -7.84697175e-01 9.16575849e-01
-2.67916113e-01 9.87477064e-01 -9.97983366e-02 -5.27979314e-01
9.17168260e-01 -6.99647889e-02 7.03208670e-02 -8.33178401e-01
3.15477610e-01 4.25267965e-01 4.23046380e-01 -6.61544740e-01
1.18136685e-02 -2.97534078e-01 1.08631992e+00 1.69844106e-01
-3.21274012e-01 -3.64121348e-02 1.87309816e-01 -2.85036176e-01
6.90752506e-01 -3.01091857e-02 3.83550078e-01 -1.11518763e-01
8.15445781e-01 -2.55949378e-01 1.33746907e-01 8.46372768e-02
-1.41595274e-01 8.90259802e-01 5.57652414e-01 -5.38135231e-01
-1.22880590e+00 -8.23133469e-01 -5.17396688e-01 -9.66323316e-02
4.81984913e-02 -2.26275120e-02 -8.18863094e-01 4.13286090e-02
-9.85665843e-02 2.82760680e-01 -1.78621948e-01 1.63562465e-02
-2.81150550e-01 -1.02957022e+00 -1.26856938e-01 -1.01407357e-01
7.04360962e-01 -6.88500047e-01 -7.60947883e-01 2.36487836e-02
5.95695376e-02 -7.95117080e-01 -7.52995536e-02 -6.12867355e-01
-1.09643006e+00 -1.38225162e+00 -8.01181734e-01 -7.43831873e-01
9.38457847e-01 4.79248405e-01 7.12187231e-01 3.43901455e-01
-1.04061759e+00 4.35290337e-02 -9.40548107e-02 -4.88564909e-01
-3.89023155e-01 -7.48862088e-01 -2.46353969e-01 4.99843210e-01
2.89127648e-01 -3.95318508e-01 -1.24080729e+00 2.15457901e-01
-7.07904160e-01 -1.09365605e-01 9.89131331e-01 6.45297885e-01
8.85516882e-01 5.96665204e-01 -4.39372053e-03 -7.42231309e-01
5.97869873e-01 2.17421934e-01 -1.04155171e+00 -1.44110978e-01
-9.41312194e-01 -2.48510644e-01 2.83476748e-02 -7.32864961e-02
-1.16123629e+00 -1.04560286e-01 1.17517829e-01 -1.37731537e-01
-4.33344692e-01 4.46056366e-01 1.58092722e-01 -5.44472575e-01
8.85129035e-01 6.45454004e-02 7.80717611e-01 -5.68594813e-01
-1.51340052e-01 8.26607049e-01 4.91517723e-01 3.28523159e-01
3.30331206e-01 8.15061927e-01 5.47457457e-01 -1.25934684e+00
-6.45400345e-01 -8.29860270e-01 -3.23848754e-01 -4.39383745e-01
8.91247571e-01 -7.77534246e-01 -8.68140459e-01 1.08342278e+00
-9.33355212e-01 -3.12737882e-01 -2.59018362e-01 7.83313394e-01
-1.97620407e-01 6.37247980e-01 -4.24741089e-01 -8.31072927e-01
-2.36207187e-01 -1.19959676e+00 6.41582966e-01 6.47166550e-01
4.22494411e-01 -8.27063203e-01 4.94777672e-02 6.94474638e-01
4.60092008e-01 1.83182836e-01 1.12881899e+00 3.80697161e-01
-8.59052002e-01 -1.15483455e-01 -6.73762858e-01 9.27897096e-01
1.00834526e-01 2.16032282e-01 -9.87131000e-01 -1.63186073e-01
1.78657472e-01 2.51061320e-01 1.07599223e+00 1.19102621e+00
6.91428959e-01 -9.21925455e-02 -9.23691541e-02 6.80404842e-01
1.79069686e+00 7.13176578e-02 1.60188603e+00 3.27922225e-01
3.41492176e-01 1.02577150e+00 6.85274839e-01 9.60143507e-02
-1.37808934e-01 6.51462257e-01 4.04232323e-01 -5.60202777e-01
-9.78968799e-01 4.63563651e-01 1.37747899e-01 2.22282633e-01
-7.02297032e-01 2.75170654e-01 -7.11899936e-01 7.18899786e-01
-1.42091000e+00 -7.36109197e-01 -9.25947964e-01 2.63312387e+00
8.16470027e-01 -2.70742297e-01 -2.89995642e-03 1.90528899e-01
5.64573050e-01 -6.74336433e-01 -2.89059848e-01 -2.48001397e-01
-2.00835988e-01 6.50307059e-01 6.20891511e-01 9.27556992e-01
-8.41760635e-01 6.30783379e-01 5.45263195e+00 5.04946411e-01
-1.36554384e+00 -1.19412221e-01 5.79079270e-01 -1.40878931e-01
1.63519487e-01 1.66935064e-02 -6.54827714e-01 5.77221513e-01
6.87649608e-01 5.23891114e-03 4.87387985e-01 -6.64958032e-04
7.33688176e-01 -8.59538972e-01 -4.33529407e-01 1.01524162e+00
6.03510533e-03 -1.10030174e+00 1.23354234e-02 3.89569521e-01
5.97175777e-01 1.40446424e-02 5.69859743e-02 -7.71567404e-01
-4.20274138e-01 -1.23458290e+00 -2.15895861e-01 1.16955662e+00
8.84964347e-01 -6.88243330e-01 1.11518610e+00 -1.43194035e-01
-4.63955939e-01 -2.90134456e-03 -5.27939081e-01 -2.10279822e-01
-6.93031400e-02 9.43036795e-01 -8.17627549e-01 1.93481743e-01
7.40264356e-01 5.99698544e-01 -8.12609792e-01 2.23601699e+00
-1.72853008e-01 4.14562583e-01 -3.85973305e-02 4.41103131e-01
-2.66197741e-01 -8.59476864e-01 8.49393129e-01 6.04009688e-01
4.62574750e-01 1.93940565e-01 -5.51629961e-01 8.76881182e-01
4.30136353e-01 1.86904162e-01 -2.60267824e-01 1.32328570e-01
-3.50409038e-02 1.19670463e+00 -6.65445149e-01 1.26454309e-01
-4.10254896e-01 5.43749750e-01 -4.55531478e-01 5.97818434e-01
-4.96425070e-02 -3.56058866e-01 7.99924552e-01 9.22012269e-01
7.19448552e-02 -6.01745471e-02 -2.87216425e-01 -6.10704303e-01
2.62464844e-02 -5.86108387e-01 -8.42830017e-02 -1.13210964e+00
-8.76468360e-01 3.53247076e-01 -3.75067294e-01 -1.12622762e+00
2.64705032e-01 -8.68550181e-01 -5.07693648e-01 1.36543548e+00
-1.99809492e+00 -1.04907322e+00 -3.70981008e-01 6.01979494e-01
1.09684490e-01 -1.67448640e-01 5.10870457e-01 2.61568666e-01
-2.76499867e-01 -1.19581550e-01 1.83177769e-01 -3.23342830e-01
8.05671036e-01 -1.20716083e+00 -1.99910432e-01 1.13531804e+00
-3.99198741e-01 4.19457614e-01 6.75813973e-01 -6.40330553e-01
-8.70557070e-01 -1.00454617e+00 8.95049870e-01 1.01821803e-01
2.64039576e-01 4.23807204e-01 -9.42310452e-01 2.32194439e-01
2.04607472e-01 1.52701139e-01 5.03999710e-01 -3.73517692e-01
2.26754189e-01 -3.95813763e-01 -1.05866456e+00 3.86196136e-01
3.41334820e-01 -1.94946021e-01 -3.97801816e-01 3.19892168e-01
1.21162087e-01 -3.15003723e-01 -7.24127412e-01 3.12802970e-01
5.30792952e-01 -1.52308071e+00 8.55611324e-01 -3.13410461e-01
2.55313933e-01 -6.80940747e-01 3.84047657e-01 -1.09499764e+00
-2.01156452e-01 -6.35720909e-01 2.09546134e-01 7.89298713e-01
4.91483770e-02 -9.89859164e-01 5.04882872e-01 3.91113997e-01
-2.43095919e-01 -4.74282950e-01 -6.62817061e-01 -2.28061870e-01
-2.52857834e-01 1.17110252e-01 -3.45849767e-02 3.12395245e-01
-5.50949097e-01 -3.38811874e-02 8.23996812e-02 4.16953743e-01
8.92774820e-01 7.28702098e-02 6.47451222e-01 -1.70591605e+00
-1.03175675e-03 -2.19896436e-01 -9.92452264e-01 -4.13540930e-01
-4.33449268e-01 -4.67786193e-01 -4.04845297e-01 -1.84902000e+00
7.74722770e-02 -2.01349989e-01 2.87135113e-02 -1.65726796e-01
-2.83815786e-02 5.47812402e-01 -3.46272625e-02 3.42176229e-01
1.66935787e-01 -1.44297257e-02 1.79583943e+00 1.30348474e-01
-6.47895873e-01 4.18590933e-01 -4.11026329e-01 6.85452878e-01
6.69298589e-01 4.45993952e-02 -4.83993083e-01 -1.23682417e-01
3.82277429e-01 -1.79361656e-01 9.31133509e-01 -1.13633406e+00
3.04228932e-01 4.35691267e-01 4.17461306e-01 -3.71663123e-01
1.27309650e-01 -5.78831494e-01 1.66083008e-01 4.20053899e-01
2.10181296e-01 -9.59865689e-01 2.04134300e-01 4.23970908e-01
-4.52691346e-01 -1.81698412e-01 1.34901905e+00 -2.15049416e-01
-3.45974237e-01 -5.93013018e-02 -5.56622744e-01 -2.75943667e-01
8.95813167e-01 -5.33973157e-01 -6.97810948e-01 -5.17024286e-02
-7.20794201e-01 -2.40166336e-01 7.21137404e-01 -3.40209305e-01
1.04327679e+00 -6.87515140e-01 -9.40782189e-01 4.51225519e-01
-1.39423743e-01 7.97917992e-02 5.54571033e-01 1.61287498e+00
-1.14230776e+00 2.95952767e-01 -5.51591218e-01 -5.96058428e-01
-1.85998094e+00 1.59745440e-01 8.57814908e-01 1.43810630e-01
-7.68534064e-01 6.52088463e-01 1.29586980e-01 5.15848517e-01
3.75845954e-02 -6.91931546e-01 -5.22022307e-01 -1.47935331e-01
9.48439121e-01 6.93936825e-01 3.40089083e-01 -5.73061526e-01
2.88604647e-01 1.08852553e+00 -2.38082841e-01 2.81986713e-01
1.22996640e+00 -5.74504316e-01 -8.16972792e-01 1.11317754e-01
6.49161100e-01 1.84476256e-01 -1.09267640e+00 -7.24536628e-02
-3.37885529e-01 -7.89107859e-01 8.75737846e-01 -1.05032098e+00
-8.45902741e-01 9.66338336e-01 1.16941369e+00 2.60968804e-01
1.62180662e+00 -2.34137207e-01 7.78338611e-01 -2.24688590e-01
9.12742838e-02 -7.45419145e-01 -4.43907946e-01 1.40137281e-02
7.49184132e-01 -1.08161175e+00 2.51618028e-01 -8.17560256e-01
-2.44622707e-01 1.17279577e+00 6.02807105e-03 -9.10473317e-02
6.36043787e-01 -1.19905375e-01 2.62338698e-01 -2.70414442e-01
-2.03971595e-01 -9.01296020e-01 7.75071025e-01 8.63033712e-01
4.19032663e-01 -2.14032501e-01 -7.14468360e-01 2.48005353e-02
2.95350179e-02 3.20206851e-01 8.86520565e-01 2.61322856e-01
-8.08662176e-01 -9.61140037e-01 -4.39075768e-01 6.73062384e-01
-6.59859598e-01 -1.16009273e-01 -1.68916151e-01 5.58363140e-01
3.89356762e-01 9.01558876e-01 1.97497398e-01 2.67408937e-01
1.06542967e-01 -1.05582118e-01 7.52188385e-01 -6.37544513e-01
-1.85316339e-01 5.11216164e-01 5.38580231e-02 -6.11278176e-01
-7.40209877e-01 -4.99425381e-01 -9.35918272e-01 7.75352567e-02
-1.99321672e-01 -4.45453793e-01 7.56439149e-01 5.96889317e-01
3.75412524e-01 4.15264100e-01 2.84310549e-01 -5.76249361e-01
1.69278502e-01 -9.32824194e-01 -1.16867459e+00 1.96691424e-01
6.43053234e-01 -5.76442719e-01 -4.96250689e-01 2.47769967e-01] | [15.795419692993164, -3.982178211212158] |
f7a289ea-d67e-4a7e-96a6-e5623d41509a | quinoa-a-q-function-you-infer-normalized-over | 1911.01831 | null | https://arxiv.org/abs/1911.01831v1 | https://arxiv.org/pdf/1911.01831v1.pdf | Quinoa: a Q-function You Infer Normalized Over Actions | We present an algorithm for learning an approximate action-value soft Q-function in the relative entropy regularised reinforcement learning setting, for which an optimal improved policy can be recovered in closed form. We use recent advances in normalising flows for parametrising the policy together with a learned value-function; and show how this combination can be used to implicitly represent Q-values of an arbitrary policy in continuous action space. Using simple temporal difference learning on the Q-values then leads to a unified objective for policy and value learning. We show how this approach considerably simplifies standard Actor-Critic off-policy algorithms, removing the need for a policy optimisation step. We perform experiments on a range of established reinforcement learning benchmarks, demonstrating that our approach allows for complex, multimodal policy distributions in continuous action spaces, while keeping the process of sampling from the policy both fast and exact. | ['Abbas Abdolmaleki', 'Martin Riedmiller', 'Nicolas Heess', 'Jost Tobias Springenberg', 'Jonas Degrave'] | 2019-11-05 | null | null | null | null | ['normalising-flows'] | ['methodology'] | [ 1.39322296e-01 2.86234409e-01 -7.42470086e-01 1.40766576e-02
-1.11982918e+00 -8.02764535e-01 8.72937322e-01 8.95090252e-02
-8.83548200e-01 1.31912959e+00 3.61025631e-01 -4.44133282e-01
-4.60055918e-01 -4.82883126e-01 -5.37089765e-01 -1.02911592e+00
-2.66390055e-01 4.40496445e-01 1.24412648e-01 -3.42465967e-01
5.09655654e-01 3.81017834e-01 -1.19438004e+00 -1.71378851e-01
7.34236062e-01 9.16596174e-01 -1.20027043e-01 1.01585984e+00
1.47094518e-01 9.36404824e-01 -7.30919182e-01 -1.59758016e-01
7.20191658e-01 -6.92967057e-01 -8.67843628e-01 -4.60920706e-02
1.53788090e-01 -5.18580019e-01 -1.75899729e-01 9.26303923e-01
5.94237566e-01 8.79474342e-01 9.22038019e-01 -1.01411998e+00
-2.54413992e-01 4.67035234e-01 -2.77208060e-01 1.95549130e-02
2.96440929e-01 7.85135984e-01 1.14177787e+00 2.27282080e-03
5.36285162e-01 1.54625690e+00 5.29690325e-01 6.10440314e-01
-1.56646550e+00 -1.62501156e-01 1.49251178e-01 -1.28231063e-01
-6.34447694e-01 -2.25910038e-01 3.90017390e-01 -4.16667908e-01
1.05597103e+00 -7.24942386e-02 1.14412689e+00 1.05421853e+00
3.98280472e-01 7.01743960e-01 1.43902516e+00 -3.14249545e-01
7.46016324e-01 -9.07208174e-02 -6.89890385e-01 6.27978742e-01
-1.68153659e-01 8.39235842e-01 8.10658261e-02 -2.95292228e-01
9.92109716e-01 -2.36649439e-01 7.77806574e-03 -9.13037896e-01
-1.15596497e+00 1.11584628e+00 3.29305232e-01 -6.73737749e-02
-4.63343978e-01 6.73202157e-01 7.59516537e-01 5.88464916e-01
3.54875743e-01 7.92744219e-01 -5.82932293e-01 -7.00971246e-01
-6.47275805e-01 8.54837239e-01 9.89399135e-01 3.92138094e-01
6.10699058e-01 3.64075691e-01 -6.56145334e-01 5.57126343e-01
1.31513461e-01 5.03552318e-01 3.74747127e-01 -1.91077244e+00
3.15727502e-01 3.06139048e-02 6.88114345e-01 -1.82986647e-01
-4.63827662e-02 -1.64550319e-01 -2.18619332e-01 1.03554428e+00
8.84063780e-01 -5.53076506e-01 -8.25330436e-01 1.67010248e+00
4.00874496e-01 1.55886561e-01 1.48639932e-01 5.98565400e-01
-3.60915869e-01 5.05466700e-01 2.73650348e-01 -4.63384211e-01
8.13832104e-01 -8.38987291e-01 -5.61755776e-01 -2.87680272e-02
7.26734698e-01 -3.43154907e-01 1.03651154e+00 3.46223056e-01
-1.24892557e+00 -2.14442387e-01 -7.87398577e-01 2.96207279e-01
-2.73410589e-01 -5.26323676e-01 4.74453270e-01 4.42664832e-01
-1.09319186e+00 1.39662695e+00 -7.31079519e-01 7.75313517e-03
6.29061460e-01 5.39958596e-01 -3.12477481e-02 5.31286895e-01
-1.11279106e+00 1.48667073e+00 8.63621354e-01 -4.60913777e-01
-1.26244462e+00 -7.70062745e-01 -7.74876595e-01 -1.10992461e-01
7.57655859e-01 -6.17137492e-01 1.95638788e+00 -1.17117274e+00
-2.45405507e+00 1.50847808e-01 3.94703597e-01 -7.17177808e-01
7.97820091e-01 -1.57811190e-03 -1.30275106e-02 6.18115589e-02
-1.32949173e-01 6.63885057e-01 1.21343946e+00 -8.66340518e-01
-7.15983152e-01 1.70729652e-01 3.33119154e-01 6.15106881e-01
1.54293969e-01 -2.31865853e-01 4.21281844e-01 -4.77000803e-01
-9.04663682e-01 -8.50384831e-01 -6.96253717e-01 1.94124699e-01
1.37461022e-01 -4.28511441e-01 5.90703785e-01 -4.65765804e-01
1.04670513e+00 -1.66178000e+00 4.31864470e-01 3.91711891e-01
-1.72853962e-01 4.00426269e-01 -1.17085867e-01 3.90788078e-01
-2.53164340e-02 -7.86611624e-03 -6.73707664e-01 1.31014615e-01
2.98182368e-01 5.53686082e-01 -4.28643912e-01 6.99966073e-01
2.54219174e-01 1.18534064e+00 -1.41389930e+00 -2.81418473e-01
4.35387731e-01 2.53926575e-01 -7.67457485e-01 2.97317117e-01
-6.86758816e-01 7.25699127e-01 -6.12150669e-01 7.29915574e-02
3.33744437e-02 2.74491906e-01 1.95100769e-01 6.34764671e-01
-1.06145605e-01 2.66683429e-01 -1.25222254e+00 1.61560023e+00
-5.12105703e-01 2.67715484e-01 2.66539514e-01 -1.35062814e+00
8.16156089e-01 2.82619685e-01 8.26615512e-01 -6.34314060e-01
2.64682472e-01 1.96287856e-01 1.72812678e-02 -3.16900223e-01
1.39730081e-01 -6.43238902e-01 -1.09048672e-02 6.24757648e-01
2.12330863e-01 -7.03827858e-01 2.41629764e-01 -2.33532086e-01
9.84079540e-01 8.11765373e-01 6.54881954e-01 -4.06644523e-01
5.89682698e-01 -1.18854061e-01 1.76495403e-01 8.01272094e-01
-5.03609896e-01 -7.97594190e-02 9.98739064e-01 -3.55528146e-01
-1.29746735e+00 -1.01519454e+00 -7.29679242e-02 1.25862300e+00
-2.90331185e-01 -2.09206313e-01 -5.52279949e-01 -1.04035425e+00
5.14333487e-01 7.23196685e-01 -7.35072196e-01 -1.67273462e-01
-7.02217877e-01 -3.07244390e-01 2.74219990e-01 4.32775319e-01
1.15977921e-01 -1.40997672e+00 -8.69728386e-01 5.91978610e-01
5.45246482e-01 -4.98586059e-01 -5.59157968e-01 5.02381086e-01
-9.87764895e-01 -9.76070523e-01 -9.11476970e-01 -2.31089070e-01
3.03080618e-01 -6.00123346e-01 1.03353381e+00 -2.91379452e-01
5.19610345e-02 7.46337891e-01 1.96412042e-01 -2.23299369e-01
-7.56339908e-01 -3.55950110e-02 -4.57298458e-02 -3.56876582e-01
-1.84746489e-01 -3.87058496e-01 -6.63396716e-01 1.67752951e-01
-7.80504167e-01 -3.95927906e-01 2.18240112e-01 1.16327345e+00
4.38938022e-01 -3.12736601e-01 7.68074453e-01 -6.65026903e-01
1.15776253e+00 -4.57060277e-01 -1.10466290e+00 5.56342825e-02
-8.24138463e-01 9.49672461e-01 9.07840729e-01 -6.22475982e-01
-1.00222600e+00 7.27299377e-02 -2.23995581e-01 -2.56701350e-01
-1.71876367e-04 -7.37514254e-03 4.26209301e-01 -2.37333789e-01
8.37527514e-01 -1.01590402e-01 6.73759401e-01 -2.96133041e-01
9.60715473e-01 2.98682779e-01 3.52336168e-01 -1.08506584e+00
5.82557023e-01 2.59343058e-01 3.03425997e-01 -2.29834542e-01
-6.67570353e-01 -2.44903207e-01 -4.00079101e-01 -4.89935093e-02
5.15020132e-01 -4.78357822e-01 -1.14880705e+00 -8.46051890e-03
-5.66743195e-01 -1.09788299e+00 -1.22095728e+00 5.24680257e-01
-1.54718471e+00 2.11841688e-01 -3.38045925e-01 -1.01005363e+00
4.20480128e-03 -1.21687293e+00 8.50732505e-01 1.37754962e-01
-7.84683898e-02 -1.55982685e+00 7.25594580e-01 -3.94916236e-01
5.52629828e-01 3.27822208e-01 6.89704120e-01 -4.16857153e-01
-1.09427482e-01 3.33997637e-01 2.88519025e-01 6.12157583e-01
-3.76649611e-02 -4.71511036e-02 -7.20122695e-01 -5.60768843e-01
-1.88398466e-01 -6.27224505e-01 8.81884515e-01 5.92801809e-01
1.09850490e+00 -6.43772304e-01 1.11276172e-01 6.15755618e-01
1.40137506e+00 3.22771817e-01 3.92015547e-01 5.17959356e-01
3.37636858e-01 3.81757617e-01 7.21929491e-01 6.76610112e-01
-1.99097469e-01 5.67351401e-01 3.92506599e-01 2.94703782e-01
3.16136718e-01 -3.41832370e-01 7.32717395e-01 2.88772006e-02
-1.68519780e-01 2.48050556e-01 -5.21031082e-01 4.49543864e-01
-2.08988428e+00 -1.22815740e+00 7.13709116e-01 2.30433393e+00
1.32767797e+00 6.48179501e-02 7.98176050e-01 -2.27812052e-01
3.25220406e-01 2.88677990e-01 -1.14977562e+00 -9.59569514e-01
4.59688842e-01 7.01344073e-01 9.36809719e-01 8.97292495e-01
-1.19598126e+00 8.84114623e-01 7.63188887e+00 9.82038140e-01
-8.96218359e-01 -1.50948584e-01 5.07762671e-01 -3.56397986e-01
-3.76078784e-01 1.05819881e-01 -3.95478934e-01 4.00378644e-01
1.23447466e+00 -4.06924397e-01 1.05104148e+00 8.43673646e-01
3.97912800e-01 -1.43130764e-01 -8.55866909e-01 5.94598651e-01
-7.34354734e-01 -1.24614930e+00 -4.78253901e-01 1.84201837e-01
1.05257618e+00 -2.07647979e-02 8.64832327e-02 7.70498276e-01
1.00196624e+00 -1.15947998e+00 4.80153382e-01 4.24202681e-01
9.29110587e-01 -1.11041296e+00 2.48253360e-01 3.05272639e-01
-9.11141098e-01 -4.85713184e-01 -2.85719812e-01 -1.85578138e-01
-6.42727539e-02 -2.60972351e-01 -7.50838876e-01 1.77917272e-01
2.70269290e-02 9.54016209e-01 5.91302523e-03 1.00130379e+00
-4.30741012e-01 4.88511980e-01 -2.99809575e-01 -2.48027116e-01
8.80928695e-01 -5.66905022e-01 6.03666067e-01 1.00266838e+00
5.17328680e-02 -3.13441068e-01 3.44806820e-01 7.36194193e-01
1.68999657e-01 1.52709678e-01 -8.73589635e-01 -9.09722149e-02
-2.93943547e-02 1.03591657e+00 -5.08577049e-01 -3.14689517e-01
-1.00006856e-01 5.43434083e-01 5.07115245e-01 6.58325434e-01
-7.76783347e-01 -3.55071992e-01 1.00952923e+00 -1.94196254e-01
5.69860578e-01 -2.47778609e-01 1.97565481e-01 -9.74080324e-01
-3.18190247e-01 -1.10144651e+00 4.71315384e-01 -1.51648909e-01
-1.05278134e+00 -8.78894553e-02 4.55442697e-01 -1.12525713e+00
-1.13895667e+00 -6.70796633e-01 -4.64511573e-01 9.16204214e-01
-1.54682720e+00 -1.90972060e-01 6.29287362e-01 6.01063967e-01
4.14130688e-01 -3.01936597e-01 7.17610598e-01 -4.57200766e-01
-1.48986414e-01 1.90371767e-01 6.47398651e-01 -3.12780648e-01
5.33795893e-01 -1.89085853e+00 2.20326677e-01 2.66326994e-01
-1.13043018e-01 1.54384613e-01 6.11333549e-01 -3.62929553e-01
-1.19747436e+00 -7.30657756e-01 -1.56254068e-01 -3.94577235e-01
1.03308737e+00 3.25838029e-02 -7.27883577e-01 5.98839879e-01
5.02102554e-01 1.88132361e-01 4.17429879e-02 -2.39729077e-01
9.51118246e-02 -4.36971895e-03 -1.31250966e+00 7.01754272e-01
7.54826605e-01 -5.14982164e-01 -5.51980138e-01 3.04949105e-01
4.55591708e-01 -5.68878770e-01 -1.08688867e+00 2.69983374e-02
5.42239308e-01 -7.24185765e-01 1.03884041e+00 -1.13109124e+00
1.62093237e-01 -1.23212054e-01 3.44043702e-01 -1.95065224e+00
-9.91460234e-02 -1.49907887e+00 -6.72942519e-01 3.49770516e-01
-6.48967400e-02 -8.80754054e-01 4.54860717e-01 4.67550606e-01
2.45082200e-01 -1.20009589e+00 -1.09131384e+00 -8.66907299e-01
6.46340430e-01 -2.02883437e-01 4.90740031e-01 4.49078768e-01
9.48694721e-02 3.08089447e-03 -3.47428024e-01 -5.38964450e-01
6.59226596e-01 -2.52664685e-01 5.00836134e-01 -8.03608477e-01
-7.19378531e-01 -1.04997444e+00 -1.07524253e-01 -1.17197430e+00
6.84887826e-01 -8.88065994e-01 3.56695428e-02 -1.30886889e+00
-2.58890718e-01 -2.99567819e-01 -4.82081890e-01 4.30896372e-01
-2.34582588e-01 -3.99321049e-01 2.62043238e-01 -1.55470118e-01
-5.04181385e-01 9.42875504e-01 1.80991220e+00 -3.11425403e-02
-4.15459871e-01 1.85676813e-01 -5.50845444e-01 6.05760515e-01
1.02464747e+00 -5.64647138e-01 -7.60754704e-01 3.23571205e-01
1.04664236e-01 3.22960317e-01 3.18237096e-01 -6.78834677e-01
-2.32444540e-01 -8.54772747e-01 2.71954983e-01 2.48602182e-01
6.93669543e-02 -6.03892624e-01 -3.75031829e-01 6.34115696e-01
-9.12899494e-01 -5.70832938e-02 2.58920521e-01 7.60091960e-01
1.33252233e-01 -5.52444994e-01 1.09763670e+00 -3.12147975e-01
-6.26265585e-01 3.19111675e-01 -3.85108829e-01 6.92270160e-01
1.15484369e+00 4.26893942e-02 -7.83045515e-02 -6.00050807e-01
-5.94947398e-01 3.48405927e-01 3.92304659e-01 3.71582881e-02
3.08479428e-01 -1.33794034e+00 -5.56752741e-01 1.35150596e-01
-4.09784824e-01 -3.18363994e-01 -4.60239112e-01 4.96459842e-01
-3.47080439e-01 2.29931518e-01 -4.36158210e-01 -2.59217590e-01
-5.32912791e-01 4.65245634e-01 9.35956836e-01 -6.76259398e-01
-6.56758964e-01 3.56830359e-01 -3.94646049e-01 -4.71320093e-01
2.79630393e-01 -5.13526082e-01 -4.83490014e-03 -1.51515417e-02
3.71533245e-01 4.78675097e-01 -4.42806780e-01 -2.42697835e-01
1.32350907e-01 4.37086523e-01 2.76202887e-01 -7.81946242e-01
1.20873213e+00 2.83929825e-01 4.14384693e-01 3.36204886e-01
1.22072840e+00 -4.39047039e-01 -2.27011085e+00 -1.24559879e-01
1.70413107e-01 -4.92503524e-01 9.29910392e-02 -6.99302375e-01
-7.53027439e-01 5.93916357e-01 5.15298486e-01 3.21642995e-01
7.26996124e-01 -4.04842079e-01 4.77641761e-01 6.77336395e-01
2.94431925e-01 -1.53350723e+00 1.40515149e-01 7.11244702e-01
7.93631852e-01 -1.09091341e+00 2.88617671e-01 6.98940158e-01
-9.62930202e-01 1.28123569e+00 2.39207491e-01 -5.59335709e-01
4.97681886e-01 2.32980609e-01 -5.23733422e-02 2.30430886e-01
-8.86284292e-01 -4.30180550e-01 2.86506057e-01 7.58853197e-01
2.24455729e-01 1.82006821e-01 -2.50690371e-01 -4.85635132e-01
5.11641093e-02 -1.14476398e-01 2.71591485e-01 1.11250281e+00
-5.04877627e-01 -1.49762118e+00 -1.08790413e-01 5.15476346e-01
-5.57577670e-01 2.25892633e-01 1.20983496e-01 9.20737565e-01
-3.68678182e-01 5.00795901e-01 -9.43478793e-02 2.06330314e-01
2.14800954e-01 3.35808009e-01 8.52540135e-01 -4.88577873e-01
-8.66388917e-01 6.80984631e-02 2.10593898e-05 -1.03737473e+00
-3.68115216e-01 -5.52256584e-01 -1.30045283e+00 -1.13671839e-01
1.48863450e-01 1.88067481e-01 5.21621108e-01 1.01073623e+00
3.93144079e-02 4.00831491e-01 7.75231361e-01 -1.01469576e+00
-1.67412257e+00 -6.27104104e-01 -6.20815635e-01 4.89992797e-01
8.87561083e-01 -9.42143440e-01 -4.31117147e-01 -3.38432103e-01] | [4.1274919509887695, 2.217689037322998] |
a1ba4d3a-f2bc-4f64-89a9-5ada00bd6d93 | 3d-pop-an-automated-annotation-approach-to | 2303.13174 | null | https://arxiv.org/abs/2303.13174v1 | https://arxiv.org/pdf/2303.13174v1.pdf | 3D-POP - An automated annotation approach to facilitate markerless 2D-3D tracking of freely moving birds with marker-based motion capture | Recent advances in machine learning and computer vision are revolutionizing the field of animal behavior by enabling researchers to track the poses and locations of freely moving animals without any marker attachment. However, large datasets of annotated images of animals for markerless pose tracking, especially high-resolution images taken from multiple angles with accurate 3D annotations, are still scant. Here, we propose a method that uses a motion capture (mo-cap) system to obtain a large amount of annotated data on animal movement and posture (2D and 3D) in a semi-automatic manner. Our method is novel in that it extracts the 3D positions of morphological keypoints (e.g eyes, beak, tail) in reference to the positions of markers attached to the animals. Using this method, we obtained, and offer here, a new dataset - 3D-POP with approximately 300k annotated frames (4 million instances) in the form of videos having groups of one to ten freely moving birds from 4 different camera views in a 3.6m x 4.2m area. 3D-POP is the first dataset of flocking birds with accurate keypoint annotations in 2D and 3D along with bounding box and individual identities and will facilitate the development of solutions for problems of 2D to 3D markerless pose, trajectory tracking, and identification in birds. | ['Máté Nagy', 'Fumihiro Kano', 'Iain D. Couzin', 'Mathilde Delacoux', 'Junran Yang', 'Alex Hoi Hang Chan', 'Hemal Naik'] | 2023-03-23 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Naik_3D-POP_-_An_Automated_Annotation_Approach_to_Facilitate_Markerless_2D-3D_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Naik_3D-POP_-_An_Automated_Annotation_Approach_to_Facilitate_Markerless_2D-3D_CVPR_2023_paper.pdf | cvpr-2023-1 | ['pose-tracking', '3d-pose-estimation', 'multiple-object-tracking', 'animal-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-1.62769154e-01 -5.57909250e-01 1.04941335e-02 -2.97659278e-01
-1.83000848e-01 -1.12849796e+00 2.57920533e-01 2.11271778e-01
-9.55150425e-01 3.68632197e-01 -2.15510339e-01 4.16381776e-01
-7.03620389e-02 -9.11874995e-02 -8.81574512e-01 -3.55621308e-01
-5.84898889e-01 5.15646935e-01 7.02426553e-01 -2.99731418e-02
2.20121920e-01 8.32921922e-01 -1.57151890e+00 -2.92518467e-01
-1.00457139e-01 6.85915530e-01 1.28780320e-01 9.41501796e-01
5.24779379e-01 2.26433039e-01 -5.28297007e-01 -3.54300469e-01
5.73982179e-01 -1.82286471e-01 -3.65967900e-01 5.34427054e-02
9.57145393e-01 -6.33917570e-01 7.16410205e-02 7.97399282e-01
2.59857118e-01 1.90384582e-01 3.49041730e-01 -1.23953199e+00
-1.88193858e-01 1.23015702e-01 -9.51526701e-01 5.14434814e-01
5.08648276e-01 1.96958646e-01 7.72266567e-01 -4.86363888e-01
7.75402427e-01 9.09891963e-01 1.17951727e+00 7.24453986e-01
-1.18319726e+00 -4.55292672e-01 -1.56979144e-01 -1.51228644e-02
-1.52866220e+00 -2.52296805e-01 4.34036463e-01 -9.84938145e-01
8.28193665e-01 2.48931244e-01 1.15695870e+00 8.11196625e-01
1.81370266e-02 1.66639879e-01 6.95468962e-01 -1.00127399e-01
6.58186823e-02 -1.28108725e-01 -1.49880767e-01 9.15048838e-01
3.78627270e-01 2.12120801e-01 -5.20901442e-01 -4.13910866e-01
1.04048073e+00 3.56458187e-01 -2.15394437e-01 -9.56034839e-01
-1.61688697e+00 7.26196349e-01 4.43957418e-01 -3.13826889e-01
-1.20665878e-01 2.62095481e-01 2.19274461e-01 -7.76751339e-02
2.86042601e-01 6.62676871e-01 -6.03049278e-01 -2.16699108e-01
-7.41798937e-01 5.42900681e-01 6.74392879e-01 1.21093249e+00
4.91721451e-01 -1.44411832e-01 3.59390169e-01 6.52995527e-01
4.14067179e-01 7.66950846e-01 4.26923603e-01 -1.03938198e+00
1.96179718e-01 7.58957684e-01 5.05948603e-01 -1.21654260e+00
-7.48044550e-01 1.05505869e-01 -2.20397085e-01 1.46215171e-01
5.38158834e-01 -3.11116844e-01 -6.18257582e-01 1.49963200e+00
9.17105377e-01 1.39604896e-01 -3.03916961e-01 1.18476295e+00
9.58955526e-01 4.65021551e-01 -2.07826465e-01 1.86035797e-01
1.53055346e+00 -8.70839119e-01 -2.05090836e-01 -4.34257299e-01
5.08152306e-01 -5.35953343e-01 3.40543330e-01 -6.36252165e-02
-7.88630426e-01 -3.42799991e-01 -1.02436888e+00 2.25275069e-01
-4.74490881e-01 2.94673353e-01 4.55561876e-01 3.91005397e-01
-8.32953036e-01 3.91089231e-01 -1.04858756e+00 -5.10086358e-01
2.34991416e-01 6.95383549e-01 -1.04052305e+00 5.53315818e-01
-4.55618590e-01 1.00667989e+00 3.27734239e-02 3.41207296e-01
-1.06871974e+00 -6.36424541e-01 -1.12590909e+00 -6.68637514e-01
5.97290158e-01 -3.09544653e-01 1.35050416e+00 -5.28741002e-01
-1.19636989e+00 1.41006339e+00 2.26051405e-01 -5.83634436e-01
5.54468989e-01 -7.33968556e-01 5.17184399e-02 2.70756274e-01
2.05751762e-01 1.09210169e+00 8.53906453e-01 -8.59812856e-01
-9.43053842e-01 -7.72314489e-01 7.35023990e-02 1.88372821e-01
2.39358187e-01 2.60888934e-01 -4.22752380e-01 -7.02235460e-01
1.92857072e-01 -1.25741017e+00 -2.51787215e-01 5.89919806e-01
1.61649182e-01 2.30064526e-01 7.61536717e-01 -6.79787815e-01
7.79799879e-01 -2.10162282e+00 3.75169128e-01 -4.66065824e-01
2.63896435e-01 5.17533243e-01 1.56853303e-01 2.24821240e-01
1.77022666e-01 -1.21704467e-01 -1.57718211e-01 -1.21004060e-01
-2.42055491e-01 3.85219872e-01 1.19913809e-01 1.11068916e+00
-3.02548092e-02 7.27315962e-01 -1.01795137e+00 -5.57400763e-01
3.83133113e-01 2.51936257e-01 -7.74808526e-01 4.19159681e-01
2.01533958e-01 4.26454723e-01 -2.27996930e-01 9.49717939e-01
3.89994979e-01 3.15930068e-01 -7.78504834e-02 -4.66140471e-02
-4.44764614e-01 -3.75387639e-01 -1.11405551e+00 1.52691913e+00
4.08382574e-03 7.98290372e-01 3.43102455e-01 -4.83671099e-01
6.10363722e-01 5.98160252e-02 7.40586340e-01 2.07183257e-01
1.48390308e-01 1.05922241e-04 -1.60752922e-01 -5.95926106e-01
6.39946580e-01 8.80313516e-02 -3.11050653e-01 2.71975137e-02
2.78715134e-01 -2.22972065e-01 2.25652650e-01 -3.83565813e-01
8.00243616e-01 4.16054070e-01 6.27289236e-01 -9.65307355e-02
1.83498263e-01 4.21922177e-01 4.94837403e-01 4.26469803e-01
-5.21407306e-01 5.67756176e-01 -2.22457722e-02 -8.53332400e-01
-1.07697737e+00 -8.18233192e-01 -1.87778771e-01 1.10187411e+00
4.46854204e-01 -3.96441370e-01 -9.86159027e-01 -4.61374372e-01
3.21843624e-01 -3.12553585e-01 -8.35312009e-01 1.14386834e-01
-8.76929820e-01 -6.55671954e-01 5.08171380e-01 6.22569859e-01
3.48950624e-01 -8.18948567e-01 -1.53732884e+00 -4.85254265e-02
1.73999146e-01 -1.22532618e+00 -6.96763277e-01 5.69599830e-02
-6.11805737e-01 -1.31178308e+00 -5.78080416e-01 -7.23901808e-01
9.53991055e-01 5.96513391e-01 5.50822914e-01 1.54346982e-02
-6.85650408e-01 3.07499826e-01 -4.96645957e-01 -3.00187647e-01
1.05652707e-02 -3.32815409e-01 5.65125167e-01 -3.04648072e-01
2.05579624e-01 -2.11683705e-01 -4.62023556e-01 8.79801571e-01
-4.48837101e-01 -5.66521287e-02 1.72373965e-01 6.18440151e-01
7.57070065e-01 -3.88923705e-01 -1.14923619e-01 -5.48838854e-01
-2.04641208e-01 -3.69537652e-01 -1.28956604e+00 3.39491963e-02
3.02373558e-01 -3.07742655e-01 5.09532273e-01 -6.74069941e-01
-1.35962099e-01 6.41377270e-01 7.15185031e-02 -4.43499982e-01
-5.76885045e-01 3.05898059e-02 2.80375183e-01 -6.20717764e-01
5.71840584e-01 -3.56520355e-01 8.90639797e-02 -6.93587601e-01
1.81948468e-01 3.76492471e-01 1.01103199e+00 -1.03054896e-01
9.12542403e-01 5.35733461e-01 2.45440692e-01 -9.42909181e-01
-7.28273153e-01 -8.24360609e-01 -1.37369394e+00 -5.35707355e-01
1.06497133e+00 -1.04574406e+00 -8.39890420e-01 7.08666325e-01
-7.87572920e-01 -1.72477588e-01 -9.54155996e-02 8.46997380e-01
-6.14067495e-01 3.39105785e-01 -3.60570371e-01 -5.97964466e-01
-2.82841682e-01 -1.03748941e+00 1.31525087e+00 4.01721656e-01
-3.19929808e-01 -6.02945507e-01 2.24338844e-01 3.72924596e-01
-5.82757331e-02 7.18832493e-01 6.80684298e-02 -7.02746272e-01
-3.12550783e-01 -6.77733064e-01 3.54151070e-01 -3.41249667e-02
-8.47709551e-03 3.94335777e-01 -4.70503420e-01 -4.68891352e-01
-1.81143299e-01 -1.57465696e-01 2.38758340e-01 5.34456015e-01
3.61693054e-01 -2.80801743e-01 -4.88831222e-01 9.44891930e-01
1.16142333e+00 2.32928693e-01 -2.77078420e-01 3.88104081e-01
6.34589970e-01 6.62789345e-01 1.02737093e+00 4.58463252e-01
3.26256037e-01 1.23339319e+00 7.49481559e-01 2.12523416e-01
2.06043527e-01 -4.77355510e-01 2.90305883e-01 4.00373250e-01
-3.24809998e-01 6.79768920e-02 -9.11277413e-01 7.60445595e-01
-1.63382995e+00 -1.03824687e+00 -1.34852737e-01 2.45898056e+00
3.93260986e-01 -2.56314278e-01 6.93274021e-01 -1.95684433e-01
1.00669706e+00 -1.31061196e-01 -5.53932965e-01 8.35464895e-02
2.95075953e-01 -4.13961619e-01 1.04630363e+00 1.62886143e-01
-1.51602244e+00 7.44465709e-01 6.72053385e+00 8.10168087e-02
-9.53004479e-01 -3.52400392e-01 -1.73974693e-01 -2.69299418e-01
6.87209666e-01 -2.86686152e-01 -1.36611068e+00 4.18998718e-01
5.52554846e-01 2.42938638e-01 3.98837149e-01 1.17225397e+00
-1.63108587e-01 -3.84212166e-01 -9.92048740e-01 1.10154510e+00
3.71553838e-01 -1.06723201e+00 -4.76100713e-01 1.14665970e-01
7.57808924e-01 3.16278264e-02 -1.78500578e-01 -3.09968144e-01
2.92986501e-02 -6.66335702e-01 1.21577573e+00 3.17443460e-01
6.94610715e-01 -5.88622928e-01 8.22817624e-01 7.96767652e-01
-1.34581542e+00 -1.19551772e-03 -4.86315608e-01 -1.33126706e-01
-9.70035885e-03 -5.72264850e-01 -1.07618964e+00 2.51596361e-01
1.06345141e+00 8.63805413e-01 -8.98374557e-01 1.54728532e+00
1.21108644e-01 3.14412475e-01 -8.14291179e-01 -2.44949564e-01
1.13770105e-01 -1.38113216e-01 8.02043319e-01 8.80061626e-01
3.48387957e-01 2.35279113e-01 1.12466896e-02 3.51637453e-01
3.45336609e-02 -2.01668367e-01 -5.98492503e-01 1.51764423e-01
5.58073819e-01 1.48031580e+00 -1.06859815e+00 3.71287800e-02
-2.21263990e-01 4.80270118e-01 7.72433132e-02 -3.84403288e-01
-1.15844679e+00 -1.51806980e-01 7.41834760e-01 3.34207237e-01
6.16871297e-01 -5.82891524e-01 6.12648427e-01 -1.01945281e+00
2.50740610e-02 -4.21218663e-01 3.10771585e-01 -5.42051077e-01
-1.05873787e+00 4.65811729e-01 6.37684047e-01 -1.50212872e+00
-3.04001957e-01 -7.23121703e-01 -6.34470731e-02 2.74895102e-01
-8.29863489e-01 -1.23143601e+00 -3.69055033e-01 3.07281703e-01
5.08928716e-01 -7.81179070e-02 5.97707570e-01 2.57602304e-01
-3.07377160e-01 2.65558571e-01 1.50631174e-01 3.53451401e-01
5.23211777e-01 -1.13129640e+00 5.45505822e-01 8.00667226e-01
4.98513371e-01 2.72990376e-01 7.04019904e-01 -6.61148071e-01
-1.62491095e+00 -1.15190399e+00 2.46287689e-01 -1.13591325e+00
5.04753411e-01 -5.80588996e-01 -3.52892578e-01 8.58474314e-01
-5.75187445e-01 5.04675984e-01 6.86001599e-01 -5.02212882e-01
2.64663815e-01 6.32259473e-02 -1.14646208e+00 2.22777247e-01
1.08122742e+00 -4.20260765e-02 -7.82614648e-01 1.77575886e-01
3.09509724e-01 -1.00646770e+00 -7.78808534e-01 4.32293653e-01
7.56468177e-01 -6.04906082e-01 1.18976998e+00 -6.28254235e-01
8.32086951e-02 -9.01512980e-01 1.44680990e-02 -1.11088526e+00
-1.15449369e-01 -4.20469016e-01 1.60041049e-01 9.11811650e-01
-1.92373097e-02 -1.44297779e-01 5.23888886e-01 3.01890522e-01
-1.69596881e-01 -2.68062949e-01 -1.15888476e+00 -7.48713613e-01
-4.79161233e-01 2.28279643e-02 4.34468508e-01 5.84173501e-01
-3.51426214e-01 -3.49358171e-02 -7.07796752e-01 3.91775280e-01
5.28700233e-01 3.45605016e-01 1.17396891e+00 -1.35047114e+00
2.17552707e-01 -8.00729841e-02 -1.10314298e+00 -1.15924740e+00
1.81484949e-02 -5.60109735e-01 3.30071688e-01 -8.75681281e-01
-6.84301332e-02 8.12691450e-02 6.08863890e-01 5.95855355e-01
8.72996151e-02 5.94863653e-01 -7.84868225e-02 1.36426881e-01
-7.85502195e-01 -5.16892076e-02 9.53778803e-01 3.29989314e-01
1.02360234e-01 1.59044847e-01 2.58945059e-02 1.23578322e+00
3.66879821e-01 -6.23006642e-01 1.81100160e-01 -4.83335406e-01
1.73644144e-02 1.88586444e-01 8.29664886e-01 -1.24779499e+00
2.81748682e-01 -2.47461110e-01 5.22677004e-01 -7.79296517e-01
5.56529462e-01 -1.15904295e+00 6.20875239e-01 4.38929141e-01
-3.00376624e-01 4.07369018e-01 2.78847098e-01 7.45634317e-01
5.11502437e-02 -5.21956265e-01 8.91993403e-01 -5.53859711e-01
-1.01027083e+00 2.64209121e-01 -4.55210119e-01 1.77838847e-01
1.63427985e+00 -2.87536412e-01 -1.60232663e-01 2.96088874e-01
-5.45122921e-01 1.88084617e-01 8.29791248e-01 5.07838011e-01
5.68241000e-01 -1.02108502e+00 -4.94110614e-01 6.01074219e-01
2.12643862e-01 3.12046856e-01 2.73314148e-01 7.60607958e-01
-1.42071700e+00 3.74967128e-01 -6.47515595e-01 -9.00948226e-01
-1.80997550e+00 5.46782017e-01 4.35319126e-01 5.05870521e-01
-6.18658602e-01 1.10533047e+00 -9.09333304e-02 -3.85505021e-01
1.01047635e-01 -5.21009803e-01 -2.67847568e-01 2.07789317e-01
5.79723299e-01 3.42354149e-01 -1.76796347e-01 -1.29350126e+00
-5.60351610e-01 1.10682929e+00 2.99368083e-01 2.79851824e-01
1.54056633e+00 -2.12585002e-01 2.21046470e-02 3.70420605e-01
8.79675448e-01 1.40831426e-01 -1.69000089e+00 1.06690444e-01
1.78679544e-02 -1.01270735e+00 -3.57984722e-01 -2.26162985e-01
-8.70713651e-01 7.73442984e-01 5.56470096e-01 1.20750569e-01
5.02216995e-01 1.30230561e-01 6.72658503e-01 2.48679191e-01
7.10287392e-01 -9.51662302e-01 -1.57124043e-01 2.35146433e-01
7.40440071e-01 -1.31556940e+00 2.58200675e-01 3.12199909e-02
-7.20249474e-01 8.78794193e-01 8.25222611e-01 -1.69528604e-01
2.65181482e-01 2.64441013e-01 5.31213917e-02 -2.92341202e-01
-5.83371520e-01 -9.26885381e-02 4.54877764e-01 7.65908122e-01
7.73703530e-02 1.92856461e-01 2.10631832e-01 3.07259291e-01
-4.30106491e-01 -4.41305071e-01 4.59353179e-01 1.15990102e+00
-4.16481495e-01 -5.37211359e-01 -5.81342816e-01 1.35164652e-02
-5.01318157e-01 2.77015746e-01 -5.67767143e-01 9.44072664e-01
4.52126443e-01 6.59999669e-01 2.14257208e-03 -3.41383219e-01
4.98731971e-01 -3.65517765e-01 5.30345738e-01 -8.04483473e-01
-6.60770118e-01 -6.28880318e-03 -2.73286123e-02 -1.72070950e-01
-8.12794268e-01 -6.71373248e-01 -8.82578135e-01 -2.75746733e-01
-5.90052724e-01 2.11077064e-01 7.33556926e-01 5.18195271e-01
2.20718104e-02 -1.41320363e-01 3.62011433e-01 -1.21001875e+00
-3.32287073e-01 -6.17513061e-01 -5.91879487e-01 1.86843470e-01
6.17286563e-01 -1.01003921e+00 -4.74484622e-01 5.27796745e-01] | [7.6869354248046875, -0.9853897094726562] |
3905b76a-c901-4340-b903-e7cf7d6eb34c | 3d-object-reconstruction-and-6d-pose | 2203.01051 | null | https://arxiv.org/abs/2203.01051v1 | https://arxiv.org/pdf/2203.01051v1.pdf | 3D object reconstruction and 6D-pose estimation from 2D shape for robotic grasping of objects | We propose a method for 3D object reconstruction and 6D-pose estimation from 2D images that uses knowledge about object shape as the primary key. In the proposed pipeline, recognition and labeling of objects in 2D images deliver 2D segment silhouettes that are compared with the 2D silhouettes of projections obtained from various views of a 3D model representing the recognized object class. By computing transformation parameters directly from the 2D images, the number of free parameters required during the registration process is reduced, making the approach feasible. Furthermore, 3D transformations and projective geometry are employed to arrive at a full 3D reconstruction of the object in camera space using a calibrated set up. Inclusion of a second camera allows resolving remaining ambiguities. The method is quantitatively evaluated using synthetic data and tested with real data, and additional results for the well-known Linemod data set are shown. In robot experiments, successful grasping of objects demonstrates its usability in real-world environments, and, where possible, a comparison with other methods is provided. The method is applicable to scenarios where 3D object models, e.g., CAD-models or point clouds, are available and precise pixel-wise segmentation maps of 2D images can be obtained. Different from other methods, the method does not use 3D depth for training, widening the domain of application. | ['Babette Dellen', 'Florentin Wörgötter', 'Tomas Kulvicius', 'Osman Kaya', 'Marcell Wolnitza'] | 2022-03-02 | null | null | null | null | ['3d-object-reconstruction', 'object-reconstruction', 'robotic-grasping'] | ['computer-vision', 'computer-vision', 'robots'] | [ 3.67632359e-01 2.50212908e-01 1.41237438e-01 -3.77235532e-01
-4.57303941e-01 -7.59627342e-01 4.16713893e-01 1.49670750e-01
-5.39673865e-01 1.64825022e-01 -6.25467062e-01 -8.49226713e-02
-6.87195212e-02 -5.14271855e-01 -8.41322780e-01 -4.75531399e-01
3.01720612e-02 1.33571768e+00 5.92113853e-01 1.51565466e-02
5.04494548e-01 1.31052232e+00 -1.52405727e+00 -1.27734140e-01
3.32483083e-01 9.96873915e-01 7.39802003e-01 3.88326079e-01
-3.22949499e-01 -1.87151179e-01 -4.04520512e-01 -2.10753918e-01
7.26988196e-01 -1.01647647e-02 -5.12070775e-01 7.86679387e-01
5.31221092e-01 -6.00628972e-01 2.26922706e-01 8.55000257e-01
5.14299758e-02 -2.13600948e-01 7.83472240e-01 -9.29078519e-01
1.04099095e-01 3.31509076e-02 -4.91570324e-01 -6.39131010e-01
6.89898670e-01 -1.84899271e-02 5.42738676e-01 -1.00145745e+00
9.51462984e-01 1.41510212e+00 5.68953335e-01 2.95297414e-01
-1.48433006e+00 -1.11574061e-01 -7.95012340e-02 -1.74318761e-01
-1.20677578e+00 -7.22987857e-03 9.30437982e-01 -6.27021849e-01
7.83323467e-01 8.85348618e-02 7.51193762e-01 5.65880418e-01
-2.78400835e-02 5.66878796e-01 1.28882289e+00 -8.30656171e-01
2.13229924e-01 4.12308544e-01 -2.38662865e-02 5.69556057e-01
3.63723844e-01 2.87072025e-02 1.19470172e-01 5.05558252e-02
1.28067565e+00 1.67516962e-01 -1.58159584e-01 -1.43999457e+00
-1.35988653e+00 3.53658557e-01 3.65370035e-01 2.12726519e-01
-5.35374105e-01 -1.88574731e-01 1.25105247e-01 -9.40891579e-02
1.86675996e-01 3.52323920e-01 -3.19549084e-01 1.04616612e-01
-6.90818667e-01 2.18271658e-01 8.45090508e-01 1.41683388e+00
9.48393762e-01 -2.69922286e-01 5.95132113e-01 5.81748068e-01
4.28920865e-01 7.60089576e-01 1.02613218e-01 -1.19300020e+00
2.69989818e-01 9.28979099e-01 3.86366278e-01 -8.77845943e-01
-5.54643095e-01 1.24107219e-01 -2.38461643e-01 7.17872500e-01
7.07540631e-01 5.66232085e-01 -1.05938971e+00 8.47397506e-01
6.02603137e-01 -4.27026391e-01 1.15508027e-01 1.06636453e+00
4.29035276e-01 2.40451857e-01 -7.61314988e-01 -1.21568397e-01
1.33830404e+00 -4.68426585e-01 -2.53082156e-01 -3.62333179e-01
1.72203675e-01 -9.25766110e-01 6.10659957e-01 6.52941227e-01
-1.07312012e+00 -4.82406884e-01 -8.88986647e-01 1.10205663e-02
-2.30093315e-01 2.74529606e-01 3.67690176e-01 3.54688823e-01
-4.44636256e-01 5.78213751e-01 -1.10856676e+00 -7.31423140e-01
2.31564984e-01 5.78791976e-01 -7.32896388e-01 -2.13653415e-01
-2.88740098e-01 1.35158634e+00 7.57032454e-01 3.00913602e-01
-6.92672849e-01 -1.61326781e-01 -7.41784573e-01 -2.28033617e-01
4.96594816e-01 -2.07762048e-01 1.12709582e+00 -6.10859931e-01
-1.64802563e+00 1.29320014e+00 1.85331970e-01 -2.13035241e-01
8.09205115e-01 -2.62199879e-01 4.04565424e-01 5.29976904e-01
-2.29095370e-01 6.67705834e-01 8.24278176e-01 -1.78214192e+00
-2.94696838e-01 -6.70887172e-01 2.89256841e-01 3.07277679e-01
3.34788233e-01 -1.79537818e-01 -6.40225887e-01 2.29423298e-04
8.92324150e-01 -1.00117803e+00 -3.47624153e-01 2.78111488e-01
-3.62885684e-01 1.90912932e-01 9.77729321e-01 -6.29067421e-01
4.46700864e-02 -2.00629354e+00 2.91124880e-01 4.25702661e-01
-9.69725177e-02 1.26766577e-01 1.18935816e-01 4.77947414e-01
1.71372846e-01 -3.52508813e-01 -3.95805359e-01 -3.08779597e-01
-1.61413066e-02 3.84992599e-01 9.22643766e-02 7.53313422e-01
8.23344439e-02 5.38785100e-01 -5.41494489e-01 -4.37569648e-01
7.37281084e-01 4.38845098e-01 -2.64609367e-01 3.74456167e-01
-3.66606593e-01 5.67964911e-01 -5.06496012e-01 5.33772290e-01
9.16670859e-01 1.24907285e-01 1.21693596e-01 -5.18753231e-01
-2.62320817e-01 7.71758333e-03 -1.53741658e+00 1.85673356e+00
-5.04259348e-01 2.67986119e-01 3.82991731e-01 -1.09993625e+00
1.47523022e+00 2.99031109e-01 6.03323340e-01 -2.63385296e-01
3.18987608e-01 4.82092112e-01 -1.54693186e-01 -5.69704175e-01
3.56667072e-01 9.29737836e-03 1.70935825e-01 3.48147780e-01
1.86968848e-01 -1.16926169e+00 1.79779381e-01 -2.15527475e-01
3.71821791e-01 7.10944533e-01 3.72348160e-01 -1.32928252e-01
4.23342288e-01 4.09116745e-01 1.31163388e-01 2.54973650e-01
3.70544553e-01 8.38251829e-01 3.62141639e-01 -4.18095082e-01
-1.54966390e+00 -9.85617399e-01 -4.11931574e-01 -1.54987231e-01
6.76654279e-01 1.98621787e-02 -7.97426581e-01 -4.86680031e-01
1.93081304e-01 5.60404360e-01 -3.57505023e-01 3.69012386e-01
-7.23279893e-01 -2.39151612e-01 -2.50150561e-01 2.64354229e-01
1.83834121e-01 -7.60715127e-01 -1.44655561e+00 1.06242344e-01
2.05799729e-01 -1.36484647e+00 1.61860377e-01 2.66048878e-01
-1.38425231e+00 -1.37093735e+00 -7.50278354e-01 -6.31531119e-01
1.11626565e+00 3.88904512e-01 8.28711987e-01 -1.98942423e-01
-4.57622141e-01 7.69231796e-01 -4.38001603e-01 -3.04304004e-01
-5.59351861e-01 -2.85201520e-01 5.67003526e-02 -1.99885190e-01
1.94565654e-01 -4.29381996e-01 -3.40233207e-01 7.65403688e-01
-7.32984364e-01 1.50012106e-01 6.84857905e-01 8.06071281e-01
8.08034003e-01 -2.18557283e-01 1.74070876e-02 -6.62960589e-01
-3.74029316e-02 2.80703157e-01 -1.26817262e+00 1.19937815e-01
-4.29171950e-01 -1.41033530e-01 2.23400325e-01 -6.02341831e-01
-1.00743687e+00 6.93327188e-01 1.85845837e-01 -7.68596947e-01
-6.64265931e-01 1.13666855e-01 -2.05234393e-01 4.26631980e-02
6.29808486e-01 -1.46133592e-02 5.17920554e-01 -7.91257918e-01
3.32693189e-01 6.47825360e-01 5.57660758e-01 -3.77968758e-01
7.66671777e-01 6.88924253e-01 1.33623198e-01 -8.92765939e-01
-2.73327768e-01 -6.45649672e-01 -1.50458384e+00 -3.16800356e-01
7.65181005e-01 -5.78845322e-01 -7.21489966e-01 2.56403595e-01
-1.52222538e+00 -6.13383204e-02 -4.09048587e-01 9.88045037e-01
-1.02758324e+00 3.00167888e-01 -2.03894064e-01 -9.95105982e-01
4.36631590e-02 -1.55863106e+00 1.36811352e+00 -3.39206941e-02
-1.30857065e-01 -7.72002518e-01 -4.53157455e-01 3.77361029e-01
-1.15943253e-01 2.77645856e-01 7.71180570e-01 -6.17074132e-01
-9.46464658e-01 -5.88207960e-01 6.80114180e-02 4.49309647e-01
1.30493969e-01 6.37758970e-02 -8.78198743e-01 -1.19760901e-01
2.00350747e-01 -1.49815127e-01 2.54686922e-01 2.81433344e-01
5.49117744e-01 1.82004988e-01 -5.32016337e-01 2.28828579e-01
1.45740557e+00 3.16300273e-01 3.74933928e-01 3.20338696e-01
5.91956913e-01 1.01890159e+00 1.03956771e+00 1.24038361e-01
2.18099691e-02 1.02922058e+00 9.57681835e-01 -4.41101566e-02
-5.13555389e-03 -1.85347330e-02 7.70336464e-02 3.94241601e-01
-2.88263112e-01 3.05105567e-01 -9.42033350e-01 3.10906023e-01
-1.41354978e+00 -4.87823814e-01 -4.86639291e-01 2.59084082e+00
2.30248109e-01 2.67984152e-01 1.24587052e-01 2.93581605e-01
6.93118155e-01 -5.22295713e-01 -6.26952648e-01 -2.08115220e-01
7.38025457e-02 -2.22698841e-02 7.74503589e-01 5.99821091e-01
-7.64184415e-01 8.39359105e-01 5.50856638e+00 2.89613992e-01
-1.08973217e+00 -2.23057553e-01 -1.34244397e-01 3.30233514e-01
1.66255478e-02 2.34989047e-01 -7.09190309e-01 -6.21940978e-02
7.11422414e-02 1.80254370e-01 2.72482455e-01 8.96290064e-01
1.31310057e-02 -5.97040296e-01 -1.45343089e+00 1.04752290e+00
2.73736835e-01 -8.40697765e-01 -2.79934943e-01 5.24183847e-02
3.21884364e-01 -3.48122828e-02 -3.86217713e-01 -3.65428209e-01
-9.85555649e-02 -5.64700663e-01 1.04056108e+00 5.07664680e-01
6.41564667e-01 -3.51653904e-01 7.72414446e-01 7.21793950e-01
-7.10188627e-01 4.64500524e-02 -4.42840964e-01 1.69595733e-01
5.11765003e-01 5.32868445e-01 -1.56485760e+00 7.19622910e-01
6.31898701e-01 4.25124049e-01 -4.51826096e-01 1.06690788e+00
-1.94009215e-01 -8.62894463e-04 -7.97069371e-01 1.28566828e-02
-3.91391627e-02 -6.05444074e-01 7.18839765e-01 9.15911317e-01
4.14585501e-01 1.68463722e-01 8.07727203e-02 1.05684900e+00
4.02051270e-01 4.69547547e-02 -7.55174100e-01 1.38043046e-01
2.21766874e-01 1.36432493e+00 -1.28854144e+00 -5.87291382e-02
-1.62263691e-01 8.86852920e-01 -1.14618182e-01 1.10802591e-01
-3.28149348e-01 -2.11385310e-01 -1.49141019e-02 1.53887033e-01
4.71814960e-01 -6.39003873e-01 -7.90553093e-02 -9.30915833e-01
1.98320836e-01 -4.45324212e-01 -2.68010736e-01 -1.03793275e+00
-8.36014271e-01 6.95800126e-01 6.35565460e-01 -1.50907075e+00
-4.08803135e-01 -1.14575016e+00 -6.82985038e-03 8.89914572e-01
-1.11364043e+00 -1.10438573e+00 -3.60833049e-01 7.62284622e-02
5.33508420e-01 6.93182722e-02 9.12606776e-01 -1.57265738e-01
1.57702073e-01 -2.16351420e-01 1.47073627e-01 -2.46460795e-01
4.54492152e-01 -1.11993074e+00 1.42631471e-01 3.97067279e-01
2.86231399e-01 4.10662651e-01 7.87383258e-01 -4.69078809e-01
-1.61505795e+00 -4.00512010e-01 2.78948188e-01 -5.74094176e-01
1.72734946e-01 -7.40670264e-01 -9.16701078e-01 6.04825139e-01
-2.74916142e-01 -8.95653442e-02 -4.29667681e-02 -3.96125555e-01
7.23412186e-02 -7.94203728e-02 -1.36269438e+00 -5.71561605e-03
7.94919729e-01 -8.53770822e-02 -7.83831120e-01 3.10246229e-01
2.87087470e-01 -8.97675157e-01 -9.12423491e-01 4.94184732e-01
9.17956650e-01 -1.00454307e+00 1.00895560e+00 -2.21011043e-03
7.00515807e-02 -4.04334366e-01 -2.16212019e-01 -1.25076032e+00
3.16979647e-01 -2.01653168e-01 3.47200900e-01 9.55959022e-01
1.99704960e-01 -5.27039826e-01 8.59778643e-01 5.02441049e-01
-1.22268230e-01 -4.27953601e-01 -8.78014088e-01 -8.89023006e-01
-2.83590406e-01 -3.14572722e-01 3.54997128e-01 5.85412085e-01
-3.48241895e-01 1.57316968e-01 1.55109674e-01 4.31892037e-01
7.60112405e-01 6.99580669e-01 1.20537710e+00 -1.67647386e+00
-1.37118325e-01 -1.85968399e-01 -7.57180274e-01 -1.26871789e+00
1.32326350e-01 -8.12119424e-01 1.54214740e-01 -1.62191749e+00
-9.79275703e-02 -7.34136105e-01 6.81520700e-01 2.36867711e-01
5.74603617e-01 1.21768601e-01 4.30423051e-01 3.66211355e-01
1.11304164e-01 2.48247147e-01 1.36291564e+00 1.01788163e-01
-2.13893324e-01 1.88730970e-01 2.54757315e-01 1.02662313e+00
5.18140614e-01 -3.35522115e-01 -1.11064978e-01 -4.30670112e-01
-2.60180652e-01 7.46818259e-02 4.07051623e-01 -8.09853554e-01
6.19682558e-02 -1.35079503e-01 3.62194389e-01 -9.80382442e-01
8.77793312e-01 -1.69116914e+00 3.77086878e-01 3.50017071e-01
-3.11761312e-02 -4.40845154e-02 7.13350177e-02 2.77363747e-01
-6.75176755e-02 -9.13611710e-01 6.99022651e-01 -3.82286459e-01
-5.03371775e-01 1.91195495e-02 -6.83399662e-02 -6.77411675e-01
1.32137489e+00 -8.18636060e-01 3.34996819e-01 -2.06735451e-02
-8.92735004e-01 3.74502577e-02 1.07145977e+00 4.92081866e-02
8.50549042e-01 -1.05125856e+00 -4.78281319e-01 4.24411803e-01
1.45918265e-01 8.36952627e-01 -7.30635822e-02 8.11551809e-01
-1.03327000e+00 4.35899347e-01 -3.74596983e-01 -1.33993971e+00
-1.28139520e+00 6.60422206e-01 3.94138694e-01 3.16444337e-01
-8.24605763e-01 2.90493518e-01 1.60308227e-01 -7.98381567e-01
1.37380376e-01 -6.29631877e-01 -4.26705927e-02 -1.60298213e-01
7.49385506e-02 3.56311888e-01 2.77199000e-01 -8.32888246e-01
-2.86614716e-01 1.05010378e+00 3.21072750e-02 -1.86997384e-01
1.49477887e+00 -1.74382910e-01 -1.20961912e-01 6.41918361e-01
9.91495788e-01 -4.10738178e-02 -1.52837145e+00 -1.78437665e-01
1.14091799e-01 -6.39886975e-01 -2.34057695e-01 -6.07780516e-01
-8.90153289e-01 1.21416306e+00 7.07626998e-01 9.03632790e-02
8.92800748e-01 4.06625301e-01 1.17171839e-01 5.40321887e-01
7.37491250e-01 -7.66974866e-01 -1.18490547e-01 2.96477884e-01
1.09569311e+00 -1.18371832e+00 2.61231542e-01 -8.36622477e-01
-4.33545560e-01 1.60817325e+00 4.10665929e-01 -2.99067199e-01
2.27216527e-01 2.32972920e-01 2.05485806e-01 -2.64873296e-01
1.82254352e-02 -7.17315227e-02 1.83398679e-01 6.91946089e-01
-8.80789682e-02 -7.84230232e-02 -1.00579664e-01 -1.60362460e-02
-4.43159230e-02 -2.61641055e-01 7.51972914e-01 1.09052682e+00
-3.86335045e-01 -1.26247847e+00 -8.93875599e-01 4.77109440e-02
-6.35209084e-02 5.90964854e-01 -3.23486269e-01 1.12272716e+00
-6.61739036e-02 4.29092526e-01 4.59223926e-01 1.26156926e-01
8.09614480e-01 2.37061810e-02 1.22272885e+00 -7.84086645e-01
-1.11128680e-01 4.63346660e-01 -1.04319640e-01 -5.04565537e-01
-6.99525177e-01 -7.22440064e-01 -1.31211817e+00 4.86562639e-01
-7.48466730e-01 7.77271912e-02 1.39711177e+00 7.48785436e-01
-1.91155188e-02 -1.18312433e-01 6.27080262e-01 -1.65474391e+00
-5.33492088e-01 -8.75220656e-01 -6.62739038e-01 3.62471819e-01
1.55942932e-01 -1.06207943e+00 -2.40894943e-01 3.96785080e-01] | [7.483335971832275, -2.6081371307373047] |
ffa27a36-8800-4802-a174-81cc06fe89b9 | caire-an-end-to-end-empathetic-chatbot | 1907.12108 | null | https://arxiv.org/abs/1907.12108v4 | https://arxiv.org/pdf/1907.12108v4.pdf | CAiRE: An Empathetic Neural Chatbot | In this paper, we present an end-to-end empathetic conversation agent CAiRE. Our system adapts TransferTransfo (Wolf et al., 2019) learning approach that fine-tunes a large-scale pre-trained language model with multi-task objectives: response language modeling, response prediction and dialogue emotion detection. We evaluate our model on the recently proposed empathetic-dialogues dataset (Rashkin et al., 2019), the experiment results show that CAiRE achieves state-of-the-art performance on dialogue emotion detection and empathetic response generation. | ['Farhad Bin Siddique', 'Genta Indra Winata', 'Pascale Fung', 'Zihan Liu', 'Peng Xu', 'Jamin Shin', 'Zhaojiang Lin'] | 2019-07-28 | null | null | null | null | ['empathetic-response-generation'] | ['natural-language-processing'] | [-3.19337964e-01 5.90225458e-01 1.16030410e-01 -6.94449604e-01
-1.08322132e+00 -3.98789465e-01 8.50999415e-01 -1.19433947e-01
-6.05236948e-01 1.08058369e+00 7.97655046e-01 2.36506373e-01
3.71018410e-01 -4.86017793e-01 3.40521872e-01 -1.63028151e-01
2.16785386e-01 1.02587914e+00 -4.49075013e-01 -1.22000253e+00
3.86666268e-01 1.41234100e-01 -6.68165147e-01 1.28277600e+00
5.36772907e-01 6.89902306e-01 -3.00272673e-01 1.30015028e+00
3.54681686e-02 1.75857043e+00 -8.21798027e-01 -7.58203804e-01
-3.54977757e-01 -7.21379876e-01 -1.70722675e+00 -4.43104476e-01
-3.17993224e-01 -3.09370130e-01 -1.53437316e-01 3.85886699e-01
1.04386377e+00 5.32804191e-01 1.02656615e+00 -1.32060659e+00
-8.01009297e-01 8.01282287e-01 -1.63978490e-03 -2.68111438e-01
9.46900785e-01 1.41850814e-01 1.04860604e+00 -1.09860408e+00
6.88558459e-01 1.60907435e+00 7.68494546e-01 1.38714325e+00
-1.07846105e+00 -4.68074501e-01 -3.58536512e-01 3.08055043e-01
-7.83784807e-01 -6.64065480e-01 9.70995605e-01 -2.14082435e-01
1.52176738e+00 1.77867860e-01 1.74177110e-01 1.88787186e+00
1.38292193e-01 1.21382761e+00 1.23248816e+00 -5.05621970e-01
5.88268153e-02 4.34080899e-01 1.27177700e-01 3.70233774e-01
-1.42952919e+00 -1.19251408e-01 -8.47203195e-01 -5.42008877e-01
1.22595824e-01 -8.86014938e-01 -1.14259198e-01 4.10562783e-01
-9.52837706e-01 1.60938680e+00 4.69070971e-02 1.13781020e-01
-7.36959279e-01 -3.25994074e-01 1.09392560e+00 9.76939738e-01
7.52335489e-01 9.46979642e-01 -4.16248620e-01 -9.24437881e-01
-1.35256991e-01 4.62992728e-01 1.48226571e+00 5.76966286e-01
9.74188298e-02 -8.71928185e-02 -4.70550984e-01 1.67123199e+00
-2.33007893e-02 1.01157181e-01 8.10922563e-01 -1.27475166e+00
4.18852061e-01 2.21431762e-01 3.96926135e-01 -5.38885295e-01
-8.51736903e-01 4.49182928e-01 -8.37643564e-01 1.22391336e-01
1.60544157e-01 -8.68623197e-01 3.39947611e-01 1.64929581e+00
1.04478352e-01 -4.78944361e-01 9.65668380e-01 8.26189935e-01
1.43774116e+00 1.10139906e+00 4.94498521e-01 -3.81307185e-01
1.17035472e+00 -1.74222326e+00 -6.50751531e-01 -2.27018178e-01
9.09864068e-01 -1.03341281e+00 1.25599325e+00 5.95653176e-01
-1.39512122e+00 -3.73424292e-01 -3.46217692e-01 -2.57089913e-01
-3.49114127e-02 3.53720039e-01 6.27552629e-01 2.26907521e-01
-9.73681629e-01 1.44125029e-01 9.09989700e-02 -6.17801011e-01
-4.23159361e-01 2.36866370e-01 -5.00719726e-01 5.93510091e-01
-1.62672007e+00 1.46498168e+00 1.01564847e-01 -2.77559131e-01
-3.05628031e-01 -4.23574269e-01 -6.04500473e-01 -1.14861794e-01
-2.70979077e-01 -7.77842522e-01 2.12786365e+00 -1.04811037e+00
-2.75276041e+00 1.32186425e+00 1.72009766e-01 -5.55164158e-01
4.64969635e-01 -4.11820322e-01 -4.94898826e-01 2.79756993e-01
-3.60847980e-01 1.09635270e+00 3.93018425e-01 -8.97608161e-01
-4.37500536e-01 1.27138883e-01 -8.28154106e-03 5.64393342e-01
-2.75339484e-01 8.27004313e-01 6.41134739e-01 -3.68748635e-01
-8.26292813e-01 -9.22892213e-01 -1.10647939e-01 -5.96690416e-01
-7.19098151e-02 -9.64376390e-01 3.74675870e-01 -5.09935498e-01
8.85093033e-01 -1.63250303e+00 1.05330095e-01 -2.49597237e-01
5.48156463e-02 1.63700193e-01 -6.86126590e-01 1.21738577e+00
-2.23074585e-01 -4.35058087e-01 -7.68013811e-03 -6.82177186e-01
3.57671112e-01 -1.14685178e-01 -5.98908484e-01 7.86271840e-02
1.06431708e-01 9.58333313e-01 -8.27413082e-01 -3.83072615e-01
7.01980591e-02 2.50921577e-01 -5.65131366e-01 1.17694700e+00
-2.32379720e-01 6.27824068e-01 -4.15111572e-01 4.57461476e-02
7.92012140e-02 1.72501177e-01 -1.28491089e-01 1.87070832e-01
-1.48748547e-01 5.82770407e-01 -3.00615281e-01 1.62966347e+00
-1.03364968e+00 5.34029067e-01 1.14867263e-01 -6.50306106e-01
1.49364877e+00 8.69737506e-01 5.70838153e-01 -6.23889804e-01
4.13044900e-01 3.48571837e-02 -3.38374032e-03 -7.09152520e-01
9.65238810e-01 -4.59976763e-01 -8.84963632e-01 1.20801139e+00
1.20423183e-01 -4.32904720e-01 -1.79249227e-01 3.32226425e-01
9.50899243e-01 -1.23434022e-01 7.71579981e-01 -8.61541182e-02
1.11175835e+00 1.60895258e-01 3.14806364e-02 6.79433286e-01
-4.20345008e-01 3.06021243e-01 4.73364115e-01 -4.65737075e-01
-6.63894892e-01 -7.24608302e-01 3.66499901e-01 2.00691652e+00
-5.18487930e-01 -3.89778137e-01 -7.41079390e-01 -6.49461269e-01
-5.32554448e-01 1.12126696e+00 -6.57503664e-01 -3.64635170e-01
-6.19514704e-01 -5.43165982e-01 1.17992866e+00 2.99520820e-01
3.44647944e-01 -2.01057196e+00 -6.68977618e-01 6.84120297e-01
-9.18026745e-01 -9.34820890e-01 -4.49681103e-01 -1.33008897e-01
-1.96146175e-01 -6.16587341e-01 -5.93199730e-01 -8.86878788e-01
-8.77157375e-02 -3.57809782e-01 1.57656157e+00 -3.86662513e-01
-5.12027480e-02 6.43814087e-01 -7.81418324e-01 -2.55513430e-01
-1.28876662e+00 9.86595750e-02 -1.38229519e-01 -1.75391451e-01
4.50338811e-01 -5.37497640e-01 -3.01705807e-01 1.30865842e-01
-6.62545264e-02 2.53506541e-01 2.26196244e-01 1.20822883e+00
-2.29620844e-01 -1.08601582e+00 1.53838432e+00 -8.58779907e-01
1.83762240e+00 -6.07747912e-01 1.99507788e-01 3.50314915e-01
-3.42875093e-01 -1.05810478e-01 9.49965537e-01 -6.57883942e-01
-1.61070514e+00 -1.54669866e-01 -8.18617940e-01 3.03112954e-01
-3.85637373e-01 2.57817179e-01 2.22995207e-01 7.49155134e-02
1.02385533e+00 4.43079919e-02 -1.00255400e-01 -2.91314691e-01
8.12912166e-01 1.39181447e+00 8.71203721e-01 -9.58084464e-01
-7.87057728e-02 -9.35244262e-02 -6.67967081e-01 -4.61772859e-01
-7.26786613e-01 -6.08221531e-01 -3.75487387e-01 -5.01134753e-01
8.49793553e-01 -8.71333599e-01 -1.38099623e+00 3.81848484e-01
-1.75979185e+00 -8.69798958e-01 -3.74838337e-02 3.58409673e-01
-1.36817861e+00 1.78879231e-01 -1.40139842e+00 -1.06846559e+00
-1.43437231e+00 -4.53044683e-01 8.25039804e-01 1.92626251e-03
-1.34577239e+00 -1.16116965e+00 9.93796527e-01 4.88631248e-01
4.54345107e-01 -2.26473622e-02 8.91252637e-01 -1.16778064e+00
8.21677923e-01 2.07579844e-02 1.32610137e-02 1.83332011e-01
-4.44703966e-01 -3.05717289e-01 -1.03473949e+00 1.26930967e-01
1.53434500e-01 -1.68837464e+00 2.51231819e-01 -2.48915806e-01
6.26412988e-01 -7.42714643e-01 4.68571752e-01 4.89617735e-02
4.53638166e-01 2.52631843e-01 5.24133623e-01 2.67785370e-01
9.56759006e-02 1.13308311e+00 9.88516569e-01 1.00660217e+00
8.71579528e-01 7.07422554e-01 -1.84959739e-01 -7.38059729e-02
3.29247385e-01 -2.33508926e-02 8.00135016e-01 9.17399824e-01
1.18792765e-01 -4.83647555e-01 -6.69688344e-01 2.33245358e-01
-2.39275050e+00 -1.35450399e+00 -2.71587282e-01 1.36121511e+00
1.31375241e+00 -5.09261549e-01 5.68125308e-01 -5.15860081e-01
3.70783150e-01 1.49598002e-01 -7.28535950e-02 -1.67861009e+00
-1.49396464e-01 3.23190749e-01 -5.95475376e-01 9.24047053e-01
-8.19356740e-01 1.46242046e+00 6.16215897e+00 5.90712488e-01
-8.09064388e-01 2.13450000e-01 5.55254936e-01 3.98594886e-03
1.83138996e-01 -5.11649966e-01 -3.64980727e-01 -5.82329743e-02
1.19737458e+00 -4.37577128e-01 4.09514368e-01 1.09905756e+00
3.01303267e-01 1.69497401e-01 -1.23449004e+00 1.14690101e+00
4.28391397e-01 -9.12757039e-01 -2.92279422e-01 -4.80674565e-01
4.24449712e-01 -9.40956846e-02 -2.04741880e-01 1.13792276e+00
9.16822851e-01 -1.19369507e+00 4.15279716e-01 4.75817472e-01
2.86244720e-01 -1.01548588e+00 8.23742390e-01 5.64871192e-01
-3.97216618e-01 -3.25037986e-02 -1.63786635e-01 -3.63976032e-01
5.75266838e-01 -3.06039453e-01 -1.19790506e+00 -1.11095682e-01
2.05145270e-01 5.48912644e-01 -1.77388303e-02 1.66249737e-01
-4.38673973e-01 5.49804986e-01 1.30743101e-01 -4.16751325e-01
4.23122436e-01 -2.28260413e-01 4.32196617e-01 1.65148771e+00
-1.09395869e-01 7.64863968e-01 3.07205439e-01 6.96366727e-01
-2.69953191e-01 7.10978091e-01 -2.34592557e-01 1.67708620e-01
4.09152001e-01 1.72687852e+00 2.91388959e-01 -3.56511354e-01
-1.05165750e-01 1.30362070e+00 6.37170851e-01 2.20319461e-02
-6.57915592e-01 -2.72533774e-01 4.05013412e-01 -7.93188512e-01
-4.30535316e-01 4.89966899e-01 -2.20969796e-01 -9.28478718e-01
-6.37633264e-01 -1.31254792e+00 7.19116509e-01 -1.04100168e+00
-1.88628733e+00 1.01773489e+00 -3.41124982e-01 -7.93347418e-01
-1.33251727e+00 -4.35893357e-01 -1.09814000e+00 9.27777112e-01
-8.44765306e-01 -1.44752133e+00 4.49413583e-02 8.88946176e-01
9.49227035e-01 -5.89775503e-01 1.75284874e+00 -3.26463953e-02
-1.47708565e-01 7.70624518e-01 -3.01428914e-01 2.58112550e-01
1.44954860e+00 -1.17266595e+00 3.52228552e-01 -3.06429952e-01
-3.43059689e-01 9.63974595e-02 8.51047456e-01 -6.16559163e-02
-7.83153832e-01 -7.20986843e-01 1.36158133e+00 -5.53453982e-01
9.39905465e-01 -3.07111472e-01 -8.06488931e-01 5.52187622e-01
1.03860903e+00 -8.53541613e-01 1.10148323e+00 3.13282877e-01
-3.81559849e-01 4.39265341e-01 -1.37776196e+00 7.37842798e-01
4.62890625e-01 -6.21854305e-01 -9.78908062e-01 6.20223641e-01
3.95983815e-01 -2.43240550e-01 -1.18081057e+00 2.98983067e-01
5.87411940e-01 -1.05016446e+00 8.41390312e-01 -1.13202834e+00
9.56256747e-01 7.65350461e-01 2.40424633e-01 -1.76461315e+00
-6.80767000e-02 -1.20951259e+00 2.71531567e-02 1.22781014e+00
3.43921244e-01 -4.69246447e-01 3.38865012e-01 8.30769718e-01
-2.53934562e-01 -6.07026339e-01 -7.28805602e-01 -1.73995480e-01
6.24626100e-01 -5.04995659e-02 1.37285143e-01 1.21672726e+00
1.19661951e+00 1.41385448e+00 -9.18868780e-01 -6.93988204e-01
-4.87174988e-02 2.23020226e-01 1.16031587e+00 -1.06638503e+00
-5.21019101e-01 -7.01143980e-01 4.03414905e-01 -9.35066700e-01
1.05588007e+00 -6.71027720e-01 3.40488881e-01 -1.36059141e+00
1.23136770e-02 -6.49263486e-02 2.42933899e-01 5.50558329e-01
-7.55563378e-02 1.33291081e-01 2.52233714e-01 2.90365480e-02
-8.76581132e-01 1.00146556e+00 8.40939879e-01 9.42329392e-02
-4.80813265e-01 1.97246388e-01 -6.49105906e-01 7.70029366e-01
1.26674557e+00 -4.56338972e-01 -2.41982296e-01 1.52629852e-01
8.61651152e-02 8.13760340e-01 7.15633929e-02 -3.37932646e-01
2.56437778e-01 -2.89101690e-01 -1.82510123e-01 -3.76660347e-01
8.70129824e-01 2.97856741e-02 -5.95418990e-01 2.77850777e-01
-1.23252439e+00 3.54216456e-01 6.32299930e-02 -1.41654179e-01
-3.22905034e-01 -4.10259455e-01 9.60335732e-01 -1.21261626e-01
-5.05342484e-01 -2.87551224e-01 -1.16122389e+00 3.98447275e-01
6.93324506e-01 4.75008488e-01 -7.11178720e-01 -1.35933840e+00
-6.07324958e-01 2.86562294e-01 -8.61288086e-02 6.71878874e-01
6.64567530e-01 -1.18499887e+00 -1.50806618e+00 -5.26649296e-01
4.32343245e-01 -7.59080648e-01 3.01744908e-01 6.86286807e-01
-1.56876341e-01 2.56425917e-01 -3.81152242e-01 7.05391914e-02
-1.69450998e+00 1.60447195e-01 4.57964242e-01 -6.48280203e-01
-5.12622356e-01 1.04983866e+00 1.87519304e-02 -1.13192630e+00
8.29176381e-02 6.13977075e-01 -5.59043586e-01 4.37039062e-02
6.62359715e-01 3.77762794e-01 -4.11384791e-01 -7.56372511e-01
-4.00049053e-03 -1.21943779e-01 -2.42717266e-01 -7.64570832e-01
1.17698836e+00 -9.28000286e-02 -3.13054115e-01 5.22022963e-01
1.15926874e+00 8.31419751e-02 -6.02840304e-01 -1.74008936e-01
6.33965610e-05 2.17591345e-01 -5.25602579e-01 -1.46037257e+00
-2.42188528e-01 9.16042149e-01 2.21327301e-02 9.20397788e-03
8.07167947e-01 -9.23837628e-03 1.33125031e+00 1.13687265e+00
7.40885437e-02 -1.56320500e+00 5.68102121e-01 1.31266868e+00
1.40369701e+00 -1.29323351e+00 -4.81799096e-01 3.47986594e-02
-1.59740484e+00 1.41107106e+00 9.72178996e-01 -1.26668155e-01
-4.89509329e-02 1.72843724e-01 8.65054667e-01 -6.11676276e-02
-1.56830609e+00 2.18168810e-01 1.75576340e-02 5.63649595e-01
8.99106205e-01 1.81537107e-01 -5.66000938e-01 1.12387550e+00
-7.10733950e-01 -1.90462157e-01 6.78008556e-01 4.65905279e-01
-3.36561799e-01 -1.25804400e+00 -1.74601153e-01 -3.08618993e-01
-2.80343920e-01 -2.46062651e-01 -1.49562967e+00 5.76732457e-01
-7.61871696e-01 1.50379229e+00 -1.70338124e-01 -4.42355216e-01
5.45852423e-01 4.46386546e-01 2.86170602e-01 -5.78701496e-01
-1.56503320e+00 -3.42252672e-01 1.08158565e+00 -3.35294574e-01
-4.03250158e-01 -3.69037628e-01 -1.53928673e+00 -3.44567806e-01
1.68678284e-01 6.46584868e-01 3.54646415e-01 8.75441015e-01
2.72887349e-01 1.93334576e-02 1.12038994e+00 -6.41229451e-01
-1.06183767e+00 -1.64265203e+00 -1.35009378e-01 6.10864937e-01
-1.83108270e-01 -9.20097064e-03 -2.44793594e-01 1.49751948e-02] | [13.140615463256836, 7.655147552490234] |
768bb3cf-af72-4b31-a35b-4e38e2bd9e1d | probabilistic-time-series-forecasting-with | 2010.07349 | null | https://arxiv.org/abs/2010.07349v3 | https://arxiv.org/pdf/2010.07349v3.pdf | Probabilistic Time Series Forecasting with Structured Shape and Temporal Diversity | Probabilistic forecasting consists in predicting a distribution of possible future outcomes. In this paper, we address this problem for non-stationary time series, which is very challenging yet crucially important. We introduce the STRIPE model for representing structured diversity based on shape and time features, ensuring both probable predictions while being sharp and accurate. STRIPE is agnostic to the forecasting model, and we equip it with a diversification mechanism relying on determinantal point processes (DPP). We introduce two DPP kernels for modeling diverse trajectories in terms of shape and time, which are both differentiable and proved to be positive semi-definite. To have an explicit control on the diversity structure, we also design an iterative sampling mechanism to disentangle shape and time representations in the latent space. Experiments carried out on synthetic datasets show that STRIPE significantly outperforms baseline methods for representing diversity, while maintaining accuracy of the forecasting model. We also highlight the relevance of the iterative sampling scheme and the importance to use different criteria for measuring quality and diversity. Finally, experiments on real datasets illustrate that STRIPE is able to outperform state-of-the-art probabilistic forecasting approaches in the best sample prediction. | ['Nicolas Thome', 'Vincent Le Guen'] | 2020-10-14 | null | null | null | null | ['probabilistic-time-series-forecasting'] | ['time-series'] | [-1.36525571e-01 -2.11264327e-01 -1.26272023e-01 -2.91549891e-01
-7.36917853e-01 -7.95220435e-01 1.21168876e+00 1.94245260e-02
2.04824001e-01 7.23661542e-01 3.68819475e-01 -2.86952287e-01
-5.74181378e-01 -7.64337778e-01 -5.47309577e-01 -1.00239444e+00
-4.32912111e-01 5.96247733e-01 2.11077943e-01 -1.53490618e-01
5.04546583e-01 6.40459597e-01 -1.78371227e+00 1.33266136e-01
9.99687374e-01 1.11942148e+00 -1.22563563e-01 6.37145817e-01
5.51166832e-02 7.64349878e-01 -3.58362556e-01 -4.89245355e-01
1.46342129e-01 -1.97528943e-01 -3.41454178e-01 4.24092077e-02
5.72477989e-02 2.15315506e-01 1.72235072e-01 5.70721686e-01
2.59344667e-01 -6.06951341e-02 1.38054073e+00 -1.44927764e+00
-4.61915970e-01 3.62733036e-01 -2.10234940e-01 1.72482327e-01
2.01588556e-01 -1.19699948e-02 1.06244600e+00 -8.67331862e-01
4.06767786e-01 1.06899273e+00 1.09017146e+00 1.20988145e-01
-1.55811989e+00 -3.68153572e-01 3.03578526e-01 7.82996714e-02
-1.32558846e+00 -3.41111064e-01 7.99081028e-01 -9.51042295e-01
5.31675637e-01 4.67630059e-01 4.78551090e-01 1.26065147e+00
6.25918210e-01 6.82308793e-01 1.05478883e+00 -1.89095989e-01
6.56238735e-01 9.68754441e-02 1.96111456e-01 2.17293844e-01
9.90872160e-02 3.03948164e-01 -3.95078689e-01 -7.65610754e-01
5.74953079e-01 1.56465873e-01 -3.07332307e-01 -7.00206339e-01
-1.37074482e+00 9.56618369e-01 -2.39281744e-01 3.72542471e-01
-7.59001553e-01 -1.26817793e-01 2.01625541e-01 2.31391773e-01
8.98658514e-01 2.50400066e-01 -5.10729551e-01 -4.09863323e-01
-1.15070701e+00 4.98742104e-01 1.06411672e+00 7.04845130e-01
3.00084263e-01 8.64254013e-02 -6.03444099e-01 5.45831025e-01
1.73055023e-01 5.99364877e-01 2.95295745e-01 -8.59580815e-01
3.15514982e-01 3.44139606e-01 5.79571962e-01 -1.05589819e+00
-4.40539628e-01 -6.41665459e-01 -1.07668996e+00 -1.67450402e-02
3.48100781e-01 -8.65774520e-04 -6.93029821e-01 1.72649121e+00
3.13891917e-01 4.97449189e-01 -3.02620372e-03 5.29689431e-01
7.05285296e-02 9.40192401e-01 1.11141346e-01 -4.22921658e-01
9.66456056e-01 -5.31809926e-01 -3.39024276e-01 4.84404713e-01
4.80014503e-01 -5.49632430e-01 8.52001667e-01 5.33620954e-01
-1.03514016e+00 -4.22527552e-01 -5.72449744e-01 4.66115743e-01
-9.00457799e-02 2.91863203e-01 4.91394103e-01 7.44853139e-01
-1.14374638e+00 7.78059363e-01 -9.61345196e-01 -2.25638580e-02
4.00606766e-02 7.98289105e-02 2.76431646e-02 3.28454494e-01
-1.20729446e+00 7.27890491e-01 1.42994717e-01 -1.02250174e-01
-5.40000200e-01 -1.06116152e+00 -5.75856745e-01 3.45046282e-01
-1.81692958e-01 -7.74021506e-01 1.00966036e+00 -6.63709760e-01
-1.48338854e+00 2.40434349e-01 -2.05080718e-01 -5.57705760e-01
8.06975067e-01 2.03807265e-01 -6.72795594e-01 9.26473811e-02
-7.03508034e-02 3.24210495e-01 1.15902209e+00 -1.21012020e+00
-6.50683403e-01 -1.51706725e-01 -3.97312850e-01 -2.51230478e-01
-2.36017942e-01 -3.30855191e-01 1.52417809e-01 -8.50823462e-01
-3.28008235e-02 -1.01765192e+00 -3.15185279e-01 -3.44643533e-01
-2.09959462e-01 -3.41937542e-01 5.33097744e-01 -6.21001840e-01
1.53528690e+00 -1.96491241e+00 2.47608677e-01 4.30224568e-01
4.45892215e-02 -8.77774730e-02 6.96479296e-03 9.17255402e-01
1.13647074e-01 2.54287086e-02 -4.33780491e-01 -5.33430994e-01
2.85285771e-01 2.42353126e-01 -1.02868795e+00 5.91348410e-01
4.27509397e-01 7.86547482e-01 -7.38469601e-01 -3.14203650e-02
2.21336391e-02 6.79355443e-01 -3.49669069e-01 -1.66387283e-04
-3.40862602e-01 5.97355783e-01 -4.20114517e-01 5.99279046e-01
6.82273924e-01 -3.51899624e-01 6.31275401e-02 2.78844744e-01
-3.80197406e-01 1.51834920e-01 -1.13254392e+00 8.54838371e-01
-3.58612806e-01 3.05020630e-01 -4.55578983e-01 -9.59449053e-01
1.16040099e+00 3.90398979e-01 5.06883502e-01 -4.00029480e-01
-1.73675269e-01 3.23405743e-01 -3.99904758e-01 7.50236446e-03
6.22716427e-01 -1.26631320e-01 -1.01928994e-01 4.42483395e-01
-2.50827640e-01 1.04348427e-02 1.39855638e-01 -1.58435926e-01
8.39018047e-01 7.66225159e-02 6.61072731e-02 -6.92866385e-01
5.34686804e-01 -1.39034420e-01 5.46664417e-01 6.28470004e-01
1.25797704e-01 6.16461039e-01 8.74335051e-01 -4.92293775e-01
-1.21866369e+00 -1.10043740e+00 -4.03513521e-01 9.31491554e-01
-1.11964844e-01 -2.02048913e-01 -2.95542300e-01 -4.23102528e-01
2.88178593e-01 1.13522542e+00 -9.78004754e-01 -3.33820470e-02
-2.77845472e-01 -9.38543081e-01 2.29414850e-01 4.10826683e-01
-2.38742128e-01 -7.01968610e-01 -5.95974565e-01 3.73036176e-01
-3.26657780e-02 -6.79835260e-01 -2.98692137e-01 3.11278878e-03
-1.03172588e+00 -7.32513547e-01 -1.03169954e+00 -1.97511300e-01
4.20280784e-01 1.14044666e-01 1.30080509e+00 -5.79474092e-01
2.87515014e-01 7.23126411e-01 -2.85701901e-01 -2.84101307e-01
-5.32266617e-01 1.92879528e-01 -5.05636260e-02 3.62655967e-01
1.45397469e-01 -9.42091346e-01 -6.78232729e-01 4.98048872e-01
-1.05428886e+00 6.55641407e-02 2.84296215e-01 8.53399873e-01
4.52068180e-01 6.23351708e-02 7.26668715e-01 -5.90493739e-01
8.58559489e-01 -9.54478502e-01 -6.38023496e-01 4.66352284e-01
-8.99919689e-01 2.98704416e-01 7.09541202e-01 -7.12079763e-01
-8.42762709e-01 -7.34497607e-02 8.86235833e-02 -4.92781639e-01
5.15805818e-02 5.68262458e-01 3.56429368e-01 2.36719534e-01
4.31750864e-01 5.50894380e-01 -8.83378088e-02 -3.96800190e-01
3.68221253e-01 3.42434138e-01 1.43048778e-01 -6.64175987e-01
3.90954703e-01 6.49983466e-01 3.16801965e-01 -6.28369570e-01
-5.31725883e-01 -3.98207277e-01 -6.10050797e-01 -1.85253590e-01
3.10635209e-01 -7.48727918e-01 -6.70979917e-01 4.24008489e-01
-9.62976038e-01 -1.44249484e-01 -3.27955067e-01 3.90380144e-01
-8.20144951e-01 1.85574278e-01 -4.81590897e-01 -1.22765279e+00
-1.47192210e-01 -9.22219992e-01 1.39490151e+00 -2.81463712e-01
-2.12614387e-01 -1.42462325e+00 5.47936499e-01 -3.94275904e-01
6.65842652e-01 4.68059778e-01 1.03750968e+00 -6.96188927e-01
-3.78688127e-01 -1.89482823e-01 6.23745658e-02 1.11383043e-01
-1.70750827e-01 4.56449062e-01 -7.49942720e-01 -8.80906880e-02
1.05216123e-01 2.94839680e-01 1.12400007e+00 5.61103702e-01
8.91040862e-01 -3.51047873e-01 -3.34913164e-01 3.77331644e-01
1.14098072e+00 2.86356211e-02 5.43035924e-01 1.44523054e-01
4.58786011e-01 8.68876934e-01 4.34809774e-01 9.70253944e-01
6.08988643e-01 6.84636950e-01 1.57252923e-01 4.57527518e-01
3.28334987e-01 -2.45982319e-01 4.95889544e-01 9.09386396e-01
-6.31314740e-02 -2.05663338e-01 -1.03610086e+00 5.70515811e-01
-2.08653283e+00 -1.28677094e+00 -3.50611508e-01 2.37803054e+00
6.90033138e-01 2.99036759e-03 4.90705788e-01 2.47370914e-01
5.49040735e-01 1.52774140e-01 -3.93125087e-01 -4.12260622e-01
-1.84134528e-01 -1.83365032e-01 3.84879559e-01 5.19681394e-01
-1.13441646e+00 4.02653873e-01 6.63480091e+00 9.61158276e-01
-1.32784605e+00 -1.25518620e-01 8.95576239e-01 1.95980340e-01
-8.69408309e-01 -1.08668864e-01 -1.01132393e+00 9.21996057e-01
1.22945774e+00 -2.64258564e-01 5.76698966e-02 7.74163485e-01
1.79188594e-01 1.57159820e-01 -1.02875984e+00 6.75170064e-01
-2.14078322e-01 -1.43589747e+00 1.05740398e-01 1.20854517e-02
9.61162329e-01 -1.88486651e-01 4.06943291e-01 3.77350479e-01
9.65707600e-02 -9.58383918e-01 9.82908130e-01 1.26956487e+00
2.91034520e-01 -9.43216860e-01 4.53969687e-01 6.71180308e-01
-1.15728462e+00 -1.39275447e-01 -2.80216783e-01 -1.19775154e-01
3.05481225e-01 1.08528244e+00 -6.12512112e-01 6.04650795e-01
3.04659396e-01 7.90445745e-01 -3.35899949e-01 1.03967977e+00
9.99803320e-02 9.09089506e-01 -3.50690722e-01 -1.02672651e-01
2.41407752e-01 -3.71031761e-01 6.56965435e-01 1.34765005e+00
9.39028144e-01 -2.60949552e-01 5.52063994e-02 8.48695278e-01
6.02211416e-01 -1.10022068e-01 -4.73552644e-01 -4.63385656e-02
4.28391576e-01 8.31065953e-01 -5.85772336e-01 -2.38091797e-01
-2.75462940e-02 7.24864006e-01 5.41298874e-02 3.91560614e-01
-7.70084798e-01 1.26646489e-01 5.69259048e-01 1.09445281e-01
7.56230175e-01 -5.09403229e-01 -5.14575243e-01 -1.23524845e+00
6.59825429e-02 -5.29338837e-01 1.73844084e-01 -3.84508699e-01
-1.73365617e+00 7.61870265e-01 1.27932370e-01 -1.53390861e+00
-7.91521788e-01 -3.90626341e-01 -6.53471768e-01 1.02775526e+00
-1.50357962e+00 -1.20431876e+00 1.94606230e-01 4.51843053e-01
2.37661719e-01 -2.46949587e-02 8.08369577e-01 -3.62804867e-02
-3.32414925e-01 2.67320603e-01 5.98354399e-01 -8.25973570e-01
2.74996996e-01 -1.27695763e+00 6.86366677e-01 5.24799347e-01
4.52858135e-02 5.43651044e-01 9.12362099e-01 -6.43768132e-01
-1.29164100e+00 -1.12903893e+00 1.19227409e+00 -7.37161398e-01
9.60655391e-01 -3.35797787e-01 -1.07909310e+00 3.21172565e-01
-3.10433865e-01 -2.06762478e-01 6.99560702e-01 1.69766739e-01
-3.92322838e-01 6.90780655e-02 -1.02032459e+00 5.35954952e-01
6.63658321e-01 -2.89972156e-01 -4.71225411e-01 2.48683393e-01
4.44742411e-01 1.46802841e-02 -9.53855813e-01 6.13734782e-01
7.24835932e-01 -1.20276928e+00 1.02184975e+00 -3.98918122e-01
4.53741342e-01 -1.95622012e-01 -1.69735953e-01 -1.28630459e+00
-6.78764284e-01 -6.67943656e-01 -5.46560168e-01 1.29721618e+00
3.58521461e-01 -9.47163701e-01 5.01679838e-01 6.76200986e-01
1.46044061e-01 -1.18079102e+00 -9.36919868e-01 -1.02574432e+00
2.74943590e-01 -6.16469979e-01 8.27248514e-01 8.86128187e-01
-2.02919021e-01 -2.85222419e-02 -6.87120080e-01 3.42748493e-01
4.83307153e-01 5.34437716e-01 5.63619792e-01 -1.64751709e+00
-2.66047716e-01 -7.52891123e-01 -3.66042614e-01 -9.02166009e-01
1.34248197e-01 -5.09373069e-01 -2.97694951e-01 -1.13528538e+00
-5.97343035e-02 -6.45312250e-01 -3.81287515e-01 -1.17660472e-02
-1.02130786e-01 2.00207997e-02 2.33792304e-03 6.87034190e-01
-4.11871344e-01 9.85377491e-01 7.98933506e-01 2.50517458e-01
-3.34062159e-01 5.57037055e-01 -5.46003580e-01 4.56614763e-01
8.48299682e-01 -3.53006393e-01 -5.61809719e-01 4.57358845e-02
3.52353662e-01 3.57454330e-01 5.52744031e-01 -8.81694555e-01
1.75446495e-01 -2.79648662e-01 2.02272758e-01 -8.37923944e-01
4.23175603e-01 -6.79710865e-01 5.47939301e-01 3.50895852e-01
-4.22032565e-01 1.79843679e-01 7.41740018e-02 7.90589571e-01
-2.54614681e-01 3.66271026e-02 4.53059733e-01 4.27751780e-01
-3.43535960e-01 3.95173728e-01 -3.75188589e-01 -2.03064069e-01
1.09249389e+00 -1.09485112e-01 -5.79379089e-02 -6.25304759e-01
-6.09598577e-01 2.01261923e-01 5.04875124e-01 2.30739757e-01
2.20147282e-01 -1.38235307e+00 -8.50944459e-01 2.26611063e-01
1.08663023e-01 -7.45603502e-01 3.15655142e-01 1.00605369e+00
-2.18608782e-01 6.61692739e-01 9.63929221e-02 -7.72386432e-01
-7.21326649e-01 6.48285747e-01 9.39171948e-03 -5.02432764e-01
-4.67147678e-01 3.47022176e-01 1.54526770e-01 -3.28851134e-01
2.80439317e-01 -6.91405177e-01 -2.71928102e-01 3.73832613e-01
4.92656678e-01 6.33452952e-01 -5.80011010e-02 -6.05150938e-01
-2.57538229e-01 5.81589878e-01 1.97873339e-01 -1.92012340e-01
1.36463177e+00 -2.12846473e-01 7.39099272e-03 8.29771757e-01
8.96217108e-01 8.29263330e-02 -1.76017189e+00 -6.05959073e-03
3.18522274e-01 -3.30140114e-01 -2.29876190e-01 -7.01016188e-01
-5.05207598e-01 8.45433772e-01 3.26592296e-01 1.02615964e+00
9.69663501e-01 -3.03528994e-01 6.48173571e-01 2.30486244e-02
4.82414424e-01 -5.72040319e-01 -4.81287688e-01 7.22570360e-01
1.05407226e+00 -8.92565608e-01 -2.69817621e-01 -2.92914122e-01
-8.62831414e-01 1.24371326e+00 -3.65859002e-01 -2.79490411e-01
9.67968225e-01 3.35989296e-02 -2.05136925e-01 7.60023817e-02
-1.11056566e+00 1.00404762e-01 7.98354626e-01 3.76930594e-01
3.36248517e-01 3.40083510e-01 -4.80317384e-01 7.33470619e-01
-2.81261414e-01 -4.54760827e-02 1.20095745e-01 5.73512077e-01
-3.48355681e-01 -1.16544878e+00 -3.55045885e-01 3.69029164e-01
-3.44185919e-01 5.53381927e-02 -9.09224059e-03 5.58637321e-01
-1.10630341e-01 7.16975272e-01 7.99383223e-02 -2.94232070e-01
8.90673324e-02 1.86862156e-01 1.54534534e-01 -5.14524952e-02
-4.13567215e-01 2.41288487e-02 -4.79825176e-02 -2.51737893e-01
-2.65985966e-01 -1.03008091e+00 -4.99457031e-01 -3.84288579e-01
3.71562913e-02 3.39629173e-01 6.44772351e-01 8.43943834e-01
8.33319962e-01 1.99407429e-01 9.35944557e-01 -1.04515529e+00
-1.05144918e+00 -7.17879117e-01 -7.77778804e-01 2.33241007e-01
3.79537433e-01 -8.56293142e-01 -5.11065900e-01 -2.79552100e-04] | [7.035571575164795, 3.407447099685669] |
c3aca083-8a14-475b-a252-75896c791143 | collecting-high-quality-adversarial-data-for | 2206.14272 | null | https://arxiv.org/abs/2206.14272v1 | https://arxiv.org/pdf/2206.14272v1.pdf | Collecting high-quality adversarial data for machine reading comprehension tasks with humans and models in the loop | We present our experience as annotators in the creation of high-quality, adversarial machine-reading-comprehension data for extractive QA for Task 1 of the First Workshop on Dynamic Adversarial Data Collection (DADC). DADC is an emergent data collection paradigm with both models and humans in the loop. We set up a quasi-experimental annotation design and perform quantitative analyses across groups with different numbers of annotators focusing on successful adversarial attacks, cost analysis, and annotator confidence correlation. We further perform a qualitative analysis of our perceived difficulty of the task given the different topics of the passages in our dataset and conclude with recommendations and suggestions that might be of value to people working on future DADC tasks and related annotation interfaces. | ['John Culnan', 'Magdalena Anioł', 'Damian Y. Romero Diaz'] | 2022-06-28 | null | https://aclanthology.org/2022.dadc-1.6 | https://aclanthology.org/2022.dadc-1.6.pdf | naacl-dadc-2022-7 | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 6.61866367e-02 6.85095191e-01 5.76660931e-01 -3.37290257e-01
-1.36876142e+00 -1.40827692e+00 7.05867827e-01 7.21728861e-01
-6.29450798e-01 4.74316061e-01 1.01293802e+00 -6.02121174e-01
-4.10753489e-02 -7.59816915e-02 -5.94878733e-01 -1.91813082e-01
1.25791162e-01 8.08654845e-01 1.96042880e-01 -5.32772362e-01
3.00310969e-01 1.59803808e-01 -6.74252689e-01 6.17292225e-01
8.23670030e-01 4.08490270e-01 -3.41673881e-01 8.85661960e-01
2.58099318e-01 1.11500776e+00 -1.21438026e+00 -1.22931790e+00
3.37147534e-01 -2.20898092e-01 -1.58517992e+00 -6.11546516e-01
7.17394233e-01 -1.39267951e-01 4.48226966e-02 1.01408052e+00
9.61911976e-01 1.60847515e-01 4.22293097e-01 -1.21097481e+00
-8.05045128e-01 1.10150278e+00 1.41677856e-01 3.82056355e-01
7.29771614e-01 7.64242709e-01 8.63501191e-01 -6.03894591e-01
8.85634959e-01 1.22632515e+00 7.94783711e-01 9.78680253e-01
-1.03858435e+00 -3.39483321e-01 -1.43067986e-01 2.41574183e-01
-5.74852169e-01 -7.59197712e-01 5.34293830e-01 -6.65957808e-01
1.05049706e+00 4.63773936e-01 1.03526421e-01 1.79667866e+00
-1.62438780e-01 5.91159225e-01 9.63163018e-01 -5.19129992e-01
2.13748753e-01 2.91423023e-01 9.52762887e-02 4.40201283e-01
-1.53405309e-01 -7.89184794e-02 -5.39730072e-01 -4.36879903e-01
-1.62187546e-01 -9.33797896e-01 -4.98145908e-01 6.95221722e-02
-1.04350162e+00 9.49105144e-01 2.34106526e-01 2.41180524e-01
-2.53105283e-01 2.05239803e-02 1.15445459e+00 5.35585105e-01
6.43888295e-01 1.37990224e+00 -6.44887030e-01 -6.46688223e-01
-2.57683218e-01 7.36158133e-01 1.14517379e+00 9.88446593e-01
-2.49251693e-01 -1.76783144e-01 -2.93977171e-01 4.96762872e-01
5.75416535e-02 4.53349590e-01 1.73196584e-01 -1.16800392e+00
9.99928236e-01 2.19103768e-01 5.47024369e-01 -9.81402099e-01
-4.23610210e-01 -2.60312837e-02 -2.44633541e-01 3.10248405e-01
8.30166280e-01 -6.67205215e-01 -3.62884372e-01 1.68456244e+00
5.18200770e-02 -7.20585465e-01 1.53085575e-01 1.02578676e+00
4.72547054e-01 2.59441495e-01 5.84474921e-01 -9.88654047e-02
1.45806527e+00 -7.01433778e-01 -8.39645803e-01 -6.15373440e-02
9.13396239e-01 -9.40373898e-01 1.51951945e+00 6.02732778e-01
-1.27236736e+00 -4.88196135e-01 -7.95002043e-01 -3.87926698e-01
-4.09291208e-01 -5.90271711e-01 1.27014756e-01 7.01362491e-01
-7.66062260e-01 4.08024460e-01 -6.10559225e-01 -9.38538387e-02
3.22704166e-01 -1.43548816e-01 -2.90888101e-01 1.32938236e-01
-1.55784190e+00 1.44980502e+00 2.61772066e-01 -1.65191233e-01
-9.88485098e-01 -9.57503021e-01 -6.51691675e-01 -1.50081664e-01
5.60048461e-01 -3.61373574e-01 1.61842167e+00 -9.11073864e-01
-9.53466058e-01 8.39987695e-01 3.77842098e-01 -6.78983033e-01
6.49724424e-01 -9.61724460e-01 -3.45081121e-01 1.24994889e-02
1.79634392e-02 4.44176555e-01 3.73620331e-01 -1.34138989e+00
-3.24412793e-01 -3.59330684e-01 3.79669666e-01 2.32677877e-01
-2.12151557e-01 4.73781794e-01 2.57751912e-01 -7.80458391e-01
-7.77920783e-01 -8.07639658e-01 -1.28333256e-01 -4.05209661e-01
-3.23635399e-01 -3.61296088e-01 7.12802052e-01 -1.20869505e+00
1.15613985e+00 -1.93432617e+00 5.85296154e-01 2.28941767e-03
3.97304118e-01 6.26428425e-01 -9.58971605e-02 7.74242222e-01
-3.93962376e-02 6.88315511e-01 5.94115518e-02 -5.04489899e-01
2.34844834e-01 -3.43643963e-01 -6.56471074e-01 9.61877108e-02
2.25287825e-01 8.01248491e-01 -1.16722023e+00 -4.44397479e-01
-3.13175261e-01 -2.73510888e-02 -3.85781735e-01 5.94227791e-01
-4.33335602e-01 5.12615860e-01 -3.66033792e-01 4.29249734e-01
2.00336367e-01 3.82153332e-01 -1.02023914e-01 -1.96678475e-01
1.00074746e-01 3.09951752e-01 -3.83986950e-01 1.81807852e+00
-4.38143760e-01 8.06147575e-01 1.96364045e-01 -4.01989520e-01
5.48314214e-01 6.86506093e-01 -1.43093735e-01 -6.02298975e-01
2.67258883e-01 6.23357110e-02 1.92312986e-01 -7.58868575e-01
7.59588838e-01 1.04365367e-02 -6.17695987e-01 7.03835547e-01
2.94954211e-01 -3.12473089e-01 -2.26787515e-02 6.43688738e-01
1.49982452e+00 -1.13617510e-01 -9.12773758e-02 -3.77552748e-01
2.13194415e-01 4.85985100e-01 2.76240647e-01 9.81822848e-01
-6.61630452e-01 4.11479741e-01 6.86667860e-01 -6.35651350e-01
-1.29059565e+00 -8.81271780e-01 3.70143026e-01 1.05615413e+00
-2.80927300e-01 -5.72357178e-01 -1.22779810e+00 -1.14576733e+00
-5.08896708e-01 1.13524687e+00 -7.20880270e-01 -3.92233729e-01
-4.51018244e-01 -2.17410624e-02 1.30529034e+00 5.03997743e-01
3.07794392e-01 -1.45950210e+00 -7.92396903e-01 5.68268299e-02
-6.83172286e-01 -1.12266600e+00 -5.12933552e-01 1.75352544e-01
-2.10330531e-01 -1.01556432e+00 2.32557431e-02 -4.99221087e-01
1.75169155e-01 -4.85987484e-01 1.50414634e+00 2.13930756e-01
-6.23138025e-02 6.41646743e-01 -9.94942427e-01 -1.10420871e+00
-1.07940114e+00 1.42744362e-01 -3.81023176e-02 -6.83026850e-01
4.61699307e-01 -1.45105734e-01 -3.98439348e-01 2.88492739e-01
-9.80343163e-01 -6.32706061e-02 -2.86577232e-02 6.69255614e-01
-6.87731802e-02 -1.29194438e-01 3.85224938e-01 -1.20746708e+00
1.36258054e+00 -5.45861304e-01 -3.66653502e-01 4.88788515e-01
-3.85082185e-01 -1.80656269e-01 8.18921447e-01 -3.99816573e-01
-1.10201001e+00 -3.28186959e-01 -1.65165588e-01 -2.81605482e-01
-2.75004745e-01 4.12825108e-01 -3.87895882e-01 1.81666836e-01
1.38284492e+00 -6.74236357e-01 -4.19786759e-02 -3.50622118e-01
5.14887452e-01 5.86934805e-01 7.75874138e-01 -9.77851868e-01
1.04379737e+00 -2.82255590e-01 -6.34139001e-01 -1.07435383e-01
-9.67481136e-01 -8.83196741e-02 -5.22594869e-01 -4.27506208e-01
1.24252796e+00 -6.69925630e-01 -8.20408583e-01 5.88875830e-01
-1.43391752e+00 -9.74403560e-01 -4.69415516e-01 1.07797356e-02
-4.72166568e-01 1.53346553e-01 -6.45921469e-01 -7.43680596e-01
-7.26479173e-01 -1.08184266e+00 7.44567573e-01 6.37709945e-02
-9.19520795e-01 -1.26549327e+00 3.74365002e-01 9.82968628e-01
6.75495684e-01 5.17328441e-01 1.06248665e+00 -1.39204597e+00
1.17062181e-01 -4.34467465e-01 2.65812933e-01 3.28943193e-01
-3.04019511e-01 -1.46550983e-01 -1.09832513e+00 -2.37060130e-01
-1.39951869e-03 -1.19357240e+00 3.93637791e-02 -4.41244632e-01
1.22661388e+00 -5.67023218e-01 2.41528183e-01 -4.07314301e-02
1.09500849e+00 4.75496054e-02 7.91969836e-01 4.18687344e-01
8.38166177e-01 9.22222257e-01 6.76624894e-01 1.39219433e-01
3.56842652e-02 6.42240345e-01 4.38686103e-01 2.19717443e-01
2.08275616e-01 -3.89784098e-01 4.69561696e-01 6.18007779e-01
-1.51803523e-01 -9.53161716e-01 -1.33417368e+00 6.25268817e-01
-1.51607096e+00 -1.04684472e+00 -1.62674904e-01 1.85927105e+00
1.01873302e+00 4.86085236e-01 1.26485154e-01 -3.59171815e-02
2.89743841e-01 6.07496798e-02 -1.08768798e-01 -1.13110495e+00
4.92818058e-02 4.22371894e-01 2.86351800e-01 5.31107843e-01
-1.07958949e+00 8.22213173e-01 6.49278164e+00 7.18569994e-01
-3.31342608e-01 2.68340498e-01 6.28003001e-01 7.89837167e-02
-5.11656880e-01 2.40265459e-01 -3.12757462e-01 5.66879451e-01
1.47879016e+00 -1.04572624e-01 4.60267305e-01 5.54490566e-01
-4.58300151e-02 -5.44365272e-02 -1.10002708e+00 6.96930662e-02
6.58792555e-02 -1.09965622e+00 -6.93575889e-02 -7.50359073e-02
3.85519952e-01 -8.19487348e-02 -6.67037293e-02 4.94891614e-01
9.27132189e-01 -1.12423837e+00 9.59968746e-01 5.11853456e-01
5.49450874e-01 -5.47218561e-01 1.09253109e+00 5.66398084e-01
-1.16924003e-01 -2.20458671e-01 2.99411975e-02 3.60393128e-03
4.37404335e-01 -1.47066444e-01 -8.77843082e-01 4.85401988e-01
8.78731310e-01 -8.59856829e-02 -1.01545632e+00 2.69974560e-01
-1.57553583e-01 8.82492244e-01 1.47954032e-01 -7.13470951e-02
-3.47063169e-02 3.32330793e-01 8.15828383e-01 1.24112880e+00
-3.72713745e-01 3.69961500e-01 2.22405493e-02 5.62973261e-01
-3.06280643e-01 -1.12164654e-01 -5.09407341e-01 -2.69019216e-01
8.03525567e-01 1.20556617e+00 -3.54845941e-01 -1.45945847e-01
-1.40621722e-01 8.51282716e-01 5.54816782e-01 -5.85971493e-03
-8.65446031e-01 -4.55640733e-01 2.54743159e-01 4.49040085e-02
-3.28517169e-01 -8.11460763e-02 -4.20848012e-01 -7.43262112e-01
5.83222732e-02 -1.78770876e+00 4.74289238e-01 -1.04973471e+00
-1.63843882e+00 7.45804191e-01 3.08969557e-01 -9.65764046e-01
-2.30896667e-01 -4.28467959e-01 -8.83268297e-01 9.55464840e-01
-3.78890634e-01 -9.62204158e-01 -4.49238032e-01 5.38913548e-01
5.78179717e-01 -4.01810318e-01 1.29215598e+00 -1.65556010e-03
-1.78933114e-01 1.01749885e+00 -3.99838954e-01 4.92675126e-01
1.13246417e+00 -1.66975820e+00 8.17352235e-01 9.87312913e-01
9.56982374e-03 6.96394920e-01 1.07299495e+00 -7.43585765e-01
-9.89784122e-01 -7.22599983e-01 7.28069127e-01 -1.36897814e+00
9.75245595e-01 -5.47104478e-01 -1.18105602e+00 5.24593472e-01
7.68116951e-01 -4.98825550e-01 9.33910489e-01 2.23486483e-01
-4.78615582e-01 3.98178577e-01 -1.30762243e+00 5.84268808e-01
1.00978410e+00 -9.56292331e-01 -9.35647190e-01 5.11678874e-01
1.25775516e+00 -8.52554142e-01 -1.14824164e+00 2.16115102e-01
2.96946824e-01 -4.26216662e-01 5.74005306e-01 -1.07950306e+00
7.44226396e-01 1.06047846e-01 1.26776677e-02 -1.56826532e+00
-1.72869578e-01 -9.64356065e-01 3.60538542e-01 1.63616788e+00
6.94981933e-01 -5.45699708e-02 3.04993927e-01 1.36454010e+00
-3.99560571e-01 -3.61959368e-01 -6.29435539e-01 -3.90141934e-01
5.43514490e-01 -4.52755272e-01 2.75028795e-01 9.42301989e-01
2.45425358e-01 4.12081540e-01 -3.28462660e-01 1.04825124e-01
2.39715531e-01 -8.90963197e-01 9.49464858e-01 -9.71106470e-01
-3.44601810e-01 1.32809225e-02 -3.10538262e-01 -1.50943443e-01
6.71948446e-03 -6.19755387e-01 2.56056070e-01 -1.04817390e+00
-1.30970120e-01 -1.83031291e-01 2.03633621e-01 6.17023051e-01
-5.31363547e-01 -1.61692947e-01 4.60154921e-01 3.14062268e-01
-7.56643891e-01 1.53883308e-01 9.32165742e-01 -1.51855290e-01
-8.94479528e-02 -2.79107392e-01 -9.12519515e-01 4.31953520e-01
8.44181895e-01 -5.85971177e-01 -5.54814219e-01 -8.66849720e-01
3.54905784e-01 -3.92551422e-02 4.86732483e-01 -9.09747243e-01
1.63159922e-01 6.32700622e-02 2.72503674e-01 -2.94564553e-02
-8.00527260e-02 -7.77334690e-01 -2.61250641e-02 3.25734884e-01
-1.01796007e+00 3.76303375e-01 4.99703020e-01 3.17105949e-01
6.93799704e-02 -5.18598616e-01 3.82723927e-01 -1.00062296e-01
-3.35120708e-01 -2.89655477e-01 -4.07224298e-01 8.70915055e-01
9.28535044e-01 3.62717777e-01 -9.11248922e-01 -6.34170413e-01
-8.84411275e-01 5.19598663e-01 5.33276677e-01 5.58845818e-01
2.80950069e-01 -9.88202572e-01 -1.02237189e+00 -4.81070042e-01
1.90757677e-01 6.57591075e-02 7.19005540e-02 3.80007893e-01
-6.05399489e-01 5.25226854e-02 -2.31903076e-01 9.18213576e-02
-1.28356731e+00 6.91116750e-01 2.68399537e-01 -6.29438639e-01
-2.51156777e-01 8.02996635e-01 -4.58126485e-01 -3.94062757e-01
4.01533306e-01 4.10881639e-01 -2.19924614e-01 8.62637162e-02
5.24412036e-01 4.09714669e-01 1.68165386e-01 -3.29278648e-01
-1.54417679e-01 -1.99306071e-01 -2.92677343e-01 -5.85084319e-01
1.05418360e+00 1.01701312e-01 1.44636378e-01 4.76874679e-01
9.49850142e-01 2.64147937e-01 -9.50756669e-01 -6.54570460e-02
2.12551549e-01 -1.41837329e-01 -5.82154870e-01 -1.50493813e+00
-3.49643052e-01 7.36069262e-01 5.69583535e-01 4.20273304e-01
1.00853586e+00 1.17030941e-01 7.38575578e-01 4.68929231e-01
2.92962193e-01 -1.18360651e+00 3.27210248e-01 5.65615058e-01
1.48282540e+00 -1.13132370e+00 -6.91087916e-03 -5.06844297e-02
-1.19746387e+00 9.86447096e-01 1.04248047e+00 8.35068151e-02
1.89436033e-01 2.14931875e-01 4.93101776e-01 -6.56367600e-01
-8.25405777e-01 3.91532242e-01 2.03570902e-01 9.59603667e-01
4.64859933e-01 -1.08954556e-01 -3.83822471e-01 6.96204185e-01
-6.38172269e-01 -5.95446587e-01 4.95712847e-01 8.19762111e-01
-2.40759104e-02 -1.11157703e+00 -2.38748729e-01 -2.91585401e-02
-7.62609601e-01 -1.40571222e-02 -1.17674601e+00 9.05573845e-01
-2.53463462e-02 1.30967796e+00 -2.09617224e-02 -7.69504368e-01
8.43846679e-01 2.37798095e-01 1.06289044e-01 -5.94128788e-01
-1.24753439e+00 -6.60908639e-01 7.82205641e-01 -5.69643974e-01
-1.02915600e-01 -7.18013227e-01 -8.73983622e-01 -3.97976518e-01
-3.37434262e-01 5.64769030e-01 4.86658394e-01 9.59152341e-01
2.89034247e-01 2.57797182e-01 4.57518101e-01 -4.28359956e-01
-7.62906730e-01 -1.46155465e+00 3.29407185e-01 1.13596368e+00
1.19645773e-02 -1.73121914e-01 -3.82495373e-01 3.36284935e-01] | [6.34376335144043, 8.2019681930542] |
5bcb10d4-13e7-42b9-aac3-5842b285c624 | embodied-multimodal-multitask-learning | 1902.01385 | null | http://arxiv.org/abs/1902.01385v1 | http://arxiv.org/pdf/1902.01385v1.pdf | Embodied Multimodal Multitask Learning | Recent efforts on training visual navigation agents conditioned on language
using deep reinforcement learning have been successful in learning policies for
different multimodal tasks, such as semantic goal navigation and embodied
question answering. In this paper, we propose a multitask model capable of
jointly learning these multimodal tasks, and transferring knowledge of words
and their grounding in visual objects across the tasks. The proposed model uses
a novel Dual-Attention unit to disentangle the knowledge of words in the
textual representations and visual concepts in the visual representations, and
align them with each other. This disentangled task-invariant alignment of
representations facilitates grounding and knowledge transfer across both tasks.
We show that the proposed model outperforms a range of baselines on both tasks
in simulated 3D environments. We also show that this disentanglement of
representations makes our model modular, interpretable, and allows for transfer
to instructions containing new words by leveraging object detectors. | ['Lisa Lee', 'Ruslan Salakhutdinov', 'Dhruv Batra', 'Devendra Singh Chaplot', 'Devi Parikh'] | 2019-02-04 | null | https://openreview.net/forum?id=r1lQQeHYPr | https://openreview.net/pdf?id=r1lQQeHYPr | null | ['embodied-question-answering'] | ['computer-vision'] | [-5.10364398e-02 1.81500882e-01 3.75547307e-03 -2.38841653e-01
-5.93636334e-01 -8.71547341e-01 9.76689398e-01 1.22537352e-01
-7.26816475e-01 5.86711168e-01 5.57420433e-01 -4.43142414e-01
7.33260512e-02 -5.50550461e-01 -1.04634202e+00 -5.57474494e-01
-1.86769500e-01 5.85563302e-01 1.81532986e-02 -4.34620142e-01
3.28841776e-01 2.63060033e-01 -1.58867729e+00 5.84142208e-01
6.00097060e-01 5.56297123e-01 7.37530828e-01 8.64033580e-01
-2.21093640e-01 9.75901306e-01 -4.56361741e-01 -2.67862622e-03
1.31409198e-01 -3.09551805e-01 -1.04967463e+00 -1.47544742e-01
6.62434340e-01 -6.89407289e-01 -5.35781860e-01 6.67530060e-01
1.97008356e-01 5.61961770e-01 9.15980279e-01 -1.34599328e+00
-1.22567296e+00 4.62732792e-01 -4.55004007e-01 7.10919872e-02
5.05508244e-01 3.33241701e-01 1.39678168e+00 -8.36974084e-01
6.31389320e-01 1.51165307e+00 1.12367973e-01 9.13882375e-01
-1.21731615e+00 -6.40038311e-01 6.75811470e-01 3.08500946e-01
-9.57838356e-01 -3.74478519e-01 3.93958658e-01 -6.25212729e-01
1.42658269e+00 -2.03220829e-01 5.42377770e-01 1.53128600e+00
2.39116937e-01 8.17554235e-01 1.25339544e+00 -4.82531697e-01
1.62718400e-01 1.46813469e-03 1.00981109e-02 1.18912947e+00
2.62061447e-01 3.07212502e-01 -9.55967665e-01 1.10754959e-01
1.14087784e+00 -1.28333382e-02 -1.34111628e-01 -9.12864327e-01
-1.47968495e+00 8.83750379e-01 7.70314395e-01 2.39204049e-01
-3.49439591e-01 6.97128415e-01 1.46747783e-01 8.65606219e-02
-8.56347606e-02 5.47300816e-01 -3.53370637e-01 3.61188799e-01
-2.01929972e-01 2.42970660e-01 4.05936480e-01 1.12799501e+00
7.68308818e-01 4.60714027e-02 -4.76712912e-01 4.78712946e-01
9.17482615e-01 8.81346166e-01 5.48463106e-01 -9.88337636e-01
6.42093599e-01 3.33735406e-01 2.35594660e-01 -7.39099741e-01
-6.13032341e-01 -7.10153803e-02 -1.77313656e-01 6.99253738e-01
4.30831730e-01 -1.03373982e-01 -1.39027035e+00 2.29286742e+00
-2.07459982e-02 -9.84012634e-02 3.72360468e-01 1.01563323e+00
9.93166387e-01 5.90461910e-01 6.87970579e-01 5.63935935e-01
1.61003685e+00 -1.17582786e+00 -6.61616862e-01 -6.92677140e-01
5.05258620e-01 -2.36893743e-01 1.21994174e+00 -1.66004851e-01
-9.01251316e-01 -6.96286440e-01 -1.22988427e+00 -5.94539225e-01
-7.03925133e-01 -8.44402611e-02 6.74913108e-01 1.53375164e-01
-1.30383337e+00 1.05760388e-01 -8.19450974e-01 -7.05431640e-01
6.02703631e-01 4.19269323e-01 -5.16509295e-01 8.97303522e-02
-1.04510152e+00 1.38829875e+00 3.83503556e-01 -4.89439853e-02
-1.63087606e+00 -1.67268395e-01 -1.28336608e+00 1.26348406e-01
2.42493525e-01 -1.39378798e+00 1.37624049e+00 -8.42844129e-01
-1.55216944e+00 1.12276149e+00 -2.46423185e-01 -2.67740667e-01
4.59674895e-02 -4.99718755e-01 7.79181123e-02 2.15478569e-01
1.90248296e-01 1.29317915e+00 8.57514679e-01 -1.68463504e+00
-4.77702916e-01 -5.50883651e-01 4.51764017e-01 7.14484990e-01
-1.79382786e-02 -5.61200976e-01 -1.23833992e-01 -3.05981129e-01
-1.71057776e-01 -8.08081567e-01 -1.71046674e-01 1.95666999e-01
-1.92154437e-01 -2.69876961e-02 4.11565602e-01 -5.71300924e-01
3.15395325e-01 -1.87501872e+00 7.84011424e-01 -3.45564373e-02
3.50195974e-01 -2.38543227e-01 -6.72420204e-01 5.35496950e-01
1.22484051e-01 -8.79108682e-02 7.49472007e-02 -5.57375610e-01
2.76735425e-01 5.60337603e-01 -4.91260558e-01 2.77380168e-01
2.66217083e-01 1.45709980e+00 -1.04674220e+00 -6.38859943e-02
2.25343361e-01 4.26799685e-01 -5.28632045e-01 5.99754512e-01
-6.07110262e-01 7.38259673e-01 -3.73922199e-01 2.82440931e-01
2.14396700e-01 -3.19724023e-01 4.27080423e-01 -1.15228370e-01
6.59376234e-02 3.33737910e-01 -6.65744722e-01 2.30451226e+00
-7.81063735e-01 7.48842597e-01 1.33285895e-01 -8.70422006e-01
5.70711315e-01 2.46599555e-01 -2.38804027e-01 -1.14850807e+00
1.88876048e-01 -9.29683596e-02 9.40034166e-02 -7.47591257e-01
4.64269280e-01 -2.40368769e-01 -1.05721146e-01 7.07003236e-01
6.68180525e-01 -8.85107443e-02 -1.17052630e-01 3.70536804e-01
5.95458388e-01 7.35144436e-01 2.81253546e-01 -1.70713678e-01
2.54468918e-01 -9.61063579e-02 -3.65280747e-01 9.79188502e-01
-7.45312944e-02 2.36000314e-01 2.15737864e-01 -1.98341042e-01
-9.50673580e-01 -1.41527748e+00 4.77807552e-01 1.79602039e+00
3.62092733e-01 -1.02686897e-01 -3.49893749e-01 -6.29540443e-01
1.34898737e-01 9.71171796e-01 -1.05023932e+00 -3.18087697e-01
-4.28816706e-01 -8.92528892e-02 4.79159802e-01 8.03215146e-01
1.16259195e-01 -1.27942061e+00 -9.77104485e-01 -2.01434150e-01
-4.04576957e-02 -1.15425932e+00 -1.26035407e-01 3.66737485e-01
-6.51050746e-01 -1.04216194e+00 -7.73006141e-01 -1.13068402e+00
7.66026914e-01 5.41718423e-01 1.22973275e+00 5.34295216e-02
3.59414294e-02 1.11182010e+00 -1.66063294e-01 -2.80003786e-01
-2.97737211e-01 7.51553178e-02 -2.18931794e-01 -1.44206956e-01
1.30715668e-01 -4.09432173e-01 -5.24239600e-01 -7.97563582e-04
-9.11935329e-01 3.45757812e-01 6.52930737e-01 9.00532365e-01
1.81730255e-01 -1.22217929e+00 4.62800145e-01 -4.90287453e-01
8.50910187e-01 -5.68743527e-01 -5.62896788e-01 3.88593584e-01
-5.95130958e-02 6.73897803e-01 8.44627842e-02 -3.02521735e-01
-1.05650449e+00 -1.59361109e-01 -5.42485639e-02 -1.16556436e-01
-4.90011841e-01 3.98654282e-01 -7.10110068e-02 4.93148081e-02
5.07294774e-01 1.03452161e-01 1.12268791e-01 -1.44432366e-01
1.27547765e+00 1.17511205e-01 4.31569993e-01 -8.06992590e-01
5.99413335e-01 6.47025704e-01 -7.76648801e-03 -6.10230148e-01
-6.60623848e-01 -1.84960917e-01 -7.18074441e-01 3.89306583e-02
1.44456625e+00 -1.10224712e+00 -1.14400864e+00 1.09017618e-01
-1.39864349e+00 -7.57491291e-01 -1.60376579e-01 4.37238008e-01
-9.17827487e-01 5.03011383e-02 -3.77455413e-01 -5.92486620e-01
1.48557901e-01 -1.28122294e+00 1.23672318e+00 3.67293656e-01
-3.75523388e-01 -1.27127981e+00 2.71772504e-01 2.22437367e-01
3.78189743e-01 -9.66192782e-02 1.34866965e+00 -7.64753103e-01
-8.25616241e-01 3.60760838e-01 -4.38921660e-01 -2.86854208e-01
5.10884784e-02 -5.06476998e-01 -1.05425429e+00 -3.30107778e-01
-3.77553135e-01 -9.57724154e-01 1.20604205e+00 2.28371546e-01
7.19874799e-01 -3.84717025e-02 -5.32855690e-01 5.21332145e-01
1.18425941e+00 2.79893488e-01 3.82146478e-01 4.77521211e-01
7.72992432e-01 7.58522928e-01 1.01319559e-01 -4.73151542e-02
8.47119689e-01 6.04919493e-01 8.80851567e-01 -1.90042689e-01
-1.68167844e-01 -3.46040726e-01 3.97128493e-01 3.72673661e-01
-4.61813137e-02 -2.84289956e-01 -1.00673568e+00 5.02264977e-01
-1.95368195e+00 -8.53245378e-01 4.41437483e-01 1.74172390e+00
6.47037923e-01 -1.58038944e-01 -6.98261037e-02 -9.50879633e-01
3.09949309e-01 3.68668765e-01 -6.05344296e-01 -4.57962871e-01
2.93044583e-03 3.64441931e-01 1.37623787e-01 9.05843079e-01
-9.83983755e-01 1.18788970e+00 6.57213306e+00 -9.35427099e-03
-7.75394738e-01 2.84760267e-01 1.01423338e-01 -5.11840247e-02
-5.22921324e-01 -1.46648765e-01 -6.32707179e-01 -4.57212925e-01
6.19341254e-01 1.59471378e-01 4.22612578e-01 4.22297150e-01
-2.29268953e-01 -2.02238128e-01 -1.51316190e+00 7.86564291e-01
5.42025387e-01 -1.32375646e+00 4.25880820e-01 -8.89362693e-02
6.32845998e-01 8.84716660e-02 4.79797333e-01 5.61751068e-01
7.35688031e-01 -1.51907361e+00 8.39156747e-01 6.36733949e-01
5.12022257e-01 -2.95194834e-01 3.70372266e-01 1.61108926e-01
-8.63083124e-01 -1.37165815e-01 6.38601109e-02 -2.42905587e-01
2.58485317e-01 -7.91078091e-01 -7.52851069e-01 4.12147403e-01
3.49765331e-01 5.44273019e-01 -5.15362501e-01 6.21224940e-01
-5.50466418e-01 -8.35949183e-02 2.82015890e-01 -2.86382049e-01
4.36096013e-01 -6.95571303e-02 3.44683319e-01 1.05063426e+00
1.46294326e-01 9.43957083e-03 2.27115378e-01 1.00982618e+00
-4.70129177e-02 -7.67780691e-02 -8.54307890e-01 -9.33075473e-02
1.37053847e-01 9.09030259e-01 -5.58411062e-01 -3.62423599e-01
-5.88829219e-01 1.14495921e+00 8.76660645e-01 9.52550352e-01
-7.56795585e-01 -1.28124624e-01 1.07495153e+00 -4.63984847e-01
5.53463459e-01 -9.03882444e-01 -3.31717938e-01 -1.17504239e+00
-4.21558768e-01 -6.14549875e-01 1.82200775e-01 -1.14800489e+00
-1.11434567e+00 8.33268821e-01 4.03384529e-02 -7.92734385e-01
-2.73806512e-01 -1.27048743e+00 -2.44563699e-01 1.12500966e+00
-1.65160894e+00 -1.59365463e+00 -3.14861983e-01 7.39426136e-01
5.66468716e-01 -2.56029695e-01 1.17764354e+00 -3.11347455e-01
-1.73239738e-01 3.75526816e-01 -1.90031797e-01 8.59455839e-02
7.28286982e-01 -1.23633075e+00 3.88237655e-01 4.45235968e-01
5.99047661e-01 9.20417190e-01 6.14446580e-01 -4.52094406e-01
-1.44776797e+00 -5.95679283e-01 2.55540907e-01 -9.02299702e-01
5.95300198e-01 -8.03171635e-01 -9.02061403e-01 1.05166507e+00
1.02670896e+00 -3.49639118e-01 8.84157956e-01 3.53905022e-01
-1.01283526e+00 3.89476597e-01 -5.55201530e-01 9.08951700e-01
1.12604809e+00 -9.91168320e-01 -1.09544337e+00 3.63252789e-01
7.98040330e-01 -5.15622497e-01 -1.02487020e-01 -2.67572273e-02
7.19044328e-01 -8.21826041e-01 1.08273649e+00 -1.44475150e+00
4.86031204e-01 -2.46173427e-01 -3.73322070e-01 -1.49825048e+00
-3.72096002e-01 -4.74038832e-02 -1.85669318e-01 6.96895361e-01
5.76064587e-01 -4.77350384e-01 4.11837727e-01 3.00501972e-01
-1.41295776e-01 -2.91475117e-01 -9.10978079e-01 -3.44117999e-01
3.36705595e-01 -1.66822389e-01 1.30777746e-01 7.59911478e-01
1.13414057e-01 9.11991954e-01 -2.88856417e-01 3.51469874e-01
2.99639970e-01 2.86037266e-01 7.06502795e-01 -8.90705228e-01
-3.61478597e-01 -5.73020339e-01 -7.84202814e-02 -1.45958221e+00
6.17664456e-01 -1.27395952e+00 2.90724069e-01 -2.08937573e+00
2.02763557e-01 1.09762950e-02 -2.69588411e-01 7.16389537e-01
-1.31647915e-01 -4.71674390e-02 4.34623271e-01 -8.16523731e-02
-7.42712080e-01 7.73140550e-01 1.39867210e+00 -3.48250002e-01
4.98258462e-03 -7.21027672e-01 -7.27460206e-01 6.05577230e-01
5.48613906e-01 -1.75532520e-01 -5.07323861e-01 -1.11172271e+00
2.36760333e-01 -2.37393901e-01 7.54772007e-01 -6.19743884e-01
1.66943654e-01 -1.69702753e-01 6.08468652e-01 -2.28855953e-01
6.15832329e-01 -8.64573419e-01 -5.95246255e-01 5.48454583e-01
-7.06906021e-01 3.39759618e-01 7.21834421e-01 8.12680244e-01
1.10194184e-01 1.10773910e-02 4.54393655e-01 -3.63031656e-01
-1.18361771e+00 -4.44247238e-02 -7.21252441e-01 6.69369325e-02
1.01620233e+00 2.34180782e-02 -6.84076011e-01 -5.54039180e-01
-1.06317067e+00 5.57672381e-01 2.32870251e-01 7.67855108e-01
8.18178296e-01 -1.25833464e+00 -4.83844459e-01 1.41235203e-01
2.90330529e-01 -3.86085719e-01 2.58551687e-01 3.99604708e-01
-4.66665477e-01 6.94234312e-01 -7.73520231e-01 -6.34481072e-01
-9.28082466e-01 7.31456995e-01 4.65976864e-01 -2.37958669e-03
-1.72257498e-01 1.00777125e+00 7.63787925e-01 -6.81208432e-01
3.60512018e-01 -3.49207699e-01 -4.60413069e-01 8.60958397e-02
4.32088554e-01 -2.09264576e-01 -3.61066461e-01 -7.29710579e-01
-3.43543291e-01 7.28405535e-01 -2.29599461e-01 -6.22077167e-01
1.09165680e+00 -3.30410153e-01 7.36069605e-02 6.54653132e-01
9.68469501e-01 -2.00419500e-01 -1.38181388e+00 -1.63193002e-01
-1.12506136e-01 -1.08201593e-01 -3.56184155e-01 -9.65590954e-01
-4.69408184e-01 1.29070580e+00 5.66763222e-01 -1.18240044e-01
6.56513512e-01 5.00910103e-01 2.16699481e-01 7.37834156e-01
2.52748638e-01 -5.94992638e-01 7.96393216e-01 9.26413596e-01
1.21224916e+00 -1.37198591e+00 -3.61123353e-01 -2.78613809e-03
-8.14874709e-01 1.05798221e+00 1.03680301e+00 -1.73577309e-01
1.86507031e-01 -2.18783304e-01 3.16803902e-01 -3.01535517e-01
-8.91051233e-01 -7.64100969e-01 4.73832607e-01 1.06171155e+00
2.85872579e-01 -1.42246574e-01 2.67444551e-01 3.56750399e-01
1.61149904e-01 -6.27788007e-01 2.20881015e-01 1.06382096e+00
-5.94494402e-01 -8.41995478e-01 -1.97704196e-01 -1.69232577e-01
1.27386376e-01 -3.17321569e-01 -3.78497839e-01 8.88247728e-01
8.11507832e-03 6.13313377e-01 2.61827618e-01 -1.10114843e-01
3.87812048e-01 3.43302816e-01 9.59746003e-01 -8.90543938e-01
-3.51621360e-01 -1.70943946e-01 2.07309108e-02 -5.21143854e-01
-4.93916839e-01 -3.70072991e-01 -1.55638850e+00 4.78299677e-01
8.42100382e-02 8.08313675e-03 7.07831860e-01 1.18250346e+00
5.02516091e-01 7.93979168e-01 -1.05149694e-01 -1.05986857e+00
-4.47242916e-01 -7.83755600e-01 -1.51097745e-01 4.49224114e-01
6.12743199e-01 -1.14358544e+00 4.31533158e-02 4.29153182e-02] | [4.536915302276611, 0.5586137771606445] |
41b31e7c-53eb-4534-bc09-3b725ac9dc75 | farewell-to-mutual-information-variational | 2104.02862 | null | https://arxiv.org/abs/2104.02862v2 | https://arxiv.org/pdf/2104.02862v2.pdf | Farewell to Mutual Information: Variational Distillation for Cross-Modal Person Re-Identification | The Information Bottleneck (IB) provides an information theoretic principle for representation learning, by retaining all information relevant for predicting label while minimizing the redundancy. Though IB principle has been applied to a wide range of applications, its optimization remains a challenging problem which heavily relies on the accurate estimation of mutual information. In this paper, we present a new strategy, Variational Self-Distillation (VSD), which provides a scalable, flexible and analytic solution to essentially fitting the mutual information but without explicitly estimating it. Under rigorously theoretical guarantee, VSD enables the IB to grasp the intrinsic correlation between representation and label for supervised training. Furthermore, by extending VSD to multi-view learning, we introduce two other strategies, Variational Cross-Distillation (VCD) and Variational Mutual-Learning (VML), which significantly improve the robustness of representation to view-changes by eliminating view-specific and task-irrelevant information. To verify our theoretically grounded strategies, we apply our approaches to cross-modal person Re-ID, and conduct extensive experiments, where the superior performance against state-of-the-art methods are demonstrated. Our intriguing findings highlight the need to rethink the way to estimate mutual | ['Lizhuang Ma', 'Yuan Xie', 'Yanyun Qu', 'Shaohui Lin', 'Zhizhong Zhang', 'Xudong Tian'] | 2021-04-07 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Tian_Farewell_to_Mutual_Information_Variational_Distillation_for_Cross-Modal_Person_Re-Identification_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Tian_Farewell_to_Mutual_Information_Variational_Distillation_for_Cross-Modal_Person_Re-Identification_CVPR_2021_paper.pdf | cvpr-2021-1 | ['cross-view-person-re-identification', 'multi-view-learning'] | ['computer-vision', 'computer-vision'] | [ 2.95350820e-01 2.97536030e-02 -4.39898908e-01 -4.37573373e-01
-8.11421573e-01 -4.30386424e-01 6.16497338e-01 1.34170562e-01
-2.28193477e-01 6.14730656e-01 5.33215404e-01 2.15674210e-02
-4.23483491e-01 -3.74833673e-01 -4.13988233e-01 -7.10347652e-01
3.45561840e-02 3.89493942e-01 -1.43712178e-01 2.39599193e-03
1.70238316e-01 1.61864534e-01 -1.69491041e+00 2.86123514e-01
9.36844528e-01 7.95299292e-01 1.70780912e-01 8.40635523e-02
-1.79777488e-01 7.96762764e-01 -3.06564718e-01 -5.87845087e-01
1.93127841e-01 -5.26657879e-01 -1.01352477e+00 2.20034085e-02
2.04684466e-01 -9.99264717e-02 -3.30603689e-01 8.76217246e-01
6.84955657e-01 3.00829023e-01 1.01845062e+00 -1.26922405e+00
-6.62043571e-01 5.13114333e-01 -8.47487688e-01 2.73717102e-02
4.72889364e-01 -2.90854484e-01 1.27095628e+00 -7.85278618e-01
5.92835724e-01 1.32221878e+00 6.61768258e-01 7.56265461e-01
-1.45363557e+00 -6.28662229e-01 3.94851983e-01 3.41588378e-01
-1.40020168e+00 -4.91260588e-01 9.81081367e-01 -6.58900917e-01
4.64014739e-01 2.63807714e-01 3.13716471e-01 1.16827822e+00
-2.11786583e-01 1.13786197e+00 1.34320557e+00 -5.74010313e-01
1.76993906e-02 4.26550448e-01 3.32032770e-01 6.55157208e-01
1.57086402e-01 1.27288640e-01 -6.70608938e-01 -1.86976358e-01
6.25357449e-01 2.08904877e-01 -4.12437588e-01 -8.08564067e-01
-1.28084397e+00 7.68326700e-01 3.63186151e-01 2.16582552e-01
-6.65054619e-02 -2.01877713e-01 5.25050223e-01 2.76919276e-01
5.77052057e-01 3.16926062e-01 -4.46338534e-01 1.49322659e-01
-8.99600327e-01 1.32138103e-01 6.72294855e-01 9.14813459e-01
8.10083568e-01 -3.94131154e-01 -4.62094396e-01 1.10270572e+00
3.00454557e-01 3.62009346e-01 4.08671141e-01 -1.06852376e+00
3.35919619e-01 6.72596335e-01 -1.07642636e-01 -1.03914750e+00
-4.91392016e-01 -5.73999941e-01 -1.17995286e+00 -2.83784658e-01
4.09217834e-01 1.50854960e-01 -4.10252035e-01 2.07748723e+00
3.31554264e-01 -1.44665718e-01 -1.54568315e-01 7.60008335e-01
8.45576704e-01 1.72381163e-01 1.17174489e-02 -5.32335818e-01
1.26629210e+00 -8.47095370e-01 -8.04815054e-01 -8.28057341e-03
7.90539086e-01 -4.17063951e-01 8.51820707e-01 1.73850030e-01
-9.24791873e-01 -5.89313030e-01 -9.11321759e-01 -1.67283133e-01
-2.10553989e-01 -1.65869445e-02 7.89293051e-01 5.33240318e-01
-9.01279151e-01 6.25812352e-01 -5.34651339e-01 -4.06177461e-01
3.88720632e-01 2.70963043e-01 -6.82526886e-01 -2.01645106e-01
-1.22569156e+00 8.67244124e-01 4.20427352e-01 -1.57465011e-01
-6.31137192e-01 -6.91236436e-01 -8.21971297e-01 -3.15444134e-02
4.88730222e-01 -9.52253044e-01 8.90489459e-01 -7.56877601e-01
-1.23469079e+00 1.04432547e+00 -5.45651078e-01 -1.12162687e-01
6.21853292e-01 -1.64572403e-01 -2.21891552e-01 3.05932909e-02
2.33008936e-01 6.09509826e-01 7.41472781e-01 -1.55733204e+00
-2.19890043e-01 -6.54094279e-01 5.10013103e-02 3.63123477e-01
-4.70204026e-01 -2.50624537e-01 -4.57216054e-01 -5.23597836e-01
4.80297476e-01 -8.98621023e-01 -1.99357588e-02 -6.52724802e-02
-4.12864268e-01 -4.57987398e-01 3.99708331e-01 -6.17053568e-01
1.36848950e+00 -2.11860609e+00 4.19046730e-01 1.24757245e-01
5.16752064e-01 1.86742339e-02 -1.41089037e-01 5.85322201e-01
-9.94870961e-02 1.73799664e-01 -2.56706297e-01 -7.59209752e-01
7.67766386e-02 2.03072891e-01 -1.44856423e-01 5.29163778e-01
-1.65207192e-01 9.66708183e-01 -9.87583458e-01 -7.16060817e-01
1.72166243e-01 5.73693931e-01 -7.68300056e-01 2.26858631e-01
1.27052590e-01 7.92552769e-01 -3.12752753e-01 4.65279609e-01
7.22830296e-01 -5.64581513e-01 4.51773494e-01 -4.19937253e-01
9.82754156e-02 2.29087532e-01 -1.11324382e+00 1.87320566e+00
-3.46677333e-01 4.22385842e-01 -1.17894262e-01 -1.28000402e+00
6.67934000e-01 8.02517831e-02 8.43450367e-01 -6.93794370e-01
-6.59462884e-02 -1.44889742e-01 -4.33683902e-01 -3.81601751e-01
2.32350960e-01 -2.53362447e-01 8.87414720e-03 6.00206137e-01
2.04798311e-01 4.81821984e-01 -6.36348175e-03 4.02446121e-01
6.52904510e-01 1.56952605e-01 5.78307629e-01 -3.14672828e-01
7.16937900e-01 -5.98308623e-01 6.65124536e-01 9.00919378e-01
-4.43835676e-01 5.67861915e-01 4.82901365e-01 -9.14680958e-02
-8.01082492e-01 -1.07836962e+00 -3.04937601e-01 1.19704878e+00
2.57101536e-01 -7.07378924e-01 -7.72272646e-01 -8.76783848e-01
-2.21937224e-02 6.60178542e-01 -7.59330869e-01 -2.25447476e-01
-2.09235579e-01 -7.62335420e-01 2.92828202e-01 3.33039999e-01
6.50090158e-01 -5.50475419e-01 -2.12308746e-02 -1.00895733e-01
-8.20297956e-01 -9.08250093e-01 -4.92067665e-01 -8.38598385e-02
-7.86320686e-01 -1.00272703e+00 -7.86340237e-01 -4.09438372e-01
5.13190508e-01 7.32349575e-01 1.22132456e+00 8.13037530e-02
-1.21204421e-01 7.52206624e-01 -2.74943233e-01 -1.34118199e-01
-1.84417024e-01 1.28997952e-01 2.42341265e-01 1.75529104e-02
5.11614025e-01 -6.94055080e-01 -7.45618403e-01 4.82661426e-01
-5.55154264e-01 2.17943683e-01 5.87477624e-01 9.38374817e-01
6.36335969e-01 -3.85401286e-02 5.68109572e-01 -9.92924750e-01
4.75315541e-01 -5.47535598e-01 -7.99677819e-02 4.50779587e-01
-9.00353193e-01 4.36730832e-01 2.43400827e-01 -1.40448183e-01
-1.18873668e+00 -1.33432969e-01 7.40868449e-02 -3.17979991e-01
1.96300689e-02 3.59495521e-01 -3.07912290e-01 9.71753895e-02
4.71245766e-01 4.55121845e-01 2.33376399e-01 -6.38325632e-01
7.13871360e-01 5.84111929e-01 3.54508907e-01 -6.11894667e-01
5.24322748e-01 7.43136764e-01 -8.28969385e-03 -7.32580125e-01
-1.43824911e+00 -8.79189134e-01 -1.08664906e+00 -2.89076746e-01
7.79547691e-01 -1.00967360e+00 -1.03969145e+00 1.11998662e-01
-1.02805185e+00 2.27163196e-01 -1.91279083e-01 3.22903872e-01
-6.78599179e-01 6.61824942e-01 -2.96028048e-01 -9.45640206e-01
-1.96746573e-01 -9.14459109e-01 1.04217279e+00 -1.36015832e-01
-1.61216617e-01 -1.25799739e+00 3.28305632e-01 7.69747913e-01
1.47822484e-01 -2.83048991e-02 8.85384560e-01 -6.50921822e-01
-3.91046941e-01 8.04327428e-02 -4.98425126e-01 3.81775767e-01
2.12865114e-01 -6.39313757e-01 -1.12451553e+00 -5.39453983e-01
2.61093006e-02 -3.18016261e-01 1.26217675e+00 3.11352998e-01
1.19334877e+00 -2.44536474e-01 -5.49866199e-01 6.77861512e-01
1.25381351e+00 -4.31988955e-01 3.83956701e-01 2.00311735e-01
9.70719457e-01 9.09376860e-01 2.63861656e-01 6.04146242e-01
7.94241369e-01 8.14824402e-01 1.55974135e-01 2.35883743e-02
-1.85384959e-01 -5.05173743e-01 2.73380131e-01 9.89777565e-01
-3.57830703e-01 -2.23059095e-02 -4.13265526e-01 2.10419774e-01
-2.03011084e+00 -1.24006283e+00 6.12560399e-02 2.52270150e+00
8.21012139e-01 -2.99534410e-01 2.83606797e-01 1.22720897e-01
7.09770739e-01 3.07167143e-01 -5.75011075e-01 2.31886491e-01
-1.65212303e-01 -4.04716015e-01 2.89110005e-01 4.10710454e-01
-1.08788073e+00 7.26269126e-01 6.95662260e+00 8.36444795e-01
-5.74425638e-01 3.05873722e-01 3.47163826e-01 -2.06213836e-02
-4.75421131e-01 -7.01979995e-02 -8.11088741e-01 3.70800793e-01
7.32569337e-01 -6.29487783e-02 4.93692726e-01 6.59883976e-01
6.14670254e-02 9.45084020e-02 -1.34541810e+00 1.28912151e+00
2.41356850e-01 -1.03510213e+00 1.59628198e-01 1.76961169e-01
7.38859594e-01 -1.98006496e-01 8.13728422e-02 4.58013415e-01
1.56847537e-01 -8.06502044e-01 4.38115805e-01 7.71089017e-01
7.25236714e-01 -7.43859649e-01 5.84391236e-01 5.28124213e-01
-1.24483073e+00 -9.77404565e-02 -5.41082621e-01 1.44498497e-02
4.35420498e-02 7.46431947e-01 -3.90587837e-01 8.08567762e-01
4.55886453e-01 1.02469754e+00 -6.42001569e-01 5.56064069e-01
-7.94970617e-02 3.22666585e-01 3.65738459e-02 3.35770547e-01
-2.45109677e-01 -2.90351719e-01 5.34065962e-01 1.19789958e+00
-4.87537235e-02 -3.06972302e-02 2.23126307e-01 8.46826434e-01
5.65333990e-03 2.55703181e-01 -6.67889714e-01 1.87006235e-01
5.15430927e-01 1.01680243e+00 -4.04447943e-01 -2.16939509e-01
-4.63290066e-01 1.23834026e+00 8.19970787e-01 3.92421663e-01
-5.29746473e-01 1.20857112e-01 6.81824565e-01 1.08886227e-01
1.52236432e-01 -1.13192610e-01 -2.00698599e-01 -1.38637578e+00
4.71439771e-02 -7.00675547e-01 6.35290861e-01 -2.24508956e-01
-1.73023677e+00 1.43987000e-01 1.82663843e-01 -1.29651451e+00
-3.25465530e-01 -3.82688075e-01 -9.76269692e-02 7.10512877e-01
-1.56679773e+00 -1.21314287e+00 -2.17979833e-01 7.51787245e-01
3.03946733e-01 -2.08866194e-01 7.85085201e-01 3.82322460e-01
-6.39481544e-01 9.68663216e-01 3.92718226e-01 -1.82002727e-02
8.70667875e-01 -1.19618344e+00 -1.94718271e-01 3.96296263e-01
1.61711290e-01 8.94556463e-01 5.94629705e-01 -3.54594201e-01
-1.28823841e+00 -9.07697976e-01 9.22167301e-01 -6.72549188e-01
4.17360812e-01 -4.68190849e-01 -8.62958908e-01 7.29182124e-01
-7.06545636e-02 -1.66283920e-01 1.12683189e+00 5.83820999e-01
-7.66040742e-01 -1.00864924e-01 -1.00507760e+00 4.49466914e-01
1.50250506e+00 -7.92243898e-01 -5.87450445e-01 4.87688690e-01
7.82770932e-01 1.39324591e-01 -9.57443535e-01 3.92541796e-01
7.43033409e-01 -1.39775920e+00 1.23247099e+00 -6.71162009e-01
2.93425828e-01 -6.89637735e-02 -3.88681740e-01 -1.18055105e+00
-6.36535645e-01 -5.32060802e-01 -3.43411207e-01 1.54798913e+00
6.90222308e-02 -5.83168328e-01 4.97954816e-01 6.04711115e-01
1.82369322e-01 -6.20685101e-01 -8.27559650e-01 -9.07466054e-01
1.37978524e-01 -4.26967889e-01 3.91935050e-01 1.17989266e+00
2.83744156e-01 4.39041764e-01 -8.46302986e-01 -8.34019929e-02
1.07049334e+00 8.47434178e-02 6.23604774e-01 -1.51436484e+00
-3.34356040e-01 -4.23594952e-01 -3.76228184e-01 -1.39775503e+00
4.95129585e-01 -1.26910377e+00 -2.46590376e-01 -1.48323405e+00
8.59088778e-01 -2.57828921e-01 -7.08599627e-01 9.43477228e-02
-4.15406793e-01 -1.35292355e-02 3.03280592e-01 7.31085539e-01
-9.39165413e-01 7.77911484e-01 1.26243472e+00 -1.49507672e-01
-9.04329047e-02 9.21131968e-02 -1.05752885e+00 6.81716979e-01
4.65230227e-01 -3.96795690e-01 -6.01650774e-01 -1.98889077e-01
1.83299109e-01 -6.83451220e-02 3.27872366e-01 -6.32982731e-01
1.81989342e-01 2.08307244e-02 3.48062098e-01 -6.41994238e-01
2.28251070e-01 -6.26072764e-01 -1.11626729e-01 2.43016914e-01
-6.88658655e-01 -2.93414176e-01 -2.67578870e-01 9.78763759e-01
-1.22283392e-01 -1.50551900e-01 6.29892886e-01 -2.33746246e-01
-6.75217271e-01 3.87455940e-01 1.15147568e-01 2.24343508e-01
7.64684856e-01 -9.51386467e-02 -1.29389241e-01 -4.49174106e-01
-7.74697781e-01 3.03659976e-01 3.91780525e-01 2.90986001e-01
4.77222532e-01 -1.57792103e+00 -7.31872559e-01 2.13152528e-01
4.59964991e-01 -4.38831627e-01 5.99881530e-01 9.32790875e-01
3.37484330e-01 6.66539252e-01 8.11246112e-02 -7.15220988e-01
-1.06337905e+00 8.98562610e-01 1.01849660e-01 -5.21742821e-01
-6.72057867e-01 7.86246121e-01 6.06211364e-01 -6.55499697e-01
2.47711644e-01 3.71040285e-01 -4.62844521e-01 4.07896221e-01
6.88115001e-01 6.46777749e-01 -1.76535964e-01 -7.53693700e-01
-3.58977079e-01 5.81598699e-01 -2.93247461e-01 -1.51691392e-01
1.01514459e+00 -7.20750868e-01 -2.18194261e-01 7.07113266e-01
1.45058191e+00 -1.99307710e-01 -1.18697548e+00 -6.61380589e-01
1.04895182e-01 -4.57931131e-01 1.27299920e-01 -5.15631258e-01
-9.26014841e-01 9.02237356e-01 6.02841377e-01 1.08875163e-01
9.51306999e-01 2.57985264e-01 4.80998516e-01 3.96832347e-01
5.80064297e-01 -1.07236910e+00 2.14826643e-01 3.20408851e-01
8.18823695e-01 -1.58913898e+00 2.17089906e-01 -4.67600971e-01
-7.46508121e-01 7.14264810e-01 3.94819051e-01 3.41132522e-01
8.36364269e-01 -2.84992158e-01 -2.67394841e-01 -1.67692930e-01
-7.54641652e-01 -3.45558524e-01 6.84604645e-01 7.89614081e-01
6.64775312e-01 1.70732930e-01 -3.19220662e-01 6.46190107e-01
-8.66013989e-02 -2.39537820e-01 -8.32653940e-02 5.86206794e-01
-3.61717016e-01 -1.06646144e+00 -6.10362925e-02 4.14004833e-01
-2.62119889e-01 -8.55026096e-02 -3.82742345e-01 7.30382562e-01
3.20505239e-02 8.86111140e-01 -1.15424097e-01 -4.07744229e-01
2.17066377e-01 2.50514358e-01 6.99069321e-01 -3.45008254e-01
-6.26685843e-02 -9.95140225e-02 -1.83010325e-01 -6.61218822e-01
-7.76474237e-01 -9.39450264e-01 -8.61131012e-01 -3.11048567e-01
-4.12100643e-01 3.38709988e-02 4.28689152e-01 1.17918527e+00
4.71119702e-01 4.20801759e-01 7.56443620e-01 -6.88472271e-01
-7.29805052e-01 -8.95735681e-01 -6.89039886e-01 7.41625130e-01
3.45426619e-01 -1.10948694e+00 -5.36718726e-01 -1.02231398e-01] | [8.529748916625977, 4.360558032989502] |
fa3742cb-0f44-46db-9fe3-3e977ad1c8ed | social-and-scene-aware-trajectory-prediction | 1909.0884 | null | https://arxiv.org/abs/1909.08840v1 | https://arxiv.org/pdf/1909.08840v1.pdf | Social and Scene-Aware Trajectory Prediction in Crowded Spaces | Mimicking human ability to forecast future positions or interpret complex interactions in urban scenarios, such as streets, shopping malls or squares, is essential to develop socially compliant robots or self-driving cars. Autonomous systems may gain advantage on anticipating human motion to avoid collisions or to naturally behave alongside people. To foresee plausible trajectories, we construct an LSTM (long short-term memory)-based model considering three fundamental factors: people interactions, past observations in terms of previously crossed areas and semantics of surrounding space. Our model encompasses several pooling mechanisms to join the above elements defining multiple tensors, namely social, navigation and semantic tensors. The network is tested in unstructured environments where complex paths emerge according to both internal (intentions) and external (other people, not accessible areas) motivations. As demonstrated, modeling paths unaware of social interactions or context information, is insufficient to correctly predict future positions. Experimental results corroborate the effectiveness of the proposed framework in comparison to LSTM-based models for human path prediction. | ['Lamberto Ballan', 'Matteo Lisotto', 'Pasquale Coscia'] | 2019-09-19 | null | null | null | null | ['social-navigation'] | ['robots'] | [-1.14618316e-02 4.72210407e-01 2.41767168e-02 -5.02632320e-01
2.56484091e-01 -1.50193796e-01 1.02259254e+00 8.39748308e-02
-7.46741891e-01 7.79119551e-01 5.69069266e-01 -3.52267981e-01
-4.52115536e-01 -1.03389299e+00 -6.72146916e-01 -3.19254398e-01
-7.41137624e-01 7.45211244e-01 3.13399017e-01 -6.33923531e-01
1.49973452e-01 7.30962098e-01 -1.76442027e+00 2.80899405e-01
5.68907082e-01 5.42279124e-01 4.62123781e-01 7.41745234e-01
-1.02770187e-01 1.03153086e+00 1.00257143e-01 -3.19560111e-01
4.10665870e-02 2.07509056e-01 -7.48676598e-01 1.60568580e-02
-2.52650142e-01 -4.65492129e-01 -6.53825283e-01 5.65711856e-01
-1.00110836e-01 5.66148937e-01 6.57400608e-01 -1.42462361e+00
-3.28214109e-01 8.71668100e-01 7.34089687e-02 -9.09242779e-03
5.12037456e-01 4.55650091e-01 7.74315774e-01 -7.46742368e-01
8.08392704e-01 1.54458785e+00 7.77058899e-01 9.42976177e-02
-8.67941797e-01 -9.44285393e-02 7.09073067e-01 4.50112939e-01
-1.16970885e+00 -3.19567978e-01 6.46781802e-01 -4.57546026e-01
1.10352278e+00 9.10443962e-02 7.68685400e-01 1.64685893e+00
7.02020824e-01 8.18634033e-01 7.60108292e-01 1.39252171e-01
1.68244705e-01 1.04319550e-01 1.46008849e-01 5.61851323e-01
1.50767148e-01 1.90112531e-01 -5.39765000e-01 -1.28972203e-01
4.90676641e-01 2.20541999e-01 2.53103644e-01 -4.23767239e-01
-1.87656796e+00 4.71236736e-01 9.13203895e-01 3.86081994e-01
-1.00785470e+00 3.26821536e-01 1.40607938e-01 5.61157241e-02
1.35264277e-01 1.83858126e-01 -4.28998649e-01 -3.38991433e-01
-4.90519941e-01 6.11855268e-01 7.42325008e-01 1.12605083e+00
8.74772131e-01 -2.31202058e-02 -1.99898332e-01 3.07773381e-01
3.88012707e-01 5.94506800e-01 3.52263451e-01 -9.74837124e-01
6.16129696e-01 6.94956303e-01 6.09089077e-01 -1.53030932e+00
-1.07827246e+00 -3.64651203e-01 -8.64442885e-01 -6.06045686e-02
6.25837624e-01 -1.98331878e-01 -7.27159798e-01 1.65938318e+00
1.68410227e-01 1.58538803e-01 1.17243640e-01 9.22690928e-01
2.82840639e-01 5.07349908e-01 5.06483793e-01 2.49754220e-01
1.28847456e+00 -8.57511401e-01 -6.10967278e-01 -5.41234255e-01
9.75272059e-01 -2.31062740e-01 6.37225926e-01 1.65348068e-01
-8.20455730e-01 -6.70333862e-01 -6.78292871e-01 -4.01804037e-03
-9.05671060e-01 1.87195912e-01 8.29842329e-01 2.67567307e-01
-1.24806201e+00 8.98679793e-01 -1.01467752e+00 -9.90450203e-01
2.97331423e-01 4.26536530e-01 -2.64534920e-01 2.13397160e-01
-1.42561865e+00 9.85166490e-01 3.34751874e-01 6.23702109e-01
-7.20679641e-01 -8.90542045e-02 -7.47255862e-01 -4.53719348e-02
1.36165619e-01 -9.33991075e-01 7.40553796e-01 -4.76887226e-01
-1.16095507e+00 4.08706754e-01 -2.35051587e-01 -8.43475580e-01
9.98148441e-01 -3.54913980e-01 -4.26634938e-01 -3.92579883e-01
1.66098639e-01 1.18519402e+00 3.89470786e-01 -1.09434533e+00
-8.01716566e-01 -4.66503501e-01 2.03785554e-01 6.07704520e-01
-1.77672356e-01 -5.45370698e-01 -6.71982840e-02 -2.01752409e-01
1.94688335e-01 -1.43437672e+00 -9.56528306e-01 4.17122210e-04
-6.61577225e-01 -1.77713796e-01 7.52447605e-01 -4.68608826e-01
9.47650671e-01 -1.70362031e+00 8.91165286e-02 4.57047641e-01
1.58138111e-01 -9.19527188e-02 -1.17881179e-01 5.95813215e-01
4.77528900e-01 -1.72713250e-01 -1.42727951e-02 -5.94628096e-01
3.46771985e-01 4.81521100e-01 -3.14498574e-01 4.39618647e-01
-1.70513406e-01 1.14113402e+00 -1.14050245e+00 -2.66324788e-01
5.60191810e-01 5.55796027e-01 -2.48296008e-01 -2.18680322e-01
-2.41124511e-01 7.25869238e-01 -8.80713701e-01 2.30319545e-01
4.75935102e-01 -8.89682025e-02 2.19611481e-01 2.11220071e-01
-4.01740313e-01 1.87241316e-01 -1.03310728e+00 1.63876975e+00
-5.20759165e-01 6.74308360e-01 -2.10909188e-01 -5.70154428e-01
7.42991269e-01 8.61005336e-02 3.86413246e-01 -7.05969155e-01
2.07358524e-01 -1.73909566e-03 -1.05708301e-01 -6.50225639e-01
9.09563959e-01 4.42884445e-01 -3.05985212e-01 4.60725993e-01
-4.25013185e-01 5.19858718e-01 -6.42366484e-02 1.88778013e-01
1.02067697e+00 4.01051074e-01 9.02163461e-02 -3.03345710e-01
6.21667027e-01 1.85767412e-01 2.03569055e-01 9.76540208e-01
-4.63821977e-01 8.65353644e-03 2.88696438e-01 -1.08215344e+00
-1.01810968e+00 -1.14222157e+00 3.57797265e-01 1.17378449e+00
4.55476284e-01 -5.82473092e-02 -5.86524069e-01 -3.17215741e-01
-4.98128384e-02 1.15366364e+00 -7.05414116e-01 -1.28754690e-01
-8.76751781e-01 -5.00121415e-01 3.86597365e-01 4.46339458e-01
3.87801677e-01 -1.37996161e+00 -1.06058729e+00 4.69940722e-01
-4.87992465e-01 -1.29612529e+00 -2.57116952e-03 -3.12096298e-01
-6.71076477e-01 -7.86532879e-01 -3.75837117e-01 -3.48635465e-01
4.10699487e-01 3.63533825e-01 7.43142188e-01 -7.40387291e-03
7.12905601e-02 5.30224264e-01 -7.80349746e-02 -1.97685048e-01
-1.15650915e-01 2.29424655e-01 4.33751196e-01 2.76778728e-01
6.02743745e-01 -8.30418229e-01 -8.64569128e-01 4.82351124e-01
-2.63085753e-01 3.51688892e-01 6.61642492e-01 3.84693682e-01
2.88735963e-02 2.70779897e-02 2.80012280e-01 -5.66765189e-01
5.88955820e-01 -9.70465839e-01 -7.73359090e-02 1.21501707e-01
-4.78048742e-01 -7.44652972e-02 5.58291972e-01 -3.92378688e-01
-1.32582605e+00 2.43038505e-01 -2.52043176e-02 -3.71365473e-02
-6.52642071e-01 3.43033910e-01 8.65644738e-02 1.99895903e-01
4.42747593e-01 3.08044165e-01 -1.99918285e-01 -6.70940280e-02
6.97087705e-01 2.53336847e-01 1.88860610e-01 -4.92272109e-01
5.39418817e-01 8.98659527e-01 5.17818667e-02 -1.13337362e+00
-1.83950335e-01 -1.69451818e-01 -1.18676496e+00 -7.58502305e-01
9.50665772e-01 -8.50388408e-01 -1.24038208e+00 4.33054715e-01
-1.33911288e+00 -4.68581408e-01 -4.92495298e-02 6.16432846e-01
-7.04895020e-01 1.69740207e-02 -5.57554543e-01 -1.20598650e+00
2.81803221e-01 -1.03481400e+00 1.01617551e+00 5.81705384e-02
-6.77328706e-01 -1.33043265e+00 -3.28762531e-01 3.30593348e-01
6.19659662e-01 4.04179394e-01 4.28294212e-01 -4.50805247e-01
-1.12507927e+00 -9.96145830e-02 -7.92512074e-02 -6.30546749e-01
-2.19026029e-01 -3.29534262e-01 -6.57098472e-01 3.11266780e-02
-3.99679929e-01 9.43778977e-02 7.56599545e-01 4.76563454e-01
6.23224676e-01 -4.56656247e-01 -9.98100638e-01 4.51239385e-02
7.84276009e-01 2.31151208e-01 6.75503433e-01 4.57797468e-01
8.38309288e-01 1.35667610e+00 7.00232983e-01 7.40342855e-01
1.31567430e+00 5.82106113e-01 4.75480884e-01 5.40230311e-02
2.65809178e-01 -5.29476881e-01 4.26245987e-01 3.78389925e-01
-2.65709281e-01 -4.21419680e-01 -1.14838493e+00 8.65664959e-01
-2.27847695e+00 -1.04166365e+00 -6.80263519e-01 2.00288296e+00
-4.78147745e-01 4.70849335e-01 1.05523191e-01 -2.48036951e-01
4.46294099e-01 1.68588102e-01 -7.36610234e-01 -2.79742002e-01
-4.52268124e-02 -8.31147730e-01 8.79750133e-01 6.69239759e-01
-9.50735509e-01 1.24972057e+00 6.06808710e+00 3.74064922e-01
-7.30419278e-01 1.02837637e-01 6.64196253e-01 5.47515750e-02
-5.45695961e-01 -3.56444046e-02 -7.57705748e-01 2.43612736e-01
1.03839934e+00 1.78421214e-01 4.56026614e-01 6.72338486e-01
7.27185309e-01 -1.60880253e-01 -9.97203767e-01 6.18199944e-01
-2.60085970e-01 -1.09978127e+00 1.42016128e-01 2.44336486e-01
3.91592383e-01 2.26652622e-01 3.21874201e-01 5.38064778e-01
7.32707739e-01 -9.79033053e-01 1.02759016e+00 1.20845318e+00
-1.18414640e-01 -4.64532316e-01 5.09470940e-01 7.56213963e-01
-1.38145602e+00 -4.85370845e-01 -7.48417154e-02 -6.90396607e-01
8.82284522e-01 9.47251990e-02 -1.05156684e+00 3.64196748e-01
7.11533904e-01 8.70890319e-01 -5.47076702e-01 4.54007000e-01
1.16157703e-01 1.70269441e-02 -3.78101826e-01 -4.18818384e-01
8.82329524e-01 -4.53309029e-01 9.48155165e-01 8.87981892e-01
5.93199313e-01 -1.15766125e-02 2.47450560e-01 9.99695122e-01
6.19614840e-01 -2.14675501e-01 -1.05661905e+00 1.75580949e-01
2.37718850e-01 9.67758179e-01 -9.25230443e-01 -2.38540933e-01
-1.91891477e-01 8.95002365e-01 3.79509777e-01 8.05883467e-01
-9.06556547e-01 8.96818042e-02 9.79520440e-01 3.89233947e-01
1.29510870e-03 -5.74020326e-01 -3.22656572e-01 -7.83644021e-01
2.26365793e-02 3.86500284e-02 9.43680033e-02 -8.76818657e-01
-7.82179356e-01 7.76305556e-01 1.43347591e-01 -1.25792897e+00
-4.46296304e-01 -5.57100058e-01 -5.62954426e-01 4.89679217e-01
-1.36811209e+00 -1.63498306e+00 -2.43890822e-01 4.67258424e-01
6.22108698e-01 -1.12052858e-01 4.51932341e-01 7.45325685e-02
-3.16521943e-01 -3.22675928e-02 -2.10814178e-01 -3.29815209e-01
5.07132821e-02 -9.34501886e-01 9.37932730e-01 6.01733387e-01
-1.22960627e-01 6.36427522e-01 9.90314543e-01 -9.80677426e-01
-1.09048283e+00 -1.26792753e+00 1.37661171e+00 -7.61191964e-01
6.84338570e-01 -3.96105707e-01 -6.31349325e-01 1.18495166e+00
-1.27933130e-01 -3.97785783e-01 3.86803418e-01 2.12588087e-01
2.26278737e-01 6.48484454e-02 -9.22359943e-01 1.26977170e+00
1.63339305e+00 -1.87473506e-01 -3.97998780e-01 3.73088390e-01
6.36564195e-01 -1.03706434e-01 -5.24831772e-01 3.63749564e-01
7.95896590e-01 -1.36358309e+00 1.20693803e+00 -5.30154526e-01
1.04523532e-01 -6.01901449e-02 -1.17268480e-01 -1.10313535e+00
-4.72019136e-01 -4.65576321e-01 4.05129679e-02 5.63543439e-01
6.62452817e-01 -8.04384172e-01 1.16153407e+00 1.05605066e+00
-1.83359250e-01 -6.38805628e-01 -9.33144569e-01 -4.63894278e-01
-1.52768567e-01 -9.26599860e-01 9.54541743e-01 6.41478002e-01
5.42726740e-02 -1.76191218e-02 -7.88746834e-01 3.22063684e-01
6.65733099e-01 -3.47171336e-01 1.03811240e+00 -1.31792450e+00
3.23746651e-01 -5.58795929e-01 -4.94368553e-01 -1.40404785e+00
2.88787723e-01 -4.54247624e-01 -4.98106368e-02 -1.85314059e+00
-2.74407655e-01 -7.85181701e-01 -1.20479546e-01 1.71625778e-01
2.51742363e-01 -2.91296244e-01 6.30477816e-02 2.81725496e-01
-8.95482183e-01 8.49307597e-01 1.25294840e+00 -8.57845619e-02
-5.02316356e-01 2.07408398e-01 -2.08135873e-01 9.57387745e-01
6.28147781e-01 -1.39299750e-01 -5.47567487e-01 -4.26281303e-01
5.57337821e-01 3.78508389e-01 6.09300315e-01 -1.33661342e+00
8.39604914e-01 -3.01541954e-01 3.10151130e-01 -9.41264272e-01
9.86163378e-01 -1.07862532e+00 5.17135024e-01 7.97328711e-01
-5.45073092e-01 1.48131475e-01 -9.49139670e-02 9.99060273e-01
1.33045480e-01 5.12606204e-01 6.91337325e-03 -3.30549926e-01
-1.15645266e+00 4.43546891e-01 -1.09212148e+00 -6.94114447e-01
1.32753313e+00 -7.71672189e-01 -2.29602996e-02 -7.37511396e-01
-1.13037121e+00 8.12893689e-01 3.54348630e-01 9.16896701e-01
7.45087147e-01 -1.36530995e+00 -4.89651322e-01 1.62352681e-01
8.71247277e-02 -4.05044466e-01 8.83157015e-01 9.17022228e-01
-6.24559700e-01 8.12172949e-01 -3.88491511e-01 -6.36926770e-01
-7.29777396e-01 5.31113029e-01 1.24974929e-01 -2.83102781e-01
-7.06831753e-01 6.34943724e-01 1.96663857e-01 -9.59693193e-01
8.25669989e-02 -5.54830194e-01 -7.31276035e-01 -1.00154482e-01
2.43602395e-01 7.89057195e-01 -3.98924828e-01 -1.32498610e+00
-3.01710367e-01 2.74804533e-01 2.13515759e-01 -1.84321955e-01
1.12299979e+00 -8.70954216e-01 2.69367099e-01 8.02906871e-01
6.34389997e-01 -6.24489725e-01 -1.29558372e+00 -2.74809629e-01
3.41285616e-01 -3.62802565e-01 -2.60660797e-01 -3.90168101e-01
-3.79259706e-01 7.84040391e-01 3.98961604e-01 1.58575788e-01
4.95567560e-01 -1.54636949e-01 1.18741882e+00 8.19875836e-01
9.52528775e-01 -1.08333564e+00 -1.33799165e-01 6.57584906e-01
6.85995400e-01 -1.28836894e+00 -3.58259827e-01 -2.23580778e-01
-1.04173827e+00 1.01228583e+00 6.58709943e-01 1.51502788e-02
8.94140840e-01 -3.64425093e-01 -7.97097161e-02 -3.03182691e-01
-9.61502612e-01 -2.87292749e-01 1.18266098e-01 6.80704534e-01
2.71560159e-02 3.98141384e-01 9.67379510e-02 3.50890189e-01
-4.81524140e-01 -2.34430835e-01 6.13977790e-01 5.82548141e-01
-5.76647758e-01 -5.49999833e-01 -2.95782149e-01 4.71913457e-01
2.61808276e-01 4.00908500e-01 -1.28112152e-01 7.60673881e-01
3.40417653e-01 1.01995015e+00 4.36589241e-01 -5.45371115e-01
3.76680315e-01 -1.55914530e-01 -9.75245163e-02 -7.43865222e-02
-2.91726440e-01 -4.91725177e-01 4.17315781e-01 -1.06226492e+00
-4.86260831e-01 -9.69689071e-01 -1.16713250e+00 -5.58389604e-01
4.15187091e-01 -2.72080749e-01 5.88281333e-01 1.29958153e+00
5.25033474e-01 4.30515945e-01 2.34952211e-01 -1.42893839e+00
-6.53489083e-02 -8.08148146e-01 -4.41130757e-01 4.79883611e-01
3.63271862e-01 -8.83769989e-01 -4.73803803e-02 -1.12625822e-01] | [5.9342756271362305, 0.7975581884384155] |
5e4939e5-b204-44f7-b6aa-e76b65fd0b2c | fine-grained-graph-learning-for-multi-view | 2201.04604 | null | https://arxiv.org/abs/2201.04604v3 | https://arxiv.org/pdf/2201.04604v3.pdf | Fine-grained Graph Learning for Multi-view Subspace Clustering | Multi-view subspace clustering (MSC) is a popular unsupervised method by integrating heterogeneous information to reveal the intrinsic clustering structure hidden across views. Usually, MSC methods use graphs (or affinity matrices) fusion to learn a common structure, and further apply graph-based approaches to clustering. Despite progress, most of the methods do not establish the connection between graph learning and clustering. Meanwhile, conventional graph fusion strategies assign coarse-grained weights to combine multi-graph, ignoring the importance of local structure. In this paper, we propose a fine-grained graph learning framework for multi-view subspace clustering (FGL-MSC) to address these issues. To utilize the multi-view information sufficiently, we design a specific graph learning method by introducing graph regularization and local structure fusion pattern. The main challenge is how to optimize the fine-grained fusion weights while generating the learned graph that fits the clustering task, thus making the clustering representation meaningful and competitive. Accordingly, an iterative algorithm is proposed to solve the above joint optimization problem, which obtains the learned graph, the clustering representation, and the fusion weights simultaneously. Extensive experiments on eight real-world datasets show that the proposed framework has comparable performance to the state-of-the-art methods. | ['Haoxi Zhan', 'Xiaobing Pei', 'Yidi Wang'] | 2022-01-12 | null | null | null | null | ['multi-view-subspace-clustering'] | ['computer-vision'] | [-1.73776448e-01 -3.47024441e-01 -3.42279583e-01 -1.77579731e-01
-5.67471206e-01 -4.92869973e-01 3.32989335e-01 6.25409856e-02
1.51261136e-01 2.40674302e-01 4.24004167e-01 1.15357332e-01
-3.90034676e-01 -6.15470290e-01 -2.85479128e-01 -1.13308167e+00
1.72448292e-01 2.04825863e-01 2.11455107e-01 6.87255263e-02
1.53401941e-01 1.70506984e-01 -1.13966191e+00 2.23662257e-01
1.07638574e+00 4.71654147e-01 2.76613116e-01 3.18669915e-01
-1.04129054e-01 7.43152142e-01 -1.29524574e-01 -2.11135939e-01
1.14679232e-01 -5.19825101e-01 -4.04250175e-01 7.68450677e-01
3.21159512e-01 2.63769507e-01 -4.14420128e-01 1.39864731e+00
4.75785971e-01 2.80626327e-01 6.58285797e-01 -1.32927084e+00
-7.33953714e-01 5.59269965e-01 -1.05234134e+00 4.81517427e-02
2.93844014e-01 -2.23286733e-01 1.21807301e+00 -9.87708211e-01
5.63598156e-01 1.26225913e+00 3.28489214e-01 1.39967635e-01
-1.28514719e+00 -3.94846946e-01 6.30062103e-01 2.80827433e-01
-1.79261625e+00 -1.65850878e-01 1.29957986e+00 -6.20494783e-01
3.48103285e-01 2.78639734e-01 6.99404538e-01 6.93745852e-01
1.24400310e-01 6.46994948e-01 1.19152319e+00 -2.74558812e-01
2.23882437e-01 -1.32292300e-01 2.36695826e-01 9.89597917e-01
4.98564154e-01 -3.77382666e-01 -3.79015625e-01 -4.44337696e-01
6.06435537e-01 4.46334720e-01 -4.97604370e-01 -1.05122423e+00
-1.27270281e+00 8.70243073e-01 4.71499711e-01 4.84687209e-01
-3.17272246e-01 -2.55060554e-01 2.45759517e-01 -9.40377451e-03
4.80384231e-01 -8.13424364e-02 6.99362997e-03 5.67681015e-01
-7.57090628e-01 -9.56978053e-02 5.40652871e-01 1.02029538e+00
1.15959573e+00 8.91581476e-02 6.80390447e-02 8.37038696e-01
6.89015388e-01 3.48614573e-01 3.53905797e-01 -8.07977855e-01
5.53247929e-01 1.20202065e+00 -3.91445786e-01 -1.47245502e+00
-3.34831893e-01 -4.63539720e-01 -1.22552669e+00 -1.88691810e-01
1.40175521e-01 -8.50323737e-02 -8.23878765e-01 1.57386696e+00
5.40280640e-01 5.17395139e-01 -1.39048547e-02 1.00776958e+00
8.57834816e-01 3.96763384e-01 -2.10429773e-01 -5.34938574e-01
1.13088179e+00 -9.51040983e-01 -8.09813559e-01 -7.86125436e-02
4.60772038e-01 -7.18414903e-01 8.20190310e-01 2.42509708e-01
-7.43378520e-01 -5.19270062e-01 -1.13929522e+00 3.83781165e-01
-2.35540554e-01 -2.64082570e-03 6.14155233e-01 4.57563102e-01
-9.28347707e-01 2.54961759e-01 -7.02636600e-01 -4.70607162e-01
-2.71611270e-02 3.95784736e-01 -4.81492490e-01 -3.86708796e-01
-8.51660788e-01 2.40272373e-01 6.71893597e-01 1.37790173e-01
-4.84769642e-01 -2.43260324e-01 -9.36198533e-01 -8.20645094e-02
7.83619523e-01 -8.28135967e-01 2.71196812e-01 -6.33116305e-01
-1.06010127e+00 6.43917739e-01 -2.11359411e-01 2.59341925e-01
-7.21331090e-02 2.80093759e-01 -6.29758894e-01 2.86278129e-01
3.37387472e-01 1.52994454e-01 1.03700280e+00 -1.80585170e+00
-4.63506579e-01 -7.28156447e-01 -1.39509052e-01 5.06007612e-01
-3.97117227e-01 -3.09040129e-01 -1.01553500e+00 -6.14115894e-01
7.94921160e-01 -9.37900901e-01 -3.64597857e-01 -6.78583086e-01
-5.70985258e-01 -2.60419399e-01 1.00767791e+00 -4.94946033e-01
1.58621669e+00 -2.23938751e+00 8.26328814e-01 5.39180875e-01
6.49602473e-01 -1.61968917e-01 -1.89057067e-02 6.40535533e-01
-1.15982853e-01 -5.86027391e-02 -2.61285573e-01 -1.35494247e-01
-1.85368568e-01 2.78589219e-01 2.25310564e-01 7.32664108e-01
-3.26941013e-01 6.83788896e-01 -1.02277267e+00 -7.88201749e-01
3.52414548e-01 3.66208106e-01 -5.53303659e-01 3.91008139e-01
2.60667384e-01 5.43720901e-01 -8.48474085e-01 6.80916250e-01
7.66293347e-01 -6.74793422e-01 7.56923914e-01 -7.90251315e-01
8.97660479e-02 -5.91640353e-01 -1.72326195e+00 1.88120902e+00
1.24176882e-01 -1.78334579e-01 4.85284328e-01 -1.29722381e+00
8.48530054e-01 2.04143032e-01 9.82022464e-01 -1.06201796e-02
9.63672698e-02 8.46000910e-02 -2.41070334e-02 -4.30878848e-01
1.46815926e-01 -1.41021058e-01 -1.34181324e-02 3.05149257e-01
2.58430243e-01 1.63310394e-01 2.36656219e-01 5.71252882e-01
8.77007663e-01 -6.36695027e-02 3.89944136e-01 -3.74714375e-01
9.94692683e-01 -2.35296264e-01 8.21034491e-01 1.93154514e-01
-1.20216720e-01 6.22267127e-01 3.26574922e-01 -2.15716690e-01
-5.95077217e-01 -9.98633027e-01 3.71305555e-01 8.71732771e-01
5.86767614e-01 -7.86699951e-01 -7.91248381e-01 -8.05772066e-01
-1.52823284e-01 6.00851849e-02 -4.70720559e-01 -3.13924909e-01
-4.75836396e-01 -9.52978909e-01 -1.24628410e-01 2.53383666e-01
4.91549939e-01 -4.82870996e-01 2.30639815e-01 1.80686638e-01
-3.78695726e-01 -1.09234488e+00 -8.25031638e-01 -1.57993063e-01
-8.69764566e-01 -1.39010680e+00 -4.82758939e-01 -8.56816709e-01
9.64718580e-01 1.04190123e+00 8.93278778e-01 2.76873201e-01
8.99848565e-02 7.50689328e-01 -4.34727013e-01 4.08487588e-01
1.07370038e-03 8.81433412e-02 1.59868315e-01 6.14928424e-01
2.86707759e-01 -7.76925445e-01 -4.61571246e-01 3.54325622e-01
-1.07340443e+00 9.78306010e-02 5.65389693e-01 8.89476597e-01
9.63048220e-01 3.45931232e-01 4.61809188e-01 -1.12066901e+00
7.17610896e-01 -4.52883542e-01 -3.83119047e-01 5.63681066e-01
-8.06014240e-01 -3.54121327e-02 7.22450614e-01 -2.40685552e-01
-8.70638788e-01 1.88360736e-01 4.64406699e-01 -9.57022309e-01
-1.20875444e-02 8.78179431e-01 -6.84260011e-01 -2.38600299e-01
1.47593960e-01 4.32206243e-01 -5.41615561e-02 -4.19717193e-01
7.58280635e-01 3.64769012e-01 2.97425359e-01 -5.99774182e-01
8.67086053e-01 6.35036409e-01 5.42656220e-02 -6.96439922e-01
-8.52204919e-01 -9.85460997e-01 -9.94056284e-01 -3.68821353e-01
1.16711962e+00 -1.13581312e+00 -5.54070532e-01 1.77934349e-01
-6.27254486e-01 2.54967123e-01 1.62917063e-01 6.38931692e-01
-5.17785609e-01 9.33884561e-01 -4.34327394e-01 -5.46706438e-01
2.38752477e-02 -1.13809669e+00 1.05967402e+00 1.08152576e-01
2.31807351e-01 -1.27254963e+00 1.88718751e-01 5.97046316e-01
-1.44003734e-01 4.49514419e-01 8.50817323e-01 -3.91585320e-01
-5.78596592e-01 1.98892318e-04 -3.43427777e-01 1.16735078e-01
4.42027390e-01 -6.72511831e-02 -5.42495906e-01 -7.49747396e-01
1.72299504e-01 1.41906515e-01 7.48078823e-01 3.81434113e-01
1.05227602e+00 -1.66181818e-01 -5.42393208e-01 7.38550842e-01
1.60317028e+00 1.03871562e-01 1.38902962e-01 -5.28241284e-02
1.56419611e+00 5.37103713e-01 3.38749528e-01 4.29014355e-01
7.27324188e-01 5.06539226e-01 2.55866587e-01 -6.04672246e-02
7.03452975e-02 -2.84394085e-01 1.48928255e-01 1.76142776e+00
-3.13792467e-01 -1.89595133e-01 -8.45366478e-01 2.98876703e-01
-2.36644340e+00 -9.37325120e-01 -3.05759847e-01 1.90731108e+00
1.13352716e-01 -1.04911618e-01 1.74100652e-01 -8.00751671e-02
1.16051829e+00 5.81892014e-01 -4.79628950e-01 3.26887012e-01
-2.27556795e-01 -5.37026167e-01 2.96893507e-01 5.08452773e-01
-1.23826706e+00 8.78636241e-01 5.92317390e+00 9.29672778e-01
-8.62478256e-01 9.14810225e-02 2.41439044e-01 3.28502357e-01
-4.97130305e-01 2.17712954e-01 -4.38837022e-01 2.81247556e-01
4.97569203e-01 -1.84419200e-01 7.26822376e-01 5.68193197e-01
1.23447925e-01 2.84872919e-01 -7.32959807e-01 1.21454298e+00
3.55617344e-01 -1.11718726e+00 3.47220182e-01 2.49553114e-01
9.17175174e-01 -1.38356194e-01 -2.52510130e-01 1.26602635e-01
4.47065294e-01 -5.07753193e-01 2.43551508e-01 6.18756711e-01
4.09298331e-01 -8.01109910e-01 4.97619063e-01 3.43959391e-01
-1.80285919e+00 -2.41875369e-02 -4.19409573e-01 2.23204464e-01
9.85720530e-02 5.96845388e-01 -2.46875897e-01 1.30363917e+00
5.43436944e-01 1.29405177e+00 -7.90950835e-01 6.96004808e-01
-2.86604073e-02 4.37753618e-01 7.47160092e-02 3.88289481e-01
3.33974391e-01 -9.72562373e-01 5.82027256e-01 8.30884695e-01
2.22263202e-01 3.10250193e-01 9.04065073e-01 7.12412298e-01
5.06814010e-02 3.88634562e-01 -7.64461398e-01 -9.96823311e-02
3.62076789e-01 1.67651975e+00 -1.00104272e+00 -3.13877255e-01
-8.96598041e-01 9.02695596e-01 6.75580919e-01 7.09139764e-01
-5.58560312e-01 2.01712698e-02 3.01552951e-01 -6.11580536e-02
2.96353847e-01 -4.57943946e-01 -3.45341489e-02 -1.68106532e+00
-1.44180311e-02 -1.07201517e+00 7.93026090e-01 -5.30970871e-01
-1.62422645e+00 4.22132820e-01 8.38265345e-02 -1.43436301e+00
4.63304929e-02 -3.16078544e-01 -6.86263740e-01 5.62998891e-01
-1.13691807e+00 -1.47938132e+00 -4.17017251e-01 1.03420126e+00
2.99228251e-01 -3.56537789e-01 4.98980492e-01 3.86660069e-01
-8.20309043e-01 2.06219733e-01 1.94180772e-01 5.58266006e-02
6.56021237e-01 -1.31206012e+00 -3.38073105e-01 1.01021111e+00
3.16707134e-01 7.56732941e-01 3.30660313e-01 -7.92063892e-01
-1.89823580e+00 -1.19879889e+00 2.11980477e-01 -3.00978035e-01
7.39518881e-01 -3.94115269e-01 -1.03648257e+00 5.47404349e-01
3.35692674e-01 2.75015116e-01 1.00877547e+00 1.80493698e-01
-4.11956608e-01 -1.22985393e-01 -8.82195771e-01 5.04915833e-01
1.06851184e+00 -6.16339982e-01 -4.10644144e-01 3.12577873e-01
7.36101210e-01 -6.18807562e-02 -1.24117243e+00 4.83920813e-01
1.94924876e-01 -1.12800550e+00 9.63225305e-01 -4.52570289e-01
-7.83991218e-02 -8.48398626e-01 -5.59082091e-01 -1.47034180e+00
-8.88869286e-01 -4.42763239e-01 -2.60758609e-01 1.42593956e+00
5.57225570e-03 -5.41091025e-01 7.24085033e-01 1.80151880e-01
-2.19852641e-01 -7.65496910e-01 -6.96183622e-01 -5.78109622e-01
-2.88466781e-01 1.20346956e-01 5.10591388e-01 1.43794787e+00
1.15975648e-01 7.83618629e-01 -4.25939262e-01 5.04884601e-01
1.14806640e+00 5.45525491e-01 7.76342988e-01 -1.29054546e+00
-2.68549323e-01 -3.42568994e-01 -5.87033331e-01 -6.02533698e-01
2.81603962e-01 -1.07785153e+00 -3.48849684e-01 -1.77420080e+00
5.98865271e-01 -4.80948053e-02 -5.15814126e-01 3.74080203e-02
-6.35675907e-01 -8.29362646e-02 2.74389297e-01 6.00898564e-01
-1.08044863e+00 5.45475245e-01 1.44644475e+00 -2.92047977e-01
-1.66659623e-01 -3.05447519e-01 -9.71063375e-01 7.76973248e-01
4.39688861e-01 -1.35367930e-01 -7.55613565e-01 -2.18446016e-01
-2.72197295e-02 1.93210900e-01 1.04507416e-01 -9.11587954e-01
5.32912791e-01 -3.46414149e-01 2.15582460e-01 -7.63161957e-01
2.18657970e-01 -9.94327068e-01 4.61952060e-01 1.39693424e-01
1.08733475e-01 1.30772172e-02 -4.06549156e-01 1.13914537e+00
-5.71482778e-01 2.78830975e-01 6.88034236e-01 -2.96968460e-01
-8.22753489e-01 7.17584193e-01 -6.00757189e-02 7.04918653e-02
1.03538740e+00 -2.55875587e-01 -9.99357924e-02 -2.40425795e-01
-1.03712821e+00 4.52532262e-01 6.50074899e-01 4.16549623e-01
7.04227507e-01 -1.82925272e+00 -5.68408370e-01 3.27137709e-01
3.28205764e-01 -1.54893193e-02 4.49865669e-01 9.26806390e-01
-3.04333121e-02 1.15725152e-01 -1.24975760e-03 -7.85158813e-01
-1.16878808e+00 1.02040207e+00 6.18978813e-02 -3.91711652e-01
-5.50428391e-01 5.04441798e-01 5.77427030e-01 -5.27918816e-01
-9.46345925e-02 1.31436631e-01 -4.75778908e-01 2.19182640e-01
8.58874246e-02 5.51591158e-01 -1.92856848e-01 -1.03606784e+00
-3.35729480e-01 1.02595496e+00 8.92030522e-02 1.90991431e-01
1.26015294e+00 -6.48713887e-01 -5.46164453e-01 5.70160627e-01
1.26400650e+00 1.63599160e-02 -1.05649388e+00 -4.32429790e-01
1.69243999e-02 -5.12326062e-01 7.66039640e-02 4.96076047e-02
-1.42123938e+00 6.90759122e-01 3.50504428e-01 4.54287410e-01
1.18835056e+00 6.37278855e-02 4.59638417e-01 1.80307686e-01
4.60153550e-01 -9.95588481e-01 2.34930053e-01 1.69871941e-01
5.81148922e-01 -1.13622057e+00 3.57361138e-01 -9.37978208e-01
-7.18056142e-01 8.16959441e-01 8.43133271e-01 -3.24931800e-01
9.41003978e-01 -2.29706690e-01 -5.47337122e-02 -7.76965976e-01
-4.62777555e-01 -3.21693063e-01 6.21507525e-01 4.61867929e-01
3.38767439e-01 2.60013968e-01 -2.49584422e-01 5.16850114e-01
3.70896369e-01 -5.35680294e-01 2.83183396e-01 6.76131666e-01
-4.91038561e-01 -1.19170380e+00 -4.84097123e-01 5.41485310e-01
-2.56587595e-01 2.16989726e-01 -5.67495227e-01 5.55860519e-01
1.29141510e-01 1.07085764e+00 -3.91047150e-01 -7.27847159e-01
3.21966738e-01 -1.71980858e-01 2.56433785e-01 -7.36355901e-01
-2.29787633e-01 7.61686027e-01 -2.73985744e-01 -6.16367817e-01
-1.02025890e+00 -5.75448275e-01 -1.12173319e+00 -1.08924851e-01
-5.96059382e-01 3.99474978e-01 6.91767260e-02 7.89037347e-01
4.30184692e-01 5.11089683e-01 9.64929938e-01 -7.75171220e-01
-1.23198085e-01 -5.80711901e-01 -1.18224025e+00 6.90471232e-01
-4.85004894e-02 -8.38372648e-01 -4.46396351e-01 1.10470161e-01] | [8.06954288482666, 4.783566474914551] |
8d929850-b59d-4798-a1db-6318f8bb5cc2 | automatic-skin-lesion-segmentation-using-semi | 1703.04301 | null | http://arxiv.org/abs/1703.04301v1 | http://arxiv.org/pdf/1703.04301v1.pdf | Automatic Skin Lesion Segmentation using Semi-supervised Learning Technique | Skin cancer is the most common of all cancers and each year million cases of
skin cancer are treated. Treating and curing skin cancer is easy, if it is
diagnosed and treated at an early stage. In this work we propose an automatic
technique for skin lesion segmentation in dermoscopic images which helps in
classifying the skin cancer types. The proposed method comprises of two major
phases (1) preprocessing and (2) segmentation using semi-supervised learning
algorithm. In the preprocessing phase noise are removed using filtering
technique and in the segmentation phase skin lesions are segmented based on
clustering technique. K-means clustering algorithm is used to cluster the
preprocessed images and skin lesions are filtered from these clusters based on
the color feature. Color of the skin lesions are learned from the training
images using histograms calculations in RGB color space. The training images
were downloaded from the ISIC 2017 challenge website and the experimental
results were evaluated using validation and test sets. | ['P. Mirunalini', 'Aravindan Chandrabose', 'S. M. Jaisakthi'] | 2017-03-13 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 6.64312124e-01 -5.40604554e-02 -2.73098916e-01 -2.41111308e-01
-5.09951353e-01 -7.03413367e-01 3.54588419e-01 4.62720066e-01
-7.62987137e-01 5.75242043e-01 -1.26360953e-01 -1.92489564e-01
-1.81954224e-02 -8.39343965e-01 -6.79585477e-03 -8.74130070e-01
2.21155435e-01 2.37058580e-01 5.71498036e-01 1.52844757e-01
2.48836696e-01 5.58602810e-01 -1.27322507e+00 4.15120125e-01
1.12548518e+00 6.43594503e-01 7.41475075e-02 1.26998246e+00
-3.36490124e-01 4.29058373e-01 -1.39618561e-01 4.42678891e-02
5.48850954e-01 -6.92379177e-01 -1.01507342e+00 5.61308622e-01
1.58684760e-01 1.03282467e-01 3.40070516e-01 1.25545561e+00
2.45692864e-01 -1.62619174e-01 1.06512058e+00 -6.85615718e-01
1.35009497e-01 -1.39131561e-01 -9.64808643e-01 -8.61321297e-03
1.88672632e-01 -1.54199690e-01 2.70221978e-01 -4.18981969e-01
6.28676355e-01 8.45477521e-01 5.10215342e-01 9.05097067e-01
-8.98403525e-01 -1.35622576e-01 -4.89626855e-01 3.63154083e-01
-1.19530296e+00 1.00797102e-01 4.31667775e-01 -3.79626453e-01
5.43488979e-01 6.92321420e-01 8.58668506e-01 2.08538428e-01
1.20882548e-01 5.30033290e-01 1.72714615e+00 -8.79669845e-01
5.16702116e-01 4.01235223e-01 3.41989934e-01 7.93042183e-01
2.35776678e-01 -7.24902079e-02 3.15549552e-01 7.73644149e-02
5.68547249e-01 1.23979989e-02 1.25177130e-01 1.31837323e-01
-4.47276562e-01 6.56319320e-01 5.35417318e-01 5.00788867e-01
-6.19426787e-01 -1.52030557e-01 5.58720708e-01 4.36263606e-02
2.07904428e-01 -1.09820373e-01 -4.50390652e-02 2.73353040e-01
-1.20039916e+00 -3.86662960e-01 6.63676381e-01 2.04267010e-01
4.21145946e-01 -5.50886571e-01 1.58337623e-01 9.39502656e-01
4.08698052e-01 4.46362168e-01 4.81338829e-01 -2.28199169e-01
-2.10406065e-01 9.69295025e-01 -1.30108818e-01 -4.88785148e-01
-3.07130724e-01 3.53670210e-01 -8.02290738e-01 4.73759472e-01
6.21301293e-01 -3.00208300e-01 -1.88415885e+00 6.63548410e-01
6.33477449e-01 4.10907790e-02 1.26037195e-01 8.70816767e-01
6.73914969e-01 5.15098095e-01 5.88956714e-01 -2.79029071e-01
1.45785928e+00 -7.81733334e-01 -7.11730480e-01 2.81874329e-01
3.18106920e-01 -1.12224901e+00 5.17345369e-01 6.31040990e-01
-8.73307943e-01 -2.30815172e-01 -8.81569803e-01 2.47618809e-01
-7.02828526e-01 4.41231281e-01 6.65661693e-01 1.20289993e+00
-1.02037740e+00 1.67853206e-01 -9.88689721e-01 -1.23560536e+00
4.95871842e-01 4.30714309e-01 -4.19669926e-01 -1.18205816e-01
-6.08460844e-01 8.64027917e-01 3.91099513e-01 -3.81125249e-02
-4.57166255e-01 -2.88411118e-02 -4.57477927e-01 -6.03902757e-01
1.14484489e-01 -1.79486066e-01 7.36883700e-01 -1.55065668e+00
-1.53023195e+00 1.17314780e+00 -2.77852863e-01 -3.23021114e-01
4.51668948e-01 2.91246980e-01 -4.12827373e-01 6.60431743e-01
-3.67430568e-01 3.71257663e-01 7.72541225e-01 -1.28036463e+00
-1.02535713e+00 -5.15635490e-01 -6.79046988e-01 3.25837314e-01
-1.18653111e-01 -1.34008750e-01 -6.43213093e-01 -8.75990912e-02
1.49498329e-01 -8.45441997e-01 -4.99653727e-01 3.72407623e-02
-8.30198288e-01 1.68327074e-02 1.00676131e+00 -1.07001531e+00
1.05789530e+00 -1.95191526e+00 -1.82920232e-01 9.83124912e-01
-6.45501018e-02 6.58521354e-01 2.15548724e-01 5.28735101e-01
-2.60160863e-01 2.83206046e-01 -2.74866909e-01 1.52762040e-01
-5.24452567e-01 1.21939695e-02 6.39454722e-01 4.99638468e-01
5.06569482e-02 3.07578057e-01 -6.92119896e-01 -9.08536613e-01
5.92091620e-01 6.28066063e-01 2.17454538e-01 -1.57549437e-02
-1.40779600e-01 1.83578372e-01 -3.77450049e-01 8.57763410e-01
9.09087956e-01 2.95824289e-01 2.31732324e-01 -3.55089039e-01
-2.05130521e-02 -4.57150191e-01 -1.13225961e+00 1.13231838e+00
-1.04312159e-01 3.22634637e-01 1.82345673e-01 -7.95671344e-01
6.53604627e-01 4.24791098e-01 5.69009244e-01 -4.45798129e-01
3.05610836e-01 1.03756495e-01 4.89296718e-03 -8.83780420e-01
9.38949138e-02 -1.84697300e-01 4.64355469e-01 2.64468700e-01
-1.66799337e-01 -1.98410451e-01 7.26868927e-01 1.66447952e-01
8.27917457e-01 -1.44090101e-01 4.69899297e-01 -2.82929868e-01
7.49444366e-01 6.54167652e-01 2.07677260e-01 9.71891880e-02
-4.32242364e-01 1.77537650e-01 2.96229959e-01 -2.18318135e-01
-9.00466979e-01 -1.23265696e+00 -3.86811286e-01 4.30835903e-01
6.46729097e-02 1.20200053e-01 -1.35449696e+00 -7.24054635e-01
-1.40535027e-01 2.90918827e-01 -8.32161009e-01 3.00344497e-01
-3.52269225e-02 -8.95179927e-01 3.04977775e-01 9.97585356e-02
7.48570085e-01 -6.89997673e-01 -3.95431548e-01 -1.90780148e-01
5.21812618e-01 -4.32308167e-01 -6.66192640e-03 1.38396323e-02
-8.96910489e-01 -1.71545923e+00 -8.07065070e-01 -1.05751336e+00
1.43971860e+00 2.49957860e-01 3.97903353e-01 4.53699201e-01
-1.32704449e+00 4.45108145e-01 -4.75932658e-01 -6.09866500e-01
-5.02153635e-01 -2.20286652e-01 -4.59814161e-01 1.20754376e-01
7.50030518e-01 -5.40081039e-03 -8.32346916e-01 -1.20892413e-01
-1.03567863e+00 -1.39148861e-01 9.63146031e-01 5.47444642e-01
7.50511527e-01 7.51092136e-01 6.29075542e-02 -1.57916725e+00
5.18475175e-01 -3.28000009e-01 -4.12264436e-01 5.58098018e-01
-3.60945433e-01 -3.82954985e-01 5.35195053e-01 -3.55740935e-02
-1.29614139e+00 6.40102863e-01 1.68698467e-02 4.17054236e-01
-5.91254294e-01 3.19994599e-01 7.89445937e-02 -3.63434583e-01
8.50739002e-01 2.23234951e-01 2.78947443e-01 -3.06456864e-01
2.42460236e-01 9.45456386e-01 2.23400936e-01 8.57153460e-02
8.25457692e-01 6.09531879e-01 2.22544804e-01 -1.27010834e+00
-3.08612347e-01 -9.80687439e-01 -7.43413150e-01 -6.13417745e-01
1.35176969e+00 -3.59353274e-01 -2.78636724e-01 8.83803308e-01
-3.90521646e-01 -2.42507562e-01 1.97886318e-01 4.31351006e-01
7.43274242e-02 6.62049294e-01 -6.69791520e-01 -1.23602593e+00
-5.82396388e-01 -6.05693161e-01 4.75302666e-01 1.10097945e+00
1.31200418e-01 -1.48046124e+00 2.17425808e-01 4.38650370e-01
2.81819165e-01 5.01129508e-01 1.00823748e+00 -4.36884522e-01
-9.15877298e-02 -6.76021755e-01 -1.69574559e-01 4.28085834e-01
6.22101426e-01 6.74905121e-01 -8.58187139e-01 -1.78197939e-02
-3.15930903e-01 -2.08725184e-01 1.05377209e+00 5.42535722e-01
9.58808005e-01 7.90571198e-02 -5.37753880e-01 3.10338944e-01
2.22939086e+00 2.61104941e-01 8.57268929e-01 -5.21187894e-02
5.79074740e-01 6.51189148e-01 6.59756362e-01 -1.04661450e-01
-3.87213938e-02 -3.14854681e-01 2.50906795e-01 -7.49475121e-01
-2.98909009e-01 6.39352575e-02 -1.32385761e-01 4.39630240e-01
-4.28054065e-01 7.06123263e-02 -8.19440722e-01 6.79034233e-01
-1.30703974e+00 -8.75271678e-01 -6.59316421e-01 2.42124891e+00
1.06123078e+00 3.20882574e-02 4.81532425e-01 4.46173102e-01
9.80027258e-01 -4.79625791e-01 -2.42916688e-01 -7.24498808e-01
2.11878106e-01 8.54926467e-01 9.56587613e-01 6.78722322e-01
-1.32861054e+00 9.28066790e-01 5.88387585e+00 8.06606829e-01
-1.48301685e+00 -2.57408500e-01 8.53187382e-01 2.97763586e-01
1.33738235e-01 -5.95874004e-02 -3.19377929e-01 2.66035676e-01
4.42590415e-01 2.20819920e-01 2.19736204e-01 5.08883834e-01
2.63998538e-01 -1.27843738e+00 -4.15684283e-01 5.78819633e-01
-6.64901957e-02 -8.51362884e-01 -2.23503232e-01 -1.35221630e-01
9.76351976e-01 -3.32368702e-01 -7.88685307e-02 -3.50092322e-01
2.83700109e-01 -1.15050924e+00 -2.73774683e-01 8.77441168e-01
7.89009690e-01 -8.46963763e-01 7.23644793e-01 1.72362641e-01
-9.52799320e-01 1.94189712e-01 -2.36052647e-01 4.32513207e-01
-3.02627236e-01 6.18119240e-01 -1.09718859e+00 3.82573456e-01
3.79713237e-01 3.42575908e-01 -9.36359882e-01 1.68627810e+00
-4.26254310e-02 9.76939201e-01 -5.40349960e-01 -3.22631210e-01
2.22011358e-01 -4.06048685e-01 1.27260447e-01 1.37803423e+00
-1.39141515e-01 -2.00312734e-02 -4.81089987e-02 3.25842589e-01
5.86755395e-01 4.89648223e-01 -1.35169730e-01 -3.24927419e-01
3.27879786e-01 1.67751896e+00 -1.42210960e+00 -1.84848279e-01
-1.39474601e-01 1.07612312e+00 -3.39733571e-01 3.22862327e-01
-3.06321055e-01 -6.84626222e-01 -1.25780776e-01 8.09971243e-02
-1.24239109e-01 1.08585984e-01 -3.89872551e-01 -4.59974051e-01
-4.29796517e-01 -3.99007708e-01 6.37695789e-01 -1.93804979e-01
-1.15743768e+00 3.24968278e-01 -1.95517287e-01 -8.07677329e-01
2.24985138e-01 -9.26204860e-01 -1.23512566e+00 9.07656789e-01
-1.50927961e+00 -1.38521612e+00 -7.03305006e-01 6.17740035e-01
3.48241061e-01 3.80612700e-03 1.10783875e+00 -1.35550871e-01
-6.61470950e-01 1.99845329e-01 3.54912043e-01 2.12919578e-01
7.52253234e-01 -1.91269577e+00 -4.62195933e-01 7.95713305e-01
-3.48596811e-01 3.84884089e-01 3.99633139e-01 -9.03922796e-01
-1.19904208e+00 -9.51801121e-01 5.49098611e-01 2.26599678e-01
2.60006309e-01 -5.27589433e-02 -4.91298914e-01 -6.97487444e-02
6.03422999e-01 -1.03283443e-01 1.13146329e+00 -3.12963098e-01
8.60410780e-02 -5.07956110e-02 -1.77176106e+00 2.80321330e-01
-1.33651331e-01 -1.93776414e-01 4.44823802e-02 7.91549504e-01
-5.92708290e-01 -1.97745517e-01 -9.12033558e-01 -4.82722521e-02
3.56552362e-01 -9.46204662e-01 4.92536247e-01 -3.59647393e-01
1.54522374e-01 -4.13054109e-01 3.30305189e-01 -1.13980460e+00
1.71103179e-01 -3.28260183e-01 7.08144486e-01 1.10501719e+00
5.14049947e-01 -1.64064914e-01 1.16847110e+00 5.63795447e-01
5.77218592e-01 -8.27633440e-01 -4.18175638e-01 -5.88510633e-02
1.84470061e-02 3.14826250e-01 -2.56040394e-01 7.98400819e-01
3.26027833e-02 -3.39377299e-02 2.34135166e-01 -1.02887616e-01
1.15727758e+00 -3.19127321e-01 3.17844808e-01 -1.00596631e+00
2.67054550e-02 -1.31043807e-01 -5.18381655e-01 1.40284196e-01
-6.02921367e-01 -5.04972875e-01 -6.51649684e-02 -1.99952650e+00
4.00301367e-01 -4.35934573e-01 -6.37844950e-02 4.43718404e-01
-3.86330545e-01 4.35456485e-01 -1.48591623e-01 -9.73111764e-02
-3.19766223e-01 -5.97209990e-01 1.15252399e+00 -2.97365952e-02
-2.40167841e-01 3.18975687e-01 -1.93224013e-01 6.32720947e-01
9.89787757e-01 -2.33627297e-02 -4.58287954e-01 3.13534766e-01
-2.91791737e-01 -1.94528893e-01 2.36970633e-01 -9.55874026e-01
3.43543202e-01 -4.35511142e-01 7.51385450e-01 -4.99060273e-01
4.55389135e-02 -9.18486178e-01 -8.09813961e-02 8.32696915e-01
-2.14761838e-01 -7.49545455e-01 5.69426417e-02 6.34833694e-01
1.32215871e-02 -5.93511701e-01 1.37675560e+00 -3.87315273e-01
-1.01563489e+00 -1.06926434e-01 -7.33825505e-01 -5.10664105e-01
1.76215684e+00 -6.06809616e-01 -2.64002681e-01 -1.15322568e-01
-1.00071156e+00 1.36180982e-01 6.00418329e-01 -3.24211299e-01
4.03049260e-01 -9.19788957e-01 -5.06786764e-01 2.87940819e-02
-3.50855291e-02 -1.25845298e-01 4.13078487e-01 9.83020246e-01
-1.36028016e+00 1.94268420e-01 -5.57257771e-01 -3.23913097e-01
-1.89657199e+00 2.47049883e-01 4.30462331e-01 -1.27341151e-01
-4.43465114e-02 1.05074012e+00 -6.32148385e-01 5.08443564e-02
2.19972521e-01 -1.46938935e-01 -6.56797171e-01 -2.75032222e-01
5.05007684e-01 4.39891309e-01 -8.84583071e-02 -6.54809296e-01
-1.75821453e-01 8.18718851e-01 -4.20337468e-01 -7.60280937e-02
1.15174174e+00 2.09954217e-01 -5.11435688e-01 9.28288549e-02
1.05370259e+00 1.49574772e-01 -7.64018536e-01 1.98253542e-01
1.08442612e-01 -4.48733389e-01 3.31042498e-01 -1.36574829e+00
-1.09831417e+00 7.64260054e-01 1.22827578e+00 3.80379945e-01
1.58296871e+00 -5.58980703e-01 7.15077579e-01 4.34355363e-02
-1.56749815e-01 -1.85063815e+00 -2.45712072e-01 -2.54028618e-01
1.27152354e-01 -1.21367943e+00 2.77950853e-01 -9.85593736e-01
-9.10860300e-01 1.36056614e+00 5.02533793e-01 -6.03413641e-01
7.86451578e-01 3.56676817e-01 4.89127129e-01 -2.39156291e-01
-3.51997077e-01 -4.92827356e-01 3.73619169e-01 8.56299222e-01
6.45174563e-01 3.22360367e-01 -9.95993674e-01 2.36302003e-01
3.97457421e-01 1.67613309e-02 3.40414047e-01 1.08495343e+00
-7.69664884e-01 -1.43544543e+00 -5.98220468e-01 9.32462811e-01
-7.87407756e-01 2.43653283e-01 -1.11121559e+00 8.61069977e-01
3.55858088e-01 1.03262603e+00 -1.91760078e-01 -2.28502706e-01
-1.79805636e-01 6.78616250e-03 7.74480820e-01 -5.59746146e-01
-7.72575617e-01 3.73804510e-01 1.36504292e-01 -1.91460073e-01
-7.07643092e-01 -5.00571370e-01 -1.58077610e+00 1.66543618e-01
-4.26564246e-01 2.57946193e-01 1.40015113e+00 7.57370412e-01
-3.80215138e-01 1.62263662e-01 6.17316246e-01 -2.39034161e-01
-2.25448444e-01 -8.02981079e-01 -8.74537885e-01 5.65665722e-01
-1.12956733e-01 -1.70986995e-01 -1.47688061e-01 5.53809524e-01] | [15.508115768432617, -3.0071685314178467] |
89a5f599-8cbb-4bea-8523-94e829f5a6d1 | automatic-vocabulary-and-graph-verification | 2107.14611 | null | https://arxiv.org/abs/2107.14611v1 | https://arxiv.org/pdf/2107.14611v1.pdf | Automatic Vocabulary and Graph Verification for Accurate Loop Closure Detection | Localizing pre-visited places during long-term simultaneous localization and mapping, i.e. loop closure detection (LCD), is a crucial technique to correct accumulated inconsistencies. As one of the most effective and efficient solutions, Bag-of-Words (BoW) builds a visual vocabulary to associate features and then detect loops. Most existing approaches that build vocabularies off-line determine scales of the vocabulary by trial-and-error, which often results in unreasonable feature association. Moreover, the accuracy of the algorithm usually declines due to perceptual aliasing, as the BoW-based method ignores the positions of visual features. To overcome these disadvantages, we propose a natural convergence criterion based on the comparison between the radii of nodes and the drifts of feature descriptors, which is then utilized to build the optimal vocabulary automatically. Furthermore, we present a novel topological graph verification method for validating candidate loops so that geometrical positions of the words can be involved with a negligible increase in complexity, which can significantly improve the accuracy of LCD. Experiments on various public datasets and comparisons against several state-of-the-art algorithms verify the performance of our proposed approach. | ['Zhengguo Li', 'Fanghong Guo', 'Wei Wang', 'Weihai Chen', 'Jinyu Miao', 'Haosong Yue'] | 2021-07-30 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [-2.10584939e-01 -3.78574163e-01 -3.16300057e-02 -1.44543424e-01
-4.36638027e-01 -6.71649337e-01 4.94792163e-01 8.13322365e-01
-2.75176018e-01 4.98841316e-01 -8.33644792e-02 -2.88743615e-01
-3.62392873e-01 -8.84511054e-01 -6.01873279e-01 -5.64509094e-01
-2.13569343e-01 2.19109952e-01 6.42616808e-01 -2.34242901e-01
6.81168318e-01 4.85050529e-01 -1.64313054e+00 -3.64559859e-01
1.07571292e+00 1.08766484e+00 3.60138655e-01 1.13858551e-01
-4.85936582e-01 2.89515406e-01 -3.79471093e-01 -1.47697451e-02
1.06047608e-01 -9.37693100e-03 -3.77099067e-01 -1.45795226e-01
4.38575834e-01 -1.75508950e-03 -1.82494432e-01 1.18485415e+00
2.12819919e-01 4.01236191e-02 5.10010362e-01 -1.07667994e+00
-6.08392417e-01 1.17101341e-01 -8.37136447e-01 1.50654078e-01
4.73570734e-01 -2.46907726e-01 1.00011837e+00 -8.53297830e-01
5.45104563e-01 1.06581342e+00 9.24498975e-01 -1.83126122e-01
-1.09434819e+00 -7.28864133e-01 4.31617141e-01 2.60042936e-01
-2.09261537e+00 -3.25806111e-01 9.76719975e-01 -3.52574706e-01
5.63334763e-01 3.54194701e-01 6.32052004e-01 2.79326946e-01
2.27510169e-01 2.81240493e-01 7.10980475e-01 -5.13359547e-01
2.04420477e-01 1.70054540e-01 -1.98391508e-02 9.36370790e-01
7.08915412e-01 -2.25718915e-01 -4.84437108e-01 -3.09879184e-01
7.09425867e-01 -4.18927930e-02 -3.39587003e-01 -9.93027031e-01
-1.10853100e+00 5.08913040e-01 8.57836545e-01 4.49848771e-01
-1.36926398e-01 4.45161946e-02 3.15528959e-01 1.64641850e-02
2.76811182e-01 2.30958462e-01 1.09336926e-02 2.23882258e-01
-9.16828573e-01 2.18552649e-01 1.64825961e-01 1.06231678e+00
9.40499961e-01 -5.21688342e-01 7.45705739e-02 6.09228849e-01
4.35547888e-01 5.11576474e-01 2.46520683e-01 -9.83231217e-02
4.21142131e-01 1.04073346e+00 2.78826147e-01 -1.87376499e+00
-4.79237467e-01 -4.62207317e-01 -7.71602273e-01 -1.09957986e-01
2.67786711e-01 5.25079429e-01 -7.20801413e-01 1.44398510e+00
5.85609972e-01 2.04621285e-01 -3.59304726e-01 8.40345025e-01
6.43721402e-01 4.68027383e-01 -6.75349534e-02 -1.89822167e-01
1.16695154e+00 -5.50607383e-01 -7.18279898e-01 -2.62311250e-01
6.53946519e-01 -7.18237817e-01 9.20788646e-01 3.07652615e-02
-5.40871501e-01 -5.80598533e-01 -1.33926356e+00 4.12584580e-02
-6.41010284e-01 1.83468655e-01 4.52987164e-01 4.23359573e-01
-9.47090805e-01 2.53616035e-01 -6.81971192e-01 -5.75556517e-01
1.74000282e-02 1.99589148e-01 -3.31499815e-01 1.98101848e-01
-1.01694858e+00 7.31405079e-01 4.22145754e-01 4.72287893e-01
-3.06706399e-01 -4.41190094e-01 -8.30793858e-01 -1.91114813e-01
3.90437990e-01 -6.64136484e-02 5.13197243e-01 -5.73583543e-01
-7.24298358e-01 6.27567172e-01 -3.46882522e-01 -2.56252229e-01
5.53190470e-01 1.05102554e-01 -4.62498248e-01 5.52664399e-02
3.90195996e-01 4.39862072e-01 7.12701917e-01 -1.41358447e+00
-7.03182220e-01 -5.20601094e-01 -1.45335421e-01 2.19767749e-01
-3.77545416e-01 -3.94815594e-01 -6.96931899e-01 -2.80726910e-01
8.11260998e-01 -7.64227688e-01 -2.92191301e-02 3.25608671e-01
-3.13206524e-01 -4.13121581e-01 8.10632169e-01 -5.35734355e-01
1.70295691e+00 -2.37604547e+00 -1.47193938e-01 7.88194478e-01
2.85274833e-01 -2.61790249e-02 1.13879010e-01 3.55621278e-01
2.91372538e-01 6.82251602e-02 2.08785385e-02 -1.66212849e-03
-2.38864511e-01 -3.01877633e-02 -2.84992784e-01 7.84169912e-01
-8.67102593e-02 5.45355618e-01 -1.00749266e+00 -7.77089417e-01
4.58408296e-01 2.90935189e-01 -1.77378863e-01 -1.56507343e-01
-2.22963281e-02 2.39854097e-01 -6.82281733e-01 5.89415312e-01
9.89212096e-01 -2.73546129e-01 1.61895081e-01 -3.29955399e-01
-4.98974681e-01 1.92217633e-01 -1.43732858e+00 1.64748883e+00
-3.06877345e-01 5.00532806e-01 -3.02303553e-01 -8.50554883e-01
1.35854602e+00 -1.58214748e-01 1.90252632e-01 -8.35960329e-01
5.38266376e-02 3.43853295e-01 -5.45105815e-01 -3.47349614e-01
6.27712309e-01 5.82680702e-01 -9.87975299e-02 8.57527554e-02
-5.37727237e-01 8.43209773e-02 4.01386991e-02 1.73167557e-01
9.28212047e-01 8.94636009e-03 5.01810551e-01 -3.75804096e-01
8.88310313e-01 1.25710025e-01 4.20845538e-01 5.82110047e-01
-2.39135966e-01 4.00288045e-01 3.02307010e-01 -6.24400675e-01
-8.41887236e-01 -8.24885845e-01 -1.78814411e-01 5.87638319e-01
9.58465099e-01 -6.28982723e-01 -4.17390734e-01 -5.18009901e-01
2.42539033e-01 4.60675865e-01 -7.61507928e-01 -1.18649146e-02
-3.98556203e-01 -1.95359766e-01 3.11025828e-01 2.85479188e-01
6.25994563e-01 -4.98893648e-01 -8.54009628e-01 8.85991473e-03
-2.28289872e-01 -8.33865523e-01 -2.68371850e-01 -1.81715980e-01
-6.35589719e-01 -1.19644940e+00 -3.78930002e-01 -9.11379695e-01
1.22296190e+00 8.60160768e-01 5.56811512e-01 6.23897195e-01
-1.56947255e-01 -4.51169498e-02 -4.34553087e-01 1.83250960e-02
9.58835613e-03 1.05923027e-01 9.03103650e-02 -1.53208703e-01
3.68111730e-01 -3.45656306e-01 -7.14937985e-01 4.56159264e-01
-4.29974169e-01 -1.32435597e-02 5.06212711e-01 5.46819150e-01
8.58903587e-01 2.71298379e-01 2.25884110e-01 -2.73468107e-01
5.21022439e-01 -2.69171178e-01 -1.02716613e+00 6.21429324e-01
-7.58071959e-01 3.31283480e-01 5.31254292e-01 -2.57565111e-01
-4.50692087e-01 1.63304031e-01 4.61763948e-01 -3.32912773e-01
1.34447396e-01 4.23315555e-01 4.42241095e-02 -6.25119865e-01
5.05243003e-01 4.57638562e-01 -1.96518138e-01 -2.89521575e-01
4.14721698e-01 7.67265022e-01 4.83638495e-01 -2.32748657e-01
9.64167058e-01 5.36419272e-01 1.58470571e-01 -8.66356969e-01
-3.05898547e-01 -7.73032129e-01 -6.62173629e-01 -3.49292308e-01
4.44064260e-01 -8.29118431e-01 -7.23956704e-01 1.26429915e-01
-1.06950855e+00 4.20204431e-01 4.32742566e-01 1.18674271e-01
-1.08433791e-01 5.40581167e-01 1.63083062e-01 -1.02815568e+00
-2.13812679e-01 -8.47771943e-01 1.01528347e+00 2.98363179e-01
-1.81449607e-01 -7.18986869e-01 6.76571429e-02 -1.05165809e-01
3.12753350e-01 3.92351896e-01 9.21476245e-01 -2.84996837e-01
-7.56762445e-01 -4.32802826e-01 -4.59526926e-01 -3.82846147e-01
1.53535858e-01 7.23727420e-02 -5.89044452e-01 -3.84304941e-01
-4.34688538e-01 2.07791805e-01 4.68319356e-01 1.75120085e-01
1.02361536e+00 -2.65868362e-02 -8.67879868e-01 4.35512125e-01
1.56534553e+00 2.25760683e-01 5.15194058e-01 6.25720918e-01
5.43768167e-01 4.24580932e-01 1.01337051e+00 5.31392276e-01
5.13488173e-01 7.24741161e-01 4.57250059e-01 -2.73912475e-02
1.45526573e-01 -6.64482713e-01 -1.54676303e-01 6.53161645e-01
2.80591488e-01 -1.04006305e-01 -1.03609169e+00 8.51832807e-01
-1.81234360e+00 -4.86844301e-01 -3.22546124e-01 2.50493240e+00
3.83110166e-01 2.63164371e-01 -5.27505018e-02 3.48560363e-01
9.23870981e-01 2.49547452e-01 -2.30894953e-01 1.38739161e-02
1.49682879e-01 -3.15273911e-01 8.82965505e-01 4.46040988e-01
-1.05054438e+00 8.56537879e-01 5.46961069e+00 7.68146157e-01
-1.19446540e+00 -1.06301419e-01 1.09535091e-01 2.60345042e-01
-2.71251470e-01 1.45594358e-01 -6.62666976e-01 4.65361208e-01
1.70177117e-01 -1.92721654e-03 3.85823637e-01 7.80190170e-01
2.37290204e-01 -4.90960419e-01 -6.60630882e-01 1.04318714e+00
1.50082350e-01 -1.17381716e+00 -4.88247983e-02 -1.12338968e-01
5.00406206e-01 -3.22505981e-01 -1.45281166e-01 -1.84276432e-01
-3.30041766e-01 -6.06113672e-01 9.96394873e-01 3.96867216e-01
8.58033240e-01 -8.24704170e-01 7.17210293e-01 1.82977095e-01
-1.85151553e+00 2.91409949e-03 -5.56534588e-01 1.20019093e-01
-5.44289276e-02 5.31489015e-01 -8.24902177e-01 6.52566493e-01
7.91689038e-01 4.63121295e-01 -9.22009706e-01 1.42166173e+00
-1.87604263e-01 6.69310391e-02 -5.44902742e-01 -4.52702314e-01
1.51676491e-01 -4.71442193e-02 5.52919805e-01 8.07487726e-01
3.86734039e-01 -1.99157819e-01 1.95759565e-01 6.82179451e-01
2.13397831e-01 5.53350568e-01 -6.84347451e-01 1.71022356e-01
9.92160201e-01 1.05534196e+00 -1.16762233e+00 -1.84431300e-02
-2.92415828e-01 6.99896276e-01 5.46676815e-01 1.50868133e-01
-8.67005587e-01 -7.77552187e-01 2.33781472e-01 4.70436335e-01
3.98747027e-01 -4.26208436e-01 -3.49303603e-01 -7.01041520e-01
4.71178651e-01 -4.04570222e-01 1.23957679e-01 -5.99969089e-01
-7.09214270e-01 4.34880197e-01 -8.88768211e-02 -1.60623193e+00
2.19235227e-01 -4.23365496e-02 -3.35396975e-01 5.59984982e-01
-1.41998971e+00 -1.10832131e+00 -8.55087459e-01 4.73836273e-01
2.28629887e-01 3.54163170e-01 5.74316382e-01 4.16248351e-01
-3.23285401e-01 6.63121819e-01 2.88577348e-01 1.47806592e-02
7.28900552e-01 -7.81878769e-01 4.01963353e-01 9.45682943e-01
9.25153717e-02 7.02573597e-01 9.32480037e-01 -1.00268328e+00
-1.29419732e+00 -9.56230283e-01 1.01463091e+00 5.08814864e-03
6.53322577e-01 -5.71179330e-01 -8.71510983e-01 3.78690481e-01
-4.81268644e-01 -2.46403069e-05 1.53104573e-01 8.58813524e-02
-3.59116197e-01 -4.69967127e-01 -9.70826209e-01 4.84725028e-01
1.12676501e+00 -4.47038412e-01 -4.85627919e-01 2.15481237e-01
4.43196237e-01 -4.43009615e-01 -3.99370164e-01 3.58723372e-01
7.18876898e-01 -9.92035031e-01 8.39329720e-01 1.86475188e-01
-2.04019710e-01 -8.79942417e-01 3.40295350e-03 -9.29421306e-01
-3.40308428e-01 -3.32675874e-01 1.76608071e-01 1.37907827e+00
1.45583704e-01 -6.87483013e-01 3.49079549e-01 1.35904819e-01
1.69481695e-01 -6.27974629e-01 -1.08004308e+00 -8.32510412e-01
-6.92017734e-01 -1.47986695e-01 8.29713345e-01 8.23926032e-01
5.92389181e-02 -7.79877231e-02 -2.18780324e-01 5.48007965e-01
6.77019417e-01 1.29756212e-01 9.47564244e-01 -1.35046911e+00
4.31182474e-01 -2.91635901e-01 -9.84188855e-01 -1.12342131e+00
-9.43883881e-02 -6.70692623e-01 3.04700106e-01 -1.61027622e+00
-3.46928239e-02 -9.91434813e-01 -2.69326925e-01 2.73262262e-01
-2.34538645e-01 3.95158716e-02 -2.40766481e-01 3.75333607e-01
-8.65355551e-01 4.92648572e-01 8.57977152e-01 -4.41148803e-02
-4.63426620e-01 -3.86553913e-01 -4.94338542e-01 4.62683380e-01
5.27111053e-01 -4.59109962e-01 -3.63420725e-01 -3.92402053e-01
5.06223917e-01 -3.95256668e-01 3.50040555e-01 -1.14400792e+00
4.75239545e-01 -1.03712529e-01 3.85911852e-01 -1.03568506e+00
-1.00048579e-01 -1.02454948e+00 3.06487203e-01 4.73588258e-01
-1.45321131e-01 1.99394256e-01 1.86432540e-01 9.90030468e-01
-5.34435138e-02 -1.25889733e-01 4.58702326e-01 1.73373774e-01
-9.54808772e-01 5.44061884e-03 8.56184214e-03 -3.16403270e-01
1.37035453e+00 -3.73247743e-01 -3.32170337e-01 -2.46543229e-01
5.69256395e-02 5.51749110e-01 7.27836192e-01 4.83944833e-01
7.03553021e-01 -1.40534329e+00 -3.43080640e-01 3.76616925e-01
5.96600413e-01 6.18076213e-02 1.72506198e-01 9.12969828e-01
-9.25058007e-01 5.01859546e-01 5.79966269e-02 -8.36247742e-01
-1.41308200e+00 7.87250459e-01 1.35410026e-01 4.38628010e-02
-6.05652571e-01 6.68090224e-01 -1.58360839e-01 2.07217913e-02
3.48097771e-01 -4.10633713e-01 -2.17720285e-01 5.50525747e-02
4.75899756e-01 3.16254020e-01 1.49692833e-01 -7.78585076e-01
-9.14061844e-01 1.10248828e+00 3.15186917e-03 2.36690953e-01
9.26114976e-01 -5.78993201e-01 -2.77368784e-01 4.34289902e-01
8.71606290e-01 3.01601559e-01 -7.76461601e-01 -1.00991532e-01
7.97384232e-02 -7.65532672e-01 2.41407901e-01 -1.96886227e-01
-6.41026318e-01 6.61639929e-01 8.88764560e-01 4.06138241e-01
9.62669015e-01 -1.52429238e-01 6.10488057e-01 2.52074540e-01
7.07924306e-01 -1.08978248e+00 -2.96859264e-01 5.21640740e-02
6.95311189e-01 -1.18306363e+00 3.45345676e-01 -4.84812945e-01
-1.73517659e-01 1.11658084e+00 4.97824490e-01 -1.01185426e-01
5.27472138e-01 -9.29031074e-02 -1.47550285e-01 -1.94484517e-01
-1.80730596e-01 -9.55155045e-02 2.48113602e-01 3.02756906e-01
3.69046591e-02 -6.88786060e-02 -6.30703151e-01 1.50064901e-01
-5.42504080e-02 -2.06973717e-01 -1.95220150e-02 1.02174044e+00
-7.18337953e-01 -6.83397710e-01 -5.48307955e-01 1.52212396e-01
1.92337066e-01 1.13463238e-01 -2.92844146e-01 7.96377003e-01
2.15612113e-01 9.17174280e-01 1.72183812e-01 -5.35075784e-01
2.95529366e-01 -2.01729253e-01 2.91109771e-01 -1.99321620e-02
-7.34723210e-02 -8.11166167e-02 -2.09464446e-01 -4.64067429e-01
-3.32457066e-01 -4.49459910e-01 -1.37531149e+00 -3.62213016e-01
-8.39990199e-01 2.61231482e-01 6.71163499e-01 7.68223286e-01
5.25871933e-01 1.66742072e-01 7.07839787e-01 -4.99818563e-01
-3.08006436e-01 -5.43911874e-01 -2.78477788e-01 2.99961358e-01
4.39281762e-01 -1.11744750e+00 -4.66594487e-01 -3.81554395e-01] | [7.574217796325684, -2.1839425563812256] |
94c80b0a-6421-4b85-a551-c55a3fbf968e | fairdisco-fairer-ai-in-dermatology-via | 2208.10013 | null | https://arxiv.org/abs/2208.10013v1 | https://arxiv.org/pdf/2208.10013v1.pdf | FairDisCo: Fairer AI in Dermatology via Disentanglement Contrastive Learning | Deep learning models have achieved great success in automating skin lesion diagnosis. However, the ethnic disparity in these models' predictions, where lesions on darker skin types are usually underrepresented and have lower diagnosis accuracy, receives little attention. In this paper, we propose FairDisCo, a disentanglement deep learning framework with contrastive learning that utilizes an additional network branch to remove sensitive attributes, i.e. skin-type information from representations for fairness and another contrastive branch to enhance feature extraction. We compare FairDisCo to three fairness methods, namely, resampling, reweighting, and attribute-aware, on two newly released skin lesion datasets with different skin types: Fitzpatrick17k and Diverse Dermatology Images (DDI). We adapt two fairness-based metrics DPM and EOM for our multiple classes and sensitive attributes task, highlighting the skin-type bias in skin lesion classification. Extensive experimental evaluation demonstrates the effectiveness of FairDisCo, with fairer and superior performance on skin lesion classification tasks. | ['Rafeef Garbi', 'Ghassan Hamarneh', 'Nourhan Bayasi', 'Ben Hers', 'Siyi Du'] | 2022-08-22 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 4.13682997e-01 2.11030141e-01 -6.29166424e-01 -9.09953117e-01
-6.39360070e-01 -2.08601311e-01 5.23979068e-01 4.96560991e-01
-6.07891738e-01 1.00700092e+00 2.32293785e-01 -5.00895120e-02
-4.08772141e-01 -8.82130921e-01 -1.53253913e-01 -8.14043701e-01
1.51023284e-01 2.78946519e-01 -3.43006194e-01 4.76924330e-02
8.54594409e-02 5.31841278e-01 -1.32339907e+00 6.08012140e-01
1.14473879e+00 1.07893658e+00 -7.13569522e-01 5.72907507e-01
3.40145058e-03 1.00775814e+00 -3.65620345e-01 -1.08596599e+00
1.65923148e-01 -1.58075288e-01 -1.05313468e+00 -2.16794178e-01
7.88740993e-01 -5.90251029e-01 -1.57901451e-01 9.18922246e-01
1.05512035e+00 -3.67824972e-01 1.14112866e+00 -1.46653438e+00
-8.23778927e-01 4.03133631e-01 -8.65688264e-01 3.95250395e-02
-1.45772591e-01 1.97012007e-01 1.12272811e+00 -5.10435760e-01
5.82019269e-01 1.22609103e+00 9.71100926e-01 9.51498568e-01
-1.23023808e+00 -7.54294515e-01 -2.17173308e-01 4.46866959e-01
-1.32305729e+00 -5.03104448e-01 4.86979812e-01 -3.05218697e-01
4.80016887e-01 7.74113774e-01 3.95706475e-01 1.48476315e+00
1.76527530e-01 7.31377840e-01 1.71925664e+00 -3.54164302e-01
2.89642304e-01 2.01063693e-01 1.76000878e-01 4.67145920e-01
3.97189260e-01 2.60492146e-01 -3.13598454e-01 -4.50014830e-01
4.74601626e-01 -9.14341882e-02 1.27242077e-02 -2.39300266e-01
-5.14526904e-01 1.00264966e+00 4.87796068e-01 -3.81615132e-01
-4.43763018e-01 -7.41514564e-02 5.99230945e-01 2.01726496e-01
7.46731281e-01 3.99841607e-01 -1.65761128e-01 3.28316122e-01
-6.82094693e-01 3.62392873e-01 6.69191658e-01 3.59702051e-01
2.46452555e-01 -2.81787246e-01 -7.86174715e-01 1.36080718e+00
-1.15979172e-01 3.33579898e-01 2.39801213e-01 -8.51540387e-01
2.20843162e-02 5.60332358e-01 -1.38661504e-01 -6.68951750e-01
-5.36750734e-01 -4.35906291e-01 -1.10011041e+00 5.41360676e-01
4.15689051e-01 -2.77903434e-02 -1.05172777e+00 1.84924245e+00
3.63931477e-01 -3.29523087e-01 -1.34275660e-01 1.07184982e+00
8.50293636e-01 -1.74992040e-01 8.51754785e-01 2.24879503e-01
1.38607800e+00 -7.44600236e-01 -5.55473745e-01 1.65972769e-01
4.72088605e-01 -5.56492805e-01 9.70302582e-01 3.91436189e-01
-1.03943253e+00 4.03914973e-02 -8.56413186e-01 -3.82214814e-01
-4.02311623e-01 -2.39807870e-02 9.54750538e-01 9.00722146e-01
-8.91366541e-01 7.46666014e-01 -2.75376648e-01 -4.21060085e-01
1.42542696e+00 1.96423367e-01 -5.72602212e-01 -2.35207140e-01
-1.43086016e+00 1.02657044e+00 -1.14582619e-02 -3.44248444e-01
-7.79194057e-01 -1.13115954e+00 -5.87671757e-01 4.92743626e-02
2.66701460e-01 -8.45456064e-01 9.41456318e-01 -1.13595629e+00
-1.08509386e+00 1.45383787e+00 2.09118351e-01 -3.98507237e-01
9.24530029e-01 9.62635502e-02 -5.83019197e-01 1.76629096e-01
-7.32022244e-03 6.25501871e-01 8.48307490e-01 -9.86289740e-01
-6.10003233e-01 -6.16279423e-01 1.43568814e-02 1.79233700e-01
-5.64105868e-01 1.48091782e-02 3.94134313e-01 -5.74503541e-01
-6.43500865e-01 -6.37213528e-01 -2.99534202e-01 7.25613236e-01
-6.70412779e-01 -2.97000945e-01 1.23906717e-01 -7.58910120e-01
1.16275203e+00 -1.93171096e+00 -2.22027585e-01 4.17749971e-01
7.42314160e-01 4.89075363e-01 -4.63479668e-01 5.73482700e-02
-4.91886914e-01 5.04444003e-01 -1.85915351e-01 -1.39771536e-01
1.85919628e-01 6.45564348e-02 9.10831094e-02 4.53856498e-01
7.83662081e-01 7.63656080e-01 -8.23471725e-01 -7.95362651e-01
1.68728188e-01 4.31906313e-01 -5.45196414e-01 -6.77087605e-02
2.37023637e-01 -1.86175555e-01 -2.53488868e-01 9.62757289e-01
1.15846193e+00 -1.51768252e-01 1.90480649e-01 -2.75547922e-01
1.88387200e-01 4.06407975e-02 -6.02520466e-01 1.14832234e+00
-3.32193404e-01 2.58336335e-01 8.39248747e-02 -4.96448070e-01
6.73078477e-01 1.20779769e-02 4.38087612e-01 -8.52331400e-01
9.79755521e-02 2.11149111e-01 2.08118543e-01 -6.38544858e-01
2.68475801e-01 -4.59444880e-01 1.43351048e-01 1.88609436e-01
-1.57679349e-01 1.12823397e-01 -3.98630202e-02 1.79086939e-01
9.46907759e-01 -2.88616598e-01 6.75843477e-01 -4.62950647e-01
2.77693152e-01 -2.14672655e-01 6.22644961e-01 6.25426888e-01
-8.74821544e-01 6.75521791e-01 9.39056516e-01 -4.56572503e-01
-1.07996285e+00 -1.20281732e+00 -7.85561621e-01 1.17730904e+00
-1.97839364e-01 -1.83572583e-02 -8.59091699e-01 -1.08555317e+00
5.06365299e-01 7.19822288e-01 -1.29673040e+00 -3.74293208e-01
-1.10452827e-04 -1.28512168e+00 1.01762664e+00 6.71392560e-01
2.73816973e-01 -5.65256774e-01 -1.83102682e-01 -3.91782731e-01
-3.82935181e-02 -4.73418742e-01 -2.96955407e-01 1.38655394e-01
-1.83448315e-01 -1.43184388e+00 -1.01148117e+00 -3.09102982e-01
6.44321322e-01 -1.00540146e-01 1.23179579e+00 -8.37691575e-02
-9.85929430e-01 9.39686149e-02 -1.18252940e-01 -7.07300365e-01
-3.58834386e-01 1.18136266e-02 -1.68619186e-01 1.44637719e-01
6.53487921e-01 -2.22836673e-01 -9.31139052e-01 1.91918001e-01
-9.88213062e-01 -1.23792559e-01 7.34225631e-01 1.13647091e+00
4.45527226e-01 -2.63319075e-01 7.77720451e-01 -1.31815088e+00
6.88783407e-01 -7.03156412e-01 2.20941901e-01 3.90119463e-01
-1.01822805e+00 -1.84045494e-01 3.88467968e-01 -2.30540231e-01
-1.09430003e+00 -3.71249199e-01 -2.00449109e-01 -1.20224774e-01
-1.64454818e-01 1.54749200e-01 -1.41965404e-01 -2.19508961e-01
1.18193042e+00 -4.21675116e-01 3.48775864e-01 -2.76627243e-01
1.44036859e-01 9.17620063e-01 1.84717193e-01 -4.05580282e-01
3.03460598e-01 5.24137259e-01 1.59998059e-01 -3.77754331e-01
-9.76801395e-01 -1.34151727e-01 -2.45504722e-01 -1.43399844e-02
6.62571907e-01 -7.99375713e-01 -7.52118289e-01 7.77036667e-01
-6.27650678e-01 -1.79246023e-01 -5.89732766e-01 2.34870359e-01
-2.22338662e-01 3.32936406e-01 -6.52219951e-01 -6.92394376e-01
-7.22039044e-01 -6.25335932e-01 8.93664539e-01 3.82140756e-01
-5.62646985e-01 -9.80090499e-01 9.19715092e-02 4.46010888e-01
6.69191182e-01 6.01013184e-01 1.36035764e+00 -8.72533500e-01
2.98417449e-01 -2.81197757e-01 -7.97621310e-01 4.44993585e-01
3.33415478e-01 2.28119597e-01 -1.48241103e+00 -1.50890380e-01
-5.32958210e-01 -9.46597159e-01 1.21619749e+00 4.56572890e-01
1.82914436e+00 -3.36605191e-01 -5.64557798e-02 8.35109651e-01
1.48609173e+00 -2.76574582e-01 8.51423919e-01 1.72753811e-01
5.66853583e-01 1.03647768e+00 3.54600668e-01 6.22373462e-01
5.28043628e-01 4.18059200e-01 6.07274175e-01 -7.30516076e-01
-5.14491379e-01 -1.34624064e-01 -2.94544488e-01 3.14192846e-02
-3.68862391e-01 -1.82567328e-01 -7.75774598e-01 4.49562043e-01
-1.52464998e+00 -8.34607780e-01 -8.46747011e-02 2.25396824e+00
1.06774449e+00 -1.68116599e-01 2.97326922e-01 4.03130474e-03
9.04735446e-01 1.87217563e-01 -9.31888282e-01 -7.70246506e-01
-5.17468393e-01 5.01515806e-01 6.44493699e-01 2.58039892e-01
-1.16316295e+00 5.15519917e-01 6.39996910e+00 1.20211875e+00
-1.10984731e+00 -2.50253044e-02 1.21463549e+00 -4.74165857e-01
-6.31081343e-01 -4.50438976e-01 -1.52099907e-01 4.38412637e-01
4.96182024e-01 -2.25147560e-01 2.35954911e-01 6.16044044e-01
-5.99727184e-02 1.30819708e-01 -1.17142904e+00 8.65770400e-01
6.96967766e-02 -1.05595195e+00 2.41395473e-01 -9.79427830e-04
6.34981573e-01 -2.92050689e-01 5.92655540e-01 2.00170964e-01
4.45436537e-01 -1.49751902e+00 2.78039366e-01 6.88263237e-01
1.37719870e+00 -8.41329753e-01 9.45605993e-01 -4.83006746e-01
-3.28723699e-01 -1.16063185e-01 -4.46806699e-01 1.57839254e-01
-4.73135233e-01 9.24857020e-01 -8.82041156e-01 5.84928572e-01
7.85767317e-01 5.78000546e-01 -8.17417502e-01 1.00925004e+00
-6.82046413e-02 3.93598646e-01 1.21865846e-01 -1.29115239e-01
-5.26162162e-02 2.75956124e-01 1.53842255e-01 1.27050257e+00
-2.14206070e-01 -9.20850486e-02 -3.36827725e-01 9.67887282e-01
-1.96002737e-01 2.71448940e-01 -1.46342963e-01 -4.61746976e-02
3.65855604e-01 1.59043801e+00 -8.15210268e-02 -1.84313759e-01
-7.65006393e-02 7.46885896e-01 4.05310303e-01 1.62400410e-01
-7.80181348e-01 -5.92889547e-01 9.71186280e-01 1.35332525e-01
-4.28205729e-01 7.83594608e-01 -6.78133011e-01 -8.89924526e-01
-2.62270659e-01 -1.01398313e+00 9.05819833e-01 -5.23905694e-01
-2.12535954e+00 3.02004337e-01 -3.48011792e-01 -8.44801009e-01
1.19501248e-01 -7.11043954e-01 -4.38069969e-01 1.22889280e+00
-1.94831252e+00 -1.49886167e+00 -4.57874358e-01 6.48390710e-01
-1.45594627e-01 -1.66544646e-01 1.06984615e+00 3.43641669e-01
-8.27834845e-01 1.24810648e+00 -4.34254529e-03 1.65153250e-01
1.38244379e+00 -1.61286807e+00 8.81070644e-02 2.10002303e-01
-6.74253047e-01 2.15900272e-01 1.50640234e-01 -4.60534424e-01
-5.91916561e-01 -1.20794392e+00 8.11030924e-01 -4.48162317e-01
4.14231658e-01 -1.26044467e-01 -8.61130476e-01 8.79628509e-02
2.50769798e-02 9.87991616e-02 1.39722693e+00 3.22171658e-01
-9.45589662e-01 -4.10076171e-01 -1.89244640e+00 6.43507063e-01
1.02178204e+00 -5.16490757e-01 -1.71889905e-02 3.59774858e-01
2.42658347e-01 -1.44600002e-02 -1.19856703e+00 5.57116210e-01
9.45085764e-01 -1.04662561e+00 9.92415786e-01 -1.24496055e+00
1.12324810e+00 4.34142709e-01 -1.78187102e-01 -1.45960665e+00
-7.74717093e-01 -1.63639054e-01 2.37509355e-01 1.17921984e+00
2.95966923e-01 -6.60659730e-01 9.27459598e-01 1.00441146e+00
2.30064616e-01 -1.10989225e+00 -8.75585914e-01 -3.44964504e-01
5.70020556e-01 8.46011713e-02 8.61850917e-01 1.42184925e+00
-1.41849220e-01 -4.57120985e-02 -4.40648407e-01 -1.99040726e-01
9.41360712e-01 -1.27740532e-01 2.24122405e-01 -1.31226647e+00
7.10646138e-02 -6.91553295e-01 -3.97061408e-01 3.36420685e-01
1.90845847e-01 -1.20437050e+00 -4.23005551e-01 -1.29150367e+00
8.29137623e-01 -5.93703985e-01 -7.73365676e-01 8.60154748e-01
-4.78816777e-01 3.17570657e-01 -1.11032195e-01 5.21018170e-03
-2.21653476e-01 3.05104971e-01 1.24725878e+00 -5.07959843e-01
3.76609623e-01 -1.06260784e-01 -1.27508569e+00 6.76747441e-01
8.75695169e-01 -3.81469935e-01 -2.57398933e-01 -3.80760729e-01
3.34100462e-02 -3.93769115e-01 7.33503699e-01 -5.37426651e-01
-2.00471684e-01 -4.72710043e-01 9.08101082e-01 1.81724980e-01
1.11493632e-01 -4.14687961e-01 -3.21195185e-01 6.93445504e-01
-1.01111746e+00 -5.33814073e-01 6.96382225e-02 4.57552314e-01
1.16026700e-01 -1.37587599e-02 1.26241529e+00 -1.42574199e-02
-6.22397840e-01 6.09120727e-01 -1.41759291e-02 2.18936428e-01
9.95696604e-01 -3.07823159e-02 -1.07582593e+00 -7.97869414e-02
-7.74349809e-01 2.17372224e-01 3.95016760e-01 2.58126736e-01
5.74018836e-01 -1.43698335e+00 -1.31235313e+00 1.67286977e-01
6.00508630e-01 -4.75381523e-01 7.98439801e-01 8.88753772e-01
-4.40293849e-01 -7.10966662e-02 -8.12745869e-01 -3.19040477e-01
-1.31298268e+00 3.26920986e-01 4.55479234e-01 -3.15225929e-01
8.85499455e-03 1.07415140e+00 2.00184762e-01 -6.53438509e-01
8.90224203e-02 1.24178857e-01 -1.64804369e-01 2.29345292e-01
5.63166440e-01 9.91735458e-01 1.56228870e-01 -1.26028866e-01
-5.35927832e-01 2.64055520e-01 -6.21125877e-01 3.71581703e-01
1.04440391e+00 1.38678059e-01 -3.10568571e-01 -1.57574669e-01
1.10223103e+00 -8.28098413e-03 -9.44130182e-01 -7.14485645e-02
-2.95249552e-01 -6.66673422e-01 1.40316918e-01 -1.62806976e+00
-9.81322706e-01 9.60309744e-01 1.03833830e+00 1.11872502e-01
1.22017932e+00 -4.58384186e-01 5.76090097e-01 -9.74812284e-02
1.08745575e-01 -1.37253451e+00 -2.73211509e-01 -2.14029536e-01
7.54594088e-01 -1.43911314e+00 2.39822254e-01 -5.40687978e-01
-8.41215312e-01 8.45401525e-01 7.78386772e-01 8.92979652e-02
3.86396557e-01 2.48618975e-01 1.86967760e-01 5.54814972e-02
-7.84223199e-01 -1.20536715e-01 3.02436888e-01 9.18771744e-01
4.14762497e-01 5.67983687e-01 -5.38781881e-01 7.98428714e-01
1.08316466e-01 1.86415672e-01 4.03082311e-01 4.83970612e-01
7.23651275e-02 -1.03890324e+00 -2.75030211e-02 1.12663734e+00
-7.97817230e-01 -1.85886219e-01 -9.30397272e-01 7.04421699e-01
3.00797224e-01 6.55151486e-01 1.26301810e-01 -3.90188485e-01
2.91961372e-01 -8.41776654e-02 5.35941184e-01 -3.23410898e-01
-7.16139436e-01 -5.39121270e-01 4.88002062e-01 -7.11578131e-01
-2.95543641e-01 -4.44479078e-01 -6.94715559e-01 -7.19259143e-01
-1.08967990e-01 -3.08489710e-01 5.03452718e-01 5.20591259e-01
4.86402243e-01 5.06920397e-01 8.27528059e-01 -7.72364587e-02
-1.02975404e+00 -6.34609222e-01 -9.42589879e-01 8.34800303e-01
3.43903244e-01 -4.47901994e-01 -3.10653836e-01 -5.58126152e-01] | [15.434767723083496, -2.6467080116271973] |
ed7f978e-cb9a-43a4-8227-74cef709954c | adversarial-attack-and-defense-of-yolo | 2202.04781 | null | https://arxiv.org/abs/2202.04781v2 | https://arxiv.org/pdf/2202.04781v2.pdf | Adversarial Attack and Defense of YOLO Detectors in Autonomous Driving Scenarios | Visual detection is a key task in autonomous driving, and it serves as a crucial foundation for self-driving planning and control. Deep neural networks have achieved promising results in various visual tasks, but they are known to be vulnerable to adversarial attacks. A comprehensive understanding of deep visual detectors' vulnerability is required before people can improve their robustness. However, only a few adversarial attack/defense works have focused on object detection, and most of them employed only classification and/or localization losses, ignoring the objectness aspect. In this paper, we identify a serious objectness-related adversarial vulnerability in YOLO detectors and present an effective attack strategy targeting the objectness aspect of visual detection in autonomous vehicles. Furthermore, to address such vulnerability, we propose a new objectness-aware adversarial training approach for visual detection. Experiments show that the proposed attack targeting the objectness aspect is 45.17% and 43.50% more effective than those generated from classification and/or localization losses on the KITTI and COCO traffic datasets, respectively. Also, the proposed adversarial defense approach can improve the detectors' robustness against objectness-oriented attacks by up to 21% and 12% mAP on KITTI and COCO traffic, respectively. | ['Qing Tian', 'Jung Im Choi'] | 2022-02-10 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [-1.38416156e-01 -9.69476178e-02 -6.09146170e-02 -1.15005597e-01
-3.28367233e-01 -6.78532124e-01 6.75445318e-01 -1.06566086e-01
-5.01092613e-01 5.67807078e-01 -4.19369608e-01 -4.64923471e-01
2.83621967e-01 -1.00160038e+00 -9.01109636e-01 -8.49861503e-01
-5.25553040e-02 -9.48126689e-02 8.33026588e-01 -5.28656960e-01
2.13128552e-01 8.47573996e-01 -1.61324060e+00 4.37369533e-02
7.47299969e-01 1.19227982e+00 -3.20339322e-01 8.62631202e-01
1.01431638e-01 1.10566604e+00 -1.01489425e+00 -6.09254718e-01
4.90531474e-01 -8.83304048e-03 -2.77209520e-01 -3.03115368e-01
7.48029292e-01 -4.82420683e-01 -8.34285378e-01 1.52373338e+00
5.59939563e-01 -2.07153648e-01 6.55346930e-01 -1.92931795e+00
-6.21153116e-01 2.76124716e-01 -5.77069759e-01 6.75481856e-01
-2.17984393e-01 6.65488958e-01 3.47960353e-01 -6.15190208e-01
8.18863437e-02 1.44433868e+00 4.33268547e-01 7.32517421e-01
-5.99502623e-01 -1.43299913e+00 9.40765440e-02 7.07430780e-01
-1.35947728e+00 -3.66900593e-01 8.18351984e-01 -5.35169721e-01
5.87008178e-01 2.30834648e-01 2.32058376e-01 1.22031057e+00
6.30634248e-01 7.46878028e-01 1.05287898e+00 -6.85292780e-02
-4.73432168e-02 4.48493868e-01 1.66137233e-01 6.86397970e-01
5.57828963e-01 9.18999374e-01 -1.53313369e-01 9.09537897e-02
2.41180867e-01 -1.77665100e-01 2.65321434e-01 -2.28122875e-01
-5.87287724e-01 1.15320289e+00 8.07883501e-01 -2.22644672e-01
7.76170194e-02 4.45371866e-01 6.62212312e-01 2.57763177e-01
4.43255864e-02 5.48376888e-02 1.01541188e-02 3.49791378e-01
-2.86730170e-01 3.48435611e-01 2.43526503e-01 1.15911472e+00
6.97869599e-01 6.92938805e-01 -3.53826672e-01 3.71530294e-01
3.17223877e-01 1.19831789e+00 6.29593208e-02 -7.04685748e-01
5.45877874e-01 3.25517535e-01 -4.74480018e-02 -1.49476027e+00
-3.47981632e-01 -4.13195491e-01 -7.63660133e-01 8.81813765e-01
2.86850214e-01 -2.68372178e-01 -1.02124262e+00 1.61282623e+00
2.43594557e-01 7.80894756e-02 2.47171104e-01 8.05396795e-01
9.70007956e-01 6.18574679e-01 4.16732341e-01 1.99130788e-01
1.41488326e+00 -7.87670553e-01 -7.15041220e-01 -4.49522465e-01
3.26094270e-01 -7.98573792e-01 6.32356822e-01 1.94721699e-01
-6.73743963e-01 -8.07471097e-01 -1.56233358e+00 2.81999350e-01
-7.08416820e-01 -6.75452352e-02 4.15715724e-01 1.45252299e+00
-3.68157506e-01 -3.49752121e-02 -5.05617738e-01 -7.85016418e-02
7.54668951e-01 2.11781979e-01 -1.46495074e-01 -6.93166116e-03
-1.57685256e+00 1.23921025e+00 4.19656992e-01 -6.36080727e-02
-1.68289471e+00 -4.48451579e-01 -8.12077701e-01 -6.97667375e-02
5.11263967e-01 -1.35731936e-01 9.51283216e-01 -7.21939206e-01
-1.00626493e+00 6.20019257e-01 1.66242152e-01 -1.06848991e+00
7.72092700e-01 -2.44874433e-01 -8.20320070e-01 1.40670523e-01
8.67625251e-02 7.75602043e-01 1.25749183e+00 -1.41284847e+00
-8.55906546e-01 -3.45652997e-01 3.52292717e-01 -1.51416212e-01
-5.34903049e-01 1.04393467e-01 7.07518458e-02 -6.59588277e-01
-4.22474027e-01 -8.76640081e-01 -1.12272217e-03 2.04155400e-01
-5.42060316e-01 -6.24151044e-02 1.47066891e+00 -3.94583941e-01
9.63533819e-01 -2.34807897e+00 -6.02777064e-01 5.94609566e-02
3.51991326e-01 8.60759318e-01 -3.52702998e-02 1.43407835e-02
1.73303306e-01 2.34147370e-01 -5.33875525e-02 2.71367282e-01
-1.65885333e-02 6.72192797e-02 -7.68305779e-01 6.10773802e-01
5.48604429e-01 7.73563385e-01 -6.57650471e-01 -4.02880311e-01
4.96066153e-01 2.69762158e-01 -3.77586842e-01 7.82262981e-02
7.47794509e-02 1.01537421e-01 -5.64726830e-01 8.38182449e-01
1.07251859e+00 6.67088866e-01 -5.45322120e-01 -1.57644987e-01
5.64552359e-02 -3.09198648e-01 -7.86597311e-01 3.71758521e-01
-2.50503182e-01 1.01696193e+00 -1.59799293e-01 -9.45465088e-01
1.17999947e+00 -2.19426062e-02 1.81093663e-02 -9.31833267e-01
4.50738251e-01 3.35423835e-02 4.22509551e-01 -4.84414965e-01
5.15636027e-01 -6.55057207e-02 -2.96996325e-01 -1.18837394e-01
-3.32187146e-01 -5.56279682e-02 -6.41251355e-02 2.83225626e-01
1.13191414e+00 -4.10518974e-01 -4.20108996e-02 -1.50521234e-01
9.45767879e-01 1.84556559e-01 6.71374679e-01 9.00898576e-01
-9.39969003e-01 1.58777624e-01 5.24642944e-01 -5.39211094e-01
-1.00675046e+00 -1.30657482e+00 -3.25919032e-01 6.74051881e-01
7.41533995e-01 1.31581396e-01 -6.99493945e-01 -1.02054298e+00
3.70128274e-01 7.71467865e-01 -7.34789133e-01 -9.69920337e-01
-6.01619899e-01 -6.11276746e-01 1.37599170e+00 7.16034532e-01
1.04721546e+00 -1.11090064e+00 -5.35791278e-01 -2.89782267e-02
2.58018464e-01 -1.39858830e+00 -3.35558355e-01 -3.58803459e-02
-2.57480919e-01 -1.21444440e+00 -4.45862174e-01 -6.35627568e-01
5.07370889e-01 6.32513583e-01 6.99492574e-01 7.27070868e-02
-5.36049247e-01 2.73993891e-02 -2.39663064e-01 -1.11188459e+00
-9.31595743e-01 -1.77530959e-01 3.22517604e-01 1.92133561e-01
4.30352360e-01 -7.71135511e-03 -3.77660424e-01 7.89144456e-01
-7.76914299e-01 -6.47219896e-01 6.37416601e-01 6.77406847e-01
2.98693813e-02 4.28643465e-01 7.60745227e-01 -6.31873310e-01
3.80419403e-01 -4.30013984e-01 -9.63910818e-01 -1.42936990e-01
-4.80978787e-01 -2.42037624e-01 7.60232151e-01 -5.78765392e-01
-8.86750162e-01 -4.90611158e-02 -3.09346467e-01 -6.38078392e-01
-2.74858624e-01 -1.84836179e-01 -4.68306690e-01 -6.89538896e-01
8.96472573e-01 1.52022004e-01 -1.48132071e-02 5.98576516e-02
2.32775167e-01 5.26839435e-01 5.07188022e-01 -1.24971628e-01
1.47553957e+00 5.64630866e-01 2.09955901e-01 -9.48695421e-01
-5.00048220e-01 -7.95111135e-02 -1.46389484e-01 -4.03873682e-01
1.05931032e+00 -9.21084344e-01 -1.03730500e+00 8.39006841e-01
-1.03995085e+00 8.96622092e-02 1.87320158e-01 2.69389421e-01
-3.00640851e-01 4.64073628e-01 -2.20468238e-01 -8.25289905e-01
-8.43139440e-02 -1.36179996e+00 5.99740565e-01 3.17368835e-01
5.05035222e-01 -5.76363444e-01 -4.04559493e-01 3.74731362e-01
5.78331530e-01 4.10297751e-01 6.19375825e-01 -5.07436275e-01
-7.66965210e-01 -7.15878785e-01 -5.63263178e-01 7.12123513e-01
-1.56048194e-01 9.91131961e-02 -1.12402391e+00 -4.31245625e-01
-4.31217775e-02 -4.53373313e-01 1.04097152e+00 1.12476453e-01
1.06817484e+00 -2.73437619e-01 -3.01203609e-01 5.48518896e-01
1.23930752e+00 5.59714198e-01 9.02122378e-01 4.95405018e-01
7.77882159e-01 4.57342446e-01 8.74805927e-01 1.77332431e-01
5.87342978e-02 6.47960126e-01 1.15752840e+00 -1.28888071e-01
-1.67228356e-01 -1.54238418e-01 6.48786128e-01 -1.20605402e-01
2.83701628e-01 -2.66054571e-01 -5.95900059e-01 4.25900161e-01
-1.41539764e+00 -1.12158060e+00 -3.59323829e-01 1.94760704e+00
1.82250917e-01 8.94439101e-01 2.87615359e-01 3.53384912e-01
8.82859826e-01 3.28714281e-01 -7.55557895e-01 -6.09803915e-01
-2.81727463e-01 -2.41713911e-01 1.26500142e+00 3.61467570e-01
-1.46430278e+00 1.24122977e+00 5.99400520e+00 1.06954968e+00
-1.05015898e+00 1.70414224e-01 5.55491924e-01 8.05708691e-02
2.45251775e-01 -2.52936333e-01 -1.14940941e+00 5.20509660e-01
8.19895744e-01 -2.27278247e-01 -3.67930606e-02 1.18901086e+00
9.18354839e-02 6.01594374e-02 -5.24053812e-01 7.16883421e-01
2.00094298e-01 -1.09957862e+00 1.46021649e-01 1.73885733e-01
4.49929446e-01 -1.46401316e-01 4.36078697e-01 6.09778345e-01
3.36305380e-01 -1.12201202e+00 9.12204504e-01 1.39060998e-02
4.88276213e-01 -1.31909132e+00 9.62228298e-01 2.88812965e-01
-1.25107253e+00 -3.78616422e-01 -6.68632627e-01 8.78789797e-02
-4.32399940e-03 1.47364795e-01 -5.25176048e-01 3.37384939e-01
7.79171944e-01 3.34709197e-01 -8.18940938e-01 7.92624950e-01
-4.67695594e-01 6.78762913e-01 -4.98301024e-03 -1.75639644e-01
4.75951582e-01 5.18250406e-01 9.13164496e-01 9.18178678e-01
-1.10638782e-01 -3.24230403e-01 2.38437608e-01 8.37141633e-01
-5.53363636e-02 -3.15122277e-01 -9.97864902e-01 1.63318649e-01
5.36316574e-01 1.10790157e+00 -4.30627465e-01 -2.86560118e-01
-2.90968627e-01 5.30202925e-01 -4.24313098e-02 1.28415555e-01
-1.57429910e+00 -6.58758879e-01 1.09885228e+00 1.86520386e-02
3.08385372e-01 -1.48190886e-01 -3.89233202e-01 -6.35795355e-01
8.05083662e-03 -9.79992211e-01 1.83203667e-01 -3.55246991e-01
-1.08733416e+00 7.47982264e-01 6.71704486e-02 -1.45365834e+00
3.50775048e-02 -8.78132999e-01 -7.89787352e-01 7.34332144e-01
-1.60658872e+00 -1.09322655e+00 -4.55692381e-01 6.31363094e-01
5.31603336e-01 -8.30020249e-01 2.23247841e-01 4.39479053e-01
-8.01550031e-01 1.22463584e+00 -2.62090772e-01 4.00899291e-01
5.09150207e-01 -7.83607304e-01 5.90580463e-01 1.31411576e+00
-1.60452232e-01 1.09164529e-01 7.85267293e-01 -4.61173922e-01
-1.32244921e+00 -1.45268154e+00 2.94391125e-01 -6.30150259e-01
5.76678455e-01 -3.01618755e-01 -7.84363627e-01 4.03314978e-01
1.70509309e-01 1.59329504e-01 -2.94989310e-02 -5.85457206e-01
-4.31876987e-01 -3.37704271e-01 -1.51319921e+00 7.05608249e-01
6.97574914e-01 -3.49819839e-01 -3.43482226e-01 1.52061224e-01
8.66391003e-01 -2.58208036e-01 -3.05276126e-01 5.81799924e-01
4.83113945e-01 -1.09249353e+00 1.22117567e+00 -5.50675213e-01
1.38304904e-01 -6.33333027e-01 -1.79274261e-01 -8.47868919e-01
-3.49159539e-01 -3.84640604e-01 -5.55449799e-02 1.18746245e+00
1.37826517e-01 -8.96425486e-01 8.61858964e-01 -6.70561148e-03
-3.07240039e-01 -2.18945384e-01 -9.93199944e-01 -1.24659753e+00
2.89227962e-01 -7.12634265e-01 4.25320476e-01 5.57131112e-01
-7.34296024e-01 6.20181076e-02 -6.04838014e-01 8.38111997e-01
8.41127336e-01 -6.46349072e-01 1.13944423e+00 -7.46759593e-01
2.25898132e-01 -5.02351522e-01 -1.17946053e+00 -5.85653961e-01
1.02927007e-01 -6.54341877e-01 1.04087137e-01 -9.04346168e-01
-1.46758497e-01 -3.31303954e-01 -3.32381040e-01 3.24838609e-01
-2.73674309e-01 8.02277207e-01 3.83190036e-01 -1.51466224e-02
-2.01383233e-01 5.07463336e-01 9.32749927e-01 -7.70511270e-01
2.53708869e-01 2.15745792e-01 -5.73154449e-01 8.55782211e-01
9.92598295e-01 -7.33764231e-01 -3.55444103e-01 -2.36331016e-01
-2.70074010e-01 -4.44720358e-01 8.28253567e-01 -1.41798735e+00
6.07127771e-02 -2.73824364e-01 3.32654566e-01 -7.68554449e-01
3.34787816e-01 -9.00190294e-01 -2.06923544e-01 1.07914472e+00
-1.01790294e-01 -5.52072823e-02 7.34201074e-01 7.32440591e-01
-1.44272342e-01 -1.06460676e-01 1.31798017e+00 2.04413310e-01
-1.17123163e+00 3.61389220e-01 -6.94361508e-01 2.59997807e-02
1.72654116e+00 -2.45736659e-01 -7.04130828e-01 -2.02757522e-01
-3.97747427e-01 3.39078516e-01 1.37709066e-01 9.08896029e-01
8.44475746e-01 -1.50069666e+00 -8.18610013e-01 4.05130237e-01
3.91934633e-01 -4.49860066e-01 3.30534428e-01 3.96558285e-01
-6.56679332e-01 4.65094209e-01 -6.42486691e-01 -6.05425239e-01
-1.49986577e+00 1.08153772e+00 4.11046892e-01 8.69305432e-02
-3.92437577e-01 7.66855478e-01 6.00809216e-01 3.69480960e-02
4.22898799e-01 5.49336411e-02 -2.40635827e-01 -2.73220688e-01
6.09907150e-01 7.06991434e-01 -1.66755076e-02 -9.89394844e-01
-6.00124359e-01 5.20735264e-01 -1.75771475e-01 2.73610562e-01
4.00496572e-01 5.99002764e-02 4.21049327e-01 -1.74325302e-01
1.02748477e+00 2.37292796e-02 -1.26170230e+00 2.60264277e-02
-2.61971921e-01 -5.77496409e-01 4.60492298e-02 -6.46018267e-01
-1.28132749e+00 1.18273830e+00 1.03459847e+00 4.46759015e-01
8.15291762e-01 -2.52077460e-01 1.01894951e+00 3.17099184e-01
3.67146760e-01 -7.64224768e-01 2.25417286e-01 5.14112413e-01
8.12955558e-01 -1.50802648e+00 -1.25006631e-01 -6.58028841e-01
-6.65060639e-01 7.37094283e-01 1.07783532e+00 -4.22379732e-01
6.27789319e-01 2.97327787e-01 3.48870099e-01 4.23619263e-02
-4.64121193e-01 -3.61572564e-01 3.19306463e-01 9.86203611e-01
-3.15949380e-01 4.78430949e-02 -1.24769108e-02 2.40947336e-01
-3.50435264e-02 -5.33962727e-01 5.67944407e-01 7.25872457e-01
-7.90719986e-01 -7.21538365e-01 -7.92132020e-01 2.88806468e-01
-5.44202030e-01 6.41169772e-02 -4.39887553e-01 8.85014057e-01
3.72251630e-01 1.28401077e+00 -1.24078058e-01 -8.16405833e-01
5.13272405e-01 -3.48439604e-01 2.82032210e-02 -1.92527860e-01
-3.45005512e-01 -4.33781087e-01 -1.45943603e-03 -4.33644921e-01
6.62875026e-02 -2.22003192e-01 -1.04816151e+00 -6.81283355e-01
-2.72450209e-01 -1.38663247e-01 6.31019413e-01 8.03591728e-01
1.09617203e-01 7.70494103e-01 1.02693832e+00 -6.44452929e-01
-8.67505550e-01 -7.22401202e-01 -4.06821638e-01 2.99457431e-01
4.60371315e-01 -1.11101854e+00 -6.38867497e-01 -3.28361303e-01] | [5.426267147064209, 7.8778767585754395] |
022e1f51-2603-4509-8985-8c7dff0baf9d | harms-a-hardware-acceleration-architecture | 2112.06772 | null | https://arxiv.org/abs/2112.06772v1 | https://arxiv.org/pdf/2112.06772v1.pdf | hARMS: A Hardware Acceleration Architecture for Real-Time Event-Based Optical Flow | Event-based vision sensors produce asynchronous event streams with high temporal resolution based on changes in the visual scene. The properties of these sensors allow for accurate and fast calculation of optical flow as events are generated. Existing solutions for calculating optical flow from event data either fail to capture the true direction of motion due to the aperture problem, do not use the high temporal resolution of the sensor, or are too computationally expensive to be run in real time on embedded platforms. In this research, we first present a faster version of our previous algorithm, ARMS (Aperture Robust Multi-Scale flow). The new optimized software version (fARMS) significantly improves throughput on a traditional CPU. Further, we present hARMS, a hardware realization of the fARMS algorithm allowing for real-time computation of true flow on low-power, embedded platforms. The proposed hARMS architecture targets hybrid system-on-chip devices and was designed to maximize configurability and throughput. The hardware architecture and fARMS algorithm were developed with asynchronous neuromorphic processing in mind, abandoning the common use of an event frame and instead operating using only a small history of relevant events, allowing latency to scale independently of the sensor resolution. This change in processing paradigm improved the estimation of flow directions by up to 73% compared to the existing method and yielded a demonstrated hARMS throughput of up to 1.21 Mevent/s on the benchmark configuration selected. This throughput enables real-time performance and makes it the fastest known realization of aperture-robust, event-based optical flow to date. | ['Ryad B. Benosman', 'Alan D. George', 'Himanshu Akolkar', 'Daniel C. Stumpp'] | 2021-12-13 | null | null | null | null | ['event-based-optical-flow', 'event-based-vision'] | ['computer-vision', 'computer-vision'] | [ 3.70070308e-01 -4.78429407e-01 6.31546676e-01 -2.37679631e-01
-6.53923452e-02 -4.68510985e-01 5.54265797e-01 3.30698133e-01
-1.02155244e+00 7.23424792e-01 -1.42270610e-01 4.17845175e-02
-5.23488298e-02 -8.74440312e-01 -4.32049066e-01 -6.34881318e-01
-1.84550762e-01 1.32679284e-01 9.46873188e-01 2.42247939e-01
4.10293072e-01 7.91584969e-01 -2.18139577e+00 3.09206724e-01
2.50928104e-01 1.28506386e+00 4.12516177e-01 1.11071825e+00
5.28921001e-02 7.65877664e-01 -8.24357212e-01 1.00521185e-01
2.03965396e-01 -5.65829039e-01 -2.05333665e-01 -3.18603605e-01
5.57443142e-01 -7.25237966e-01 -4.00566459e-01 6.04939401e-01
7.21778989e-01 -4.47385125e-02 9.95186195e-02 -1.18248153e+00
3.42872113e-01 1.31777942e-01 -2.17891574e-01 7.77835369e-01
4.20910567e-01 4.74313617e-01 3.64804208e-01 -5.44226766e-01
9.11713719e-01 8.16348135e-01 5.75563550e-01 4.68453944e-01
-1.31399035e+00 -4.28806335e-01 -4.13980067e-01 3.55534017e-01
-1.16176689e+00 -6.05501711e-01 3.95298928e-01 -3.49837214e-01
1.52605486e+00 3.16124052e-01 1.16377127e+00 7.94418812e-01
7.14742839e-01 1.03183545e-01 1.21855891e+00 -1.72334000e-01
7.89477170e-01 -3.00775737e-01 5.41150570e-02 6.10576093e-01
5.28694332e-01 2.38836601e-01 -8.88415396e-01 1.95046291e-01
1.03817427e+00 -2.06762880e-01 -3.15120161e-01 -1.96352620e-02
-1.40341783e+00 3.07620853e-01 2.37554744e-01 2.56231964e-01
-4.50179100e-01 6.85463011e-01 5.93453407e-01 7.28471652e-02
-1.77088991e-01 2.09139138e-01 -1.19044073e-01 -7.76034117e-01
-1.03781319e+00 2.01299608e-01 8.98441315e-01 5.23861825e-01
4.46972907e-01 2.65954763e-01 -1.26969680e-01 -1.34286165e-01
2.30873153e-01 6.30682409e-01 8.38158488e-01 -1.14480197e+00
-1.34712338e-01 4.41328347e-01 1.52428786e-03 -7.13165462e-01
-7.70124018e-01 -4.92765531e-02 -3.79636079e-01 8.01423907e-01
8.70487809e-01 -1.44033015e-01 -6.49543285e-01 1.46182001e+00
2.46302038e-01 2.71052659e-01 -2.70966231e-03 9.70125377e-01
5.20296454e-01 8.71960223e-01 -2.26591125e-01 -5.71603954e-01
1.53917432e+00 -2.76839912e-01 -7.23324776e-01 3.86316106e-02
1.58416450e-01 -7.85986364e-01 6.43154979e-01 6.29470050e-01
-1.14534450e+00 -4.33736771e-01 -1.32779026e+00 -3.16872895e-02
-2.43055478e-01 -5.87548278e-02 7.59504497e-01 8.47380877e-01
-1.29114676e+00 5.94251215e-01 -1.23661411e+00 -4.36566561e-01
1.93286166e-01 6.52701080e-01 -8.51205811e-02 4.74212110e-01
-5.54850757e-01 6.11337662e-01 2.44768903e-01 -7.62700513e-02
-5.66719830e-01 -7.09634542e-01 -3.71250987e-01 -5.49468817e-03
-1.74807355e-01 -7.30375886e-01 1.33277094e+00 -9.87284541e-01
-1.93531954e+00 5.40229201e-01 -2.50194907e-01 -9.47525203e-01
4.78165269e-01 -6.33017067e-03 -2.49533653e-01 7.05701828e-01
-2.37008736e-01 6.15467370e-01 8.40121686e-01 -4.27688003e-01
-5.22630692e-01 -1.75065339e-01 1.49773769e-02 -1.77867234e-01
-3.17847073e-01 -5.90463653e-02 -9.04435441e-02 -1.21818334e-01
-2.29460374e-01 -7.58530736e-01 -3.67616522e-05 4.86948192e-01
3.98332804e-01 1.53078303e-01 8.32339406e-01 6.73019215e-02
1.00962901e+00 -1.87743902e+00 -3.94749105e-01 -1.72392085e-01
9.02769640e-02 4.92320031e-01 2.71569431e-01 4.75839257e-01
1.68315396e-01 -6.26959503e-01 -1.20359855e-02 -6.32029772e-02
-4.32794213e-01 2.53406674e-01 -2.91153103e-01 6.15970671e-01
7.55050778e-02 4.46487337e-01 -8.77241969e-01 -4.63781506e-01
7.02533185e-01 7.34214306e-01 -4.72726047e-01 6.04703575e-02
6.14695251e-02 3.17550957e-01 -6.29524961e-02 2.78177559e-01
3.93585116e-01 -1.47294059e-01 2.97233444e-02 -4.90607262e-01
-8.76779199e-01 3.47805172e-01 -1.51705897e+00 1.76247156e+00
-3.97578925e-01 1.13191056e+00 -6.21465407e-02 -4.91814107e-01
9.66533601e-01 3.04863572e-01 7.58035779e-01 -1.11622334e+00
4.96073395e-01 2.60140777e-01 8.84357765e-02 -4.92904127e-01
5.34027815e-01 1.88339744e-02 3.53116125e-01 3.54728907e-01
1.18180528e-01 -1.08729014e-02 5.89237273e-01 6.56315610e-02
1.57020235e+00 3.96661580e-01 3.72283876e-01 -3.24327499e-01
3.30969334e-01 3.17225344e-02 5.41543186e-01 5.67886651e-01
-5.12284219e-01 5.03407121e-01 3.54689866e-01 -5.44717848e-01
-9.91804361e-01 -1.26635444e+00 -4.28733766e-01 4.07051623e-01
3.87447387e-01 -7.28186786e-01 -6.97950780e-01 1.58925295e-01
-3.02674830e-01 2.61444807e-01 -6.30561113e-02 1.19937800e-01
-6.74616873e-01 -9.41648185e-01 4.55616862e-01 3.39034230e-01
5.85910380e-01 -1.05665600e+00 -2.28327060e+00 7.19025016e-01
3.56604695e-01 -1.38467252e+00 1.15002625e-01 9.72372815e-02
-1.10408103e+00 -1.03354084e+00 -1.39558151e-01 -2.21854597e-01
3.96860063e-01 -2.03949083e-02 8.26076448e-01 -4.83140588e-01
-9.60211039e-01 5.70590377e-01 -1.27970055e-01 -3.68762106e-01
-3.31984125e-02 -5.88090479e-01 6.63225055e-02 1.87882587e-01
8.96511897e-02 -7.71848321e-01 -9.21084523e-01 -8.80407095e-02
-9.85881507e-01 1.10932060e-01 3.81084919e-01 4.76116091e-01
5.63301980e-01 -2.31406718e-01 4.33799088e-01 -3.44743848e-01
1.38728902e-01 -1.66750383e-02 -1.14268649e+00 -3.57272774e-01
-5.41822851e-01 2.28874028e-01 9.30431306e-01 -4.81073022e-01
-1.00206757e+00 4.88189042e-01 -6.00909553e-02 -1.81342602e-01
-5.95862642e-02 -2.30176359e-01 4.62963939e-01 -3.54666382e-01
7.44645536e-01 1.25650540e-01 6.44040480e-02 1.60436705e-01
-3.30638252e-02 3.79063785e-01 7.05161631e-01 -1.75708607e-01
2.06057206e-01 1.02274013e+00 6.43944204e-01 -1.07391953e+00
1.42566934e-01 -3.44626278e-01 -3.49542767e-01 -7.15949535e-01
7.95714736e-01 -7.97880292e-01 -1.23725379e+00 7.86508977e-01
-1.28278291e+00 -2.36687884e-01 -4.73781407e-01 8.04049730e-01
-7.10284591e-01 3.37189250e-02 -5.57407260e-01 -1.01701748e+00
-5.42097688e-01 -1.10201287e+00 7.81037152e-01 7.40125716e-01
-2.31720388e-01 -6.62778795e-01 1.55475467e-01 -2.51641154e-01
7.69901574e-01 6.34473920e-01 6.63395450e-02 1.83101237e-01
-9.15827751e-01 -1.57916680e-01 -2.95205712e-01 -2.07611680e-01
-2.56806970e-01 4.49115127e-01 -1.30671036e+00 -1.87051445e-01
2.32735485e-01 4.12787721e-02 5.85459173e-01 5.41524947e-01
6.30365908e-01 2.86966264e-01 -2.26824090e-01 5.46861231e-01
2.05373359e+00 3.62637341e-01 7.23825812e-01 1.91544980e-01
2.44935527e-01 3.06937218e-01 3.10905993e-01 9.09592748e-01
1.13055278e-02 7.79272616e-01 6.16863310e-01 3.90092254e-01
-5.28189540e-01 2.66220480e-01 7.07824588e-01 6.86282158e-01
-2.38633543e-01 -1.03812896e-01 -6.03909552e-01 5.57042062e-01
-1.68737876e+00 -1.08513951e+00 -7.62654543e-01 2.62840796e+00
5.14744580e-01 3.62534761e-01 6.87238052e-02 2.11072579e-01
4.42527622e-01 -6.12841323e-02 -3.69287550e-01 -7.47201324e-01
4.83441949e-02 7.39050746e-01 8.65607083e-01 3.10144335e-01
-6.27721429e-01 4.17465150e-01 5.69115114e+00 1.97000429e-01
-1.51340997e+00 1.73689306e-01 -1.30739331e-01 -8.63179147e-01
2.11361706e-01 2.39306659e-01 -8.97047579e-01 7.03218997e-01
1.63213265e+00 -2.28327081e-01 3.16293329e-01 3.41628790e-01
4.94097829e-01 -9.14576888e-01 -1.01735210e+00 1.24162662e+00
-2.55754106e-02 -1.35279942e+00 -3.17835778e-01 6.23282380e-02
1.03647709e-01 1.39012516e-01 -5.43503702e-01 -4.59848434e-01
-2.95031041e-01 -5.02768695e-01 8.82933974e-01 7.49344349e-01
7.30582118e-01 -5.66155553e-01 3.84606063e-01 7.79477283e-02
-1.33418739e+00 -7.65525848e-02 -2.63864368e-01 -6.73590779e-01
5.15035808e-01 8.38820517e-01 -7.33715951e-01 1.20666526e-01
8.38683307e-01 4.26231772e-01 -5.17423272e-01 1.27617657e+00
3.25830102e-01 4.81149614e-01 -8.38528275e-01 -5.03540695e-01
1.74556114e-02 9.18255746e-02 8.36888611e-01 1.29286003e+00
4.77317154e-01 -1.59423336e-01 -4.04617518e-01 6.68266475e-01
3.62901986e-01 -3.10988039e-01 -5.27240694e-01 2.36064062e-01
3.55399400e-01 1.54528403e+00 -1.09765828e+00 -3.10482144e-01
-5.17778873e-01 9.49433088e-01 -3.59746516e-02 -1.66735932e-01
-8.39079022e-01 -8.52205157e-01 7.49281943e-01 2.23850906e-01
2.45653272e-01 -5.20071507e-01 -2.64021277e-01 -9.40949678e-01
1.22832775e-01 -7.93719143e-02 1.69179135e-03 -8.37085605e-01
-6.22363865e-01 5.84101439e-01 -1.61678731e-01 -1.21258044e+00
-3.89723659e-01 -7.87056446e-01 -4.69549239e-01 5.10863781e-01
-1.29916906e+00 -5.45914769e-01 -5.57782829e-01 6.82303250e-01
3.75646472e-01 2.06406772e-01 8.83612156e-01 3.00465614e-01
-3.99369717e-01 6.57033846e-02 1.41322864e-02 -2.67064184e-01
7.45292783e-01 -1.11727381e+00 1.31419793e-01 1.24677992e+00
1.16535723e-01 2.10231960e-01 8.37978661e-01 -4.54742044e-01
-1.96580434e+00 -6.73608601e-01 6.73289299e-01 -9.69398692e-02
7.38116205e-01 -4.10544485e-01 -5.59349775e-01 1.14218920e-01
3.19423527e-01 3.20731819e-01 2.82720089e-01 -8.20876300e-01
-2.95572057e-02 -7.53828824e-01 -1.27874219e+00 3.67482096e-01
9.01964962e-01 -4.03993905e-01 -3.50416631e-01 1.30843557e-02
1.95402294e-01 -4.35641825e-01 -7.62709379e-01 -3.71086858e-02
7.64965177e-01 -1.51902878e+00 6.99446261e-01 3.31439108e-01
5.53414412e-02 -7.02874422e-01 1.07667014e-01 -5.38027823e-01
1.88159279e-03 -8.79610777e-01 -2.64731407e-01 1.03888142e+00
-1.74133375e-01 -8.79851699e-01 5.73725522e-01 4.42242980e-01
-1.53391272e-01 -3.37132037e-01 -1.38264596e+00 -7.08607674e-01
-8.14070940e-01 -5.60227334e-01 7.03692734e-02 2.62482673e-01
1.02574371e-01 2.06924170e-01 1.99057028e-01 1.55339822e-01
7.45463133e-01 2.06401795e-01 3.27777684e-01 -1.16873050e+00
-4.17328238e-01 -3.42939973e-01 -1.22772479e+00 -6.27467036e-01
-5.65397501e-01 -5.40543199e-01 -4.05993611e-02 -1.31286776e+00
-2.70869672e-01 1.00213207e-01 8.04275498e-02 1.44134849e-01
4.04971093e-01 6.18923485e-01 2.01579705e-01 4.49603871e-02
-4.37551171e-01 -9.28884465e-03 7.92683899e-01 5.18432915e-01
-2.08973631e-01 -4.32763398e-01 2.46511415e-01 5.58878660e-01
4.98863727e-01 -3.29329997e-01 -4.84969884e-01 -4.16529149e-01
3.39988172e-01 1.15154743e-01 7.41210520e-01 -1.79234624e+00
7.70740092e-01 3.05670738e-01 5.81458867e-01 -2.92183310e-01
4.81462628e-01 -8.33006203e-01 4.42061931e-01 9.36383605e-01
9.01130140e-02 4.12583709e-01 4.36723292e-01 4.58251953e-01
2.60304436e-02 -2.97518849e-01 9.18479085e-01 -2.96609644e-02
-1.02111757e+00 -2.14524120e-02 -8.12107384e-01 -1.78251594e-01
1.32881927e+00 -5.15657485e-01 -6.49274349e-01 1.74097404e-01
-3.68590921e-01 -5.45911729e-01 5.01686811e-01 5.20712417e-03
5.62437475e-01 -9.46406543e-01 -2.72034615e-01 3.30984652e-01
-1.65683255e-01 -2.69944638e-01 1.62122950e-01 8.16974401e-01
-1.19630957e+00 3.64275128e-01 -9.14559186e-01 -8.66804421e-01
-1.21266639e+00 1.85497016e-01 3.91929775e-01 1.88234374e-01
-9.93715584e-01 5.72222888e-01 -3.95367652e-01 6.40936494e-01
-1.85935602e-01 -4.95161623e-01 -2.45329142e-02 1.69039786e-01
8.58502626e-01 6.10755265e-01 3.23875129e-01 -3.11873734e-01
-5.99913180e-01 7.05855072e-01 3.31722111e-01 -4.62202072e-01
1.09870565e+00 -1.84097029e-02 -2.20940057e-02 7.25196898e-01
1.00385237e+00 -7.51481345e-03 -1.65728033e+00 3.21996272e-01
-2.30354652e-01 -3.95722747e-01 4.18952882e-01 -5.23013771e-01
-9.00753558e-01 8.62055480e-01 1.11975253e+00 3.37050706e-01
1.57159424e+00 -5.83845377e-01 8.80025506e-01 1.77543595e-01
6.17296159e-01 -1.01572251e+00 -2.19791591e-01 2.78079003e-01
3.02566290e-01 -5.51585197e-01 1.26646757e-01 -2.15648726e-01
1.78485438e-02 1.70742440e+00 5.53233922e-01 -3.68340522e-01
3.20237339e-01 1.03358114e+00 -1.39570281e-01 2.37306990e-02
-9.82519925e-01 -2.32978970e-01 -2.94463038e-01 6.70577586e-01
3.26495677e-01 -1.24648198e-01 -6.43039703e-01 -2.03126997e-01
1.59650184e-02 5.31787157e-01 1.04472506e+00 1.12554193e+00
-5.09466827e-01 -8.74823093e-01 -3.41960996e-01 4.24706399e-01
-2.76151359e-01 1.72908112e-01 9.05593410e-02 5.32514215e-01
1.94235131e-01 7.74300933e-01 6.73102796e-01 -1.04025871e-01
3.91219229e-01 -1.26561761e-01 9.05185461e-01 -8.30476135e-02
-9.01969552e-01 -7.97962397e-02 -4.07022424e-03 -1.29525161e+00
-5.49368978e-01 -7.59131730e-01 -1.74085951e+00 -3.33821028e-01
1.88202634e-01 -4.51624542e-01 1.13485718e+00 7.72070527e-01
7.26068318e-01 7.49373615e-01 2.68768519e-01 -8.49273503e-01
3.72418500e-02 -4.34527218e-01 -2.88790822e-01 1.19093940e-01
3.01610559e-01 -4.53209281e-01 -1.91252232e-01 2.47445196e-01] | [8.68730354309082, -1.2646853923797607] |
0261e9de-183d-48a0-ae7a-d4ce78258928 | when-complexity-is-good-do-we-need-recurrent | null | null | https://openreview.net/forum?id=u6ybkty-bL | https://openreview.net/pdf?id=u6ybkty-bL | When Complexity Is Good: Do We Need Recurrent Deep Learning For Time Series Outlier Detection? | Outlier detection is a critical part of understanding a dataset and extracting results. Outlier detection is used in different domains for various reasons; including detecting stolen credit cards, spikes of energy usage, web attacks, or in-home activity monitoring. Within this paper, we look at when it is appropriate to apply recurrent deep learning methods for time series outlier detection versus non-recurrent methods. Recurrent deep learning methods have a larger capacity for learning complex representations in time series data. We apply these methods to various synthetic and real-world datasets, including a dataset containing information about the in-home movement of people living with dementia in a clinical study cross-referenced with their recorded unplanned hospital admissions and infection episodes. We also introduce two new outlier detection methods, that can be useful in detecting contextual outliers in time series data where complex temporal relationships and local variations in the time series are important. | ['Payam M. Barnaghi', 'David J. Sharp', 'Mazdak Ghajari', 'Samaneh Kouchaki', 'Alexander Capstick'] | 2021-09-29 | null | null | null | null | ['home-activity-monitoring'] | ['miscellaneous'] | [-7.36257881e-02 -3.92667621e-01 -5.35980314e-02 -2.04837695e-01
-3.91354799e-01 -1.30962715e-01 4.82523829e-01 6.89745188e-01
-4.49563026e-01 5.77213943e-01 6.44550800e-01 -3.68710458e-01
-2.39265874e-01 -7.28081286e-01 -5.07938206e-01 -5.15464842e-01
-9.45239723e-01 2.62289822e-01 -3.98764051e-02 8.82481262e-02
1.45295173e-01 6.54452980e-01 -1.09614658e+00 3.57524723e-01
4.84747410e-01 8.00141752e-01 -6.73549652e-01 4.19906884e-01
-1.29557341e-01 8.38174105e-01 -1.11047947e+00 -5.53292520e-02
2.22895853e-02 -4.17554617e-01 -2.57861882e-01 -2.78828382e-01
-2.93835729e-01 -4.20152128e-01 -5.66303074e-01 7.93027818e-01
5.92301726e-01 4.31004800e-02 5.51001966e-01 -1.64348888e+00
-2.84914106e-01 5.59431493e-01 -4.16833103e-01 8.36468637e-01
5.48292458e-01 7.08541945e-02 5.63168347e-01 -3.71088803e-01
2.56837755e-01 7.29225457e-01 1.35478759e+00 1.59031451e-01
-1.05878663e+00 -6.11306250e-01 1.08869068e-01 4.12715852e-01
-9.35237706e-01 -1.43105656e-01 9.39270198e-01 -4.04161781e-01
1.22023833e+00 2.52108783e-01 7.60266125e-01 2.20316839e+00
8.36593986e-01 6.83065891e-01 5.84255040e-01 8.70733857e-02
6.67110980e-01 -7.04147160e-01 3.07065666e-01 -4.87449132e-02
7.69615054e-01 3.26019645e-01 -6.43097937e-01 -9.72475052e-01
5.18717289e-01 1.20520055e+00 -3.33659723e-02 6.79017827e-02
-1.42214537e+00 9.11729515e-01 2.63137192e-01 6.02993071e-01
-7.04736352e-01 2.27636516e-01 1.08444786e+00 8.75678778e-01
5.85096121e-01 3.42916429e-01 -6.25318050e-01 -4.86072779e-01
-9.77771461e-01 3.14488649e-01 9.39345717e-01 5.28056502e-01
-2.04653308e-01 4.25124705e-01 -1.04211040e-01 4.34706956e-01
1.23918511e-01 1.57212615e-01 1.20312107e+00 -4.02597129e-01
4.19905484e-01 5.40206313e-01 2.65706152e-01 -8.93858671e-01
-9.99562085e-01 -1.13144360e-01 -1.21482289e+00 1.79573551e-01
4.82294112e-01 -1.79113254e-01 -9.03329968e-01 1.40353203e+00
-1.66540712e-01 6.06377959e-01 7.33065307e-02 3.27000856e-01
3.73041183e-01 3.18503410e-01 3.18736682e-04 -5.50660014e-01
1.14670229e+00 -2.63669670e-01 -9.41818953e-01 -2.29721248e-01
6.73786640e-01 -3.33008915e-01 6.64367437e-01 4.69539583e-01
-5.14968038e-01 8.23290320e-04 -8.94086361e-01 5.10809302e-01
-7.68740535e-01 -6.98061466e-01 5.69230735e-01 6.67744219e-01
-6.05693579e-01 1.02108431e+00 -1.47358596e+00 -5.61206341e-01
7.43473351e-01 2.98949212e-01 -2.57621258e-01 2.43770614e-01
-1.20659602e+00 6.08876050e-01 1.63459331e-01 6.49728626e-03
-5.63400090e-01 -4.82866287e-01 -7.74597526e-01 -1.53654277e-01
-5.31810075e-02 -4.05942678e-01 1.23247087e+00 -7.91036308e-01
-7.10714042e-01 5.69017351e-01 4.49928874e-03 -8.86685491e-01
9.50375497e-01 -4.29101467e-01 -1.11069858e+00 -5.16318202e-01
2.55599499e-01 -5.99449456e-01 9.58367705e-01 -1.62742615e-01
-2.02763557e-01 -7.73189902e-01 -8.76683652e-01 -4.69383448e-01
-1.17836624e-01 1.96431205e-02 4.35488462e-01 -1.17940485e+00
3.11487526e-01 -7.06385136e-01 -4.50993091e-01 -4.77076292e-01
-5.02059340e-01 -2.31609985e-01 1.20488203e+00 -6.05235755e-01
1.42360497e+00 -2.08038425e+00 -5.03800154e-01 2.50662327e-01
3.51122350e-01 4.86534722e-02 2.16454968e-01 8.76738191e-01
-4.13454384e-01 9.75250006e-02 -3.80610675e-01 -2.02245399e-01
-8.13339874e-02 1.63365319e-01 -6.09685838e-01 7.97634065e-01
1.04806826e-01 9.39897358e-01 -1.02713251e+00 2.67982274e-01
1.40570298e-01 2.68385708e-01 -1.30953610e-01 1.50518209e-01
-5.22580594e-02 5.46826959e-01 -4.90139455e-01 9.50179815e-01
2.51668036e-01 -2.07756329e-02 -1.52396038e-01 2.58705348e-01
2.01873764e-01 1.84755147e-01 -9.93201494e-01 1.35615492e+00
-2.32152939e-02 8.15819502e-01 -8.46754074e-01 -1.08108068e+00
7.76030302e-01 5.50236821e-01 1.06492269e+00 -1.06004500e+00
1.08039640e-01 3.98761123e-01 -1.26705155e-01 -7.66445696e-01
1.95144638e-01 6.05824366e-02 -3.78665715e-01 8.89856219e-01
-3.32587212e-01 6.70198619e-01 -2.82793213e-02 -2.47303650e-01
1.91932976e+00 -2.65033931e-01 6.59206688e-01 1.73569858e-01
-9.23149660e-02 -2.91215032e-01 6.81880236e-01 9.48715210e-01
-5.12039721e-01 6.98425412e-01 8.53315890e-01 -1.25765371e+00
-1.30308676e+00 -1.34240401e+00 -2.69962847e-02 7.01468706e-01
-5.61576426e-01 -3.20972264e-01 -1.07153058e-01 -6.47366464e-01
1.89577714e-01 5.68678498e-01 -6.41766310e-01 -5.82634628e-01
-8.23383749e-01 -1.18687677e+00 6.78501070e-01 9.05561745e-01
1.59205794e-01 -1.73646557e+00 -1.12452137e+00 7.64118671e-01
-7.29909912e-02 -8.75643730e-01 -2.75519162e-01 6.72286153e-01
-1.23618519e+00 -1.25174367e+00 -5.57524681e-01 -3.23705524e-01
2.84972876e-01 -2.29212075e-01 1.15122068e+00 -1.45034105e-01
-5.92164695e-01 3.01707685e-01 -3.74719948e-01 -7.93439448e-01
-1.49260387e-01 6.68998510e-02 4.92427677e-01 -6.18249290e-02
1.06839514e+00 -8.37191582e-01 -5.04598558e-01 -1.53821176e-02
-1.15243578e+00 -1.00227833e+00 1.36688679e-01 8.59384477e-01
3.24923843e-01 1.53056338e-01 8.15534532e-01 -7.04724312e-01
1.11320901e+00 -1.17824972e+00 -1.86295941e-01 -2.84160197e-01
-5.11393964e-01 -1.24862842e-01 7.24836290e-01 -5.70844233e-01
-1.80159062e-01 -1.13301411e-01 -7.98168108e-02 -3.91268849e-01
-3.26881766e-01 3.30342174e-01 4.65104058e-02 7.48976111e-01
7.96033263e-01 8.32538586e-03 -1.41907409e-01 -3.84176135e-01
-2.73187041e-01 5.78657269e-01 5.19046962e-01 -1.42996714e-01
4.04348850e-01 7.27420986e-01 -2.56701976e-01 -8.83064747e-01
-4.05611128e-01 -7.02076614e-01 -6.52427137e-01 2.03670636e-01
4.91420031e-01 -6.72572017e-01 -6.68483675e-01 9.80042040e-01
-1.03087544e+00 -4.14337546e-01 -7.02675402e-01 5.30974627e-01
-7.62225807e-01 8.98787305e-02 -5.99952757e-01 -7.68685222e-01
-3.41009974e-01 -6.21987641e-01 9.39616919e-01 6.92005455e-02
-8.72460663e-01 -1.23759830e+00 6.00153625e-01 -5.71357012e-01
5.00364900e-01 1.07948363e+00 8.39332700e-01 -1.45387506e+00
-1.39122009e-01 -7.16212928e-01 1.63630366e-01 -6.43239869e-03
5.49404085e-01 -2.51191944e-01 -9.71282780e-01 -4.66815442e-01
3.21706861e-01 1.12405293e-01 8.41907501e-01 6.11927629e-01
1.22740471e+00 -5.37892580e-01 -4.55967069e-01 7.77448952e-01
1.22470176e+00 4.09224272e-01 8.04661989e-01 8.08684289e-01
6.13926113e-01 1.18360572e-01 3.60798091e-02 9.97857809e-01
-2.86591332e-03 3.77879590e-01 5.74834466e-01 2.38362119e-01
7.04193592e-01 -2.32757558e-03 7.69851446e-01 4.22856003e-01
8.75834674e-02 -2.35230505e-01 -1.35012126e+00 6.46834552e-01
-2.18709111e+00 -1.52231288e+00 -2.91627049e-01 2.37854433e+00
2.02731118e-01 2.04952613e-01 6.30533695e-01 4.48869079e-01
5.95553219e-01 2.77646303e-01 -1.16530573e+00 -4.50578153e-01
-3.94040011e-02 9.31873620e-02 4.25608754e-01 -1.28599152e-01
-1.21926796e+00 3.25205386e-01 6.83835602e+00 -1.56166881e-01
-9.98114645e-01 9.03908908e-02 7.10784674e-01 -4.26383764e-01
7.85783455e-02 -6.31855607e-01 -2.05310397e-02 8.88214529e-01
1.50686264e+00 -2.02763423e-01 6.33219406e-02 7.73534596e-01
4.46105361e-01 1.81993186e-01 -1.49312305e+00 1.26494753e+00
-4.47193123e-02 -9.75245237e-01 -2.76476413e-01 1.92374274e-01
4.18322116e-01 4.51370806e-01 -4.72882800e-02 2.64243305e-01
3.23660374e-01 -1.29689360e+00 2.57247388e-01 6.99848890e-01
3.14618558e-01 -8.64169180e-01 1.04894209e+00 1.23333357e-01
-9.02741849e-01 -5.25166392e-01 -4.05619629e-02 -2.38574609e-01
3.00776988e-01 9.44567204e-01 -6.58498883e-01 1.33234784e-01
1.16729593e+00 1.07311463e+00 -3.74045968e-01 1.31230581e+00
2.32538298e-01 4.83394384e-01 -5.12143731e-01 2.12935403e-01
1.51261151e-01 -1.12060308e-01 6.92158520e-01 1.11778390e+00
6.76850677e-01 -4.20097917e-01 -3.88276204e-02 5.23262441e-01
1.65738165e-01 -2.65350729e-01 -1.36492193e+00 -1.99932247e-01
2.07016796e-01 5.82444847e-01 -8.40257347e-01 -5.50814793e-02
-4.50995326e-01 1.20823658e+00 -1.49446115e-01 5.31152248e-01
-9.24551725e-01 -6.15688920e-01 1.05774999e+00 2.04940170e-01
8.19715485e-02 -2.85830140e-01 -2.84286380e-01 -1.35580480e+00
4.29654986e-01 -8.68416905e-01 9.12783504e-01 -3.65720153e-01
-1.79852605e+00 4.68378097e-01 -2.59699941e-01 -1.55786800e+00
-7.77453899e-01 -4.93976235e-01 -1.29247379e+00 4.02110428e-01
-9.42978621e-01 -5.52341342e-01 -3.29493940e-01 1.02840483e+00
4.47320879e-01 -4.60193366e-01 8.05759251e-01 2.26719663e-01
-6.66153491e-01 3.00979733e-01 4.66970235e-01 5.30201197e-01
6.47814631e-01 -1.19244039e+00 1.08483946e+00 8.60064685e-01
1.21285193e-01 5.09144902e-01 5.89141548e-01 -9.83238876e-01
-1.03265536e+00 -1.37697971e+00 8.36970448e-01 -7.75986493e-01
1.00081599e+00 -3.07480842e-01 -1.19665635e+00 1.13890469e+00
-1.63681671e-01 1.48074970e-01 9.17725146e-01 5.29212467e-02
-2.54141152e-01 -2.18180455e-02 -1.26942968e+00 6.27564967e-01
1.02643466e+00 -5.89583457e-01 -8.14322233e-01 4.45804745e-01
5.32097518e-01 -1.51236400e-01 -7.00013995e-01 2.84533948e-01
5.78869820e-01 -1.12662208e+00 1.01195836e+00 -9.92664218e-01
-1.55723080e-01 1.00970663e-01 -4.02609408e-02 -1.48930061e+00
-2.55951881e-01 -9.95698631e-01 -7.01785803e-01 7.07031071e-01
-7.17624202e-02 -1.00911963e+00 8.00719619e-01 3.63276511e-01
-3.58887948e-02 -5.10657310e-01 -1.23137760e+00 -1.01601338e+00
-1.00916974e-01 -8.21973562e-01 9.16450977e-01 1.16610014e+00
1.97094768e-01 -1.67396799e-01 -4.20611799e-01 7.22558871e-02
5.03416955e-01 -3.68701905e-01 3.64880502e-01 -1.61544216e+00
2.83354700e-01 -5.97307265e-01 -8.62033844e-01 7.10839108e-02
-2.46981177e-02 -5.04546702e-01 -3.79591644e-01 -1.19931996e+00
-1.91389725e-01 -5.86459339e-02 -7.68279672e-01 3.77247512e-01
8.33897367e-02 -1.02157503e-01 -3.53910983e-01 3.63537073e-01
-3.51883888e-01 4.45584059e-01 1.66067615e-01 -1.89350188e-01
-6.02802098e-01 5.89008987e-01 -2.23388195e-01 9.73462999e-01
1.17999172e+00 -8.65394831e-01 -2.06436545e-01 -4.40515168e-02
2.76646167e-01 -1.83386784e-02 5.62394500e-01 -1.24317634e+00
2.02258959e-01 -6.42578825e-02 8.20086479e-01 -7.17790008e-01
-2.08748251e-01 -1.03998363e+00 1.80206940e-01 8.54096174e-01
-5.53558813e-03 1.00054014e+00 1.16053782e-01 9.75887597e-01
-1.20376229e-01 1.70885861e-01 4.23262954e-01 -2.72850215e-01
-5.00581145e-01 5.90642214e-01 -7.52578318e-01 1.61345527e-01
1.11272061e+00 -4.03839380e-01 -2.08786443e-01 -4.21323627e-01
-1.07766783e+00 5.34986891e-03 2.95881182e-01 8.31445575e-01
7.45384693e-01 -1.66396177e+00 -4.89103675e-01 5.50386965e-01
3.66985917e-01 -4.17276204e-01 -1.67489126e-02 8.48918140e-01
-4.96953487e-01 2.58609354e-01 -4.62726653e-01 -6.63406312e-01
-8.20200026e-01 6.76157176e-01 4.35098082e-01 -2.79467553e-01
-1.17907238e+00 -2.68222508e-03 -5.72445512e-01 -2.71618634e-01
5.86982846e-01 -7.86104202e-01 3.21309790e-02 3.95906359e-01
7.50554144e-01 7.11672008e-01 3.65922928e-01 -2.02359349e-01
-6.22834265e-01 4.85461131e-02 -1.53575331e-01 1.56281695e-01
1.63080347e+00 2.77804345e-01 -2.71908846e-02 1.27037191e+00
1.26629627e+00 -4.14902091e-01 -9.55311477e-01 -1.04852989e-01
8.71442556e-01 1.64981745e-02 -4.38183099e-01 -3.63967538e-01
-8.91518176e-01 6.27723873e-01 1.03135014e+00 7.63147414e-01
1.00107384e+00 -1.81001872e-01 1.14289999e+00 6.32003069e-01
3.90600085e-01 -1.11535919e+00 2.86286443e-01 5.57215214e-01
6.87423229e-01 -1.33856320e+00 -1.08779691e-01 7.69258142e-01
-4.12166834e-01 1.24302912e+00 2.06561506e-01 -3.62555414e-01
9.11209822e-01 2.38090649e-01 5.69966584e-02 -2.73210615e-01
-5.20462871e-01 5.53978421e-02 -2.40133345e-01 8.42299104e-01
5.20457745e-01 8.92199278e-02 9.84277055e-02 5.18637657e-01
-1.35530204e-01 5.31341545e-02 7.66326547e-01 1.03582036e+00
-9.52895731e-02 -9.31113899e-01 -4.51555640e-01 1.04307294e+00
-8.30637932e-01 1.05682857e-01 -3.29239905e-01 7.23099589e-01
-1.98535964e-01 6.93129301e-01 5.07789373e-01 -6.13740943e-02
5.37810504e-01 6.11832440e-01 -1.34984255e-01 -3.41301322e-01
-5.85566998e-01 -2.36821026e-01 -2.22531363e-01 -1.05709994e+00
-6.61894958e-03 -1.02400267e+00 -1.11141336e+00 -3.57286841e-01
3.60291421e-01 -2.14913532e-01 3.88718784e-01 9.22242582e-01
3.53925794e-01 7.61441529e-01 5.05699873e-01 -6.64806068e-01
-2.97282428e-01 -9.41696227e-01 -7.37536430e-01 7.04957783e-01
1.12392056e+00 -1.89498335e-01 -6.83308423e-01 -2.06452101e-01] | [7.4303789138793945, 2.603548526763916] |
7d7b0322-9bd4-4bf8-9a40-b62c8e170320 | skoltechnlp-at-semeval-2020-task-11-exploring | null | null | https://aclanthology.org/2020.semeval-1.234 | https://aclanthology.org/2020.semeval-1.234.pdf | SkoltechNLP at SemEval-2020 Task 11: Exploring Unsupervised Text Augmentation for Propaganda Detection | This paper presents a solution for the Span Identification (SI) task in the {``}Detection of Propaganda Techniques in News Articles{''} competition at SemEval-2020. The goal of the SI task is to identify specific fragments of each article which contain the use of at least one propaganda technique. This is a binary sequence tagging task. We tested several approaches finally selecting a fine-tuned BERT model as our baseline model. Our main contribution is an investigation of several unsupervised data augmentation techniques based on distributional semantics expanding the original small training dataset as applied to this BERT-based sequence tagger. We explore various expansion strategies and show that they can substantially shift the balance between precision and recall, while maintaining comparable levels of the F1 score. | ['Alexander Panchenko', 'Igor Markov', 'Daryna Dementieva'] | 2020-12-01 | null | null | null | semeval-2020 | ['text-augmentation', 'propaganda-detection'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.52760607e-01 3.70785147e-02 -4.90034640e-01 -2.43221596e-01
-9.99429405e-01 -9.09884870e-01 1.18799663e+00 5.64091444e-01
-8.99273455e-01 7.31867731e-01 7.06231415e-01 -6.74416006e-01
-7.86853582e-03 -5.28000951e-01 -6.07095957e-01 -4.67993766e-01
3.15805860e-02 5.59346259e-01 4.65613484e-01 -4.79664952e-01
7.05555618e-01 3.29515755e-01 -1.17151725e+00 6.98774457e-01
6.37996972e-01 6.23534262e-01 -6.35460243e-02 7.88119972e-01
-2.30614096e-01 9.96393681e-01 -9.15425181e-01 -6.36185408e-01
1.84145302e-01 -3.91276151e-01 -9.34525013e-01 -4.71280485e-01
5.92044055e-01 8.93593431e-02 -2.17388913e-01 1.13661790e+00
4.04616475e-01 8.41976479e-02 9.81462955e-01 -7.62416005e-01
-3.56729031e-01 1.35656381e+00 -5.15457928e-01 8.18538964e-01
5.94400644e-01 -4.97495890e-01 1.00381207e+00 -5.99957645e-01
1.14161825e+00 7.26644039e-01 1.02357614e+00 4.76220399e-01
-1.27121627e+00 -2.08906278e-01 -3.87435764e-01 2.38283828e-01
-1.12474585e+00 -4.30224657e-01 7.57617235e-01 -6.69603527e-01
1.03425431e+00 4.64514554e-01 3.98490220e-01 1.39590538e+00
1.35930330e-01 8.29913914e-01 1.06322956e+00 -9.97270942e-01
2.05043271e-01 1.93102464e-01 7.12726772e-01 6.44665658e-01
1.81932345e-01 -6.96952567e-02 -4.58073378e-01 -5.90077698e-01
1.64658502e-01 -5.61536133e-01 8.69612992e-02 1.51419610e-01
-1.09246397e+00 1.18145275e+00 -1.44528449e-01 7.17972398e-01
-2.94062108e-01 4.79383282e-02 1.04212046e+00 2.57421672e-01
8.29862118e-01 7.77800798e-01 -3.74828398e-01 -3.88163120e-01
-1.37435257e+00 6.26143217e-01 9.03060496e-01 7.63545871e-01
2.29955167e-01 -2.99529970e-01 -5.18377960e-01 5.87468505e-01
-2.98224330e-01 2.97380686e-01 3.78498137e-01 -4.57060695e-01
6.64494693e-01 2.90504605e-01 9.91818458e-02 -9.27744031e-01
-6.00308180e-01 -5.62395036e-01 -4.76129144e-01 -2.88149655e-01
3.07451427e-01 -3.54817927e-01 -8.12847674e-01 1.62627602e+00
2.47345656e-01 -2.70546019e-01 -2.84883082e-01 3.28743309e-01
5.97907603e-01 6.27891064e-01 3.35352927e-01 -4.49814826e-01
1.49059391e+00 -9.34086561e-01 -7.99130380e-01 -9.46362540e-02
1.05674911e+00 -9.91310000e-01 8.12988162e-01 1.36650488e-01
-7.72074521e-01 -1.73328087e-01 -9.32794392e-01 1.30800709e-01
-3.21986765e-01 1.87344342e-01 4.44186151e-01 9.27790105e-01
-5.90637386e-01 4.57301587e-01 -5.02924740e-01 -4.37634856e-01
7.94923753e-02 -1.48706719e-01 -1.63788155e-01 3.65448117e-01
-1.51707888e+00 1.21005201e+00 7.63631046e-01 -6.07624054e-01
-6.43017113e-01 -5.91067672e-01 -7.58261383e-01 -8.99908468e-02
3.91797960e-01 -1.71398669e-01 1.30138004e+00 -8.10285926e-01
-8.30997169e-01 1.21093214e+00 7.38457814e-02 -1.11217570e+00
3.96294892e-01 -5.50072968e-01 -6.93903387e-01 2.00697765e-01
3.96187454e-01 5.39712161e-02 5.58013201e-01 -7.51065314e-01
-5.49338162e-01 -8.74947608e-02 -2.49888167e-01 -1.28189728e-01
-3.04853380e-01 1.00999451e+00 2.36671884e-02 -1.19822097e+00
-4.90953714e-01 -8.82003903e-01 -9.70074534e-02 -9.16342080e-01
-4.62580770e-01 -2.97964931e-01 4.82324213e-01 -1.18882167e+00
1.96764779e+00 -1.93496430e+00 -1.39909517e-02 3.49788785e-01
-6.32263646e-02 4.86826360e-01 -5.75379655e-02 8.39740753e-01
-7.60474131e-02 4.32017773e-01 -5.70531525e-02 1.02222547e-01
-9.80942026e-02 -5.49428090e-02 -5.11590064e-01 5.02310097e-01
-1.91537544e-01 8.14305186e-01 -7.68182456e-01 -5.38172960e-01
-2.44445831e-01 -7.58930296e-02 -4.11108166e-01 -1.83870494e-01
-5.64885736e-01 3.76501456e-02 -3.74756843e-01 1.59901500e-01
2.55018026e-01 9.40170884e-02 2.37606362e-01 1.18995206e-02
-3.83033067e-01 7.68355310e-01 -6.44831061e-01 1.48123312e+00
1.19980536e-01 7.41264760e-01 -2.45502695e-01 -1.04963517e+00
7.29485571e-01 2.27142587e-01 4.72160339e-01 -4.98861551e-01
3.49015534e-01 1.47402614e-01 2.67042965e-03 -6.15767121e-01
5.04450738e-01 -4.44948494e-01 -5.33407211e-01 4.99712139e-01
2.22291738e-01 4.84843493e-01 6.25816464e-01 4.27772462e-01
1.41436732e+00 1.02472737e-01 7.70160198e-01 -5.60244083e-01
5.76773584e-01 5.46437919e-01 3.23430866e-01 1.07654858e+00
-6.20379783e-02 2.77865827e-01 5.16856015e-01 -3.75656843e-01
-1.41930819e+00 -7.10994124e-01 -1.59914464e-01 1.51995409e+00
-3.96731764e-01 -7.66082883e-01 -7.59880185e-01 -1.19245470e+00
-2.77521580e-01 1.20410097e+00 -1.02962315e+00 3.00744828e-02
-9.06409860e-01 -8.75558496e-01 1.25660729e+00 4.46365148e-01
2.06149533e-01 -9.01350737e-01 -6.58142030e-01 2.94062942e-01
-6.33075058e-01 -9.33967531e-01 -3.60225499e-01 3.59977722e-01
-3.23921233e-01 -9.85325158e-01 -6.19125187e-01 -8.60702813e-01
1.00328028e-01 -3.11731160e-01 1.00132883e+00 -1.41967967e-01
-1.64800286e-01 -4.57879268e-02 -9.41080987e-01 -6.18123710e-01
-9.56727445e-01 2.72398531e-01 2.19804183e-01 -4.87769842e-01
5.83458900e-01 -1.45204484e-01 -5.53017408e-02 -1.61439434e-01
-7.70540893e-01 -7.93242604e-02 3.46061647e-01 7.38155365e-01
9.84513666e-03 -1.36798337e-01 5.51661611e-01 -1.21950996e+00
8.22775841e-01 -4.50230837e-01 -2.78378457e-01 4.38740104e-01
-5.43189883e-01 1.37531042e-01 4.55885172e-01 -5.12368679e-01
-1.13189662e+00 -1.20044582e-01 -5.53032935e-01 2.49314636e-01
6.29879832e-02 8.83889675e-01 1.76458493e-01 1.70623243e-01
1.36548769e+00 2.67055720e-01 -3.21155749e-02 -7.81155527e-01
5.37545502e-01 6.74384654e-01 6.04102075e-01 -4.91433710e-01
5.83470225e-01 2.71343882e-03 -4.06549990e-01 -7.25605309e-01
-1.24202740e+00 -9.82156694e-01 -6.23053312e-01 -1.26473650e-01
6.88598335e-01 -5.27108669e-01 -2.05476582e-01 2.37921387e-01
-1.13997900e+00 -8.77971202e-03 -3.85375053e-01 4.03697222e-01
-6.09693587e-01 7.54593194e-01 -8.82123590e-01 -7.95230031e-01
-5.74637294e-01 -4.55140650e-01 7.80108452e-01 -2.03829676e-01
-5.99453330e-01 -8.90173435e-01 5.46871662e-01 2.17389256e-01
5.86677752e-02 1.94407657e-01 1.11297703e+00 -1.53901422e+00
5.97756505e-01 -4.05232668e-01 1.03282362e-01 2.17415795e-01
-2.52100974e-01 -1.15340151e-01 -4.46918547e-01 -4.26149108e-02
1.43469483e-01 -2.19353095e-01 1.04490376e+00 2.69527793e-01
6.33715510e-01 -8.98217201e-01 -3.41708183e-01 -2.99452636e-02
1.29209101e+00 2.00978130e-01 6.62914574e-01 9.23851967e-01
3.05558920e-01 4.99778301e-01 7.32367516e-01 4.04789507e-01
-4.10365552e-01 8.71945560e-01 -2.18166381e-01 3.77666593e-01
-1.77073985e-01 -3.91002864e-01 5.75112164e-01 4.77361679e-01
-1.42868921e-01 -4.54161853e-01 -1.20501566e+00 5.72058797e-01
-1.77583122e+00 -1.76187742e+00 -2.31819838e-01 1.93052423e+00
1.16742873e+00 3.32092583e-01 6.05579436e-01 3.28012556e-01
7.32832491e-01 1.34410962e-01 3.86693805e-01 -7.86953270e-01
-2.36079529e-01 2.93564200e-01 7.13239610e-01 4.43987846e-01
-1.65603459e+00 8.85841608e-01 7.23057127e+00 1.34229279e+00
-6.36645496e-01 2.85014153e-01 2.27473646e-01 7.73428157e-02
-2.20482782e-01 -1.44470662e-01 -1.06611264e+00 6.50278986e-01
1.40815091e+00 -3.17728311e-01 -5.95895723e-02 8.64285171e-01
1.63337924e-02 -3.32128219e-02 -6.40924096e-01 3.13740104e-01
6.14211082e-01 -1.80786526e+00 8.29940438e-02 -2.26778314e-02
6.38208091e-01 1.35888234e-01 -2.71310389e-01 4.82972324e-01
4.07286823e-01 -7.13865161e-01 1.02760243e+00 2.51194477e-01
6.83009207e-01 -6.79170489e-01 9.42081749e-01 7.53305554e-01
-5.72222412e-01 -4.35448475e-02 -4.68747830e-03 7.59358192e-03
6.24526262e-01 7.21243203e-01 -9.89259601e-01 5.09500027e-01
2.98177868e-01 1.83733448e-01 -6.39384270e-01 1.05988598e+00
-3.36640239e-01 1.35471272e+00 -2.99751788e-01 -3.78944248e-01
5.33255875e-01 2.38085598e-01 1.02773178e+00 1.88505507e+00
-3.44633823e-03 -6.09906092e-02 1.64025113e-01 2.09378690e-01
4.69354615e-02 3.95453930e-01 -4.73425925e-01 -3.15121561e-01
2.93300807e-01 8.18198621e-01 -7.36185849e-01 -5.38557708e-01
-1.28210813e-01 7.56468475e-01 3.48650217e-01 -9.74621922e-02
-1.11855197e+00 -2.57636130e-01 6.62361383e-02 3.94286364e-01
2.41479486e-01 -1.20260432e-01 -2.99750686e-01 -9.67950344e-01
-2.32983992e-01 -8.88958037e-01 7.84006357e-01 -3.31664413e-01
-1.19658542e+00 4.92745370e-01 2.35150486e-01 -8.11023116e-01
-4.96514201e-01 -4.70610499e-01 -3.04055244e-01 4.88559157e-01
-6.20744228e-01 -1.18506551e+00 5.70785701e-01 -3.59474793e-02
5.90977490e-01 -3.92788202e-01 7.79667616e-01 2.80413121e-01
-8.80056098e-02 5.65947413e-01 4.79606420e-01 4.70056504e-01
6.72937751e-01 -9.89294350e-01 3.65306735e-01 1.16355896e+00
4.46629494e-01 8.16849351e-01 1.30695581e+00 -1.09877717e+00
-6.18838608e-01 -9.49975789e-01 1.49927914e+00 -6.66698337e-01
1.03132379e+00 -4.28997099e-01 -6.42482936e-01 7.74704754e-01
1.84412196e-01 -8.30795884e-01 9.07227933e-01 4.49690849e-01
-6.55814946e-01 6.72686398e-01 -1.10069740e+00 3.81807983e-01
9.63263512e-01 -4.28549886e-01 -1.24965870e+00 5.61365366e-01
7.03607440e-01 -5.69777824e-02 -5.67779958e-01 4.29463774e-01
5.12735963e-01 -5.52695930e-01 9.35382843e-01 -1.18926668e+00
7.59007335e-01 1.28215373e-01 -1.38232425e-01 -8.41060936e-01
-5.85642695e-01 -5.41338742e-01 -2.60967523e-01 1.44375682e+00
5.13201118e-01 -1.24524198e-01 6.54422164e-01 -1.46896243e-02
5.32096811e-02 -4.64683890e-01 -1.06837285e+00 -1.01019776e+00
1.12422220e-01 -4.44243103e-01 -7.58889690e-02 9.61150229e-01
4.26881015e-01 7.44466901e-01 -6.26879871e-01 -3.30596894e-01
4.46997762e-01 -8.61401111e-02 3.19233745e-01 -9.82055187e-01
-2.95985639e-01 -5.50441802e-01 -5.38860857e-01 -7.86339521e-01
1.32216197e-02 -1.02738214e+00 1.21593364e-01 -1.27908623e+00
6.63511395e-01 -2.02010721e-01 -1.39947593e-01 4.00657088e-01
-7.28093758e-02 1.79569274e-01 7.97258839e-02 2.77819335e-01
-5.78106105e-01 1.80493265e-01 2.33234748e-01 -1.45450249e-01
-5.25143519e-02 3.63761839e-03 -5.74927151e-01 8.40788543e-01
8.69791627e-01 -7.94142425e-01 -1.45363407e-02 -1.35968700e-01
5.56350827e-01 -2.00109318e-01 2.38573611e-01 -8.22969496e-01
-2.41610706e-02 -1.06385671e-01 7.49087781e-02 -7.79777944e-01
-3.04046124e-01 -2.59498000e-01 1.35540375e-02 7.11904287e-01
-8.29012156e-01 1.71514843e-02 2.79211819e-01 6.18337989e-01
2.73808483e-02 -8.30409110e-01 7.09542632e-01 -1.30208597e-01
-8.38700354e-01 -2.94637173e-01 -8.41111481e-01 3.46716940e-01
1.03299677e+00 1.25431925e-01 -6.22194946e-01 4.25743088e-02
-8.20678294e-01 -2.65290499e-01 2.66762227e-01 3.72254491e-01
8.46502259e-02 -1.01965427e+00 -1.02958441e+00 -3.78837645e-01
2.97924310e-01 -9.74430382e-01 -1.13379713e-02 8.91009033e-01
-7.06336141e-01 5.51072955e-01 -2.01198533e-01 9.35371667e-02
-1.71395242e+00 9.38673258e-01 -2.19112426e-01 -7.61981666e-01
-7.36609578e-01 9.91191268e-01 -3.40391785e-01 -2.29644015e-01
6.58896789e-02 4.27410930e-01 -5.64425349e-01 2.71198034e-01
8.81507456e-01 4.71517146e-01 2.39887953e-01 -7.54815876e-01
-5.81069708e-01 -1.81804150e-01 -2.67017424e-01 -5.00967205e-01
1.38757789e+00 1.40458584e-01 -1.27249777e-01 2.94223964e-01
1.06164205e+00 5.53283095e-01 -7.76946619e-02 -1.84108734e-01
5.28284013e-01 -3.71278264e-02 -1.43399447e-01 -1.20155013e+00
-2.24118724e-01 3.63746643e-01 1.96564630e-01 2.53271371e-01
6.20341420e-01 1.94970459e-01 6.14645064e-01 2.44110525e-01
4.21095461e-01 -1.30926692e+00 -3.12275499e-01 7.91929722e-01
9.66969132e-01 -5.93527079e-01 2.66841263e-01 -5.09465456e-01
-6.60652518e-01 8.52945805e-01 1.12811156e-01 -3.43378484e-02
3.74216735e-01 3.84460658e-01 -1.74171194e-01 -3.96499068e-01
-3.92846912e-01 -2.39089891e-01 4.66049582e-01 4.88715768e-01
5.44595838e-01 -1.08311512e-01 -1.32891715e+00 6.91816270e-01
-2.93161690e-01 -1.46846548e-01 2.46691212e-01 1.10059726e+00
-8.49891305e-01 -9.50111926e-01 -2.26467058e-01 6.31590784e-01
-9.70122099e-01 -5.03586292e-01 -7.65851021e-01 6.49868548e-01
1.89392373e-01 7.90405691e-01 -2.35445425e-01 -5.03324449e-01
2.64755674e-02 4.54281360e-01 5.71623206e-01 -7.38898575e-01
-1.04033482e+00 1.47994637e-01 9.49838161e-01 -3.18681635e-02
-5.36895692e-01 -1.01194119e+00 -9.48378623e-01 -2.93811768e-01
-4.02553916e-01 6.46143496e-01 4.69272733e-01 1.04086626e+00
-4.20289449e-02 2.20054328e-01 1.30630344e-01 -3.26030672e-01
-9.86750543e-01 -1.20233142e+00 -4.67211485e-01 7.06251562e-01
-1.44024312e-01 -5.67856133e-01 -1.68131962e-01 2.37520725e-01] | [8.504135131835938, 10.721508979797363] |
a554eb14-0147-4b2c-b471-744c35fad533 | break-the-ceiling-stronger-multi-scale-deep | 1906.02174 | null | https://arxiv.org/abs/1906.02174v3 | https://arxiv.org/pdf/1906.02174v3.pdf | Break the Ceiling: Stronger Multi-scale Deep Graph Convolutional Networks | Recently, neural network based approaches have achieved significant improvement for solving large, complex, graph-structured problems. However, their bottlenecks still need to be addressed, and the advantages of multi-scale information and deep architectures have not been sufficiently exploited. In this paper, we theoretically analyze how existing Graph Convolutional Networks (GCNs) have limited expressive power due to the constraint of the activation functions and their architectures. We generalize spectral graph convolution and deep GCN in block Krylov subspace forms and devise two architectures, both with the potential to be scaled deeper but each making use of the multi-scale information in different ways. We further show that the equivalence of these two architectures can be established under certain conditions. On several node classification tasks, with or without the help of validation, the two new architectures achieve better performance compared to many state-of-the-art methods. | ['Doina Precup', 'Xiao-Wen Chang', 'Mingde Zhao', 'Sitao Luan'] | 2019-06-05 | break-the-ceiling-stronger-multi-scale-deep-1 | http://papers.nips.cc/paper/9276-break-the-ceiling-stronger-multi-scale-deep-graph-convolutional-networks | http://papers.nips.cc/paper/9276-break-the-ceiling-stronger-multi-scale-deep-graph-convolutional-networks.pdf | neurips-2019-12 | ['node-classification-on-non-homophilic'] | ['graphs'] | [ 9.09594744e-02 2.08229348e-01 -1.48605675e-01 -8.74840021e-02
-2.08378374e-03 -5.13219118e-01 4.61005688e-01 8.36677253e-02
-3.04495394e-01 5.60260057e-01 -4.24340228e-03 -6.04628980e-01
-3.06316167e-01 -9.05739844e-01 -5.57922959e-01 -5.91107726e-01
-4.78282958e-01 3.22777003e-01 2.00136781e-01 -3.49644244e-01
-2.10584477e-01 6.49764597e-01 -9.89547670e-01 9.83440578e-02
7.94541001e-01 8.40274274e-01 -4.82624955e-02 6.97295010e-01
-1.72722310e-01 6.07818723e-01 -1.14159927e-01 -3.30341190e-01
3.96324039e-01 -1.78590804e-01 -8.73419464e-01 -1.36519626e-01
4.48105603e-01 -1.18545830e-01 -7.87418604e-01 1.07778394e+00
3.91811043e-01 1.09897837e-01 3.90240431e-01 -1.24958622e+00
-8.04487586e-01 6.43675447e-01 -4.30252492e-01 1.32582217e-01
-1.29336447e-01 -9.44169238e-02 1.20638049e+00 -6.74282253e-01
5.58619022e-01 1.09635592e+00 1.08371186e+00 4.08963740e-01
-1.51295769e+00 -5.58446407e-01 2.63873726e-01 1.48912638e-01
-1.47189069e+00 -1.83290511e-01 1.00093055e+00 -2.15511128e-01
1.17200077e+00 1.08036324e-02 6.78623259e-01 1.08652246e+00
-2.25811414e-02 6.55772686e-01 8.40361536e-01 -3.33042413e-01
2.41405088e-02 -8.98540094e-02 2.84155369e-01 1.06128430e+00
4.28804547e-01 8.07421952e-02 -1.58602133e-01 -1.37289613e-01
7.93834984e-01 -4.99819219e-02 -4.51498419e-01 -8.16504419e-01
-9.63130713e-01 1.08106387e+00 1.00550354e+00 8.15658510e-01
-1.32330507e-01 4.20551270e-01 3.87000680e-01 4.38025177e-01
5.34184337e-01 4.99034941e-01 -2.95070797e-01 3.02218825e-01
-9.38892543e-01 -1.54662384e-02 1.01889181e+00 6.86046004e-01
7.37241507e-01 3.75294566e-01 1.26988575e-01 4.11538243e-01
1.23634167e-01 3.32449041e-02 1.46600217e-01 -6.32443070e-01
3.50449830e-01 7.92776585e-01 -5.40115654e-01 -1.13332653e+00
-9.74385917e-01 -8.75023127e-01 -1.33174539e+00 6.29284754e-02
4.55901235e-01 -1.78841233e-01 -9.72571254e-01 1.77119362e+00
-2.83893896e-03 3.65844816e-01 8.60903710e-02 7.67360389e-01
8.84159744e-01 4.80652124e-01 -2.78198551e-02 1.58073276e-01
9.23037767e-01 -1.02656412e+00 -5.22174239e-01 -2.73218483e-01
9.03413475e-01 -1.55499220e-01 8.39428067e-01 1.22259676e-01
-9.37658787e-01 -3.17837805e-01 -1.29059625e+00 -1.43156230e-01
-6.14625633e-01 3.59148681e-02 1.13177800e+00 8.16274941e-01
-1.62996852e+00 8.81546259e-01 -8.56730044e-01 -3.42196584e-01
5.33828020e-01 5.19792616e-01 -5.31965077e-01 -8.48678127e-02
-1.32088673e+00 5.97811043e-01 3.72113496e-01 3.32092643e-01
-6.70751035e-01 -7.03834116e-01 -1.02371109e+00 5.96181989e-01
4.82623011e-01 -8.53184700e-01 6.76375568e-01 -9.10547435e-01
-1.16590774e+00 5.52371979e-01 3.46475154e-01 -6.37341559e-01
3.41186434e-01 1.97971299e-01 -1.98447317e-01 1.88803375e-01
-2.23234490e-01 5.65425575e-01 5.51230431e-01 -7.79300690e-01
-2.78401785e-02 -4.42832559e-01 3.61752868e-01 6.21174090e-02
-7.97303796e-01 -1.53002709e-01 -4.67384815e-01 -5.47923028e-01
-9.24690068e-02 -1.07050073e+00 -5.95205903e-01 -3.48678455e-02
-3.42550844e-01 -1.60999775e-01 7.24211812e-01 -4.15723294e-01
1.09944975e+00 -2.11935449e+00 4.57038581e-01 3.87226552e-01
5.93995929e-01 5.94654858e-01 -4.45018828e-01 6.56244636e-01
-1.67399719e-01 3.72689545e-01 -2.58466452e-01 -2.81249195e-01
-4.22378592e-02 1.77364290e-01 8.91721919e-02 4.35808003e-01
2.56327063e-01 1.21899378e+00 -7.43955553e-01 -2.10729912e-01
8.66854191e-02 7.55578339e-01 -5.17584443e-01 -2.01779753e-01
-1.06932774e-01 1.54189780e-01 -4.46577847e-01 2.78420389e-01
5.54460108e-01 -7.39363909e-01 5.89296162e-01 -2.49319196e-01
2.95011312e-01 -3.17271054e-02 -1.18300915e+00 1.90101862e+00
-2.55898505e-01 7.42336452e-01 4.36696589e-01 -1.51794577e+00
5.10972857e-01 2.50422239e-01 4.31884974e-01 -5.55687428e-01
4.20587473e-02 1.02345981e-01 3.06601584e-01 -7.43663684e-02
1.92499027e-01 -5.26081547e-02 2.54174501e-01 2.45460033e-01
2.28292480e-01 3.40018451e-01 3.76639336e-01 4.95820135e-01
1.38849342e+00 -1.36830330e-01 1.43402725e-01 -5.81441998e-01
4.43985939e-01 -1.60448313e-01 2.33282879e-01 8.75137091e-01
6.49629235e-02 3.77811074e-01 7.86435723e-01 -5.73006749e-01
-9.00786579e-01 -7.13590503e-01 1.59490094e-01 9.43519354e-01
-1.28531769e-01 -5.70938170e-01 -7.23582029e-01 -7.36952960e-01
3.40297706e-02 1.97688684e-01 -7.78212905e-01 -2.06128821e-01
-7.34141111e-01 -8.91267478e-01 8.50391269e-01 8.37077320e-01
4.93952423e-01 -8.03301036e-01 -3.49712580e-01 2.04110831e-01
2.38102347e-01 -1.38367736e+00 -8.99107903e-02 2.38672331e-01
-9.09121931e-01 -1.08721781e+00 -7.56229997e-01 -6.91619039e-01
5.22485375e-01 4.79995638e-01 1.15122604e+00 4.72081333e-01
-2.78968781e-01 2.43956551e-01 -2.25785270e-01 -1.39034279e-02
-2.48670608e-01 6.35562658e-01 -1.88438118e-01 -6.30427450e-02
-5.77008612e-02 -9.26228166e-01 -4.98583645e-01 -5.45226596e-03
-1.08750582e+00 1.55555829e-01 7.16981947e-01 8.43177140e-01
-8.19808543e-02 1.67697161e-01 5.29050112e-01 -1.04709589e+00
8.64323676e-01 -2.42516562e-01 -5.31398773e-01 2.18419045e-01
-8.50492299e-01 4.47803229e-01 8.29499841e-01 -1.39728218e-01
-5.23884356e-01 -2.33378261e-01 -1.30359277e-01 -5.62502980e-01
2.90743019e-02 1.01806045e+00 6.12858869e-03 -4.99535173e-01
5.44370413e-01 -1.27602875e-01 1.86302140e-01 -3.97843480e-01
4.65912253e-01 5.91160879e-02 1.50416732e-01 -4.26688582e-01
8.45144510e-01 5.69825709e-01 5.31834006e-01 -8.43047261e-01
-7.05988348e-01 -2.64166355e-01 -7.03042686e-01 9.60002393e-02
8.20930183e-01 -7.83467889e-01 -5.68275392e-01 4.42896515e-01
-9.01038229e-01 -4.93390262e-01 -5.88358194e-03 3.06414276e-01
-1.77703083e-01 6.46548510e-01 -9.54259396e-01 -5.94283938e-01
-3.39429617e-01 -1.01968122e+00 7.08901942e-01 -2.51596868e-02
3.04513156e-01 -1.34851241e+00 -2.06189707e-01 -1.02289382e-03
7.60177195e-01 3.25845629e-01 1.15708899e+00 -8.04486394e-01
-5.46396852e-01 -2.99128056e-01 -5.99692404e-01 2.99132973e-01
-2.20828578e-01 -1.99488565e-01 -6.83627903e-01 -7.64793694e-01
-3.80107373e-01 -1.99448943e-01 1.25017357e+00 3.05139393e-01
1.10115862e+00 -1.63235396e-01 -4.34343636e-01 9.09002960e-01
1.52725995e+00 -3.26492101e-01 4.52813268e-01 8.69816393e-02
1.02231789e+00 5.62870145e-01 -2.91491359e-01 -1.61108613e-01
2.40067199e-01 5.06244481e-01 6.74484015e-01 -3.62853587e-01
-1.90715164e-01 -5.59521541e-02 1.27751231e-01 8.31735015e-01
-1.84346005e-01 -3.73676896e-01 -1.05518317e+00 5.17683387e-01
-2.14053369e+00 -7.30725288e-01 -1.40908435e-01 1.69685543e+00
2.39317104e-01 1.82210118e-01 8.89006704e-02 6.69535100e-02
4.88350481e-01 4.89300579e-01 -4.75590497e-01 -3.16883743e-01
-2.10044473e-01 4.48911637e-01 6.52178168e-01 3.99181694e-01
-1.08924937e+00 9.77289736e-01 6.67518520e+00 7.37743199e-01
-1.07107651e+00 -3.75316106e-03 4.18577820e-01 1.76629171e-01
-3.62109214e-01 9.87710431e-02 -3.61546159e-01 -5.96691333e-02
1.08599567e+00 2.98253417e-01 7.58827686e-01 5.53727746e-01
-2.62210608e-01 3.78180891e-01 -1.05342185e+00 7.17368484e-01
7.65922591e-02 -1.57510948e+00 -1.01245269e-01 2.91207820e-01
6.19957983e-01 3.89498770e-01 3.63270864e-02 3.57630938e-01
5.21989703e-01 -1.36874592e+00 2.63673306e-01 1.75625771e-01
6.07457817e-01 -6.10725939e-01 6.57482922e-01 3.17036361e-01
-1.43455982e+00 -1.22491091e-01 -3.71337622e-01 -3.38098288e-01
-1.26653567e-01 5.48168004e-01 -5.65884411e-01 9.95747864e-01
5.72486579e-01 7.98607349e-01 -7.20764041e-01 6.71206892e-01
-7.41821826e-02 6.58636212e-01 -3.59867692e-01 -2.36158594e-01
7.60278583e-01 -3.62701446e-01 3.79753500e-01 1.06128657e+00
1.85122877e-01 -2.76037782e-01 2.66860038e-01 1.10417438e+00
-8.55148137e-02 8.82276967e-02 -7.43906975e-01 -4.45526540e-01
5.63159445e-03 1.52071786e+00 -7.84495533e-01 -8.68000984e-02
-7.29650557e-01 7.91752338e-01 8.22624326e-01 6.76618099e-01
-5.76389313e-01 -3.46398830e-01 4.55602884e-01 9.50364247e-02
4.27106082e-01 -6.03757024e-01 -4.01965417e-02 -1.30106938e+00
-5.49318194e-02 -6.87912107e-01 4.88119423e-01 -5.20804346e-01
-1.19501662e+00 5.86949646e-01 -1.62849635e-01 -5.06651282e-01
-1.49616808e-01 -1.09663999e+00 -5.55136323e-01 7.10770249e-01
-1.56705678e+00 -1.42865181e+00 -2.26466745e-01 6.30090475e-01
-4.33429927e-02 -1.19322911e-01 8.42251718e-01 3.95637214e-01
-6.12522364e-01 5.52258968e-01 1.19795904e-01 4.78608996e-01
1.21768102e-01 -1.24526262e+00 5.62367082e-01 8.67552221e-01
4.40428317e-01 5.63661695e-01 3.27507824e-01 -3.82953286e-01
-1.53130126e+00 -1.05304193e+00 5.02781689e-01 -4.31332625e-02
9.11612988e-01 -7.18282044e-01 -9.32050288e-01 7.14929342e-01
2.99269974e-01 3.47680479e-01 6.04039431e-01 5.16366363e-01
-6.27214432e-01 3.66058089e-02 -8.18880260e-01 6.78300142e-01
1.29034221e+00 -5.77420712e-01 2.86658597e-03 1.86077416e-01
8.36721539e-01 -8.51933435e-02 -8.11186016e-01 6.20622516e-01
4.07567948e-01 -1.05688059e+00 1.02350128e+00 -1.05705571e+00
1.24494791e-01 -4.54501361e-02 -4.97644395e-02 -1.33207214e+00
-6.27451062e-01 -5.46735287e-01 -1.42713919e-01 7.03381062e-01
4.12143677e-01 -8.55159521e-01 9.83507931e-01 2.86530197e-01
-2.42473304e-01 -8.85337353e-01 -8.78714025e-01 -7.73118019e-01
3.87129873e-01 -2.37616926e-01 4.91465330e-01 1.05491960e+00
-2.58223601e-02 6.70372367e-01 -3.19346905e-01 1.09124653e-01
4.01684850e-01 2.54173249e-01 6.20275021e-01 -1.47057652e+00
-2.61130095e-01 -7.28249609e-01 -5.71650982e-01 -7.50119388e-01
5.29150784e-01 -1.31927264e+00 -5.58661222e-01 -1.64525604e+00
2.01074377e-01 -2.96501756e-01 -4.80590463e-01 6.09257817e-01
-8.46449882e-02 3.06645036e-01 2.10545942e-01 1.25273177e-02
-4.24641848e-01 3.78164679e-01 1.06409514e+00 -2.37946033e-01
4.14617360e-02 -2.69304067e-01 -6.43514454e-01 5.02572715e-01
8.07825446e-01 3.33223008e-02 -5.27762353e-01 -5.05988359e-01
5.63289165e-01 8.28607846e-03 6.71388865e-01 -1.12548029e+00
2.35265791e-01 3.06100130e-01 2.46330693e-01 -2.27521092e-01
2.17104673e-01 -8.25457811e-01 2.17682064e-01 6.24721527e-01
-3.00218731e-01 2.73556888e-01 2.08305836e-01 8.30160558e-01
-8.94120112e-02 -8.23353454e-02 6.24025822e-01 -2.30006248e-01
-5.35530806e-01 5.83034277e-01 -1.62301213e-01 -1.40766397e-01
8.09567392e-01 -1.50086045e-01 -2.90871352e-01 -4.66541886e-01
-6.90256000e-01 1.63826853e-01 3.11074376e-01 3.22768569e-01
4.02531952e-01 -1.27513254e+00 -6.69969678e-01 2.72480994e-01
-1.15044095e-01 -1.61766723e-01 2.73265898e-01 1.00093091e+00
-5.56351900e-01 6.13793314e-01 -1.41009256e-01 -4.57911640e-01
-8.94698262e-01 9.02621865e-01 4.45597112e-01 -7.68267155e-01
-8.06597590e-01 6.28690958e-01 3.22971642e-01 -5.59675515e-01
1.62348971e-01 -3.15110981e-01 -7.47896656e-02 -1.01456828e-01
-3.70152146e-02 1.09872967e-01 2.33080089e-01 -4.49475914e-01
-2.70230442e-01 4.11601961e-01 -3.14676277e-02 3.09696257e-01
1.40784204e+00 1.26354232e-01 -1.74517378e-01 5.03782406e-02
1.33061922e+00 -2.54812598e-01 -9.14983392e-01 -3.73316020e-01
4.76957299e-02 3.96990255e-02 1.53577626e-01 -4.18004453e-01
-1.50108445e+00 1.20121837e+00 3.23840886e-01 6.14754021e-01
9.69735861e-01 -4.51703072e-02 7.32281029e-01 4.15931493e-01
1.76072940e-01 -8.22003663e-01 4.39837575e-03 5.62284291e-01
5.03795743e-01 -9.75175917e-01 -7.98303410e-02 -6.51199579e-01
-2.87833698e-02 1.22200489e+00 4.56519812e-01 -3.79362106e-01
8.57212484e-01 2.63664216e-01 -2.81451672e-01 -5.43290019e-01
-6.57408357e-01 -4.47148979e-01 4.27971244e-01 6.14088178e-01
3.80778462e-01 2.28141230e-02 -1.83527693e-01 2.87693113e-01
9.20815244e-02 -3.18968862e-01 3.28623444e-01 5.16010344e-01
-1.66472465e-01 -9.30192411e-01 2.28181005e-01 4.23089921e-01
-4.63037342e-01 -3.44025195e-01 -5.62888682e-01 1.08629394e+00
-3.93615812e-01 6.37490511e-01 -1.48026779e-01 -3.96978378e-01
6.74538082e-03 -7.01412372e-03 4.64569747e-01 -4.74374890e-01
-5.27896583e-01 -1.42463923e-01 2.20741525e-01 -6.46369398e-01
-4.69744354e-01 -1.54034480e-01 -1.02923501e+00 -3.14963162e-01
-5.02836049e-01 1.22499391e-01 5.64270198e-01 7.92607009e-01
6.00553989e-01 7.92216003e-01 1.90489903e-01 -7.54788160e-01
-6.59970820e-01 -9.23825562e-01 -5.47428668e-01 2.43838161e-01
2.06575572e-01 -5.16130507e-01 -2.33376265e-01 -6.82748973e-01] | [6.878276824951172, 6.095010280609131] |
ed4db732-71ae-4cf0-8784-e9a718ba1118 | multiple-source-domain-adaptation-via | 2201.1187 | null | https://arxiv.org/abs/2201.11870v2 | https://arxiv.org/pdf/2201.11870v2.pdf | Multiple-Source Domain Adaptation via Coordinated Domain Encoders and Paired Classifiers | We present a novel multiple-source unsupervised model for text classification under domain shift. Our model exploits the update rates in document representations to dynamically integrate domain encoders. It also employs a probabilistic heuristic to infer the error rate in the target domain in order to pair source classifiers. Our heuristic exploits data transformation cost and the classifier accuracy in the target feature space. We have used real world scenarios of Domain Adaptation to evaluate the efficacy of our algorithm. We also used pretrained multi-layer transformers as the document encoder in the experiments to demonstrate whether the improvement achieved by domain adaptation models can be delivered by out-of-the-box language model pretraining. The experiments testify that our model is the top performing approach in this setting. | ['Payam Karisani'] | 2022-01-28 | null | null | null | null | ['cross-domain-text-classification'] | ['natural-language-processing'] | [ 2.67235309e-01 4.21765357e-01 -4.95783061e-01 -6.21428013e-01
-9.54978466e-01 -7.16983020e-01 8.51762474e-01 2.93122053e-01
-6.67730987e-01 8.84932101e-01 1.37959346e-01 -2.54113138e-01
-7.75802732e-02 -6.11493409e-01 -7.86973894e-01 -3.77627134e-01
7.24648014e-02 9.64808643e-01 2.82509238e-01 -4.08498853e-01
1.48938134e-01 2.46736854e-01 -1.32197022e+00 6.78319752e-01
7.53757596e-01 5.77431381e-01 1.65272176e-01 8.72652650e-01
-4.25527364e-01 6.75311625e-01 -8.36297870e-01 -4.52273458e-01
3.78154576e-01 -2.90431798e-01 -9.18183506e-01 6.08823039e-02
9.45442095e-02 -2.51035094e-01 -2.09664330e-01 8.49557638e-01
2.72265524e-01 1.52398199e-01 1.07053828e+00 -1.17762637e+00
-6.76182866e-01 9.03713703e-01 -1.90901011e-01 3.17037731e-01
3.15516829e-01 -2.74934411e-01 7.98883438e-01 -7.57177472e-01
8.29452813e-01 1.23230493e+00 5.29949367e-01 5.01103818e-01
-1.39987123e+00 -6.58360362e-01 1.83021262e-01 -6.09586686e-02
-1.13184834e+00 -5.47139287e-01 6.29435718e-01 -3.84501040e-01
1.23211920e+00 -3.62840831e-01 -1.58798084e-01 1.39094257e+00
2.47598633e-01 6.32566512e-01 1.01705050e+00 -1.06346607e+00
3.30832154e-01 8.39972317e-01 2.81632751e-01 4.85560834e-01
3.64221483e-01 1.67626724e-01 -6.60986841e-01 -3.54694873e-01
5.31533897e-01 -3.79940987e-01 7.57470652e-02 -5.38071394e-01
-9.06987846e-01 1.04374003e+00 1.04289658e-01 5.38692236e-01
-1.43177733e-01 6.71624904e-03 5.15788317e-01 6.73172474e-01
6.84819460e-01 5.45757353e-01 -9.33513105e-01 -1.00154459e-01
-9.38935339e-01 8.06095153e-02 9.31166053e-01 1.19013727e+00
5.67628384e-01 -5.87359518e-02 -4.15383205e-02 9.74888623e-01
2.59050161e-01 3.55311185e-01 9.72485602e-01 -7.05740690e-01
7.03558087e-01 5.98607779e-01 1.36502475e-01 -3.77641410e-01
-1.68298796e-01 -2.28194937e-01 -3.72436762e-01 -1.74402948e-02
4.69179541e-01 -2.56814420e-01 -8.77804577e-01 1.71181715e+00
-3.42283621e-02 -6.33454695e-02 5.28335094e-01 2.69269824e-01
1.86537296e-01 6.01534605e-01 3.28170717e-01 -6.53755814e-02
1.16307354e+00 -6.01051569e-01 -5.29081821e-01 -3.59400123e-01
1.13498950e+00 -6.62162542e-01 1.05243814e+00 2.79314160e-01
-7.76458621e-01 -6.90518618e-01 -1.30048299e+00 6.33139207e-05
-7.16162860e-01 1.46210924e-01 3.04294705e-01 8.84810150e-01
-7.96372294e-01 5.68277478e-01 -8.09809387e-01 -7.10532546e-01
2.22055078e-01 4.22076941e-01 -3.56307626e-01 2.72616651e-02
-1.37575459e+00 1.07650423e+00 8.91811788e-01 -8.72860849e-01
-6.69615686e-01 -5.81782043e-01 -6.33363545e-01 8.14547390e-02
-5.35290353e-02 -4.87857997e-01 1.57889485e+00 -1.20778549e+00
-1.57383525e+00 8.08880806e-01 -2.12546766e-01 -7.46456027e-01
3.53158057e-01 -1.07385613e-01 -4.95603651e-01 7.03459606e-02
3.03237289e-02 6.53554797e-01 8.92557263e-01 -1.12152708e+00
-8.85681868e-01 -2.56180108e-01 -2.28849858e-01 1.76666334e-01
-7.97263086e-01 6.72601983e-02 1.13061726e-01 -6.34232163e-01
-2.00716034e-01 -7.56249666e-01 7.04304278e-02 -4.43688571e-01
1.96088538e-01 -2.85790205e-01 7.31164753e-01 -6.13752723e-01
1.05081141e+00 -2.09196305e+00 -2.21423894e-01 2.59789616e-01
-2.55222857e-01 2.77437717e-02 -1.58581480e-01 4.55353200e-01
-1.22705966e-01 1.25649959e-01 -2.41802469e-01 -3.78556281e-01
-1.72790512e-02 2.84717947e-01 -6.28205955e-01 9.45957825e-02
3.83315891e-01 5.03085673e-01 -7.75393009e-01 -5.09254277e-01
-3.11923414e-01 3.77073646e-01 -5.27585685e-01 3.13691705e-01
-3.54044437e-01 -5.53499386e-02 -3.24690104e-01 1.26288489e-01
3.29750776e-01 -1.15776099e-01 4.97850120e-01 1.89967513e-01
3.33516330e-01 5.30360401e-01 -1.18201804e+00 1.86432254e+00
-5.07755637e-01 6.38398468e-01 -2.99576610e-01 -1.15093839e+00
1.13839030e+00 4.52844113e-01 2.80030280e-01 -6.01475418e-01
-1.06122270e-02 3.21860582e-01 4.73414455e-03 -1.65526867e-01
6.49357259e-01 -2.72271425e-01 -2.78116226e-01 6.52038038e-01
6.50429428e-01 3.12390756e-02 1.16659842e-01 2.73396701e-01
9.39472616e-01 3.64665747e-01 3.93125951e-01 -5.03308415e-01
3.19650203e-01 3.52031648e-01 9.37765315e-02 7.31170356e-01
-2.47554630e-02 1.33680403e-01 3.67435336e-01 -1.52139500e-01
-1.15821958e+00 -1.08114505e+00 -2.65175879e-01 1.67729092e+00
-2.96133101e-01 -2.10763946e-01 -6.60339653e-01 -1.20798993e+00
7.91317725e-04 1.18030012e+00 -4.75216508e-01 -4.64095861e-01
-3.51022542e-01 -8.42800856e-01 7.99386263e-01 8.11117291e-01
3.46534640e-01 -5.84431648e-01 -5.96943557e-01 4.48957890e-01
2.86359526e-02 -1.02406836e+00 -1.52189583e-01 8.62883925e-01
-1.02836406e+00 -8.01534772e-01 -5.49716473e-01 -9.21320498e-01
6.84105456e-01 -1.72908410e-01 1.06034017e+00 -5.43923259e-01
2.84316987e-01 5.20672858e-01 -4.72013980e-01 -6.61959827e-01
-1.13221633e+00 5.70655406e-01 2.13028505e-01 -3.39606345e-01
8.80397022e-01 -3.65615159e-01 -2.82774549e-02 2.15552628e-01
-9.29724991e-01 -2.97580302e-01 4.64407355e-01 9.99228776e-01
1.11859031e-01 2.96345174e-01 8.47576559e-01 -1.19963801e+00
9.99602675e-01 -5.93538940e-01 -5.76104879e-01 3.38369012e-01
-9.13952649e-01 7.36514330e-01 7.71794677e-01 -5.87225735e-01
-1.42846870e+00 2.19965458e-01 9.60024744e-02 2.00329013e-02
-3.70826721e-01 5.31435490e-01 -8.01818967e-02 3.74798447e-01
1.20550334e+00 1.55606672e-01 -7.94010386e-02 -5.63713610e-01
4.59496021e-01 1.13391078e+00 4.52302635e-01 -8.02631795e-01
6.00024998e-01 1.51522666e-01 -4.34787422e-01 -5.00520587e-01
-7.05132186e-01 -3.70695084e-01 -1.18070996e+00 3.53141844e-01
3.98238927e-01 -1.06674504e+00 -1.37929264e-02 1.34688377e-01
-1.21022522e+00 -4.71339732e-01 -4.23671395e-01 3.55528653e-01
-5.64570367e-01 8.59918371e-02 -3.66419256e-01 -8.11228454e-01
-2.26088494e-01 -8.60783815e-01 7.09048629e-01 4.14913408e-02
-3.64207715e-01 -1.38613224e+00 1.84434876e-01 -2.00383678e-01
4.29453135e-01 -3.01673830e-01 1.12665451e+00 -1.61665547e+00
1.72000796e-01 -2.02486798e-01 -2.74889525e-02 4.04859751e-01
2.18877777e-01 -2.94659168e-01 -1.29808807e+00 -3.58099133e-01
-1.42323449e-01 -2.88426042e-01 6.27017677e-01 3.77476104e-02
9.39281344e-01 -1.71746656e-01 -5.06788731e-01 3.16824377e-01
1.28587484e+00 4.03690666e-01 2.89304167e-01 5.68185449e-01
2.38469541e-01 4.72698927e-01 4.96528774e-01 2.61399090e-01
3.25014919e-01 7.04748273e-01 -4.41507310e-01 1.06382698e-01
-6.94271624e-02 -5.67056239e-01 6.29771411e-01 6.43163204e-01
4.47333157e-01 -3.92703295e-01 -1.22586799e+00 6.41357780e-01
-1.56239891e+00 -7.45503306e-01 4.17742580e-01 2.20797968e+00
1.22402704e+00 6.60327911e-01 1.99773282e-01 4.84213158e-02
5.82031071e-01 -5.05468547e-01 -4.03636187e-01 -7.61004925e-01
4.56222817e-02 4.67910409e-01 7.45149076e-01 5.79551041e-01
-1.06559741e+00 9.66220677e-01 7.52193260e+00 4.57528353e-01
-9.15689468e-01 1.45484164e-01 4.72343296e-01 1.88665360e-01
2.83613596e-02 -5.13781868e-02 -1.19057846e+00 5.42271852e-01
1.80907333e+00 -5.67087054e-01 2.37284884e-01 1.14545417e+00
-3.00252736e-01 8.00458714e-02 -1.51761627e+00 4.78829741e-01
9.05485228e-02 -9.72086072e-01 1.92988515e-01 -1.67251956e-02
6.05253339e-01 8.01631808e-02 -2.52949521e-02 5.68040073e-01
8.74606609e-01 -7.50352323e-01 4.28708017e-01 1.89570665e-01
8.85037780e-01 -8.54159176e-01 7.25808084e-01 5.22084773e-01
-6.84982300e-01 -2.66036689e-01 -5.64653277e-01 5.64925931e-02
-3.33111435e-01 3.52256715e-01 -1.73710513e+00 3.16059023e-01
2.94929713e-01 4.98048216e-01 -7.42637455e-01 5.41990399e-01
-7.84477070e-02 7.66740739e-01 -4.64350700e-01 4.74725738e-02
-8.67624953e-02 3.87862206e-01 2.97658026e-01 1.73855650e+00
3.16060424e-01 -3.67536098e-01 6.46521151e-02 5.41034997e-01
-1.65166482e-01 9.35731679e-02 -8.01392734e-01 -7.29073137e-02
6.30042493e-01 5.90691268e-01 -4.91054773e-01 -6.95777833e-01
-2.78104782e-01 1.01574600e+00 2.94064641e-01 3.33441287e-01
-5.09682238e-01 -5.66100538e-01 4.81129557e-01 1.63741544e-01
2.58139431e-01 -1.68438911e-01 -4.23471212e-01 -1.16808128e+00
-2.87826896e-01 -8.76716137e-01 7.11824477e-01 -5.62098145e-01
-1.44632256e+00 5.89800119e-01 5.15007854e-01 -1.04116035e+00
-8.46537352e-01 -8.64426494e-01 -1.07850596e-01 9.07016277e-01
-1.46858621e+00 -7.12393284e-01 1.16884634e-01 6.63921654e-01
7.41562366e-01 -6.50389612e-01 1.22731900e+00 -2.72355182e-03
-2.99809843e-01 1.05928981e+00 6.50934815e-01 3.38001192e-01
1.10815060e+00 -1.31581688e+00 4.81030583e-01 7.63377845e-01
2.42004573e-01 6.13240659e-01 6.96340442e-01 -5.55778742e-01
-8.62910628e-01 -9.79807854e-01 1.02880180e+00 -8.08648944e-01
6.81736410e-01 -6.65778935e-01 -1.07249641e+00 9.34004605e-01
3.42042983e-01 -3.05873752e-01 9.12427604e-01 2.09727049e-01
-6.75277293e-01 -4.41423543e-02 -1.30587173e+00 2.15140045e-01
6.66402996e-01 -5.28930545e-01 -1.16062641e+00 2.66800344e-01
6.69981897e-01 -2.84011066e-01 -8.97758842e-01 -4.06483486e-02
3.74207050e-01 -4.68077928e-01 6.96870863e-01 -8.86944234e-01
5.37618816e-01 2.01868549e-01 -2.35103354e-01 -1.63556409e+00
-2.85487831e-01 -1.06262550e-01 -1.38988242e-01 1.47984028e+00
7.17315078e-01 -8.37237835e-01 6.15412295e-01 7.28217065e-01
2.76878506e-01 5.09392545e-02 -8.92526448e-01 -1.04645145e+00
7.26205349e-01 -4.18075562e-01 5.91429830e-01 8.94036770e-01
2.56980747e-01 6.52384937e-01 1.97831839e-01 7.78152719e-02
4.74502921e-01 -2.65740693e-01 6.35171771e-01 -1.42557263e+00
-4.31857735e-01 -2.72723138e-01 -2.88450241e-01 -9.28584695e-01
4.61082190e-01 -1.11386216e+00 1.72663648e-02 -1.01807570e+00
8.66966993e-02 -4.37118322e-01 -6.28362656e-01 5.80554485e-01
-2.91645457e-03 -3.35029125e-01 4.59094942e-02 2.11466625e-01
-3.12567204e-01 1.42442554e-01 4.25933212e-01 -1.65833354e-01
-4.24042851e-01 -2.64538914e-01 -8.66676807e-01 5.25984466e-01
1.05720043e+00 -9.36953902e-01 -6.98533118e-01 -5.09539783e-01
-9.98375043e-02 -3.08556855e-01 3.05019375e-02 -9.63452935e-01
2.02020124e-01 -5.14672808e-02 6.72705889e-01 -3.99400815e-02
3.66133987e-03 -1.09388697e+00 -4.15948778e-01 3.09290826e-01
-9.05602098e-01 1.31381840e-01 5.63950717e-01 5.20838022e-01
-8.30442756e-02 -4.43566352e-01 9.37509775e-01 -2.46923640e-02
-7.80949175e-01 -3.70061249e-01 -4.79118437e-01 1.95110574e-01
7.64168739e-01 -1.11227915e-01 -3.26875567e-01 -1.68399259e-01
-4.96308565e-01 -9.72562283e-02 4.37594652e-01 5.59144616e-01
1.80410340e-01 -1.12577558e+00 -8.32276344e-01 4.78118271e-01
4.09306884e-01 -3.00649107e-01 -4.97383833e-01 8.28054398e-02
-9.23206657e-02 5.96846163e-01 -2.70563126e-01 -5.71413636e-01
-1.03097701e+00 5.31562209e-01 3.74358565e-01 -6.58192754e-01
-3.14720273e-01 7.14493215e-01 -6.59416467e-02 -4.79085714e-01
2.46520370e-01 -1.59922019e-01 4.95030992e-02 -1.00459829e-02
4.73803937e-01 -2.48454083e-02 3.75296921e-01 -1.10070676e-01
-3.34181070e-01 -4.14845906e-02 -5.23020089e-01 -4.84157950e-01
1.20975971e+00 -1.99485525e-01 5.71762085e-01 6.62865758e-01
1.26072073e+00 -2.40216568e-01 -1.10227394e+00 -4.60505009e-01
4.80907410e-01 -1.60750851e-01 1.17016867e-01 -1.25815618e+00
-3.48278493e-01 8.09229732e-01 7.46252537e-01 -2.92494446e-02
1.06011832e+00 -1.68911785e-01 2.94666409e-01 6.98292971e-01
3.37285340e-01 -1.51816416e+00 -2.74645872e-02 7.19104111e-01
5.62085330e-01 -1.14441359e+00 1.00731049e-02 6.39175102e-02
-8.16481173e-01 1.17501378e+00 5.45473754e-01 -1.17830381e-01
7.56456017e-01 5.58884263e-01 1.75881490e-01 2.53866792e-01
-1.13210285e+00 1.86359271e-01 5.41700423e-02 8.43777180e-01
7.02080548e-01 -1.52837202e-01 -4.99336459e-02 6.88894272e-01
-2.74822831e-01 1.86978921e-01 4.72044080e-01 1.09049726e+00
-4.04605567e-01 -1.46271992e+00 -4.74669904e-01 3.66747350e-01
-3.17017734e-01 -2.70037688e-02 -5.10823309e-01 9.19090331e-01
-6.13274872e-02 8.77366364e-01 1.60141066e-01 -3.18520635e-01
3.96467000e-01 1.00868177e+00 3.74201924e-01 -8.77490699e-01
-4.88286763e-01 -1.42780215e-01 9.26606059e-02 1.22972995e-01
-2.11374834e-01 -6.13383055e-01 -1.31473029e+00 1.54702783e-01
-2.32769951e-01 2.97234297e-01 7.89526105e-01 9.58630800e-01
6.10186875e-01 3.93628746e-01 5.24620533e-01 -2.71148771e-01
-9.63046312e-01 -1.23571420e+00 -4.46487784e-01 5.01212478e-01
8.49124268e-02 -7.51811862e-01 -3.45682681e-01 6.08661890e-01] | [10.752206802368164, 7.976846694946289] |
430013a4-492a-4204-b14e-93bf4a555f55 | joint-adaptive-feature-smoothing-and-topology | 2006.07988 | null | https://arxiv.org/abs/2006.07988v6 | https://arxiv.org/pdf/2006.07988v6.pdf | Adaptive Universal Generalized PageRank Graph Neural Network | In many important graph data processing applications the acquired information includes both node features and observations of the graph topology. Graph neural networks (GNNs) are designed to exploit both sources of evidence but they do not optimally trade-off their utility and integrate them in a manner that is also universal. Here, universality refers to independence on homophily or heterophily graph assumptions. We address these issues by introducing a new Generalized PageRank (GPR) GNN architecture that adaptively learns the GPR weights so as to jointly optimize node feature and topological information extraction, regardless of the extent to which the node labels are homophilic or heterophilic. Learned GPR weights automatically adjust to the node label pattern, irrelevant on the type of initialization, and thereby guarantee excellent learning performance for label patterns that are usually hard to handle. Furthermore, they allow one to avoid feature over-smoothing, a process which renders feature information nondiscriminative, without requiring the network to be shallow. Our accompanying theoretical analysis of the GPR-GNN method is facilitated by novel synthetic benchmark datasets generated by the so-called contextual stochastic block model. We also compare the performance of our GNN architecture with that of several state-of-the-art GNNs on the problem of node-classification, using well-known benchmark homophilic and heterophilic datasets. The results demonstrate that GPR-GNN offers significant performance improvement compared to existing techniques on both synthetic and benchmark data. | ['Pan Li', 'Olgica Milenkovic', 'Jianhao Peng', 'Eli Chien'] | 2020-06-14 | null | https://openreview.net/forum?id=n6jl7fLxrP | https://openreview.net/pdf?id=n6jl7fLxrP | iclr-2021-1 | ['node-classification-on-non-homophilic'] | ['graphs'] | [ 1.75105155e-01 2.06145346e-01 -4.63181794e-01 -2.14848489e-01
-2.61792485e-02 -4.53475744e-01 6.93848193e-01 4.53466028e-01
-2.84853160e-01 6.00547135e-01 -1.01565346e-01 -3.83702695e-01
-5.97900331e-01 -1.18190479e+00 -5.77186406e-01 -9.48973954e-01
-6.71425998e-01 7.85484433e-01 3.05391490e-01 -2.87403017e-01
1.43325225e-01 7.03100681e-01 -1.31201482e+00 -2.66211271e-01
7.84358382e-01 9.27746177e-01 -1.03815347e-01 8.06167901e-01
9.59511101e-03 5.16725600e-01 -2.72536606e-01 -3.72645855e-01
4.76865441e-01 -1.66172788e-01 -6.76417172e-01 -4.22381386e-02
3.92049909e-01 2.71920949e-01 -5.50564051e-01 1.15441871e+00
3.06952238e-01 -4.80162539e-02 9.45885241e-01 -1.23978686e+00
-6.32594466e-01 7.74995685e-01 -6.87270164e-01 3.10077131e-01
-1.55709952e-01 -1.16075985e-02 1.60405862e+00 -6.04142785e-01
8.40371549e-01 1.06248689e+00 8.37729812e-01 1.80893525e-01
-1.46422076e+00 -3.30217481e-01 2.12222084e-01 2.61653084e-02
-1.37695646e+00 -2.18389809e-01 1.07488954e+00 -2.77519286e-01
5.20502508e-01 2.24940360e-01 7.97153592e-01 1.07861626e+00
3.40367734e-01 5.43597102e-01 8.48870456e-01 -3.42673004e-01
3.50014985e-01 -6.97200894e-02 3.50857943e-01 9.92146730e-01
7.06976771e-01 1.31108925e-01 -4.15958554e-01 -2.79670447e-01
8.92949581e-01 -3.39535475e-02 -2.41901219e-01 -1.10469568e+00
-1.01420331e+00 9.01436567e-01 8.48533630e-01 3.47449094e-01
-4.12477016e-01 2.21666440e-01 4.89799708e-01 6.61783993e-01
3.10022593e-01 4.91554379e-01 -3.69328290e-01 3.69124055e-01
-7.02516079e-01 -7.32700527e-02 1.10288787e+00 6.39021695e-01
1.03784466e+00 9.72249731e-02 1.77605465e-01 7.20400572e-01
2.97946721e-01 2.21312359e-01 2.68842936e-01 -4.31560516e-01
3.94269854e-01 9.80935931e-01 -3.47771853e-01 -1.48546910e+00
-6.44736767e-01 -7.67865300e-01 -1.33202958e+00 4.18062657e-02
4.69992757e-01 1.29319280e-01 -1.09369373e+00 1.93174136e+00
3.64386827e-01 6.18020669e-02 -7.61414692e-02 5.95852375e-01
8.41387630e-01 3.18773061e-01 -7.92904943e-02 2.38698591e-02
1.06700468e+00 -7.22694457e-01 -3.86165231e-01 -1.40539855e-01
7.02902615e-01 -1.31294847e-01 1.02584136e+00 7.62257278e-02
-7.33874798e-01 -9.38157365e-02 -1.25131071e+00 2.91774899e-01
-6.75444245e-01 -1.84376985e-01 8.65578413e-01 7.29789376e-01
-1.34664786e+00 9.02324498e-01 -6.94157720e-01 -4.23680902e-01
3.87119651e-01 6.74522102e-01 -5.11819065e-01 1.10242061e-01
-1.21986163e+00 5.81652701e-01 4.62565482e-01 2.19334483e-01
-6.64930940e-01 -5.03137589e-01 -7.87238717e-01 3.28411490e-01
4.68916893e-01 -6.84409142e-01 5.65630734e-01 -9.32664454e-01
-1.17109179e+00 8.00166607e-01 3.08416486e-01 -6.06729567e-01
4.45157290e-01 5.36792576e-01 -3.68585438e-01 3.15770179e-01
-2.72782054e-02 5.46154797e-01 7.70004213e-01 -1.23784840e+00
-1.99233070e-01 -3.60532999e-01 -8.04766193e-02 1.49171129e-01
-5.54773748e-01 -5.86728454e-01 -3.63831729e-01 -6.12551749e-01
2.77471006e-01 -9.30705011e-01 -4.02976692e-01 -1.15518207e-02
-6.44102991e-01 -1.70112669e-01 8.43911469e-01 -4.18749973e-02
1.12396777e+00 -1.83698869e+00 1.14191316e-01 9.86247659e-01
7.40552485e-01 6.19488284e-02 -3.82763863e-01 6.62724316e-01
-1.01772189e-01 1.97337851e-01 -3.21462780e-01 -1.65990312e-02
-1.46598637e-01 3.75958860e-01 6.06892742e-02 8.02015126e-01
-6.55315816e-03 9.46817875e-01 -9.01215255e-01 -5.60274959e-01
7.05244765e-02 3.95797759e-01 -3.73397499e-01 -1.08648695e-01
-9.07538533e-02 9.96077433e-02 -5.70770144e-01 6.38111413e-01
4.20689940e-01 -8.04973662e-01 6.52609348e-01 -2.96991039e-02
3.30622643e-01 1.74482614e-01 -1.23219275e+00 1.11552644e+00
-2.45278731e-01 5.81031799e-01 2.67256826e-01 -1.29376495e+00
1.02222013e+00 3.82823348e-02 5.74153662e-01 -3.91436547e-01
2.55657047e-01 1.30078852e-01 2.48720646e-01 -5.78999594e-02
2.84680724e-01 1.77477941e-01 5.82340434e-02 5.00209630e-01
1.07941590e-01 3.56601506e-01 5.16601264e-01 5.50979793e-01
1.58012378e+00 -5.68796635e-01 4.24850434e-01 -8.80160391e-01
3.98450792e-01 -2.57764280e-01 5.84435165e-01 9.97383833e-01
-6.70347512e-02 4.29358274e-01 9.81610656e-01 -5.79839945e-01
-8.51078391e-01 -1.09013391e+00 -1.00741938e-01 1.12971056e+00
3.33798349e-01 -4.62925851e-01 -3.91541064e-01 -9.77966547e-01
1.16720133e-01 8.94304141e-02 -1.05233598e+00 -2.20713735e-01
-4.88217145e-01 -1.08489275e+00 4.31778312e-01 3.68666649e-01
2.30476439e-01 -9.79471684e-01 -2.00495809e-01 1.54259458e-01
2.51411766e-01 -9.28866684e-01 -1.63928509e-01 5.04556000e-01
-9.61607397e-01 -1.42832339e+00 -3.68593276e-01 -7.16346979e-01
1.05881786e+00 3.04557770e-01 1.43385434e+00 5.88203192e-01
-1.84676364e-01 3.70826095e-01 -2.51249552e-01 1.87684119e-01
-3.60732079e-01 6.03374779e-01 2.19624816e-03 1.52180910e-01
1.85548469e-01 -1.17955065e+00 -6.68885350e-01 3.62053484e-01
-8.26265097e-01 -9.53015760e-02 8.52279425e-01 1.03278172e+00
3.47214371e-01 1.21340126e-01 5.45090437e-01 -1.36534035e+00
6.69066668e-01 -6.05868697e-01 -6.60203576e-01 2.98478782e-01
-1.09980822e+00 3.12069088e-01 7.10681438e-01 -3.06976140e-01
-5.64068913e-01 -5.26775792e-02 1.60005897e-01 -1.24050073e-01
1.75508231e-01 6.52376115e-01 -7.44490325e-02 -5.95148027e-01
6.91798151e-01 -8.28229412e-02 -2.99818944e-02 -2.85968810e-01
3.28187555e-01 2.12228879e-01 3.77072424e-01 -6.09319150e-01
8.39542627e-01 5.49211800e-01 6.14727080e-01 -9.15416539e-01
-2.23599270e-01 -3.75374824e-01 -7.58055151e-01 -1.47360608e-01
3.62915725e-01 -4.94965702e-01 -7.45169044e-01 2.80772388e-01
-6.35554492e-01 -4.11711544e-01 -2.48602405e-01 6.89875754e-03
-4.32693601e-01 5.35326719e-01 -7.97336340e-01 -5.51036537e-01
-2.63072819e-01 -8.67027044e-01 6.62947118e-01 3.00430898e-02
-1.49305752e-02 -1.53777575e+00 1.15304515e-01 -2.72115976e-01
4.23518181e-01 3.65739793e-01 1.26422429e+00 -9.45717812e-01
-6.58220768e-01 -3.80113542e-01 -4.59294766e-01 -1.29989147e-01
5.25923781e-02 1.70059338e-01 -5.14294028e-01 -4.16956604e-01
-5.16729712e-01 -1.17749333e-01 1.09820330e+00 3.55226159e-01
8.81852746e-01 -4.25960660e-01 -5.19186556e-01 6.63217962e-01
1.54645360e+00 -4.01574761e-01 3.42624545e-01 2.49926582e-01
8.49183917e-01 6.75412774e-01 -6.85560405e-02 8.66931975e-02
3.68821293e-01 4.31623876e-01 7.35329866e-01 -1.65992185e-01
-1.42741613e-02 -4.34243560e-01 -9.32682604e-02 7.94628322e-01
-1.62154175e-02 -5.30926287e-01 -1.02688980e+00 5.30784249e-01
-1.94043112e+00 -6.04667902e-01 -2.07102105e-01 2.07082343e+00
5.63403070e-01 3.08084667e-01 1.50451839e-01 1.52671650e-01
9.62162256e-01 5.38156569e-01 -4.64132905e-01 -2.01900825e-01
-2.90746450e-01 -8.10398236e-02 7.41987586e-01 4.44156975e-01
-1.11132812e+00 7.97946870e-01 6.32385349e+00 7.24476099e-01
-9.53991592e-01 -1.99621037e-01 7.17972457e-01 2.62215704e-01
-4.57684040e-01 6.63062856e-02 -5.11634707e-01 2.73833036e-01
8.05792928e-01 2.95824069e-03 4.35045451e-01 8.16253722e-01
-1.56840235e-01 1.19332500e-01 -1.11243808e+00 6.36289835e-01
-2.15223700e-01 -1.44550979e+00 4.49255556e-02 3.55524927e-01
7.47906208e-01 2.99711168e-01 1.84066948e-02 4.68592197e-02
8.32060397e-01 -9.67067897e-01 1.46980524e-01 2.16400981e-01
6.66434109e-01 -5.23530185e-01 7.00987279e-01 1.80140764e-01
-1.47288585e+00 -1.70664936e-01 -3.37943256e-01 1.14380091e-01
-4.07162569e-02 8.40407670e-01 -8.04663360e-01 5.90669811e-01
5.27068555e-01 7.76129961e-01 -8.26172948e-01 9.78052199e-01
-1.60723701e-01 5.08820415e-01 -6.22205138e-01 -1.82620510e-01
4.04230177e-01 -3.36141765e-01 6.69423044e-01 1.13301075e+00
-1.65693294e-02 -3.78629178e-01 1.23323306e-01 6.39948964e-01
-2.90031374e-01 2.70268768e-01 -9.32037771e-01 -6.55591935e-02
5.35102606e-01 1.53526366e+00 -1.29591858e+00 -7.14310780e-02
-2.50880986e-01 4.77663845e-01 7.96389461e-01 4.20550704e-01
-2.46179864e-01 -3.55107456e-01 4.20196325e-01 3.44512314e-01
3.90483141e-01 -1.34053394e-01 -3.01071793e-01 -1.03839052e+00
-6.15843795e-02 -6.02992654e-01 6.82775974e-01 -1.32015675e-01
-1.66193402e+00 7.36037076e-01 -8.77725035e-02 -9.33169007e-01
-2.45721027e-01 -8.01500499e-01 -8.27790916e-01 2.76789665e-01
-1.40698993e+00 -1.01668036e+00 -2.16454595e-01 4.38220471e-01
-1.81129187e-01 -8.48430470e-02 4.66106147e-01 2.01759040e-01
-6.09612584e-01 5.35031199e-01 1.50064334e-01 2.81427652e-01
3.97603333e-01 -1.48484147e+00 5.12780786e-01 6.99750066e-01
3.83530885e-01 7.30105579e-01 6.74935460e-01 -7.25254238e-01
-1.41610193e+00 -9.81657684e-01 5.64549387e-01 -1.78945929e-01
9.03456926e-01 -5.95507920e-01 -9.30586636e-01 5.20577073e-01
-1.18393444e-01 4.98389304e-01 4.70960170e-01 5.59800446e-01
-5.15484035e-01 -2.42583677e-01 -1.02384806e+00 7.87913144e-01
1.29411507e+00 -2.81304568e-01 -1.44603789e-01 3.48476350e-01
3.47677618e-01 8.84310976e-02 -8.13762903e-01 6.74016058e-01
4.73033130e-01 -1.14509702e+00 9.98846591e-01 -6.04085565e-01
1.06080785e-01 -6.85597584e-02 -2.83456966e-02 -1.38999391e+00
-4.32033479e-01 -5.85287213e-01 -2.50833809e-01 9.22159076e-01
6.07997119e-01 -1.00121188e+00 1.20970142e+00 2.93914050e-01
1.67385265e-01 -9.76287544e-01 -9.91686583e-01 -6.95881724e-01
-1.19514816e-01 1.38979580e-03 4.64525044e-01 1.07180917e+00
-7.04727098e-02 6.34453475e-01 -2.79543966e-01 5.02103455e-02
7.85896778e-01 1.24858938e-01 7.15935111e-01 -1.90727460e+00
-1.50415599e-01 -8.09272647e-01 -5.97386301e-01 -8.54610085e-01
3.01422179e-01 -1.13789499e+00 -1.38096407e-01 -1.36331534e+00
1.51706010e-01 -7.56553352e-01 -4.91638243e-01 3.88565361e-01
-5.01155965e-02 2.90592253e-01 -1.17936589e-01 2.63112217e-01
-6.60697401e-01 5.54797530e-01 1.02980173e+00 -8.58918652e-02
-3.58152568e-01 9.80776548e-02 -7.09754288e-01 6.48054779e-01
7.38189816e-01 -5.34263611e-01 -4.88481909e-01 -5.79549707e-02
6.17414773e-01 1.21813715e-02 4.56845731e-01 -8.88416350e-01
3.63253236e-01 5.35906339e-03 3.13736856e-01 -3.65375549e-01
1.76928923e-01 -8.52139890e-01 -1.24651911e-02 3.70113492e-01
-4.39749360e-01 3.11367035e-01 -3.49957079e-01 1.12333691e+00
-1.07701579e-02 -8.44955817e-02 7.76606619e-01 -1.70846090e-01
-4.17410254e-01 5.84221601e-01 -2.80622870e-01 2.34893374e-02
7.83321977e-01 -2.61331230e-01 -5.05073786e-01 -5.45537651e-01
-5.00592232e-01 3.98699671e-01 5.71143985e-01 1.44661620e-01
3.63367110e-01 -1.20856345e+00 -5.22465408e-01 3.59593958e-01
1.04198068e-01 -2.63470531e-01 9.26901996e-02 9.01442230e-01
-4.82043654e-01 2.42194057e-01 -1.88605607e-01 -6.83051407e-01
-1.04836845e+00 6.75570190e-01 3.42693508e-01 -7.97547221e-01
-8.68964076e-01 5.66574097e-01 2.48789325e-01 -6.53416455e-01
1.70104012e-01 1.58332244e-01 -3.04672867e-01 6.17434084e-03
-3.78625281e-02 3.25341016e-01 1.28943950e-01 -4.89620686e-01
-1.94840074e-01 4.20147836e-01 -2.25148737e-01 2.13911891e-01
1.31240320e+00 -6.41870946e-02 -1.91676199e-01 2.64998764e-01
1.09238124e+00 -4.09579128e-02 -1.07515585e+00 -3.92511606e-01
2.27476612e-01 -2.01449364e-01 -4.55787033e-02 -4.06429499e-01
-1.38214672e+00 7.18902707e-01 2.00464830e-01 7.90577292e-01
8.88155580e-01 2.35523307e-03 4.05895114e-01 6.83700979e-01
2.45326996e-01 -1.01458108e+00 1.23421684e-01 4.46469188e-01
5.24997056e-01 -1.12351465e+00 1.98188230e-01 -5.90193152e-01
-2.73512483e-01 1.07757211e+00 5.99274933e-01 -4.92860883e-01
8.99971724e-01 1.44192606e-01 -1.90442979e-01 -6.27523661e-01
-8.04208398e-01 -2.01243415e-01 2.61710584e-01 5.35703659e-01
-6.54506981e-02 1.08727939e-01 -2.49731094e-01 1.19077712e-01
-7.47586340e-02 -6.34875178e-01 4.17825729e-01 6.61548138e-01
-4.14476216e-01 -1.01499212e+00 3.74947377e-02 8.40160966e-01
-3.10813576e-01 -1.07425265e-01 -7.36162782e-01 1.16508758e+00
-4.95514870e-01 4.84384835e-01 -4.89674471e-02 -4.42925543e-01
-1.58732846e-01 -1.94082990e-01 1.93840504e-01 -3.46085012e-01
-5.01794279e-01 -1.82550207e-01 1.94237664e-01 -3.32158983e-01
-2.19065100e-01 -2.34550580e-01 -8.93677592e-01 -4.41600472e-01
-4.48800445e-01 3.63947481e-01 4.82124954e-01 7.27671385e-01
3.66784126e-01 3.56622279e-01 6.56613052e-01 -8.09482694e-01
-5.20509720e-01 -7.74917841e-01 -7.65771031e-01 2.63102263e-01
2.98980206e-01 -8.76004875e-01 -6.59630895e-01 -6.54947281e-01] | [6.946630954742432, 5.992105484008789] |
c0e131d2-0735-4ff6-8867-53a4a18038a9 | a-data-augmentation-method-and-the-embedding | 2303.12801 | null | https://arxiv.org/abs/2303.12801v1 | https://arxiv.org/pdf/2303.12801v1.pdf | A Data Augmentation Method and the Embedding Mechanism for Detection and Classification of Pulmonary Nodules on Small Samples | Detection of pulmonary nodules by CT is used for screening lung cancer in early stages.omputer aided diagnosis (CAD) based on deep-learning method can identify the suspected areas of pulmonary nodules in CT images, thus improving the accuracy and efficiency of CT diagnosis. The accuracy and robustness of deep learning models. Method:In this paper, we explore (1) the data augmentation method based on the generation model and (2) the model structure improvement method based on the embedding mechanism. Two strategies have been introduced in this study: a new data augmentation method and a embedding mechanism. In the augmentation method, a 3D pixel-level statistics algorithm is proposed to generate pulmonary nodule and by combing the faked pulmonary nodule and healthy lung, we generate new pulmonary nodule samples. The embedding mechanism are designed to better understand the meaning of pixels of the pulmonary nodule samples by introducing hidden variables. Result: The result of the 3DVNET model with the augmentation method for pulmonary nodule detection shows that the proposed data augmentation method outperforms the method based on generative adversarial network (GAN) framework, training accuracy improved by 1.5%, and with embedding mechanism for pulmonary nodules classification shows that the embedding mechanism improves the accuracy and robustness for the classification of pulmonary nodules obviously, the model training accuracy is close to 1 and the model testing F1-score is 0.90.Conclusion:he proposed data augmentation method and embedding mechanism are beneficial to improve the accuracy and robustness of the model, and can be further applied in other common diagnostic imaging tasks. | ['Chi-Chun Zhou', 'Bin Wang', 'Si-Jing Li', 'Xin-Hui Li', 'Chen-Xin Qin', 'Yue-Jie Hou', 'Yang Liu'] | 2023-03-02 | null | null | null | null | ['pulmonary-nodules-classification'] | ['medical'] | [ 1.81508482e-01 5.74228227e-01 -1.53347194e-01 3.04521263e-01
-2.54393369e-01 -1.53487056e-01 4.71549749e-01 -4.82414395e-01
-1.00908414e-01 4.51271057e-01 2.68696934e-01 -4.49507922e-01
1.10136732e-01 -1.28808987e+00 -4.19778556e-01 -1.09656107e+00
1.91384405e-01 5.15149474e-01 5.13045430e-01 1.52586818e-01
-3.01726639e-01 7.02236831e-01 -1.04175520e+00 4.34533715e-01
7.38608539e-01 9.60956454e-01 1.74795568e-01 7.19236732e-01
-2.65070885e-01 1.02973330e+00 -4.88856286e-01 -1.07956585e-02
4.48413283e-01 -8.46662164e-01 -6.50497913e-01 1.46984249e-01
-2.82776594e-01 -5.57180285e-01 -5.03137648e-01 7.98292756e-01
6.84052110e-01 -3.12883705e-01 8.35668683e-01 -9.23367083e-01
-5.81651449e-01 2.27197990e-01 -4.74162489e-01 2.60615855e-01
-3.79594356e-01 3.83763492e-01 2.47823104e-01 -5.66454530e-01
3.54924828e-01 9.33034658e-01 7.91327953e-01 7.58299530e-01
-7.82172978e-01 -5.45307755e-01 -5.82875967e-01 7.26907775e-02
-1.07679343e+00 2.92647719e-01 5.51835418e-01 -5.63489735e-01
3.15115392e-01 4.36315507e-01 1.15728652e+00 9.00282502e-01
4.79726911e-01 6.20582759e-01 1.05124342e+00 -3.94865066e-01
-1.42624125e-01 4.68785763e-01 -5.37230849e-01 1.03445923e+00
5.14457345e-01 5.48402369e-01 4.84273583e-01 -8.15838426e-02
1.48549354e+00 3.18318933e-01 -6.72833622e-02 -6.38974980e-02
-1.20818067e+00 1.09062791e+00 1.05145776e+00 8.33811522e-01
-4.77793366e-01 3.40265751e-01 2.53471583e-01 -1.49480730e-01
2.68109590e-01 3.70425284e-01 2.90125329e-02 3.73289973e-01
-6.83432162e-01 -1.90828249e-01 5.62513173e-01 4.19939935e-01
2.53375620e-01 5.45830607e-01 -6.63446069e-01 6.79239333e-01
4.41227198e-01 6.17387950e-01 1.17694378e+00 -4.62958097e-01
5.04024047e-03 9.00444984e-01 -3.51706564e-01 -7.09523022e-01
-2.62837440e-01 -7.02723265e-01 -1.28661597e+00 2.54331082e-01
-9.86350030e-02 -2.44097561e-01 -1.36461782e+00 1.23059118e+00
3.76908153e-01 4.59580868e-01 1.12642862e-01 7.05522239e-01
1.37090516e+00 7.27216959e-01 3.78707498e-02 -2.88063958e-02
1.46844268e+00 -1.26502502e+00 -7.75770843e-01 4.81898822e-02
7.50972748e-01 -6.24386072e-01 8.12702119e-01 -1.40948653e-01
-9.87452507e-01 -8.26871216e-01 -9.16502833e-01 3.01782638e-01
-6.16722554e-02 6.04329228e-01 3.40292841e-01 9.21404243e-01
-8.40997934e-01 3.02436888e-01 -1.01239872e+00 -2.95863479e-01
6.14342809e-01 4.49272245e-01 -2.00959072e-01 -1.85424965e-02
-1.09307563e+00 8.25172365e-01 5.77254474e-01 -8.64325613e-02
-1.06011784e+00 -6.71225011e-01 -4.57251847e-01 1.76978886e-01
9.77093279e-02 -1.13292086e+00 1.12306964e+00 -1.01550710e+00
-1.35118055e+00 7.04498291e-01 1.23039469e-01 -4.79719758e-01
6.75892234e-01 2.89948016e-01 -1.65065199e-01 3.22082847e-01
4.47658412e-02 5.70202529e-01 7.71072268e-01 -1.05657530e+00
-5.72280467e-01 -9.06121507e-02 -5.14783561e-01 1.68639317e-01
-2.76421726e-01 -4.64842051e-01 -2.46338844e-01 -7.76404560e-01
6.85253665e-02 -9.60046291e-01 -4.51471925e-01 2.76195556e-01
-2.72059113e-01 1.32123649e-01 1.22615194e+00 -9.97843444e-01
9.63124812e-01 -2.05466080e+00 -3.71561438e-01 4.80822384e-01
4.50839579e-01 6.64208114e-01 1.42967626e-01 -1.13375410e-01
-1.52406678e-01 7.16403902e-01 -2.91978389e-01 4.48979795e-01
-5.65794885e-01 3.01095307e-01 3.73790115e-01 -2.30370671e-03
3.86552036e-01 1.48360610e+00 -4.80743408e-01 -7.93303609e-01
4.21889454e-01 5.47391593e-01 -4.45071161e-01 4.26465541e-01
8.38945732e-02 6.37339950e-01 -6.02653563e-01 4.59589928e-01
5.56607127e-01 -4.65757787e-01 -1.08928852e-01 -3.59108269e-01
2.52651840e-01 -1.13243453e-01 -8.79820764e-01 7.72543192e-01
-4.50704694e-01 2.67424136e-01 -3.62507313e-01 -5.45313895e-01
1.05463696e+00 8.67279768e-01 6.08826220e-01 -4.33178663e-01
3.49412143e-01 2.68969178e-01 5.57118237e-01 -8.71790290e-01
-3.59150767e-01 -3.60292673e-01 6.90941453e-01 4.86973286e-01
-1.93596303e-01 -3.03780973e-01 -3.09670985e-01 -8.24187398e-02
1.12306511e+00 -2.70888984e-01 4.93079394e-01 6.49061427e-02
7.27084517e-01 1.07013553e-01 2.02150300e-01 4.80894089e-01
-4.19446640e-02 5.04125237e-01 4.85245258e-01 -3.37430090e-01
-1.31740010e+00 -7.90362895e-01 -1.46813825e-01 4.01130226e-03
-1.85344800e-01 4.19228852e-01 -5.70016623e-01 -9.76674497e-01
-1.91002995e-01 3.64973664e-01 -9.62224424e-01 -3.97872895e-01
-5.35904765e-01 -9.90624547e-01 6.74802601e-01 8.94470632e-01
1.27177382e+00 -1.21664989e+00 -2.79788285e-01 1.39447421e-01
-2.72452414e-01 -7.15568006e-01 -5.15380427e-02 -1.51358992e-01
-1.25941300e+00 -1.03455484e+00 -1.02901459e+00 -1.03279924e+00
9.48793173e-01 1.53229218e-02 6.55191481e-01 5.79296768e-01
-4.37112451e-01 2.00257555e-01 -3.83256048e-01 -5.19769132e-01
-1.06688929e+00 -1.13820970e-01 -5.45135736e-01 -1.58891618e-01
-3.30339037e-02 -1.79485425e-01 -7.43814826e-01 3.64604533e-01
-1.09405911e+00 1.52964652e-01 1.43675637e+00 1.12888801e+00
8.03863704e-01 3.39981288e-01 3.35904002e-01 -1.17480695e+00
3.87738317e-01 -5.52359581e-01 -2.06963241e-01 -1.07116811e-02
-6.44727588e-01 -1.38696656e-01 4.49027866e-01 -3.99080575e-01
-1.10726118e+00 -4.75388058e-02 -5.13222218e-01 -4.48963553e-01
-5.03346883e-02 2.75020301e-01 2.63365209e-02 -3.28200907e-01
7.82551169e-01 3.64464194e-01 4.06322092e-01 -1.54574290e-01
-8.17192420e-02 4.68238533e-01 1.77048579e-01 6.12381138e-02
1.01659572e+00 4.52448964e-01 3.91036183e-01 -5.68488598e-01
-4.13381279e-01 -2.93728203e-01 -4.58177567e-01 -9.69148651e-02
1.23558748e+00 -6.45466030e-01 -6.91394582e-02 1.63774759e-01
-8.98678839e-01 -2.18994036e-01 -8.25353563e-01 9.38215256e-01
-3.28548253e-01 2.44093373e-01 -7.54345536e-01 -4.43827420e-01
-6.48742080e-01 -1.16574562e+00 6.41031325e-01 2.50205755e-01
1.89977452e-01 -1.19636250e+00 5.18102571e-02 2.16487393e-01
7.13853180e-01 5.21070302e-01 1.27025211e+00 -8.69497657e-01
-7.16954172e-01 -4.94697064e-01 -3.71419936e-01 6.03566170e-01
3.35648060e-01 4.76974770e-02 -9.26729083e-01 -5.51970415e-02
5.16120553e-01 1.26093939e-01 7.46257901e-01 6.54483795e-01
1.53856373e+00 -5.22518456e-01 -5.10564864e-01 5.73620975e-01
1.44532502e+00 6.31903589e-01 1.08842385e+00 2.12434500e-01
6.38915002e-01 1.76729441e-01 3.17753971e-01 -7.67452270e-02
-4.56801176e-01 1.18259743e-01 7.51568258e-01 -7.96538413e-01
-7.53143787e-01 -3.10816497e-01 -1.22767381e-01 7.95593262e-01
-4.43770558e-01 -2.74879098e-01 -8.52799475e-01 4.28905606e-01
-1.32951379e+00 -1.05738378e+00 -4.87100244e-01 1.88067186e+00
3.62759620e-01 -2.87909759e-03 -1.29993901e-01 2.85721719e-01
7.19389260e-01 -1.15078494e-01 -4.25685912e-01 -1.39378294e-01
1.65526241e-01 5.25107861e-01 3.98704171e-01 2.22003967e-01
-9.02186573e-01 3.14784855e-01 5.91175413e+00 9.05329108e-01
-1.27874100e+00 4.20323342e-01 9.07279611e-01 4.63121712e-01
-4.13095653e-01 -3.15326601e-01 -3.97230566e-01 4.23025995e-01
4.73705679e-01 -7.42020980e-02 -1.05638482e-01 8.91238749e-01
1.92788512e-01 9.59871858e-02 -7.34096467e-01 6.90219462e-01
5.85039370e-02 -1.40052700e+00 3.68839890e-01 4.70158398e-01
8.46725047e-01 -1.14197627e-01 1.29133806e-01 2.55142748e-01
-1.80808440e-01 -1.20967352e+00 -1.83164418e-01 5.56743920e-01
9.64490175e-01 -4.34747726e-01 1.63483143e+00 3.37357819e-01
-1.11097670e+00 8.27030241e-02 -2.54578233e-01 5.48170924e-01
-2.38946557e-01 4.35167193e-01 -1.71199584e+00 5.64509571e-01
4.65800822e-01 3.80289912e-01 -9.12443459e-01 1.19997728e+00
-3.56500626e-01 1.00964940e+00 -2.16028973e-01 -2.59868383e-01
2.05495134e-01 -4.11397442e-02 3.97404730e-01 8.97734880e-01
7.09326088e-01 2.03879505e-01 -1.49356797e-01 1.15790737e+00
-1.69089269e-02 2.32100248e-01 -7.68229246e-01 -4.15101834e-02
2.83192545e-01 1.33936286e+00 -9.77259040e-01 -5.08594573e-01
-2.55353600e-01 8.58401418e-01 -5.74490190e-01 2.31150519e-02
-1.21464252e+00 -1.21089712e-01 -4.16940629e-01 6.62671864e-01
2.77167290e-01 4.09250200e-01 -3.05454850e-01 -6.91971183e-01
-2.31321692e-01 -6.02498293e-01 2.11851686e-01 -8.46031427e-01
-1.02141476e+00 7.40081489e-01 -3.37564111e-01 -1.41151822e+00
-2.25792706e-01 -7.15430975e-01 -1.04877162e+00 7.48849213e-01
-1.29265118e+00 -1.33985746e+00 -8.25410843e-01 4.38947678e-01
3.44355404e-01 -4.58453208e-01 8.29368472e-01 1.39448822e-01
-2.15949193e-01 3.41265202e-01 1.02967240e-01 3.67104709e-01
2.68417984e-01 -1.07057059e+00 -2.17871331e-02 7.49853373e-01
-2.18188971e-01 4.36204970e-02 -3.05402894e-02 -8.17648888e-01
-6.82442963e-01 -1.50492501e+00 5.03119886e-01 -2.28997394e-01
1.33500695e-01 3.42026800e-01 -9.64869976e-01 6.67760134e-01
1.50861815e-02 1.72976479e-01 7.02330351e-01 -8.25877070e-01
3.98045868e-01 -3.02844830e-02 -1.67364478e+00 3.43894690e-01
3.35799217e-01 2.08892394e-03 -4.19738561e-01 5.37184417e-01
7.02733755e-01 -3.02177280e-01 -1.00140417e+00 8.46730947e-01
4.09182131e-01 -7.54179120e-01 1.03198075e+00 -3.92906904e-01
7.44989038e-01 -3.73934090e-01 2.44663164e-01 -8.75977337e-01
-6.68360591e-01 3.49881828e-01 1.22753077e-03 7.92845309e-01
3.98987353e-01 -7.89259374e-01 1.11053312e+00 2.46249601e-01
-2.25903168e-02 -1.01678491e+00 -6.83542550e-01 -5.63861668e-01
6.77248230e-03 1.94801256e-01 4.98046786e-01 8.67848814e-01
-9.57387388e-01 6.71880990e-02 2.48374119e-02 6.65930882e-02
1.37429759e-01 -4.04550523e-01 6.10802233e-01 -1.11218107e+00
-5.08152008e-01 -3.91874760e-01 -6.19287431e-01 -7.17625260e-01
-6.08060181e-01 -1.02343476e+00 -4.59648579e-01 -1.88269281e+00
3.62194300e-01 -4.82191592e-01 -2.06279263e-01 5.09502113e-01
-2.24400759e-01 3.63914996e-01 9.06630307e-02 4.39854443e-01
3.42361063e-01 4.97675180e-01 1.96029127e+00 -2.15415537e-01
-9.23463926e-02 4.87069935e-01 -4.29761112e-01 6.75956249e-01
9.66285348e-01 -5.28448880e-01 -3.43514979e-01 -1.47129357e-01
-2.86783576e-01 1.55188143e-01 7.92982340e-01 -1.17899120e+00
-2.23266035e-01 -2.76339855e-02 9.22404647e-01 -5.58662772e-01
1.68588370e-01 -1.18122816e+00 4.49227929e-01 1.54696202e+00
9.63569060e-02 -3.36422205e-01 2.85080463e-01 5.16854703e-01
-2.62081683e-01 -5.94695926e-01 1.00413954e+00 -5.76826036e-01
-2.81008750e-01 4.42575842e-01 -5.12537897e-01 -2.59247184e-01
1.42230141e+00 -4.80451256e-01 -6.53574914e-02 -2.91798621e-01
-9.35485125e-01 -2.82256067e-01 3.37201208e-02 -1.35095432e-01
6.60424650e-01 -1.66324699e+00 -7.60874510e-01 4.70202774e-01
-3.14485043e-01 2.30066687e-01 1.48139730e-01 1.11379528e+00
-1.21860218e+00 3.56745839e-01 -2.88072318e-01 -6.61135614e-01
-1.31582224e+00 3.80448341e-01 9.73902702e-01 -8.33974421e-01
-3.87200296e-01 7.77675629e-01 5.01172841e-01 -3.50849986e-01
-2.18925327e-01 -4.12661016e-01 -4.47372377e-01 -5.89482009e-01
9.98975784e-02 4.39441293e-01 1.34928182e-01 -3.01890016e-01
1.32413998e-01 4.85918730e-01 1.05207078e-01 2.30189189e-02
1.03968465e+00 4.28470373e-01 -1.48377657e-01 7.25561678e-02
9.09832895e-01 8.08314458e-02 -5.41907787e-01 4.03454490e-02
-6.56791925e-01 -3.81229311e-01 9.21706110e-02 -1.10826278e+00
-1.42792380e+00 9.77366507e-01 1.21230400e+00 3.33733648e-01
1.14549267e+00 -4.91125584e-02 8.02411854e-01 5.63724115e-02
-3.06749821e-01 -3.10075313e-01 5.21900177e-01 -3.98994870e-02
9.54285741e-01 -1.21560276e+00 3.58284675e-02 -6.35390997e-01
-6.28700435e-01 1.25875604e+00 8.66555631e-01 -2.40982860e-01
7.29507744e-01 2.08461508e-01 6.47857040e-02 -1.53585017e-01
-3.20628583e-01 -1.07891046e-01 3.74799609e-01 5.73637128e-01
4.14515465e-01 1.03103399e-01 -3.97817731e-01 6.79560661e-01
-1.15378596e-01 1.51555583e-01 2.84156829e-01 7.12398231e-01
-6.55379057e-01 -1.02783823e+00 -6.26030147e-01 9.92317557e-01
-4.12739128e-01 -3.84544134e-02 -3.91914785e-01 1.24313819e+00
5.93845069e-01 3.93008113e-01 4.49737087e-02 -2.91491479e-01
1.59686700e-01 -1.39850108e-02 3.80778581e-01 -6.53809667e-01
-7.02892363e-01 2.34104082e-01 -3.16876441e-01 1.42323207e-02
-3.98590833e-01 -2.03199074e-01 -1.21875501e+00 -7.87869766e-02
-7.38909304e-01 1.15994185e-01 5.62941909e-01 6.01030469e-01
4.05388400e-02 1.20261872e+00 7.31657326e-01 -2.52621680e-01
-3.55083197e-01 -1.07909071e+00 -4.57533896e-01 1.63485840e-01
-1.16444044e-02 -3.50306123e-01 -3.89807343e-01 5.53487465e-02] | [15.221870422363281, -2.1408073902130127] |
03b02298-1624-40f6-93d4-bfc5f37c4a1e | out-of-distribution-detection-in-satellite | 2104.05442 | null | https://arxiv.org/abs/2104.05442v1 | https://arxiv.org/pdf/2104.05442v1.pdf | Out-of-distribution detection in satellite image classification | In satellite image analysis, distributional mismatch between the training and test data may arise due to several reasons, including unseen classes in the test data and differences in the geographic area. Deep learning based models may behave in unexpected manner when subjected to test data that has such distributional shifts from the training data, also called out-of-distribution (OOD) examples. Predictive uncertainly analysis is an emerging research topic which has not been explored much in context of satellite image analysis. Towards this, we adopt a Dirichlet Prior Network based model to quantify distributional uncertainty of deep learning models for remote sensing. The approach seeks to maximize the representation gap between the in-domain and OOD examples for a better identification of unknown examples at test time. Experimental results on three exemplary test scenarios show the efficacy of the model in satellite image analysis. | ['Xiao Xiang Zhu', 'Anna Kruspe', 'Sudipan Saha', 'Jakob Gawlikowski'] | 2021-04-09 | null | null | null | null | ['satellite-image-classification'] | ['computer-vision'] | [ 9.78204682e-02 1.29800200e-01 2.81855494e-01 -7.53712118e-01
-7.51933873e-01 -5.41510165e-01 7.75757849e-01 3.29491273e-02
-2.12797880e-01 9.37218368e-01 5.31458333e-02 -4.39145416e-01
-5.77435851e-01 -9.07523990e-01 -6.98983967e-01 -8.78246546e-01
-2.56837130e-01 9.38674927e-01 -1.64987132e-01 2.81023830e-01
2.73364242e-02 4.61894810e-01 -1.48214579e+00 2.27200881e-01
9.23239768e-01 9.00726318e-01 2.65251309e-01 2.39788935e-01
-3.04655194e-01 2.24304676e-01 -8.41178358e-01 -9.10054147e-02
5.62482834e-01 5.22421412e-02 -4.60522383e-01 2.92773157e-01
1.25872970e-01 -3.85108143e-01 -5.54114878e-02 1.53643131e+00
7.40682125e-01 5.50844111e-02 1.25268233e+00 -1.40142155e+00
-7.21049666e-01 7.50155449e-01 -6.55518770e-01 3.63531470e-01
-2.02607036e-01 -1.70379300e-02 5.38631380e-01 -1.04377866e+00
9.58249643e-02 1.40923893e+00 7.38457739e-01 1.19992554e-01
-1.33075762e+00 -6.91413879e-01 -1.38667971e-01 2.17387512e-01
-1.83405828e+00 2.28083599e-02 5.97219825e-01 -7.76954353e-01
4.77598488e-01 -4.60961759e-02 3.00346434e-01 9.18543279e-01
2.31358588e-01 6.58317208e-01 1.25776565e+00 -9.93637294e-02
6.66192234e-01 4.51862276e-01 -1.75917357e-01 -2.35573500e-01
2.22954392e-01 4.33031470e-01 8.03514719e-02 -1.85882464e-01
3.62188101e-01 2.80155502e-02 -3.74274492e-01 -2.28039250e-01
-6.02410376e-01 8.73062968e-01 3.81165385e-01 1.90684944e-01
-5.35384238e-01 -1.69304878e-01 1.32339343e-01 2.55484164e-01
1.00633347e+00 1.28838152e-01 -6.06411695e-01 2.21051738e-01
-1.24260283e+00 2.12855414e-01 6.36487365e-01 7.26735294e-01
8.46598327e-01 3.10040563e-01 -1.31108999e-01 8.09568822e-01
5.92592418e-01 6.99967384e-01 6.62007451e-01 -2.35004976e-01
2.69174129e-01 4.80564028e-01 2.79925406e-01 -1.05610955e+00
-1.08011186e-01 -6.38977230e-01 -9.42591310e-01 2.77117997e-01
3.84992063e-01 -3.94408733e-01 -1.23317218e+00 1.36209679e+00
3.75881135e-01 4.64103431e-01 3.52340668e-01 9.88208175e-01
7.36789346e-01 8.11814189e-01 2.13247418e-01 1.77042872e-01
8.39634836e-01 7.00853094e-02 -3.81119102e-01 -3.61967802e-01
4.81606632e-01 -3.85143012e-01 9.66018140e-01 2.33909473e-01
-3.39578599e-01 -4.48438168e-01 -8.67480934e-01 7.39031494e-01
-5.71666837e-01 -5.61091751e-02 1.65310636e-01 6.57044947e-01
-1.01538503e+00 5.48225999e-01 -4.07536924e-01 -3.22392255e-01
6.34229839e-01 1.50416687e-01 -1.57060146e-01 -1.69689685e-01
-1.52929366e+00 7.47951269e-01 9.25363541e-01 4.27620620e-01
-1.29850030e+00 -9.37251449e-01 -6.99077189e-01 1.94709346e-01
5.29555641e-02 -1.74252883e-01 9.90807235e-01 -1.25790441e+00
-9.64378476e-01 8.72461498e-01 4.01244611e-01 -5.11526108e-01
6.36050284e-01 1.99671969e-01 -6.03486240e-01 -4.33269739e-01
6.94773793e-02 5.52966833e-01 9.94248748e-01 -1.45331275e+00
-5.16600430e-01 -6.06971383e-01 -3.53130579e-01 6.84967190e-02
-8.74603540e-03 -4.77574259e-01 3.61055434e-01 -7.06386864e-01
2.69560456e-01 -7.39013374e-01 -2.87176311e-01 -2.51348883e-01
-3.40952784e-01 -1.99891567e-01 8.55086982e-01 -3.05222511e-01
7.59271801e-01 -2.30012012e+00 -4.95169938e-01 5.90498507e-01
-9.65115279e-02 2.97055036e-01 -8.17554668e-02 2.21736073e-01
-3.54555458e-01 5.33700109e-01 -4.49606240e-01 3.22859079e-01
2.82951921e-01 5.38587570e-01 -7.32972682e-01 6.08785808e-01
5.42234242e-01 3.38742405e-01 -7.63835609e-01 -2.93327242e-01
1.50430873e-02 3.36622685e-01 -1.23423338e-01 1.70917436e-01
-4.12430376e-01 5.23082852e-01 -6.77346647e-01 6.87205613e-01
1.25580609e+00 5.27447835e-02 -6.55352101e-02 -1.36160050e-02
6.49483353e-02 -2.90698916e-01 -1.24043345e+00 9.64360416e-01
-3.69261950e-02 6.84145093e-01 -3.94089103e-01 -1.17492747e+00
1.15324461e+00 1.44848377e-01 1.37991346e-02 -4.46187347e-01
1.99841503e-02 1.25547260e-01 1.62944153e-01 -6.27375305e-01
4.84406084e-01 -6.34271264e-01 7.75817186e-02 2.26610988e-01
-8.55228677e-02 -3.25291693e-01 -2.68703043e-01 -3.59585136e-01
5.12635648e-01 -1.32248492e-03 1.82540089e-01 -5.86981058e-01
4.77282852e-02 4.61617578e-03 7.36470759e-01 9.33398485e-01
-4.13222909e-01 5.22879362e-01 5.88868678e-01 -5.56116879e-01
-1.00528300e+00 -1.18090653e+00 -9.41684783e-01 6.14297509e-01
-9.59065706e-02 4.61941719e-01 -4.62480426e-01 -6.13902450e-01
2.27648303e-01 9.92538154e-01 -6.81302547e-01 2.01913565e-02
2.69741982e-01 -1.22264707e+00 6.89928770e-01 3.70366186e-01
5.64829290e-01 -7.63833880e-01 -3.82359415e-01 1.84228927e-01
1.47914603e-01 -6.08366609e-01 3.69955897e-01 2.55221248e-01
-8.51623237e-01 -8.84293973e-01 -8.49648833e-01 -4.47824508e-01
7.42263675e-01 -1.92856684e-01 1.23594570e+00 -5.22200406e-01
3.08951102e-02 1.95595846e-01 -2.70025969e-01 -8.45808268e-01
-5.12138844e-01 -4.14909542e-01 4.18943912e-02 8.06459486e-02
9.39289689e-01 -6.46973014e-01 -5.63049853e-01 4.25141126e-01
-1.46948218e+00 -7.20532596e-01 6.15657747e-01 7.69692957e-01
7.28700638e-01 7.41177022e-01 8.37395549e-01 -5.88493228e-01
6.13469243e-01 -1.24670219e+00 -7.55258203e-01 2.06133008e-01
-6.44845247e-01 1.15886986e-01 2.48094589e-01 -5.42122781e-01
-1.11908150e+00 -1.63567513e-01 2.94335008e-01 -3.99378210e-01
-6.27062261e-01 1.22488368e+00 -2.47679144e-01 4.87079769e-01
8.12020004e-01 2.86062390e-01 -1.36707604e-01 -3.71421903e-01
-1.12553954e-01 1.12292564e+00 2.35928267e-01 -6.53836012e-01
7.01574147e-01 3.56205940e-01 -1.79552644e-01 -1.01002729e+00
-7.33273208e-01 -1.92220911e-01 -5.64674497e-01 -3.90910864e-01
5.80117404e-01 -1.16005015e+00 7.26666600e-02 4.68390048e-01
-8.58633220e-01 -4.31634098e-01 -2.13020250e-01 7.60258079e-01
-1.64798766e-01 2.01864243e-01 2.36678019e-01 -1.03240848e+00
1.18082069e-01 -1.09776664e+00 1.03012931e+00 3.42275172e-01
-2.96984632e-02 -1.34918106e+00 1.03709280e-01 -2.81888306e-01
4.33845788e-01 4.33469296e-01 1.08316898e+00 -1.29459667e+00
-2.67710596e-01 -3.36185008e-01 -2.33278781e-01 6.05926931e-01
-7.13650361e-02 -3.51631432e-03 -1.19005275e+00 -2.56519765e-01
2.27798253e-01 -3.38183016e-01 7.32378185e-01 6.87289119e-01
1.18629396e+00 -1.44459412e-01 -2.01065451e-01 3.65961105e-01
1.57718372e+00 1.11171402e-01 9.30718839e-01 3.84672105e-01
9.37775299e-02 8.14353287e-01 7.16406047e-01 7.98824728e-01
-6.65766969e-02 2.07006522e-02 7.17464030e-01 1.56314164e-01
5.27839482e-01 -7.40926415e-02 2.06432715e-01 -1.01505248e-02
6.16375923e-01 -5.69872141e-01 -1.55946863e+00 1.00025213e+00
-1.61241007e+00 -9.41780567e-01 -1.35405526e-01 2.37716150e+00
6.80342734e-01 -1.12078838e-01 -4.53705937e-01 -5.36432043e-02
1.14129567e+00 1.80430233e-01 -8.93579900e-01 -3.12342420e-02
-3.78812730e-01 -8.55824426e-02 5.78103185e-01 3.05394650e-01
-1.25565434e+00 6.18503690e-01 6.08744335e+00 1.07217634e+00
-1.07531536e+00 -2.18767300e-01 9.97005224e-01 2.04150781e-01
-6.73696935e-01 -1.52302375e-02 -6.64466679e-01 5.82155645e-01
8.62897933e-01 -3.23951602e-01 -1.77389190e-01 9.77363467e-01
4.24366683e-01 -3.13524812e-01 -1.20091665e+00 8.60762417e-01
-8.40526223e-02 -8.16914380e-01 3.06495219e-01 2.27029830e-01
8.30808878e-01 4.57701206e-01 4.57403630e-01 4.66495603e-01
4.95503157e-01 -1.12352610e+00 6.04036987e-01 6.80439472e-01
5.99481702e-01 -7.16847599e-01 1.13753808e+00 6.31336510e-01
-4.83468056e-01 5.27208000e-02 -8.91183496e-01 -1.07587231e-02
-2.85445243e-01 1.11273432e+00 -1.34113216e+00 2.21159458e-01
9.33526635e-01 5.22238851e-01 -4.25584137e-01 1.17991018e+00
2.04387289e-02 9.13913012e-01 -4.79490846e-01 3.17672011e-03
4.91953373e-01 -3.02598327e-01 8.38976383e-01 8.46195638e-01
7.77756095e-01 2.54315753e-02 2.92664617e-02 1.27797818e+00
4.56462838e-02 -2.29419112e-01 -1.06330693e+00 -1.51619883e-02
6.93850458e-01 7.07442880e-01 -4.38498288e-01 -1.00019790e-01
1.10204443e-01 4.53444660e-01 -2.12053776e-01 6.09871745e-01
-7.79647827e-01 -4.91025187e-02 4.63009119e-01 -4.65500578e-02
3.20208669e-01 1.10521480e-01 -7.69412669e-04 -8.68641496e-01
-7.97947049e-02 -6.78477108e-01 2.84996688e-01 -9.69740331e-01
-1.88089669e+00 6.26564145e-01 3.63216102e-01 -1.61160588e+00
-3.01162958e-01 -6.93227649e-01 -6.82404041e-01 1.11704278e+00
-1.62596071e+00 -7.67000437e-01 -1.31778091e-01 4.37552124e-01
3.10827404e-01 -4.69327450e-01 6.92217350e-01 9.72757190e-02
-1.80127576e-01 2.93544710e-01 8.30954969e-01 8.34275261e-02
5.76656818e-01 -1.28448391e+00 -6.25937209e-02 7.79147387e-01
-2.09043920e-01 2.37202737e-02 1.00610185e+00 -7.04111695e-01
-6.08015597e-01 -1.49743032e+00 5.50667286e-01 -2.62090325e-01
7.51328826e-01 2.57574320e-01 -1.17704809e+00 5.29036760e-01
-4.08334732e-02 3.31247412e-02 1.11090159e+00 -1.49969533e-01
-1.18005060e-01 -1.45094693e-01 -1.52276552e+00 3.52584451e-01
7.27252066e-02 -4.50452060e-01 -7.85864949e-01 3.75758141e-01
1.78627014e-01 -4.84411232e-02 -9.40618873e-01 4.45488721e-01
2.96710759e-01 -8.98373961e-01 5.54106116e-01 -7.40342915e-01
4.25393820e-01 -4.42931563e-01 -7.03980207e-01 -1.74491572e+00
-1.67494044e-01 1.06013440e-01 6.61369681e-01 1.42985415e+00
5.36209166e-01 -6.56783223e-01 6.98362947e-01 8.66550565e-01
3.25153172e-02 -3.30449700e-01 -1.02495420e+00 -7.67120481e-01
4.62166488e-01 -5.57270527e-01 9.95119274e-01 1.13361681e+00
-4.76752460e-01 -2.05556750e-01 -2.84669809e-02 9.46023405e-01
6.45232797e-01 -1.22549310e-01 5.96834004e-01 -1.61905527e+00
1.65166125e-01 -1.70664951e-01 -7.05659568e-01 -5.74952483e-01
4.33729529e-01 -8.17780077e-01 2.58615732e-01 -1.16474617e+00
9.33062285e-02 -6.00607753e-01 -2.45199084e-01 1.89941123e-01
2.49053203e-02 -1.90211032e-02 -1.99042052e-01 3.28920007e-01
-9.45748389e-02 9.04833496e-01 6.75305843e-01 -4.26705241e-01
2.94185150e-02 1.58347160e-01 -4.89517480e-01 9.13986385e-01
8.81358504e-01 -8.98198128e-01 -5.60456693e-01 -6.56602263e-01
2.19693199e-01 -9.79298819e-03 5.64679027e-01 -1.06117427e+00
-5.68199232e-02 -2.14793742e-01 5.42849422e-01 -8.55450630e-01
-2.06176028e-01 -1.20950103e+00 4.53019023e-01 1.62347764e-01
-3.36242348e-01 -1.41593933e-01 4.09759462e-01 1.06084287e+00
-5.71395397e-01 -5.07207036e-01 7.29116440e-01 -7.10174590e-02
-9.87520874e-01 4.75847274e-01 -5.91256499e-01 1.27403289e-01
1.04285014e+00 -5.27703464e-01 -2.22787559e-01 -4.62768316e-01
-8.89936328e-01 3.58277917e-01 4.59497571e-02 2.19112620e-01
6.45713985e-01 -1.33243895e+00 -1.23284721e+00 3.19439858e-01
2.40991473e-01 2.82721311e-01 5.46556711e-01 4.64064866e-01
-4.57843184e-01 1.33465081e-01 -2.22187825e-02 -1.07982171e+00
-6.45411968e-01 7.20834881e-02 6.98390841e-01 -8.93156826e-02
-2.85261832e-02 6.23477519e-01 3.55004072e-01 -7.90858150e-01
1.29847318e-01 -3.76910627e-01 -2.15509191e-01 3.13910097e-01
3.89948845e-01 1.27978504e-01 3.56262550e-02 -5.76361418e-01
-2.64582545e-01 1.70509983e-02 1.94140702e-01 -3.28873396e-02
1.49941194e+00 -3.24232250e-01 -6.86830804e-02 6.74453259e-01
1.01416874e+00 -5.22652507e-01 -1.44382584e+00 -3.99314970e-01
2.92667747e-01 -6.85768604e-01 3.62681299e-01 -9.47731674e-01
-8.22428882e-01 9.18640435e-01 1.09947026e+00 5.12134790e-01
8.64264309e-01 -6.93607181e-02 7.95469247e-03 3.65240097e-01
1.29073225e-02 -1.14866281e+00 -3.45403463e-01 4.91239786e-01
1.12932765e+00 -1.52857625e+00 -3.00570816e-01 4.73087221e-01
-5.20157039e-01 1.01908708e+00 4.19008911e-01 -1.11390855e-02
1.23241651e+00 2.07040817e-01 1.26367182e-01 -1.27716631e-01
-4.51886475e-01 -8.04186463e-02 2.33292043e-01 1.09359396e+00
1.58549577e-01 2.80520141e-01 2.39395976e-01 6.36574745e-01
1.70067418e-03 -6.42363280e-02 5.07610261e-01 5.73805511e-01
-5.40471911e-01 -5.46958387e-01 -6.35979712e-01 6.22518241e-01
-5.44610560e-01 -3.40168148e-01 -2.33612835e-01 9.18507755e-01
5.28487191e-02 7.56033361e-01 4.29712534e-01 -2.95676947e-01
1.44053593e-01 2.40121916e-01 -2.06864342e-01 -4.53784168e-01
-2.13090647e-02 -1.47303669e-02 -1.71069726e-01 7.27370754e-02
-4.67297763e-01 -5.41431725e-01 -9.29497957e-01 -1.39008731e-01
-3.96727681e-01 3.03346783e-01 7.15968013e-01 1.09108520e+00
4.22774434e-01 1.50328949e-01 5.75738251e-01 -9.17830288e-01
-1.03403389e+00 -1.15113509e+00 -1.25863922e+00 2.72090256e-01
2.85412520e-01 -6.30245566e-01 -8.44544351e-01 -1.87807292e-01] | [7.5638604164123535, 3.346796989440918] |
4bc2bb2c-1263-4fbd-b702-1c034e24a986 | learning-medical-image-denoising-with-deep | null | null | https://www.mdpi.com/2227-7390/8/12/2192 | https://www.mdpi.com/2227-7390/8/12/2192/pdf | Learning Medical Image Denoising with Deep Dynamic Residual Attention Network | Image denoising performs a prominent role in medical image analysis. In many cases, it can drastically accelerate the diagnostic process by enhancing the perceptual quality of noisy image samples. However, despite the extensive practicability of medical image denoising, the existing denoising methods illustrate deficiencies in addressing the diverse range of noise appears in the multidisciplinary medical images. This study alleviates such challenging denoising task by learning residual noise from a substantial extent of data samples. Additionally, the proposed method accelerates the learning process by introducing a novel deep network, where the network architecture exploits the feature correlation known as the attention mechanism and combines it with spatially refine residual features. The experimental results illustrate that the proposed method can outperform the existing works by a substantial margin in both quantitative and qualitative comparisons. Also, the proposed method can handle real-world image noise and can improve the performance of different medical image analysis tasks without producing any visually disturbing artefacts. | ['Mithun Biswas', 'Rizwan Ali Naqvi 2', 'S M A Sharif'] | 2020-12-09 | null | null | null | null | ['medical-image-denoising'] | ['computer-vision'] | [ 3.35452586e-01 -2.54411131e-01 3.62067431e-01 -2.35013977e-01
-6.54277563e-01 1.52204130e-02 3.13280761e-01 5.00045307e-02
-5.75654685e-01 6.43980026e-01 3.68538827e-01 5.45447730e-02
-4.11420435e-01 -7.16291964e-01 -2.47405753e-01 -1.26037061e+00
1.98042989e-01 -2.93133616e-01 -5.06084040e-02 -2.69809902e-01
1.46695212e-01 3.56756985e-01 -1.36230350e+00 3.16664070e-01
1.11116886e+00 1.06128538e+00 5.06707430e-01 2.19555616e-01
4.33290154e-02 6.20143175e-01 -8.50170791e-01 -3.00822169e-01
2.10172847e-01 -4.31623280e-01 -3.20885390e-01 2.28440791e-01
1.35025382e-01 -3.44344079e-01 -4.87155288e-01 1.43405151e+00
1.09020662e+00 2.00193658e-01 2.24873006e-01 -5.73826671e-01
-9.05032992e-01 1.54550597e-01 -7.68289268e-01 5.53386509e-01
-1.04348175e-01 3.29831600e-01 3.41231674e-01 -7.10890830e-01
3.85083288e-01 1.22316194e+00 7.97115982e-01 3.98902893e-01
-9.66058075e-01 -6.33250952e-01 -1.21789426e-01 2.91667819e-01
-1.12904680e+00 -2.28596002e-01 1.20940030e+00 -1.25667632e-01
3.78990144e-01 2.86436528e-01 6.22301280e-01 9.69828367e-01
5.29607475e-01 7.32659578e-01 1.29189372e+00 -3.60681973e-02
4.16818149e-02 -2.84269571e-01 3.43599282e-02 5.47164738e-01
3.09720457e-01 7.40108788e-02 -2.17350677e-01 6.31435513e-02
8.56305897e-01 4.03383702e-01 -6.31652594e-01 2.47693241e-01
-1.24491310e+00 6.26826167e-01 7.08738446e-01 6.36055470e-01
-7.13803291e-01 -3.13977227e-02 7.06423342e-01 2.76955962e-01
6.43787861e-01 2.40231603e-01 -1.17631428e-01 1.58552364e-01
-9.04638410e-01 -1.62557453e-01 1.22031942e-01 2.69034058e-01
2.95955658e-01 3.85680854e-01 -2.80946761e-01 9.91377592e-01
9.33041871e-02 3.04806024e-01 7.61176705e-01 -1.03042054e+00
2.27632880e-01 4.27674472e-01 -1.95809454e-01 -1.32987034e+00
-4.74907845e-01 -9.54625964e-01 -1.63095498e+00 4.20402348e-01
2.44238839e-01 -6.59673885e-02 -1.07896638e+00 1.50143671e+00
4.02000248e-01 8.87085870e-02 9.44002252e-03 1.07217014e+00
1.25054121e+00 5.07141531e-01 2.90332884e-01 -5.08837342e-01
1.35472858e+00 -8.84116352e-01 -1.45561743e+00 -1.61203772e-01
1.51613712e-01 -8.86828065e-01 9.64256346e-01 7.33924270e-01
-1.20494485e+00 -9.41026032e-01 -1.04780388e+00 -5.50040044e-02
-6.97921589e-02 1.67617902e-01 7.15962827e-01 6.14894569e-01
-1.07460892e+00 6.84357584e-01 -8.64311397e-01 -2.89762300e-02
8.56664538e-01 2.88877040e-01 -5.14510930e-01 -4.88040298e-01
-1.11521447e+00 7.63863385e-01 9.54280049e-02 7.65207410e-01
-7.12914109e-01 -8.68964493e-01 -8.14199030e-01 -5.34180030e-02
1.51153848e-01 -8.74170065e-01 7.46375799e-01 -1.07288516e+00
-1.14951217e+00 7.57705867e-01 5.06661758e-02 -2.29577005e-01
7.21073806e-01 -3.13879728e-01 -5.94853342e-01 3.54054660e-01
5.58446720e-02 4.44048524e-01 9.76155400e-01 -1.31176472e+00
-4.07525957e-01 -5.50722003e-01 -3.43234390e-01 1.41294807e-01
-3.73776495e-01 -3.14080298e-01 -5.00480831e-01 -1.20610249e+00
3.49135607e-01 -3.44956577e-01 -5.71183503e-01 2.50198901e-01
-1.51969388e-01 3.40835974e-02 9.14199352e-01 -8.79177034e-01
1.11180282e+00 -2.43373919e+00 9.20840651e-02 2.17833593e-02
5.20930767e-01 5.54247081e-01 -1.82928562e-01 1.93754181e-01
-1.33225292e-01 1.66058660e-01 -4.50355142e-01 -3.22536081e-01
-5.13894081e-01 1.72145024e-01 1.59588918e-01 8.07442546e-01
2.17830855e-02 8.32773805e-01 -9.04522896e-01 -4.87173200e-01
4.86781120e-01 8.99469674e-01 -1.88745975e-01 1.72803715e-01
2.79706389e-01 8.07269573e-01 -4.61164951e-01 5.31054795e-01
1.00161612e+00 -4.75146957e-02 -1.08265176e-01 -5.61503947e-01
2.61697143e-01 -3.66551399e-01 -1.05102718e+00 1.84330046e+00
-2.84373790e-01 4.74863589e-01 5.09961605e-01 -1.42344785e+00
8.17322552e-01 4.06380087e-01 5.70334494e-01 -1.04950082e+00
1.78726360e-01 1.91625684e-01 2.55188733e-01 -1.03971350e+00
7.72930309e-02 -5.25998473e-01 3.86595249e-01 -6.93752915e-02
-1.40423998e-01 1.87921166e-01 3.19046341e-02 -5.40727265e-02
9.70371127e-01 -2.79963017e-01 2.84372985e-01 -2.87860423e-01
6.39301598e-01 -3.36005598e-01 7.75267780e-01 7.40904748e-01
-5.76288402e-01 7.84943104e-01 1.35979429e-01 -5.19572854e-01
-7.86941171e-01 -9.79106843e-01 -2.89647996e-01 4.66291368e-01
3.48724812e-01 7.69680589e-02 -5.53499639e-01 -5.71798861e-01
-1.47425145e-01 1.80658072e-01 -8.71984124e-01 -4.78854209e-01
-5.54828882e-01 -1.16559756e+00 3.62756133e-01 4.88893628e-01
8.67695630e-01 -1.41174102e+00 -4.27999645e-01 2.05926850e-01
-3.23641211e-01 -8.00860882e-01 -3.31414640e-01 6.36343360e-02
-1.15454626e+00 -9.73142207e-01 -9.70793009e-01 -1.00505543e+00
8.37033451e-01 6.55363619e-01 1.04424512e+00 5.74974120e-01
-7.02196360e-01 1.05570838e-01 -2.90407658e-01 -2.98200995e-01
-3.90545309e-01 -3.60273421e-01 -3.59298885e-01 -1.75567959e-02
2.60326058e-01 -6.05552316e-01 -1.12745655e+00 3.80137525e-02
-1.39494908e+00 -6.02493919e-02 8.78180385e-01 1.26029110e+00
6.15329683e-01 7.57397175e-01 9.96258736e-01 -9.55289900e-01
9.02841568e-01 -4.25542474e-01 -1.19926162e-01 -3.12711522e-02
-5.18319070e-01 -1.83935508e-01 6.39657438e-01 -2.62805253e-01
-1.36788607e+00 -2.65450627e-01 -5.04471898e-01 -1.62459359e-01
-3.34008455e-01 5.97788930e-01 -1.02799781e-01 -1.92470159e-02
4.39827651e-01 2.62351155e-01 4.09919053e-01 -5.74535012e-01
-1.87429842e-02 3.84536564e-01 6.39792562e-01 -8.96975324e-02
6.73856616e-01 7.53202021e-01 1.24087594e-01 -7.64568567e-01
-7.13119030e-01 -5.25283635e-01 -2.47308791e-01 -2.29930043e-01
9.02575970e-01 -1.02467346e+00 -6.26802981e-01 5.95412433e-01
-9.22910750e-01 1.99564919e-02 -1.25003546e-01 4.30086315e-01
-1.31665885e-01 8.29283476e-01 -8.75503719e-01 -6.71869695e-01
-5.78639567e-01 -1.33432162e+00 8.67551446e-01 2.31185332e-01
8.68449062e-02 -1.04313135e+00 -3.45877290e-01 4.40266877e-01
5.76963723e-01 4.90574628e-01 1.04175615e+00 -1.78404212e-01
-2.38228470e-01 -1.17107078e-01 -2.58666545e-01 6.77183211e-01
4.80571032e-01 -4.73542005e-01 -9.11141217e-01 -3.46058071e-01
6.93473041e-01 -5.41568287e-02 1.14645970e+00 8.48128080e-01
1.50500929e+00 4.60608713e-02 3.78341472e-04 7.34859824e-01
1.60336173e+00 3.81760985e-01 9.61121261e-01 4.81306314e-01
5.29986262e-01 6.50259435e-01 4.38361079e-01 1.25533760e-01
-1.78704068e-01 1.07597433e-01 7.62763560e-01 -8.69992375e-01
-5.92915356e-01 3.36478442e-01 -1.02562211e-01 8.94294262e-01
-1.55489326e-01 3.25106457e-02 -5.75786054e-01 7.20110238e-01
-1.68445969e+00 -8.21833074e-01 -3.43505859e-01 1.73292255e+00
9.35917020e-01 -4.79354002e-02 -4.62634414e-01 4.73618895e-01
6.95000887e-01 9.36249122e-02 -4.87625033e-01 1.44019350e-01
-3.68615299e-01 2.72011817e-01 2.42596895e-01 2.07213804e-01
-1.14501953e+00 3.39623600e-01 6.41857910e+00 1.10155416e+00
-1.03691709e+00 1.64964885e-01 9.72064197e-01 3.24135795e-02
-2.53907926e-02 -7.81552553e-01 -6.20126836e-02 5.95894754e-01
4.16467845e-01 3.55941057e-01 1.82640031e-01 5.53475320e-01
6.84576035e-01 -2.27900028e-01 -5.95620036e-01 1.24596703e+00
6.22849949e-02 -1.27215648e+00 1.12288192e-01 -1.34213611e-01
7.16406345e-01 -4.22559708e-01 4.77247715e-01 -1.14725437e-02
-1.93645313e-01 -1.30479586e+00 1.84583291e-01 6.71595335e-01
5.47704995e-01 -8.84147048e-01 1.36824334e+00 2.66473413e-01
-6.43942177e-01 -1.97913006e-01 -3.53527039e-01 2.48747058e-02
1.46560669e-01 1.07033181e+00 -9.31318104e-02 7.42602170e-01
9.65247214e-01 8.39504719e-01 -7.01862395e-01 1.21437967e+00
-1.83297768e-01 6.69745982e-01 2.32102662e-01 5.05134106e-01
2.63856411e-01 -4.72655565e-01 4.60566670e-01 1.10638714e+00
3.11688513e-01 2.70093888e-01 -1.13234513e-01 4.43470120e-01
-1.45951957e-02 1.53144477e-02 -4.85734820e-01 3.87325794e-01
-9.90288481e-02 1.46660972e+00 -9.19031560e-01 -3.59740973e-01
-3.69077444e-01 9.92482603e-01 -2.83191979e-01 5.42669296e-01
-5.68240881e-01 -4.49507415e-01 3.49873841e-01 4.77327779e-02
2.08675399e-01 1.75605223e-01 -7.49672830e-01 -7.93298900e-01
1.52693182e-01 -1.19488454e+00 2.75778085e-01 -7.50361025e-01
-1.53442919e+00 6.88931704e-01 -4.98179764e-01 -1.14286184e+00
4.23976928e-01 -4.42330867e-01 -5.37345588e-01 8.20646524e-01
-1.58632874e+00 -9.65616584e-01 -5.33574700e-01 5.98886430e-01
6.85063481e-01 -1.09650388e-01 5.82505345e-01 7.58542717e-01
-6.30124688e-01 2.79845595e-01 3.52642775e-01 3.65556474e-03
8.41802657e-01 -1.22758412e+00 -1.03212066e-01 9.04894173e-01
-3.65420491e-01 7.45308995e-01 7.74480820e-01 -7.15976238e-01
-1.09297061e+00 -9.86516118e-01 3.14758927e-01 1.36772409e-01
3.27129275e-01 1.24029517e-01 -1.18712687e+00 1.16134495e-01
5.94892025e-01 2.96222568e-01 5.11368811e-01 -2.23556861e-01
5.55209480e-02 -4.11034584e-01 -1.48112202e+00 5.95221639e-01
8.32315445e-01 -1.20737031e-01 -5.91157556e-01 2.17328161e-01
5.04967093e-01 -2.53910303e-01 -1.00339639e+00 5.30180693e-01
2.16736138e-01 -1.07730150e+00 1.19873738e+00 -2.89748043e-01
7.20563591e-01 -4.27731007e-01 1.82705477e-01 -1.39803934e+00
-5.56369424e-01 -3.90181065e-01 -2.04817429e-02 1.18131769e+00
-1.14466675e-01 -4.95885044e-01 4.83286917e-01 1.96737453e-01
-4.02705640e-01 -8.94889772e-01 -9.56677616e-01 -2.69642889e-01
-4.59209494e-02 -1.25828624e-01 2.93752491e-01 9.16482270e-01
-5.47357678e-01 -5.74731268e-02 -4.76726949e-01 1.44098952e-01
8.88668001e-01 -2.94377059e-01 2.60803759e-01 -1.03576303e+00
-3.78738460e-03 -4.61823106e-01 -4.59168166e-01 -7.10665882e-01
-2.50501871e-01 -6.78950250e-01 1.56150544e-02 -1.76712191e+00
2.33220831e-01 -1.40088066e-01 -6.08874381e-01 1.44211069e-01
-6.74471200e-01 7.00372159e-01 -8.90877172e-02 1.50619596e-01
-5.06336868e-01 5.74512184e-01 1.80483961e+00 -3.47125053e-01
2.54279971e-01 -1.21497080e-01 -1.12993073e+00 8.60484540e-01
7.33864188e-01 -3.87580425e-01 -5.00277996e-01 -6.39619827e-01
1.46267889e-02 -7.73984715e-02 4.44690019e-01 -1.03370488e+00
1.75109103e-01 2.79422760e-01 9.23766077e-01 -5.61528325e-01
1.53085766e-02 -9.83501911e-01 1.28308892e-01 7.95910954e-01
-2.18305290e-01 -1.44339772e-02 2.39671484e-01 8.58523369e-01
-5.83054781e-01 -1.21170796e-01 1.10778737e+00 -3.02143633e-01
-6.85673118e-01 5.93500882e-02 -3.63583297e-01 -1.04135029e-01
9.07390893e-01 -3.90550137e-01 -1.32420525e-01 -2.67009854e-01
-7.93649197e-01 1.24675166e-02 3.31571288e-02 2.00776532e-01
8.24033558e-01 -1.20596588e+00 -6.37073755e-01 2.50295043e-01
-2.79289216e-01 1.35248750e-01 8.92847419e-01 1.17066431e+00
-5.27526259e-01 -9.88216996e-02 -1.97177097e-01 -7.03727126e-01
-1.19730163e+00 6.44244254e-01 3.85242790e-01 -4.13334191e-01
-1.08452749e+00 7.27513075e-01 5.03644586e-01 -2.08946764e-02
3.65648806e-01 -1.09631978e-01 -3.28123808e-01 -1.25071093e-01
7.12134421e-01 4.89098847e-01 3.07034969e-01 -4.96316016e-01
-8.35710019e-02 5.42307258e-01 -1.83304846e-01 3.51015896e-01
1.68491745e+00 -3.03815991e-01 -2.52438933e-01 1.65946931e-01
1.08064985e+00 -9.93071645e-02 -1.46052670e+00 -1.89325228e-01
-3.96742254e-01 -5.39212048e-01 4.77569163e-01 -1.02909386e+00
-1.75636208e+00 8.81274879e-01 1.13226843e+00 1.20529532e-01
1.74460042e+00 -3.72071862e-01 1.03292763e+00 2.42736153e-02
-6.19393587e-02 -1.05834556e+00 3.61758769e-01 -1.87243998e-01
8.26118469e-01 -1.65504980e+00 2.17535123e-01 -3.90510023e-01
-6.05000794e-01 1.03337657e+00 4.19535488e-01 -2.71442205e-01
5.59835970e-01 2.74419039e-01 5.56440651e-01 -4.03992027e-01
-1.28912985e-01 7.79444873e-02 2.81449556e-01 8.29028130e-01
4.70731348e-01 -3.89481425e-01 -7.72964656e-01 7.47583926e-01
2.55511075e-01 -6.58573136e-02 3.09410393e-01 7.10049331e-01
-3.48048061e-01 -7.70323277e-01 -6.42555714e-01 5.20837724e-01
-1.13837540e+00 -1.56442583e-01 1.99293330e-01 7.97933400e-01
4.27715003e-01 1.08554256e+00 -2.23240420e-01 1.58422731e-03
3.31011534e-01 -4.84721005e-01 1.49446517e-01 -1.56534910e-01
-9.09353912e-01 5.34410417e-01 -3.37000370e-01 -4.28292572e-01
-7.05433846e-01 -2.80201942e-01 -1.00103164e+00 -9.94925946e-02
-1.48786277e-01 -2.34254003e-02 5.11826158e-01 8.53850961e-01
2.86664814e-01 1.17278421e+00 4.63634193e-01 -6.70217872e-01
-5.29090643e-01 -1.05728471e+00 -6.91086173e-01 8.86014104e-01
6.37449145e-01 -5.13866782e-01 -4.51808393e-01 2.00997949e-01] | [13.277234077453613, -2.4761669635772705] |
0b2b2b4b-750e-4e01-9270-e81e336e9478 | hanet-hierarchical-alignment-networks-for | 2107.12059 | null | https://arxiv.org/abs/2107.12059v2 | https://arxiv.org/pdf/2107.12059v2.pdf | HANet: Hierarchical Alignment Networks for Video-Text Retrieval | Video-text retrieval is an important yet challenging task in vision-language understanding, which aims to learn a joint embedding space where related video and text instances are close to each other. Most current works simply measure the video-text similarity based on video-level and text-level embeddings. However, the neglect of more fine-grained or local information causes the problem of insufficient representation. Some works exploit the local details by disentangling sentences, but overlook the corresponding videos, causing the asymmetry of video-text representation. To address the above limitations, we propose a Hierarchical Alignment Network (HANet) to align different level representations for video-text matching. Specifically, we first decompose video and text into three semantic levels, namely event (video and text), action (motion and verb), and entity (appearance and noun). Based on these, we naturally construct hierarchical representations in the individual-local-global manner, where the individual level focuses on the alignment between frame and word, local level focuses on the alignment between video clip and textual context, and global level focuses on the alignment between the whole video and text. Different level alignments capture fine-to-coarse correlations between video and text, as well as take the advantage of the complementary information among three semantic levels. Besides, our HANet is also richly interpretable by explicitly learning key semantic concepts. Extensive experiments on two public datasets, namely MSR-VTT and VATEX, show the proposed HANet outperforms other state-of-the-art methods, which demonstrates the effectiveness of hierarchical representation and alignment. Our code is publicly available. | ['Jing Liu', 'Yiliang Lv', 'Mingqian Tang', 'Xiangteng He', 'Peng Wu'] | 2021-07-26 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [ 1.50267389e-02 -5.32153368e-01 -4.77963030e-01 -4.01862711e-01
-5.07553875e-01 -3.31429183e-01 7.95941949e-01 2.64095634e-01
-3.98467630e-01 6.02701679e-02 8.48501861e-01 2.55225450e-01
-2.21426953e-02 -5.10495484e-01 -4.65481162e-01 -5.94117820e-01
3.26249510e-01 6.60505220e-02 2.47935057e-01 7.61801973e-02
1.24331445e-01 -1.52141288e-01 -1.55392909e+00 6.38595402e-01
4.05563444e-01 1.12789547e+00 3.41762990e-01 3.39539170e-01
-4.68753457e-01 8.69739950e-01 -1.61559895e-01 -2.56115556e-01
9.14331600e-02 -5.56785643e-01 -4.37842995e-01 1.79070577e-01
7.97825515e-01 -5.35013199e-01 -9.99866784e-01 1.23190272e+00
3.64377707e-01 6.15364499e-02 5.04168153e-01 -1.53180170e+00
-1.00723171e+00 4.59620178e-01 -7.02074647e-01 1.97303817e-01
5.06327689e-01 -4.91118897e-03 1.51277113e+00 -1.02535903e+00
5.92208505e-01 1.55234838e+00 4.33441430e-01 2.97121614e-01
-8.49381506e-01 -7.00980961e-01 5.57610154e-01 6.09727979e-01
-1.52852571e+00 -3.48623782e-01 6.57272816e-01 -6.50392056e-01
7.70883501e-01 1.05964988e-01 7.85023570e-01 1.32447529e+00
1.62782282e-01 1.22685754e+00 7.02684999e-01 -1.16399541e-01
-2.22373396e-01 -5.77202179e-02 2.27942809e-01 7.58774936e-01
3.08509678e-01 -1.47774205e-01 -7.53262579e-01 1.05622008e-01
7.41310060e-01 4.64821279e-01 -4.91565824e-01 -4.68111664e-01
-1.65776277e+00 7.68039167e-01 2.20306814e-01 6.03885472e-01
-1.77138954e-01 1.77768081e-01 7.03909278e-01 1.51597738e-01
3.24616224e-01 -5.15965894e-02 -2.50168771e-01 -1.89540535e-01
-9.90754664e-01 8.27990547e-02 4.09872323e-01 1.11925292e+00
8.08353186e-01 -9.07100812e-02 -5.46086848e-01 7.68756330e-01
6.57800496e-01 5.42060077e-01 8.04354250e-01 -6.06149375e-01
7.90662050e-01 6.89154267e-01 -3.08528870e-01 -1.53098249e+00
2.44308021e-02 2.65660957e-02 -8.37723970e-01 -4.64599520e-01
-1.07873064e-02 2.78882116e-01 -8.49529922e-01 1.74838424e+00
1.33377969e-01 5.27420223e-01 -6.69956356e-02 9.23270106e-01
1.34474289e+00 8.15619171e-01 1.40586689e-01 -2.32490122e-01
1.67186868e+00 -1.19192708e+00 -1.06601310e+00 -3.32140952e-01
4.69579130e-01 -7.45374084e-01 1.06330764e+00 -2.48960346e-01
-8.51585925e-01 -7.29987264e-01 -9.32995200e-01 -4.50396031e-01
-4.36522216e-01 1.53374985e-01 2.36403033e-01 1.55214086e-01
-8.80417705e-01 2.09574550e-01 -5.53585947e-01 -5.48114777e-01
2.61338502e-01 -1.84404682e-02 -5.98504722e-01 -2.98825115e-01
-1.37170768e+00 4.81366783e-01 4.85202849e-01 -2.01232154e-02
-4.85862464e-01 -5.66987395e-01 -1.22743857e+00 1.18651509e-01
3.91450107e-01 -8.06446195e-01 7.40973830e-01 -1.10928547e+00
-1.15634310e+00 9.54021215e-01 -4.41470236e-01 -1.95547238e-01
4.14171666e-01 -3.36287975e-01 -4.35395777e-01 4.07006383e-01
2.54249573e-01 8.28162730e-01 1.17466545e+00 -8.92947257e-01
-7.30325878e-01 -4.91827518e-01 1.98221996e-01 5.10007143e-01
-8.12899530e-01 9.74006355e-02 -1.06204915e+00 -1.02623951e+00
1.87158123e-01 -6.09704316e-01 3.69529009e-01 3.25039536e-01
-9.89457220e-02 -3.60035777e-01 1.09349656e+00 -8.42077434e-01
1.44723535e+00 -2.42415285e+00 4.92733717e-01 -3.05681854e-01
4.62684810e-01 2.51736678e-02 -3.90797138e-01 4.83193904e-01
-1.95903733e-01 1.32175192e-01 1.61729202e-01 -4.24728394e-01
7.41957203e-02 2.12650180e-01 -2.66773373e-01 5.01948357e-01
-5.84185384e-02 1.14170516e+00 -1.04679370e+00 -9.40290749e-01
3.97294015e-01 5.08910596e-01 -5.48615217e-01 1.74190938e-01
-7.06520304e-02 1.15270533e-01 -7.80944049e-01 6.80456519e-01
3.15606415e-01 -3.17458123e-01 5.82891516e-02 -8.81438017e-01
6.07271120e-02 1.44726992e-01 -1.07675660e+00 1.94010258e+00
-1.73758492e-01 9.81524885e-01 -1.06426403e-01 -1.19084656e+00
5.46203256e-01 4.01710927e-01 7.41646349e-01 -6.84425533e-01
1.56869516e-01 -1.66688144e-01 -3.18131626e-01 -9.07428384e-01
4.87447530e-01 1.45141631e-01 -7.41324108e-03 3.21899742e-01
2.43974090e-01 2.05338091e-01 6.31501824e-02 3.50448608e-01
7.59562135e-01 2.43739456e-01 5.91155171e-01 -9.02174320e-03
6.56839967e-01 -4.17306244e-01 5.22439480e-01 4.81870770e-01
-4.71579254e-01 5.77859104e-01 4.21337754e-01 -3.80541265e-01
-7.83987880e-01 -8.42066824e-01 4.24014851e-02 1.19650674e+00
6.73794448e-01 -9.69893754e-01 -5.25998235e-01 -6.84425473e-01
-1.97434388e-02 3.00331831e-01 -7.77319431e-01 -1.62652135e-01
-5.47012627e-01 -3.19723666e-01 3.45817834e-01 6.10768914e-01
6.69282436e-01 -8.63029242e-01 -8.37537870e-02 -1.35035247e-01
-6.88122988e-01 -1.61671734e+00 -1.07735980e+00 -6.36192441e-01
-6.20016575e-01 -9.49440598e-01 -5.99141300e-01 -9.25761938e-01
4.24621105e-01 8.50476623e-01 9.90746975e-01 2.77022809e-01
-3.27438742e-01 7.60782242e-01 -6.90551460e-01 7.98217654e-02
6.83066398e-02 -3.45223308e-01 3.42302732e-02 4.88229007e-01
7.18569636e-01 -3.65054011e-01 -6.34034395e-01 2.95109749e-01
-1.03779435e+00 2.42383242e-01 4.53816086e-01 8.94657731e-01
6.62370861e-01 -7.93852285e-02 4.07833569e-02 -2.78462708e-01
3.56317639e-01 -5.85561931e-01 -1.81485280e-01 4.65981603e-01
-1.58656120e-01 -2.06174944e-02 3.60815167e-01 -5.77428043e-01
-7.46363640e-01 -8.63333941e-02 1.84551165e-01 -9.13776994e-01
-1.12132087e-01 5.70148766e-01 -4.48711842e-01 1.27019659e-01
-8.49651024e-02 5.82376003e-01 -1.76338330e-01 -1.90788463e-01
4.77992803e-01 6.96422517e-01 3.49392623e-01 -6.29545033e-01
7.58966982e-01 5.41399181e-01 -2.34476686e-01 -9.71695960e-01
-1.00259697e+00 -7.88630486e-01 -6.90739095e-01 -2.13947460e-01
1.26495647e+00 -1.16540992e+00 -4.54842001e-01 3.11283588e-01
-1.10320568e+00 2.41059050e-01 -1.13402873e-01 6.90660894e-01
-4.92422462e-01 8.89584601e-01 -5.59755504e-01 -2.92325228e-01
-1.16459854e-01 -1.23014259e+00 1.44136870e+00 9.12424251e-02
-4.60029803e-02 -1.10248876e+00 -1.81572244e-01 5.03847301e-01
-1.35365836e-02 4.24229726e-02 9.20349121e-01 -6.75718248e-01
-6.17798984e-01 -2.29377940e-01 -5.96005380e-01 2.31191859e-01
3.13291103e-01 1.26228005e-01 -6.80233359e-01 -4.16781664e-01
1.24479786e-01 -3.16175997e-01 9.77135777e-01 3.48806590e-01
1.21533358e+00 -2.85468519e-01 -3.57470840e-01 5.91064811e-01
1.31510401e+00 -1.23001613e-01 6.11653447e-01 3.80824327e-01
1.17973614e+00 6.70487285e-01 6.56217217e-01 2.39998519e-01
7.14354694e-01 8.71882975e-01 2.57760137e-01 1.14769027e-01
-3.09092104e-01 -4.94436592e-01 7.26336539e-01 1.16823590e+00
8.65134746e-02 -2.31433988e-01 -5.67368507e-01 5.24359584e-01
-2.18666101e+00 -1.23685312e+00 4.48323041e-02 2.01927471e+00
6.78807735e-01 -9.41879153e-02 -7.56947398e-02 -2.19510958e-01
8.74123812e-01 6.88058496e-01 -3.31523091e-01 2.53991753e-01
-1.96171433e-01 -3.37510556e-01 1.82102174e-01 2.70021081e-01
-1.25607586e+00 9.58943009e-01 5.21081972e+00 1.07004309e+00
-9.92511988e-01 1.36250913e-01 2.11005971e-01 -2.19451532e-01
-2.23789379e-01 -3.86573598e-02 -7.96533525e-01 6.11179769e-01
4.98706937e-01 -2.75546968e-01 1.71224028e-01 5.80760181e-01
1.73016161e-01 2.98246175e-01 -1.50106525e+00 1.39224052e+00
5.35230100e-01 -1.39053512e+00 6.78080916e-01 -1.27082244e-01
5.10669529e-01 -8.49087536e-02 -1.09105036e-01 3.71645093e-01
-2.73842961e-01 -9.01953697e-01 8.84028494e-01 4.70326513e-01
7.89376259e-01 -3.26179415e-01 6.54514730e-01 8.84398222e-02
-2.01581168e+00 5.65264635e-02 -2.88276166e-01 2.75335580e-01
1.50541544e-01 2.12135404e-01 1.73432566e-02 8.48434925e-01
8.76837492e-01 1.70954156e+00 -5.39335489e-01 6.23976469e-01
-1.01946823e-01 1.80759549e-01 1.96815785e-02 5.64521886e-02
3.91624391e-01 -4.17772740e-01 5.94364703e-01 1.35727382e+00
2.47664288e-01 1.27130270e-01 5.51027179e-01 7.16629624e-01
-1.58643350e-01 2.49826372e-01 -5.79830527e-01 -3.91314209e-01
4.97645855e-01 1.09659898e+00 -4.56393510e-01 -5.27009308e-01
-1.01694930e+00 1.14304173e+00 1.53133184e-01 5.10500968e-01
-8.69262993e-01 -2.23161429e-01 1.00428998e+00 -2.51027763e-01
4.38339949e-01 -2.84133613e-01 8.85754600e-02 -1.67121863e+00
3.50285947e-01 -9.14876938e-01 5.28463185e-01 -8.32477808e-01
-1.43166757e+00 3.51094216e-01 8.56078491e-02 -1.57554245e+00
3.55623178e-02 -6.64494872e-01 -3.63559693e-01 4.75107402e-01
-1.42811823e+00 -1.32736886e+00 -6.02515578e-01 9.90510881e-01
1.09767950e+00 -2.41266102e-01 5.50624549e-01 5.35350502e-01
-6.15030646e-01 7.70229757e-01 6.12187646e-02 4.88531977e-01
8.21292460e-01 -7.65167356e-01 6.65043667e-02 7.47726798e-01
5.54799974e-01 6.20158553e-01 3.90862018e-01 -6.20258451e-01
-1.56836939e+00 -1.12007332e+00 9.08069074e-01 -5.11593521e-01
1.09548056e+00 -3.97429198e-01 -9.62909937e-01 7.59571910e-01
1.64731473e-01 1.03244081e-01 6.04593992e-01 -1.56179950e-01
-7.77335405e-01 -1.67713985e-01 -6.58821642e-01 8.39013398e-01
1.13413894e+00 -1.09745634e+00 -9.61917341e-01 3.78800482e-01
1.13146996e+00 -2.51307040e-01 -8.05568218e-01 3.49838316e-01
7.24725008e-01 -9.53539252e-01 1.12360168e+00 -6.36324763e-01
7.81254053e-01 -3.50383282e-01 -6.28294110e-01 -8.46700728e-01
-3.14854771e-01 -2.80971795e-01 -2.59591401e-01 1.41235399e+00
-2.17117190e-01 -3.34027052e-01 3.50746930e-01 7.11156353e-02
9.30297151e-02 -7.85757065e-01 -8.37375283e-01 -6.66014314e-01
-1.81719452e-01 -4.09065187e-01 4.12570685e-01 1.34308720e+00
-8.01871270e-02 5.70441067e-01 -5.90442240e-01 1.98001206e-01
4.59150404e-01 2.80624092e-01 5.09665906e-01 -1.04730201e+00
-1.04379795e-01 -6.69820666e-01 -8.92360210e-01 -1.40780604e+00
4.58596021e-01 -8.78860295e-01 -7.02865934e-03 -1.41243947e+00
7.16735840e-01 2.20074117e-01 -4.87151921e-01 2.67534167e-01
-3.81727755e-01 9.82337072e-02 2.92895019e-01 5.79088271e-01
-7.57795274e-01 8.25626314e-01 1.23333204e+00 -3.77998680e-01
2.70941436e-01 -6.38144910e-01 -4.57771778e-01 8.52448463e-01
2.76553065e-01 -2.90835232e-01 -4.32693869e-01 -7.59938121e-01
4.95212211e-04 -3.84976044e-02 4.61069584e-01 -7.51076996e-01
3.72249901e-01 -2.31593013e-01 2.70871907e-01 -6.07029617e-01
3.87984127e-01 -9.89231229e-01 -2.56170034e-01 1.76215664e-01
-4.93023545e-01 1.16045758e-01 -2.97871605e-02 1.07819045e+00
-7.16768503e-01 -4.51421179e-02 4.29859847e-01 3.44638713e-02
-1.00372815e+00 8.12728405e-01 -7.66479075e-02 2.37249509e-01
9.43801582e-01 -4.02745396e-01 -3.53913605e-01 -4.58804220e-01
-3.97013694e-01 4.29353207e-01 4.58542705e-01 1.05126762e+00
8.41271996e-01 -1.65718687e+00 -6.81469619e-01 1.57656848e-01
5.77628374e-01 -3.68212968e-01 3.73740852e-01 9.01199043e-01
-1.62912592e-01 5.14798760e-01 -1.41183048e-01 -8.30299973e-01
-1.51418746e+00 6.00064397e-01 2.54883170e-01 -6.50677383e-02
-7.96399355e-01 6.89197600e-01 8.53257239e-01 -2.85977684e-03
4.40347254e-01 -3.10433477e-01 -3.97608370e-01 3.40710938e-01
7.11249113e-01 -2.45107729e-02 -4.18551803e-01 -1.12236679e+00
-3.99127275e-01 1.12619472e+00 -1.46252781e-01 1.76458880e-01
9.05927598e-01 -5.14990807e-01 -3.15668553e-01 6.15900695e-01
1.52757537e+00 -2.51435131e-01 -9.64640915e-01 -6.90328419e-01
-1.74089283e-01 -7.86526144e-01 1.82341173e-01 -6.10448048e-02
-1.13629997e+00 1.09001672e+00 4.19664353e-01 -1.33395987e-02
1.08046114e+00 1.58249468e-01 1.00598168e+00 3.84556860e-01
-1.82184894e-02 -9.56921697e-01 5.02490163e-01 5.43025315e-01
7.75454462e-01 -1.36531126e+00 1.44090727e-01 -3.24224561e-01
-6.61367416e-01 1.16199613e+00 6.66440725e-01 9.73105133e-02
5.99131763e-01 -3.24418694e-01 -1.96589798e-01 -2.16144681e-01
-6.94988847e-01 -2.53812015e-01 7.90774465e-01 3.84883106e-01
4.54888523e-01 -1.00740306e-01 -1.63899809e-01 6.09093428e-01
2.00795904e-01 -2.10547164e-01 1.41054630e-01 8.11721146e-01
-4.22668725e-01 -1.00182223e+00 -7.18503520e-02 2.95253664e-01
-4.16804582e-01 -4.22402024e-01 -4.17198181e-01 9.27570581e-01
1.37660027e-01 7.64247656e-01 3.45824301e-01 -4.36059207e-01
5.45755327e-02 -1.09804235e-01 3.11722785e-01 -5.14579356e-01
-1.19217433e-01 2.42657855e-01 -1.08465344e-01 -8.82744133e-01
-8.01716626e-01 -6.25151396e-01 -1.18007112e+00 -1.97786301e-01
-1.67865194e-02 -1.03477217e-01 2.76108801e-01 1.25001156e+00
3.34929466e-01 4.38660711e-01 4.60664630e-01 -8.03767502e-01
-1.44106671e-01 -7.35128462e-01 -4.89693850e-01 7.16652513e-01
4.66698289e-01 -7.90145695e-01 -4.69553977e-01 3.33431602e-01] | [10.465185165405273, 1.0933457612991333] |
1f7b6179-2c8e-4863-b9a2-f0ea8c3f0191 | evident-a-development-methodology-and-a | 2211.10291 | null | https://arxiv.org/abs/2211.10291v1 | https://arxiv.org/pdf/2211.10291v1.pdf | Evident: a Development Methodology and a Knowledge Base Topology for Data Mining, Machine Learning and General Knowledge Management | Software has been developed for knowledge discovery, prediction and management for over 30 years. However, there are still unresolved pain points when using existing project development and artifact management methodologies. Historically, there has been a lack of applicable methodologies. Further, methodologies that have been applied, such as Agile, have several limitations including scientific unfalsifiability that reduce their applicability. Evident, a development methodology rooted in the philosophy of logical reasoning and EKB, a knowledge base topology, are proposed. Many pain points in data mining, machine learning and general knowledge management are alleviated conceptually. Evident can be extended potentially to accelerate philosophical exploration, science discovery, education as well as knowledge sharing & retention across the globe. EKB offers one solution of storing information as knowledge, a granular level above data. Related topics in computer history, software engineering, database, sensor, philosophy, and project & organization & military managements are also discussed. | ['Samer Haidar', 'Gao', 'Mingwu'] | 2022-11-09 | null | null | null | null | ['general-knowledge', 'logical-reasoning'] | ['miscellaneous', 'reasoning'] | [-3.87506723e-01 2.26876453e-01 -2.35350758e-01 -2.35594109e-01
3.09033394e-01 -1.65413395e-01 2.06174284e-01 4.04788435e-01
1.98691472e-01 8.32414210e-01 -1.73994333e-01 -3.50065142e-01
-9.40107107e-01 -9.19606805e-01 -4.57910955e-01 -7.90505856e-02
-8.47749263e-02 1.37717620e-01 -1.21687494e-01 -3.58383685e-01
1.07635522e+00 4.73562419e-01 -1.72620392e+00 3.17331642e-01
4.98298556e-01 9.06478286e-01 1.07106335e-01 -1.54251948e-01
-5.47561467e-01 1.10685658e+00 -5.87167799e-01 -4.65354711e-01
2.37408560e-02 -3.35887447e-02 -1.19262815e+00 -3.76931190e-01
-1.13336511e-01 2.65872087e-02 2.30747133e-01 6.72586501e-01
1.28681764e-01 -3.26818913e-01 5.19585907e-02 -1.59204197e+00
-8.15211594e-01 6.22173190e-01 -6.61621571e-01 -5.18542267e-02
6.66710079e-01 -3.73708725e-01 2.59772688e-01 -9.61279809e-01
7.80336320e-01 9.45474625e-01 1.08136201e+00 3.07378799e-01
-7.17675269e-01 -9.13367093e-01 -3.26198786e-01 7.18743026e-01
-1.57871461e+00 -1.51368365e-01 5.56199610e-01 -7.71944702e-01
1.37101805e+00 2.38664255e-01 1.16871822e+00 3.40202302e-01
5.51765501e-01 1.48545980e-01 1.15961111e+00 -1.10101330e+00
3.54367435e-01 9.10862565e-01 5.63739181e-01 4.51643288e-01
8.98465633e-01 -1.13400437e-01 -1.49090838e+00 -7.40360394e-02
6.72654033e-01 1.01555668e-01 -6.45535439e-02 -2.67802000e-01
-7.83098161e-01 3.58860970e-01 -5.34562886e-01 5.16388774e-01
-7.16494083e-01 -1.62020192e-01 3.03623945e-01 6.01800919e-01
-1.39612099e-02 6.88368261e-01 -7.45985210e-01 -7.19080865e-01
-8.42610419e-01 2.56124616e-01 1.22051609e+00 1.21117961e+00
8.87535572e-01 3.92537229e-02 4.62411642e-01 2.69116789e-01
7.99888432e-01 -4.17906158e-02 2.54022062e-01 -1.04311526e+00
-2.12738156e-01 1.26416349e+00 2.37994641e-02 -1.17024159e+00
-1.84645757e-01 -1.61754325e-01 -2.06375033e-01 4.82980818e-01
-2.07600713e-01 8.74329135e-02 -4.51720953e-01 1.01883650e+00
4.31519061e-01 -1.52768016e-01 2.49893442e-01 3.10520917e-01
8.08320582e-01 2.34730914e-01 -3.22103575e-02 -4.18522716e-01
1.16411495e+00 -2.87368387e-01 -1.01439214e+00 9.72262248e-02
2.37988144e-01 -8.58313262e-01 5.61293006e-01 1.30798566e+00
-1.40199530e+00 -2.78414726e-01 -1.18142605e+00 2.02745602e-01
-8.94324243e-01 -4.91968811e-01 1.26101732e+00 9.98153925e-01
-1.08631694e+00 6.02120042e-01 -7.12334037e-01 -7.53407836e-01
5.39144516e-01 2.12305978e-01 -7.40767866e-02 -1.16696484e-01
-1.02560377e+00 1.60264134e+00 7.34680951e-01 1.62317231e-01
-4.80906367e-01 -1.05044210e+00 -3.21528018e-01 -1.70529842e-01
1.57578871e-01 -7.65080571e-01 6.64899468e-01 -4.25959468e-01
-1.22827423e+00 5.40191472e-01 3.23030025e-01 -2.72081703e-01
-6.78521842e-02 -3.56953233e-01 -8.31911683e-01 -3.57975870e-01
-1.49832424e-02 -7.09723458e-02 1.13999523e-01 -1.03796375e+00
-8.70548308e-01 -5.27461946e-01 7.06134066e-02 -1.67529389e-01
-5.92450619e-01 3.48293304e-01 3.18252221e-02 -9.97413844e-02
2.67676294e-01 -2.93282330e-01 -7.21527636e-02 -2.10981220e-01
2.59460419e-01 -2.17159256e-01 7.77111292e-01 -7.06136525e-01
1.74076307e+00 -1.79408169e+00 -2.52931267e-01 6.26513660e-01
-5.20643778e-02 -1.45981118e-01 7.48628855e-01 1.26142812e+00
3.67374462e-03 3.25879812e-01 5.40260851e-01 7.46748984e-01
-1.74600352e-02 1.18873008e-01 -1.97385579e-01 1.40894658e-03
-2.12945715e-01 3.10241580e-01 -6.70953989e-01 -5.03513098e-01
3.51894822e-04 4.05071765e-01 -8.00230056e-02 -4.21926111e-01
1.55059723e-02 -1.71260640e-01 -4.53896224e-01 1.16209114e+00
7.12170899e-01 -2.59105653e-01 5.28538525e-01 7.86739215e-02
-8.68987858e-01 1.07868358e-01 -1.43192458e+00 1.85275090e+00
-3.69812757e-01 8.00524533e-01 -1.12682633e-01 -9.61705744e-01
1.29217303e+00 5.49681187e-01 5.21166205e-01 -5.62524199e-01
-3.24587971e-01 2.43536204e-01 -9.99943241e-02 -1.10678184e+00
5.34521461e-01 -2.25260273e-01 3.56588900e-01 3.70060384e-01
-2.66091585e-01 1.11156046e-01 3.93687263e-02 1.74511746e-01
1.21537197e+00 6.22905314e-01 5.36929250e-01 -3.88783365e-01
2.24016443e-01 5.94190657e-01 7.83427358e-01 2.07325399e-01
2.44623065e-01 -2.24433303e-01 1.96445718e-01 -6.56799793e-01
-7.43127227e-01 -7.58121908e-01 -5.39874136e-01 6.78906500e-01
1.76141247e-01 -9.12393153e-01 -3.84142011e-01 1.09419925e-02
2.72887528e-01 8.18690360e-01 -2.63074309e-01 9.06826407e-02
-9.82595491e-04 -3.75282049e-01 3.35921407e-01 3.28967720e-01
3.22526425e-01 -1.11865580e+00 -9.47038889e-01 5.37438691e-01
3.93533468e-01 -5.76270998e-01 1.08125055e+00 -2.42438428e-02
-1.00401223e+00 -1.04647791e+00 3.39291543e-01 -4.64282781e-01
4.78623599e-01 1.40931010e-01 1.17221248e+00 4.25603867e-01
-7.24702597e-01 6.10398531e-01 -7.06484318e-01 -1.28459620e+00
-2.22081542e-01 -6.21466219e-01 8.87405425e-02 -9.00230050e-01
9.62400019e-01 -8.51283371e-01 -4.88923073e-01 -2.73181610e-02
-4.10314590e-01 3.60052995e-02 8.17691445e-01 4.18807983e-01
3.32540780e-01 7.81145334e-01 1.10408223e+00 -7.74717927e-01
5.05505562e-01 -8.42096746e-01 -1.61390424e-01 6.81348205e-01
-1.45547581e+00 -4.24988538e-01 -1.93415254e-01 1.22776218e-01
-1.32573497e+00 -5.94029903e-01 6.20667517e-01 9.57494006e-02
-1.79353818e-01 1.14292264e+00 -6.44205064e-02 -2.07653150e-01
7.84929872e-01 1.39093876e-01 4.17815536e-01 -4.73002225e-01
-3.27810571e-02 1.01884615e+00 -2.11584102e-02 -6.52097583e-01
4.11349177e-01 3.78446579e-01 -1.36090428e-01 -9.80209112e-01
-1.87389582e-01 -3.05869669e-01 -3.49344671e-01 -5.21668851e-01
3.23922783e-01 -7.73192525e-01 -9.10611331e-01 7.19150389e-03
-1.02442038e+00 4.07029688e-01 -2.47249037e-01 6.44892871e-01
-2.06798509e-01 2.39154935e-01 -5.38230054e-02 -1.15169621e+00
-4.99212265e-01 -2.51275450e-01 -5.31308651e-02 5.62510729e-01
-6.03846371e-01 -7.14372575e-01 -4.90676276e-02 8.05520415e-01
5.66117167e-01 7.10055530e-01 1.04721320e+00 -3.73324782e-01
-8.69801223e-01 -1.60227045e-01 2.06170395e-01 3.16133410e-01
1.90877825e-01 5.24103224e-01 -3.39008778e-01 -1.87702291e-02
-1.70941472e-01 -2.78148413e-01 -2.42680222e-01 -1.68206006e-01
9.03068185e-01 -1.97942615e-01 -4.98167813e-01 -1.09238669e-01
1.60619724e+00 8.37393045e-01 9.69363034e-01 1.05624175e+00
-1.78985387e-01 1.02010727e+00 1.13786244e+00 8.97346139e-01
4.25203145e-01 2.02950969e-01 4.86407802e-02 5.48933983e-01
1.90500468e-01 3.07962298e-01 2.28060707e-01 9.28626359e-01
-5.92249215e-01 6.67123854e-01 -1.63544917e+00 7.11269557e-01
-1.90675747e+00 -1.15995133e+00 -4.49467033e-01 1.98139226e+00
1.05475008e+00 4.34461594e-01 -2.52111852e-01 4.11034793e-01
2.89737701e-01 -9.82877314e-01 -2.18945131e-01 -6.61480367e-01
2.64763355e-01 4.89003897e-01 4.19923551e-02 -8.44907016e-02
-1.39111325e-01 6.24172032e-01 6.33203506e+00 4.45427597e-01
-7.26110518e-01 8.58944058e-02 -2.48699293e-01 -1.59470037e-01
-4.10363376e-01 5.44435978e-01 -7.67844677e-01 3.38312179e-01
8.06431293e-01 -7.82203674e-01 2.77447939e-01 1.21831131e+00
2.29407221e-01 -4.26867783e-01 -1.01691318e+00 7.19086349e-01
-5.33673242e-02 -1.97994411e+00 -4.98349033e-02 -9.41683166e-03
7.70189941e-01 -3.95352155e-01 -3.72697711e-01 2.45286301e-01
4.22292389e-02 -1.06928086e+00 6.41482055e-01 1.07551527e+00
2.24616215e-01 -9.53512371e-01 6.77080095e-01 2.46752366e-01
-6.41587198e-01 -4.97072302e-02 -3.53777707e-01 -7.79467285e-01
-9.03664455e-02 4.99210268e-01 -1.04452431e+00 1.11243045e+00
1.45031047e+00 4.81534511e-01 -3.25853795e-01 1.08234286e+00
2.21484274e-01 2.45150775e-01 -7.45705888e-02 -1.83983922e-01
-3.79032582e-01 -2.18278974e-01 5.06619066e-02 1.12794721e+00
2.76847810e-01 1.37290552e-01 -2.72480965e-01 8.41049194e-01
7.72953093e-01 4.28469665e-02 -6.33524537e-01 -2.37060547e-01
1.16399550e+00 1.05303240e+00 -4.65063721e-01 -2.66570784e-02
-7.30286300e-01 -4.29657139e-02 -2.03680024e-01 2.12422431e-01
-4.04363513e-01 -6.42698944e-01 5.45489669e-01 4.39119250e-01
-5.07047214e-02 -7.93894604e-02 -9.34366465e-01 -4.66775328e-01
2.58620977e-01 -1.00853229e+00 1.65632725e-01 -5.88667393e-01
-1.13901401e+00 5.89940837e-03 2.54972786e-01 -1.00857151e+00
-8.93607363e-03 -5.15273452e-01 -4.73780453e-01 8.20936978e-01
-1.05647695e+00 -1.39936626e+00 -5.51576138e-01 3.14122319e-01
4.35753837e-02 -4.90146458e-01 1.02483404e+00 2.17842966e-01
-3.87403816e-01 1.21073410e-01 7.16878399e-02 -4.82931376e-01
4.00235623e-01 -9.81651545e-01 -4.95800257e-01 4.26361948e-01
-3.52873772e-01 1.36939549e+00 6.02834940e-01 -9.92470205e-01
-1.93630552e+00 -3.59631836e-01 7.95829356e-01 -4.93760347e-01
7.31756032e-01 5.87142743e-02 -9.88729715e-01 7.06073105e-01
3.25478792e-01 -6.86646938e-01 1.31320417e+00 5.79769373e-01
3.62378918e-02 -5.22610009e-01 -1.32371616e+00 1.17957935e-01
7.67730296e-01 -2.34216943e-01 -8.80984664e-01 1.70362487e-01
1.71210632e-01 -2.10323073e-02 -1.59379399e+00 1.51303038e-01
1.06213474e+00 -1.08164060e+00 7.62257099e-01 -6.33830428e-01
2.32940108e-01 -3.41485560e-01 7.40206167e-02 -3.12629879e-01
-2.36069307e-01 -7.90498674e-01 -3.23522955e-01 1.68823886e+00
4.86520588e-01 -5.87200582e-01 9.61503208e-01 1.09940898e+00
-1.73453808e-01 -8.18175793e-01 -8.39617252e-01 -8.40886831e-01
-2.35166505e-01 -6.00637615e-01 6.92705274e-01 1.55505574e+00
9.80224431e-01 5.55320717e-02 -7.51540670e-03 -2.80572325e-02
7.12539017e-01 1.67263091e-01 7.22256064e-01 -1.75844193e+00
7.71919033e-04 -1.86350137e-01 -7.96260357e-01 3.29988934e-02
-5.31461298e-01 -4.91218805e-01 -5.56446731e-01 -2.06639290e+00
2.32420057e-01 -5.72817028e-01 -4.15283859e-01 6.10898614e-01
5.55494428e-01 -4.67310190e-01 -9.79876071e-02 3.36870641e-01
-2.93575317e-01 -2.13636234e-01 8.49990845e-01 5.08527935e-01
-2.70711660e-01 -5.82372904e-01 -1.02132511e+00 8.69548380e-01
7.59113908e-01 -4.07876432e-01 -4.87133354e-01 -7.04155564e-02
1.27018583e+00 -8.68439749e-02 3.00865859e-01 -9.84168172e-01
7.33648241e-01 -6.19198322e-01 4.65811223e-01 -7.86592424e-01
1.49664702e-02 -1.09829998e+00 1.19086492e+00 5.68309784e-01
3.59036744e-01 -1.21296272e-01 2.45197877e-01 2.33793169e-01
-2.46786058e-01 -7.01990604e-01 2.27556765e-01 -1.51048660e-01
-1.14806163e+00 -3.49929482e-01 -1.60209805e-01 -4.32559520e-01
1.58067226e+00 -1.07107830e+00 -4.78795439e-01 4.50224251e-01
-6.10760868e-01 3.22416514e-01 4.96930748e-01 3.36694956e-01
8.05445135e-01 -9.52042758e-01 -3.61476153e-01 -2.74896212e-02
-6.20403774e-02 -1.59597751e-02 2.90776938e-01 7.74205387e-01
-8.90446842e-01 3.98710161e-01 -8.01239371e-01 1.73580099e-03
-1.06065571e+00 5.02072990e-01 -1.90226465e-01 3.20074141e-01
-5.74464738e-01 6.03374004e-01 -7.08699167e-01 1.71725526e-01
3.43324751e-01 1.51161849e-01 -4.28096116e-01 1.08472206e-01
6.34469092e-01 1.06807721e+00 -1.16530582e-02 2.81339198e-01
-7.36300528e-01 4.57804650e-01 -2.61877567e-01 8.83210972e-02
1.73233211e+00 -1.28440157e-01 -6.31385326e-01 7.20065892e-01
5.78097776e-02 -5.68386838e-02 -3.27284813e-01 2.75902450e-01
7.35686421e-01 -5.63530862e-01 9.68393236e-02 -1.39927042e+00
-4.26905751e-01 3.84460777e-01 4.86834556e-01 6.26699090e-01
1.05407917e+00 -9.24080983e-02 6.57109618e-02 3.56143713e-01
7.87602901e-01 -1.72954428e+00 -7.60382190e-02 8.35779160e-02
1.00378716e+00 -8.60856771e-01 6.06269062e-01 -4.63510424e-01
-4.19363886e-01 1.56527102e+00 8.32408667e-01 4.71377820e-01
1.00238264e+00 6.76705062e-01 -9.51080024e-02 -7.51829088e-01
-9.75223184e-01 2.92323083e-01 -4.99062508e-01 8.84997904e-01
7.81531811e-01 -2.18517259e-01 -8.97744358e-01 1.00493920e+00
-1.99196771e-01 1.12959373e+00 6.97227418e-01 1.67348957e+00
-5.68237603e-01 -1.24850953e+00 -6.31908059e-01 1.75560474e-01
-7.56326318e-01 -1.96167435e-02 -3.40286136e-01 1.34872508e+00
6.92988098e-01 8.70029271e-01 -1.96774393e-01 -6.02477968e-01
4.61233467e-01 2.42789581e-01 6.01337314e-01 -5.39391637e-01
-6.76755071e-01 -5.64114988e-01 4.63967204e-01 -2.72443801e-01
-6.16516113e-01 -7.96836436e-01 -1.58270049e+00 -7.61022151e-01
-5.71117401e-01 5.75975001e-01 1.25129628e+00 7.67497182e-01
7.62606561e-01 4.44847673e-01 9.74062756e-02 1.95797130e-01
-2.05995277e-01 -8.95372987e-01 -8.07837784e-01 -2.86565095e-01
-5.23181319e-01 -8.26696336e-01 2.15885833e-01 4.04262036e-01] | [8.911514282226562, 6.821671485900879] |
e948f0aa-f962-4c1e-9e31-da89b32f8df1 | 1-d-convolutional-graph-convolutional | 2211.0293 | null | https://arxiv.org/abs/2211.02930v1 | https://arxiv.org/pdf/2211.02930v1.pdf | 1-D Convolutional Graph Convolutional Networks for Fault Detection in Distributed Energy Systems | This paper presents a 1-D convolutional graph neural network for fault detection in microgrids. The combination of 1-D convolutional neural networks (1D-CNN) and graph convolutional networks (GCN) helps extract both spatial-temporal correlations from the voltage measurements in microgrids. The fault detection scheme includes fault event detection, fault type and phase classification, and fault location. There are five neural network model training to handle these tasks. Transfer learning and fine-tuning are applied to reduce training efforts. The combined recurrent graph convolutional neural networks (1D-CGCN) is compared with the traditional ANN structure on the Potsdam 13-bus microgrid dataset. The achievable accuracy of 99.27%, 98.1%, 98.75%, and 95.6% for fault detection, fault type classification, fault phase identification, and fault location respectively. | ['Rob Hovsapian', 'Mayank Panwar', 'Thai-Thanh Nguyen', 'Tuyen Vu', 'Bang L. H. Nguyen'] | 2022-11-05 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-8.23622704e-01 -3.29922974e-01 3.64962161e-01 -6.08367585e-02
-2.03479722e-01 -3.08780283e-01 1.28004953e-01 3.40187699e-02
4.81247634e-01 7.71838784e-01 -2.02713311e-01 -5.82859755e-01
-2.50499457e-01 -1.00213933e+00 -3.95787448e-01 -7.09240675e-01
-9.99289751e-01 2.70916015e-01 -1.58807144e-01 -3.28751385e-01
4.62372266e-02 8.47393513e-01 -9.87060547e-01 2.69419588e-02
7.84377337e-01 1.30351877e+00 -2.13124707e-01 7.90808022e-01
5.09775877e-01 1.11257708e+00 -1.50066423e+00 4.73223746e-01
-5.46196997e-02 -1.34253249e-01 -6.49014413e-01 1.60539195e-01
-3.66924077e-01 -3.61957580e-01 -1.05047965e+00 1.10087049e+00
7.36510038e-01 2.10479349e-01 7.15709329e-01 -1.62482655e+00
-8.18344533e-01 5.56220651e-01 -3.81250262e-01 7.15552747e-01
4.66830045e-01 -2.37092115e-02 3.76379550e-01 -7.56629169e-01
9.27873421e-03 8.89118314e-01 1.11564028e+00 -1.92368418e-01
-5.14695585e-01 -6.13448620e-01 -1.74115300e-01 6.56364024e-01
-1.59766996e+00 3.79383951e-01 8.55213284e-01 -4.33429629e-01
1.97226179e+00 -2.15592206e-01 8.73892546e-01 6.50705338e-01
9.64846492e-01 4.52328324e-01 4.55757052e-01 -2.75523633e-01
3.59614074e-01 -8.37584019e-01 6.83028638e-01 5.21207750e-01
5.54476559e-01 -8.04895908e-02 -3.70066345e-01 8.01242739e-02
6.98860765e-01 2.42232457e-01 -6.21221542e-01 4.34476197e-01
-5.79389513e-01 7.16666341e-01 8.43410730e-01 3.51297736e-01
-4.88405079e-01 2.00778097e-01 8.72433960e-01 7.99241841e-01
6.79221094e-01 2.02440694e-01 -5.37322640e-01 1.04360245e-02
-7.31189489e-01 2.57253367e-03 7.86269963e-01 1.02014041e+00
7.29086399e-01 1.28973734e+00 7.32412413e-02 3.30491662e-01
3.72245967e-01 3.99584919e-01 8.08067560e-01 -5.37648648e-02
1.01961657e-01 1.07958210e+00 -1.24391757e-01 -8.70495915e-01
-1.08643937e+00 -7.91671932e-01 -1.27161515e+00 1.27515614e-01
-3.03678155e-01 -7.42000103e-01 -1.14406848e+00 1.01550174e+00
-2.05849081e-01 3.86531144e-01 2.02315763e-01 6.37655795e-01
8.74536812e-01 8.15463901e-01 -6.90588415e-01 5.50659597e-02
1.05022907e+00 -6.62173092e-01 -1.24122155e+00 -7.37592503e-02
9.78044987e-01 -2.68821418e-01 2.01803878e-01 2.42075890e-01
-8.10262263e-01 -5.66680074e-01 -1.63380158e+00 3.75149429e-01
-8.66411209e-01 4.17334735e-01 3.41406018e-01 4.58560646e-01
-1.26172471e+00 5.59435368e-01 -8.36628795e-01 -3.00062060e-01
1.60263002e-01 8.13858211e-01 -3.62134784e-01 -1.29978815e-02
-1.60969067e+00 9.44854379e-01 5.35773873e-01 6.74658895e-01
-1.08265686e+00 -3.42322797e-01 -1.04807222e+00 8.80231261e-02
-7.30368420e-02 -3.38390172e-01 9.51881826e-01 -4.62346852e-01
-1.06220925e+00 2.90405396e-02 4.95611131e-01 -1.10587037e+00
-1.29578367e-01 -2.35479876e-01 -1.13703775e+00 1.96262419e-01
4.40615602e-02 -4.52666730e-01 6.06018662e-01 -6.27231836e-01
-3.21757227e-01 -3.64943177e-01 -4.70202059e-01 -4.48672241e-03
1.34928152e-01 -2.36031100e-01 2.01896876e-01 -7.42182016e-01
1.48677140e-01 -3.30164254e-01 1.32938847e-01 -8.50813448e-01
-6.45356417e-01 -3.92875254e-01 1.76750338e+00 -9.84864593e-01
1.10522735e+00 -1.93183196e+00 -3.84794116e-01 3.01888704e-01
1.97421849e-01 4.50049490e-01 4.32091653e-01 9.15429473e-01
-4.55346435e-01 -3.40348363e-01 -2.18447164e-01 -1.37681141e-01
-4.13824357e-02 3.25315595e-01 6.98108748e-02 9.92347181e-01
6.87617123e-01 1.14782023e+00 -6.99678063e-01 4.62231725e-01
4.55102205e-01 5.88940322e-01 -7.03445747e-02 5.09217858e-01
1.49242938e-01 -3.77168022e-02 -2.66968876e-01 8.44644070e-01
8.76028836e-01 -3.03036600e-01 2.77719110e-01 -4.68623400e-01
4.89210598e-02 3.16243678e-01 -1.26066315e+00 1.13585806e+00
-5.39784804e-02 9.17670488e-01 -2.12096572e-01 -1.64995921e+00
1.42009628e+00 7.03586400e-01 4.42965001e-01 -8.20517898e-01
4.26071763e-01 6.77364990e-02 -1.16504310e-02 -5.81962526e-01
2.65840981e-02 4.15617108e-01 -5.37342392e-02 5.19393861e-01
8.68611872e-01 2.26269066e-01 1.93931177e-01 6.88540861e-02
1.13642418e+00 -3.74739915e-01 8.59498233e-02 -7.94935584e-01
3.75741452e-01 -1.31634951e-01 6.31858885e-01 3.70161235e-01
-9.56143141e-02 4.35711682e-01 6.27172172e-01 -8.68202269e-01
-3.97603482e-01 -7.17773616e-01 7.51246735e-02 -3.26587906e-04
7.03390092e-02 -3.02958488e-01 -5.78874826e-01 -5.06288409e-01
3.26933622e-01 5.26512086e-01 -8.51773441e-01 -8.07529569e-01
-5.37647665e-01 -1.22298872e+00 6.81709528e-01 9.70223188e-01
8.78339589e-01 -9.37566519e-01 -2.57944107e-01 4.22301292e-01
5.03096700e-01 -7.69476831e-01 -7.61104301e-02 1.08984458e+00
-4.84187603e-01 -1.47941124e+00 -4.67474908e-01 -1.35680485e+00
6.40797317e-01 -5.02644889e-02 1.28526151e+00 1.31876439e-01
-3.67255211e-01 9.84918624e-02 -5.47664702e-01 -1.59011319e-01
8.62276275e-03 -9.58134755e-02 2.99814731e-01 -6.73277602e-02
3.95965159e-01 -8.48057330e-01 -4.15042996e-01 1.48652270e-01
-5.23027003e-01 -6.34473205e-01 3.09836715e-01 1.19985795e+00
1.76367741e-02 6.30482137e-01 1.19260621e+00 -4.25093055e-01
8.60694945e-01 -8.66373122e-01 -7.39424109e-01 3.07826940e-02
-7.51852989e-01 -3.61073941e-01 6.31120145e-01 -2.40685064e-02
-5.13972878e-01 -3.58992577e-01 -1.79202348e-01 -7.58423865e-01
-8.76256749e-02 9.36578214e-01 4.44731303e-03 -3.33428085e-01
4.76400942e-01 7.45448619e-02 8.10794607e-02 -4.03939247e-01
-1.95201010e-01 7.37394691e-01 7.83611953e-01 1.80254340e-01
5.26656091e-01 -2.13337585e-01 -8.15417394e-02 -8.31754267e-01
-9.08205509e-02 -3.47348243e-01 -7.50917912e-01 -4.09029156e-01
7.71280468e-01 -1.34199655e+00 -8.38480055e-01 1.21680450e+00
-9.35348690e-01 -6.47047937e-01 -1.03844933e-01 4.17388558e-01
-8.64630099e-03 2.27691129e-01 -1.27521598e+00 -7.94683218e-01
-7.74725318e-01 -1.09707201e+00 8.01016331e-01 3.26775670e-01
2.58364201e-01 -1.45853102e+00 -2.32745603e-01 -5.06672263e-01
4.96248990e-01 7.62289524e-01 7.84647822e-01 -9.24642682e-01
-2.32625678e-01 -6.97615564e-01 -2.17502803e-01 4.51692402e-01
4.07404900e-01 2.92248651e-02 -6.66034341e-01 -7.52176881e-01
8.53127465e-02 -4.86075953e-02 2.66769141e-01 6.49631202e-01
5.57704687e-01 -1.64967641e-01 -1.22317709e-01 6.34554982e-01
1.65566301e+00 7.93765128e-01 5.81486106e-01 1.15604505e-01
7.62619734e-01 -2.53003448e-01 -6.71096668e-02 5.68291306e-01
5.51203549e-01 1.62184298e-01 6.28558874e-01 -1.18371017e-01
-2.25512341e-01 3.08005035e-01 3.04657400e-01 1.15229571e+00
4.54974473e-01 -6.97697997e-01 -1.30709970e+00 7.21711576e-01
-1.73726058e+00 -3.69037360e-01 -4.71524626e-01 1.40225542e+00
-6.11578599e-02 3.34684074e-01 7.98109993e-02 9.57650423e-01
8.20757926e-01 -7.92478547e-02 -4.78088737e-01 -3.72362256e-01
-6.24393284e-01 1.47552311e-01 5.69072962e-01 3.11151147e-01
-1.17743778e+00 2.88343459e-01 6.05937052e+00 1.39343113e-01
-1.05302930e+00 -2.49193043e-01 4.80717421e-01 5.32695830e-01
4.08530533e-01 -5.31577706e-01 -3.82862210e-01 4.20934051e-01
1.28535581e+00 -2.80632794e-01 2.52912641e-01 5.36018193e-01
4.89812978e-02 -1.05140857e-01 -7.64682055e-01 9.40968573e-01
2.36691460e-01 -1.29857969e+00 -1.02605112e-01 -3.70960742e-01
9.76315618e-01 3.87957722e-01 -3.81788194e-01 2.61989176e-01
4.45199430e-01 -1.03383648e+00 3.73824894e-01 4.18020427e-01
4.84911978e-01 -1.43745697e+00 1.62796056e+00 -2.59801447e-01
-1.44667673e+00 -4.49176133e-01 -5.37151583e-02 -3.00019115e-01
3.21208686e-01 9.15093899e-01 -9.53127205e-01 1.47979808e+00
1.26347423e+00 1.40462410e+00 -5.87754309e-01 1.17127919e+00
-2.06609979e-01 8.06479275e-01 -1.86172761e-02 -1.81231387e-02
4.02268738e-01 -6.09703586e-02 1.59111172e-01 1.04563415e+00
3.44802618e-01 -3.52458209e-01 3.28358263e-01 5.89548826e-01
-2.93806866e-02 -6.72679067e-01 -6.21568203e-01 6.62042648e-02
4.11548197e-01 9.02057111e-01 -6.31245673e-01 -2.05844298e-01
-3.93727452e-01 7.55437195e-01 -6.40817219e-03 7.02649415e-01
-4.43793565e-01 -1.08049560e+00 8.18893075e-01 -5.58551908e-01
2.92236209e-01 -3.70227158e-01 -3.63972306e-01 -8.75762403e-01
-3.31190377e-01 -5.47728062e-01 6.88616097e-01 -1.08729362e+00
-1.58629262e+00 8.56912673e-01 -3.40377420e-01 -1.18876481e+00
-6.40081167e-01 -7.94607282e-01 -1.24984360e+00 1.18308640e+00
-1.45288324e+00 -9.70728993e-01 -4.89975154e-01 8.86836350e-01
3.33497792e-01 -3.74153912e-01 7.63292313e-01 4.39083070e-01
-1.22422719e+00 5.04349291e-01 1.15129299e-01 8.01426232e-01
1.40974134e-01 -1.56112158e+00 8.37399065e-01 1.34002888e+00
-4.91573304e-01 -8.12514126e-02 3.55669409e-01 -9.22114849e-01
-1.96709967e+00 -1.52140892e+00 4.91908759e-01 2.72420257e-01
7.42935121e-01 -3.78195852e-01 -1.19742179e+00 1.10423541e+00
6.28644943e-01 3.35159689e-01 2.88781285e-01 -3.58863473e-01
-3.53175327e-02 -7.67097846e-02 -1.28702950e+00 1.40509427e-01
2.58029431e-01 -7.87736118e-01 -3.57111037e-01 3.75924140e-01
5.00450671e-01 -5.11895716e-01 -1.49846172e+00 5.90318739e-01
-1.54818013e-01 -4.19031203e-01 6.08603299e-01 -2.21774921e-01
-4.01404262e-01 -7.09821045e-01 -1.57648787e-01 -1.70433402e+00
-4.01924610e-01 -6.64722085e-01 -5.77147841e-01 7.78589129e-01
2.86336601e-01 -1.09406722e+00 3.06199461e-01 -3.88386041e-01
-1.04726183e+00 -6.69046581e-01 -8.30294132e-01 -6.11384213e-01
-7.98598230e-02 -1.66095138e-01 9.29600298e-01 1.15691423e+00
2.32880235e-01 3.63817543e-01 -1.15758337e-01 9.92024124e-01
4.09560263e-01 1.94702391e-03 1.79314524e-01 -1.28505659e+00
4.16035354e-01 -2.48219356e-01 -9.75809991e-01 -2.88337886e-01
-4.09252718e-02 -5.34943461e-01 -2.62032121e-01 -1.88249600e+00
-8.94298196e-01 1.19839318e-01 -6.38542712e-01 7.15067387e-01
3.81492674e-02 1.76432505e-01 -6.43245935e-01 -1.10003613e-01
-2.16381595e-01 5.75300336e-01 6.46277964e-01 -3.44488204e-01
1.81507200e-01 -1.78882569e-01 2.04470888e-01 -9.87421349e-02
1.29535186e+00 1.80354893e-01 -3.99918884e-01 -4.18032885e-01
5.79957478e-03 4.38607037e-01 1.39835835e-01 -1.48599470e+00
5.34781396e-01 8.90496850e-01 7.76961744e-01 -1.20604455e+00
-1.93932995e-01 -4.61527288e-01 2.73965567e-01 8.74141932e-01
5.82619488e-01 1.01259720e+00 5.77693105e-01 3.81992549e-01
-6.57407939e-01 3.33900630e-01 4.97425884e-01 9.57637429e-02
-9.32526767e-01 3.00100356e-01 -1.10790265e+00 -4.59822536e-01
9.13418710e-01 1.85619500e-02 -7.04220831e-01 -3.83979380e-01
-7.53733337e-01 5.94341636e-01 -2.98216753e-02 4.87653017e-01
8.81189942e-01 -1.57373357e+00 -6.33784771e-01 9.24311578e-01
-8.05509984e-02 4.53716442e-02 3.09900761e-01 6.40060484e-01
-8.23736966e-01 5.57005882e-01 -5.11212289e-01 -4.85330015e-01
-4.23113018e-01 3.42004776e-01 1.13908815e+00 -4.41407748e-02
-7.51203716e-01 7.00956047e-01 -6.25639915e-01 -1.86129496e-01
2.12375015e-01 -6.65914655e-01 -3.14793974e-01 1.15796082e-01
3.35771173e-01 6.82390928e-01 1.08089507e+00 -3.50777775e-01
-4.83640611e-01 2.59875000e-01 1.68892577e-01 6.75309002e-01
1.49160063e+00 1.05075866e-01 -3.53829533e-01 5.39490283e-01
1.28009582e+00 -1.22976971e+00 -1.05656111e+00 2.32216403e-01
2.37626925e-01 5.77596486e-01 3.80617887e-01 -8.84396732e-01
-1.58825338e+00 7.52260029e-01 7.47879267e-01 1.03418481e+00
1.34344733e+00 -4.95066315e-01 5.25041044e-01 3.47436249e-01
2.82211185e-01 -8.05533051e-01 -1.30593419e-01 8.43969464e-01
7.43369818e-01 -6.84388280e-01 -2.31011420e-01 -6.51187438e-04
1.59946159e-01 1.47242701e+00 5.72175980e-01 -8.54615629e-01
1.14778757e+00 8.39856803e-01 -7.13715702e-02 -8.23717117e-01
-8.57734859e-01 2.23003123e-02 1.06267415e-01 8.86246264e-01
1.76502854e-01 1.43145919e-01 3.30759257e-01 5.53671479e-01
-9.86005366e-02 -3.38707864e-01 6.33483708e-01 1.31999218e+00
-1.64623663e-01 -2.79565006e-01 -2.48507246e-01 8.15530598e-01
-4.28516954e-01 2.34649703e-01 -1.28645927e-01 1.07110858e+00
-1.99546278e-01 1.23500061e+00 6.00643873e-01 -8.12122226e-01
4.34334487e-01 1.02266021e-01 7.04534538e-03 -3.23563695e-01
-9.25656259e-01 -2.05103785e-01 1.13699868e-01 -3.58998984e-01
-1.03564806e-01 -3.17055285e-01 -1.31431413e+00 -1.76924050e-01
-6.81704402e-01 5.76530218e-01 3.97287786e-01 9.30327952e-01
4.10618991e-01 1.40654373e+00 9.39849794e-01 -8.12984586e-01
6.95677325e-02 -1.41598165e+00 -1.22631884e+00 -1.80117637e-01
7.28228450e-01 -7.41818905e-01 -5.28872788e-01 -3.39636862e-01] | [6.277884006500244, 2.502114772796631] |
8da5cb8f-2f7f-4167-81da-083fb769cef6 | co-training-an-unsupervised-constituency | 2110.02283 | null | https://arxiv.org/abs/2110.02283v2 | https://arxiv.org/pdf/2110.02283v2.pdf | Co-training an Unsupervised Constituency Parser with Weak Supervision | We introduce a method for unsupervised parsing that relies on bootstrapping classifiers to identify if a node dominates a specific span in a sentence. There are two types of classifiers, an inside classifier that acts on a span, and an outside classifier that acts on everything outside of a given span. Through self-training and co-training with the two classifiers, we show that the interplay between them helps improve the accuracy of both, and as a result, effectively parse. A seed bootstrapping technique prepares the data to train these classifiers. Our analyses further validate that such an approach in conjunction with weak supervision using prior branching knowledge of a known language (left/right-branching) and minimal heuristics injects strong inductive bias into the parser, achieving 63.1 F$_1$ on the English (PTB) test set. In addition, we show the effectiveness of our architecture by evaluating on treebanks for Chinese (CTB) and Japanese (KTB) and achieve new state-of-the-art results. Our code and pre-trained models are available at https://github.com/Nickil21/weakly-supervised-parsing. | ['Shay B. Cohen', 'Nickil Maveli'] | 2021-10-05 | null | https://aclanthology.org/2022.findings-acl.101 | https://aclanthology.org/2022.findings-acl.101.pdf | findings-acl-2022-5 | ['constituency-grammar-induction'] | ['natural-language-processing'] | [ 1.65969551e-01 7.60899186e-01 -2.98364967e-01 -7.99621105e-01
-1.22258747e+00 -8.84383142e-01 2.43158564e-01 3.60209256e-01
-4.84882683e-01 8.83368850e-01 2.78364927e-01 -7.95499623e-01
4.19278473e-01 -7.63173044e-01 -9.34878469e-01 -3.47058624e-01
8.13295916e-02 5.32474279e-01 4.84080166e-01 -3.58844548e-01
-1.39612451e-01 9.71693397e-02 -1.04475582e+00 7.20486999e-01
6.92571104e-01 6.17316902e-01 1.77514315e-01 8.22970152e-01
-1.81260824e-01 1.00351882e+00 -6.10946417e-01 -8.94976556e-01
1.38528496e-01 -3.42126757e-01 -1.25107408e+00 -3.21157694e-01
2.85878003e-01 -1.67141214e-01 1.59374923e-01 8.00025165e-01
9.90190506e-02 -3.41981798e-01 2.52241790e-01 -7.49439538e-01
-1.58320367e-01 1.48864222e+00 -3.14213008e-01 4.01030451e-01
1.53334647e-01 3.63683961e-02 1.45825994e+00 -4.27415818e-01
7.51318872e-01 9.74263608e-01 9.34490502e-01 7.00152516e-01
-1.28416657e+00 -5.57015955e-01 2.21834108e-01 -4.30641294e-01
-6.26848936e-01 -6.96881652e-01 6.19813561e-01 -2.32995227e-01
1.12274086e+00 6.44857511e-02 3.98476452e-01 1.10250735e+00
-5.99094899e-04 8.72593164e-01 1.27255213e+00 -1.02685928e+00
1.70169145e-01 1.45297155e-01 6.63305819e-01 8.16431642e-01
9.19886827e-02 1.50597110e-01 -5.62131584e-01 -4.27958667e-02
1.88607439e-01 -6.50170982e-01 -4.95555513e-02 7.75097534e-02
-7.07996309e-01 8.63212645e-01 1.88830525e-01 4.08636838e-01
-4.43113446e-02 -5.33154532e-02 4.77566212e-01 2.35048652e-01
3.86289060e-01 4.24677968e-01 -9.82871234e-01 -3.13834459e-01
-8.27837408e-01 -1.94866166e-01 1.14307094e+00 1.05328572e+00
8.01078379e-01 -3.63457114e-01 1.12459667e-01 8.29446316e-01
-4.12951000e-02 2.61922985e-01 2.40891039e-01 -9.00003254e-01
8.32947552e-01 5.70970953e-01 -6.37303218e-02 -3.43903631e-01
-5.00303924e-01 -2.45259702e-01 -1.96939394e-01 1.25627900e-02
9.34956074e-01 -4.01712954e-01 -9.37044442e-01 2.09549809e+00
3.28995615e-01 -3.33695829e-01 1.81395128e-01 4.25727129e-01
4.92259204e-01 4.50945556e-01 4.13385004e-01 -1.88639969e-01
1.49903834e+00 -9.89030242e-01 -4.45662022e-01 -5.57794392e-01
1.15596664e+00 -5.71712554e-01 1.09615338e+00 2.93437123e-01
-1.24879467e+00 -4.38385218e-01 -1.01312292e+00 -1.80944294e-01
-1.87061250e-01 1.99051872e-01 7.20508277e-01 7.57966280e-01
-1.01126826e+00 5.27033448e-01 -1.05889583e+00 -2.26203784e-01
2.29637668e-01 2.48566657e-01 -4.03178513e-01 2.05512773e-02
-1.23012245e+00 8.42924058e-01 3.82721573e-01 -5.52401505e-03
-8.64035785e-01 -3.81313711e-01 -8.07363212e-01 -1.35188913e-02
2.49122232e-01 -1.65479273e-01 1.62298071e+00 -1.04202521e+00
-1.41797698e+00 1.18682039e+00 -2.80477583e-01 -6.02718532e-01
4.47643787e-01 -3.88567984e-01 -1.48145258e-01 2.45812237e-01
3.77345055e-01 7.49410748e-01 4.60249603e-01 -1.13550806e+00
-9.57288384e-01 -3.35478038e-01 3.09137613e-01 -1.19423352e-01
-2.52793431e-01 4.68734056e-01 -3.74703825e-01 -3.00028235e-01
1.05821803e-01 -7.59887576e-01 -1.23723589e-01 -7.44212151e-01
-5.87133586e-01 -3.17389548e-01 2.22910762e-01 -1.06030142e+00
1.30534387e+00 -2.06516147e+00 -1.22763343e-01 -6.38676924e-04
-5.87944463e-02 7.98494667e-02 -1.43732717e-02 4.26546514e-01
-1.47934407e-01 4.27221835e-01 -5.07841825e-01 -3.56467187e-01
-3.11205655e-01 3.67425054e-01 -3.41868043e-01 9.92752686e-02
5.45359612e-01 6.70996189e-01 -9.98012066e-01 -6.30918205e-01
-3.34497303e-01 -1.30660325e-01 -5.34697950e-01 3.10856700e-01
-3.48839581e-01 3.74963164e-01 -2.54632622e-01 6.39901936e-01
3.86944562e-01 -5.13388254e-02 8.69034410e-01 1.42829910e-01
-2.06317797e-01 1.10953641e+00 -8.38139951e-01 1.43662012e+00
-5.59371054e-01 2.41583124e-01 4.49292660e-01 -1.07320321e+00
7.97851384e-01 2.78861731e-01 -1.31141827e-01 -4.42138165e-01
6.97329566e-02 4.28717375e-01 2.75837213e-01 -3.10180724e-01
2.66483516e-01 -1.03708088e-01 -3.76714408e-01 6.35098338e-01
3.63518059e-01 5.24951657e-03 5.52111268e-01 3.61172080e-01
1.55935919e+00 3.70243579e-01 3.94653589e-01 -4.57682163e-01
4.06967610e-01 2.26982147e-01 9.73830938e-01 9.80666339e-01
-2.14044213e-01 3.04050237e-01 1.01682031e+00 -3.54415119e-01
-1.04596674e+00 -8.27615082e-01 -2.02554777e-01 1.49605107e+00
-4.54442173e-01 -4.20681626e-01 -1.06132388e+00 -1.22692025e+00
-2.42081657e-01 9.18493509e-01 -6.96759880e-01 2.16152921e-01
-1.14653432e+00 -5.73203325e-01 8.55173528e-01 9.03224945e-01
1.64924398e-01 -1.20915377e+00 -6.52461827e-01 3.32290560e-01
-2.92728305e-01 -1.29704988e+00 -8.59321281e-02 9.70212996e-01
-8.86687040e-01 -1.12194848e+00 -3.40908673e-03 -1.09542358e+00
6.52378380e-01 -3.07683468e-01 1.47205472e+00 3.65482062e-01
1.92921311e-01 4.24826331e-02 -6.97424531e-01 -3.21894944e-01
-9.57923293e-01 5.79600573e-01 -3.49720746e-01 -5.09520710e-01
3.89024258e-01 -4.70295012e-01 -3.30791995e-02 1.84271097e-01
-2.32768402e-01 2.25163758e-01 5.14081240e-01 9.81818736e-01
2.30936125e-01 -2.69408286e-01 4.13550705e-01 -1.41828871e+00
2.08173588e-01 -4.22314763e-01 -7.15578556e-01 3.84184122e-01
-2.68138111e-01 1.41330123e-01 7.43291497e-01 -1.72766984e-01
-1.10633206e+00 3.23313773e-01 -2.94592589e-01 4.09343928e-01
-3.17825466e-01 4.97402519e-01 -3.19462121e-01 5.56972563e-01
7.18152165e-01 -1.99887142e-01 -1.62618876e-01 -6.80574596e-01
4.39412713e-01 7.49431670e-01 4.89639819e-01 -1.04422593e+00
5.67513704e-01 3.08013290e-01 -3.64232123e-01 -5.09130895e-01
-1.22058415e+00 -2.37683609e-01 -9.40330625e-01 1.10174544e-01
8.41515601e-01 -9.19855952e-01 -3.46960157e-01 3.22863430e-01
-1.16344619e+00 -9.66174245e-01 -1.40512139e-01 2.04326674e-01
-2.55025595e-01 1.80351034e-01 -1.08052886e+00 -8.34828675e-01
-3.23306441e-01 -9.31682587e-01 8.82024109e-01 1.85343578e-01
-3.51739287e-01 -8.75083506e-01 -7.10774660e-02 6.17745221e-01
-7.64324218e-02 -1.09150469e-01 1.10272086e+00 -1.19843769e+00
-2.62249798e-01 -1.11943975e-01 6.17956072e-02 5.16126752e-01
-9.34836119e-02 3.54026482e-02 -1.07159519e+00 -4.26588729e-02
-9.03786905e-03 -6.98073328e-01 7.96112299e-01 1.22805178e-01
6.36646271e-01 -1.83404595e-01 -3.00250024e-01 4.25127774e-01
1.01974297e+00 1.42716048e-02 3.64533126e-01 4.48015690e-01
4.06727195e-01 8.61727655e-01 5.49229264e-01 -1.27819508e-01
5.45937181e-01 2.69311488e-01 7.39123970e-02 1.12677533e-02
-4.13367637e-02 -4.22385275e-01 4.65805203e-01 1.01283276e+00
1.29977241e-01 2.61840783e-03 -1.24481153e+00 7.49742568e-01
-1.53174651e+00 -5.59831500e-01 -1.27718225e-01 1.95198417e+00
1.57491112e+00 7.09938705e-01 1.26092613e-01 1.94578320e-01
8.02946389e-01 -3.27760428e-02 -1.44492030e-01 -7.93628395e-01
-2.99458094e-02 6.73961461e-01 5.08224905e-01 8.04707825e-01
-1.32395446e+00 1.38990188e+00 6.50072527e+00 4.98387605e-01
-9.17923212e-01 2.71742195e-01 9.11218762e-01 2.53169984e-01
-2.67214030e-01 4.75242108e-01 -1.29040647e+00 3.65901053e-01
1.26428771e+00 3.95046234e-01 3.59163523e-01 9.37301636e-01
-1.28483310e-01 -2.47492328e-01 -1.20340669e+00 -4.07861844e-02
-1.90110162e-01 -1.15027499e+00 -5.14533103e-01 -1.38420522e-01
3.51995856e-01 2.93822974e-01 -6.04437232e-01 6.03558540e-01
9.68538940e-01 -7.99258888e-01 1.01511478e+00 -7.85374716e-02
6.55442595e-01 -4.05890942e-01 7.60933936e-01 5.82941353e-01
-1.02543283e+00 -1.80047080e-01 -1.53296903e-01 -3.09976339e-01
2.14262977e-01 6.46788597e-01 -9.18027163e-01 3.93996596e-01
7.59027660e-01 2.94196188e-01 -6.10809326e-01 3.07855517e-01
-1.22474420e+00 1.41424477e+00 -4.92103457e-01 -6.98122382e-02
1.36813670e-01 4.71572466e-02 3.35880339e-01 1.53314471e+00
-3.11764210e-01 2.47453272e-01 2.32142970e-01 5.90894520e-01
-1.03589818e-01 1.10499509e-01 -2.30391845e-01 1.06438294e-01
6.61117673e-01 1.38560247e+00 -9.64811921e-01 -3.62318516e-01
-4.60848659e-01 4.22907233e-01 1.05260932e+00 1.27228990e-01
-6.92910254e-01 -3.16278309e-01 2.10774988e-01 -4.92797159e-02
6.20954394e-01 -1.46345139e-01 -6.65500224e-01 -1.01137114e+00
1.54698521e-01 -8.98234606e-01 6.90359533e-01 -4.26529199e-01
-1.14915657e+00 6.49917185e-01 -5.87091111e-02 -5.90276182e-01
-2.13409737e-01 -8.44523668e-01 -6.61108613e-01 8.18244815e-01
-1.43833184e+00 -1.28130484e+00 2.28199791e-02 3.57199125e-02
2.87502557e-01 7.78402239e-02 9.47222412e-01 -4.28239033e-02
-7.10818470e-01 7.66081154e-01 -3.91515195e-01 8.19575667e-01
5.83070278e-01 -1.45151663e+00 6.88320994e-01 1.18246698e+00
2.42990240e-01 8.06475222e-01 4.25893664e-01 -6.57604337e-01
-9.65846241e-01 -7.71503925e-01 1.31238651e+00 -6.94076061e-01
8.13574314e-01 -7.27386713e-01 -7.66314805e-01 1.01427448e+00
2.67107099e-01 -4.44433726e-02 7.24483550e-01 7.88323462e-01
-6.99150860e-01 -3.97946239e-02 -8.98018897e-01 3.06793362e-01
1.13117325e+00 -4.18275863e-01 -9.64466333e-01 2.10433111e-01
8.55596364e-01 -5.49591124e-01 -8.46901596e-01 3.42829496e-01
4.36819762e-01 -1.25478089e+00 3.69754106e-01 -9.18072939e-01
7.61591315e-01 5.39504737e-02 -2.59569615e-01 -1.06395888e+00
-2.85538342e-02 -5.86775303e-01 1.34304374e-01 1.66354835e+00
1.06297588e+00 -6.73228145e-01 7.70683527e-01 5.95601022e-01
-3.83986771e-01 -8.21774423e-01 -8.17458332e-01 -7.23382354e-01
5.44824839e-01 -5.31616449e-01 4.14479703e-01 8.56851876e-01
3.05232406e-01 4.66664255e-01 2.16060385e-01 1.73930213e-01
4.26424146e-01 6.90585524e-02 6.60344124e-01 -9.88334477e-01
-5.16320646e-01 -1.20019734e-01 2.68558506e-02 -9.59526956e-01
4.50742662e-01 -1.02620077e+00 3.55826199e-01 -1.13817847e+00
1.92037985e-01 -9.51164246e-01 -1.84512779e-01 1.12669694e+00
-2.83594102e-01 -8.25466067e-02 5.54480925e-02 8.94826055e-02
-3.97028387e-01 -1.84408516e-01 5.98917305e-01 1.98099792e-01
-1.41455606e-01 -1.23090565e-01 -9.84494150e-01 8.93457830e-01
1.05726445e+00 -8.17374408e-01 7.54903704e-02 -6.64326370e-01
2.08286151e-01 4.59009446e-02 -4.65244204e-02 -7.54106045e-01
3.70259359e-02 1.23080418e-01 1.77882925e-01 -2.98351586e-01
-6.18648119e-02 -2.75681525e-01 -4.87532049e-01 4.50397104e-01
-4.94300634e-01 6.95036948e-02 1.69413611e-01 1.32924601e-01
-1.22176267e-01 -7.78770983e-01 6.98399127e-01 -4.30418879e-01
-4.52763110e-01 -3.84858698e-01 -1.97237059e-01 5.91157794e-01
6.48124814e-01 5.08967601e-02 -5.14213264e-01 -7.95550123e-02
-7.22630620e-01 2.47094393e-01 4.46234792e-01 -3.37762721e-02
-9.09873694e-02 -5.76433837e-01 -7.11430609e-01 2.26148799e-01
-1.66805133e-01 1.03015348e-01 -3.47942382e-01 7.28355289e-01
-6.05788291e-01 2.01458111e-01 7.64534771e-02 -4.17791724e-01
-1.06436372e+00 1.59889281e-01 6.61385730e-02 -7.60189652e-01
-5.12904108e-01 1.18843997e+00 -1.90015048e-01 -7.86885917e-01
1.52155191e-01 -3.49343508e-01 -2.69910339e-02 5.82090160e-03
2.00665891e-01 -2.99216304e-02 1.02773152e-01 -4.43836570e-01
-5.61762512e-01 1.03128359e-01 -1.94273293e-01 -4.03832525e-01
1.41826952e+00 2.34404013e-01 -2.81303406e-01 4.65177655e-01
8.21820259e-01 5.11752069e-01 -1.13232315e+00 1.36482194e-02
5.94655573e-01 8.30806047e-02 -3.59059483e-01 -1.13751018e+00
-6.90721154e-01 7.68222511e-01 -2.68616587e-01 3.43566239e-01
9.06068146e-01 2.74629384e-01 7.23179460e-01 3.25542510e-01
4.21003252e-01 -1.22472274e+00 -2.03012958e-01 8.63850534e-01
4.16046232e-01 -1.15350676e+00 -1.32313564e-01 -7.47711718e-01
-6.62690878e-01 1.07949960e+00 6.39068067e-01 -6.41727149e-02
5.06269872e-01 7.49144077e-01 3.68843228e-01 9.77662876e-02
-1.03263211e+00 -2.17309102e-01 -3.08478296e-01 4.30067867e-01
8.13046992e-01 2.39801049e-01 -4.59577590e-01 1.07366896e+00
-6.63082182e-01 -3.28217149e-01 4.05793220e-01 1.29707253e+00
-3.98913801e-01 -1.46886075e+00 -1.50004551e-01 3.56647015e-01
-7.62595296e-01 -3.77582550e-01 -3.14564645e-01 1.01215398e+00
1.70388743e-01 1.23087215e+00 3.00012957e-02 -2.51048356e-01
1.49452344e-01 5.82454622e-01 5.30915856e-01 -9.57691669e-01
-9.99469340e-01 1.58303510e-02 9.44169164e-01 -3.48144859e-01
-2.84453183e-01 -8.92753184e-01 -1.28565824e+00 1.79600209e-01
-5.04427075e-01 4.32639480e-01 5.72537303e-01 9.25953507e-01
1.38626536e-02 1.77639514e-01 5.46733081e-01 -4.01894003e-01
-8.87677848e-01 -1.26835775e+00 -2.18201891e-01 3.69943202e-01
-1.15147509e-01 -2.76678413e-01 -5.73116899e-01 2.27373555e-01] | [10.39371109008789, 9.615803718566895] |
68c4a696-3595-44d5-b0b1-8dc376b89575 | towards-multi-scale-style-control-for | 2104.03521 | null | https://arxiv.org/abs/2104.03521v1 | https://arxiv.org/pdf/2104.03521v1.pdf | Towards Multi-Scale Style Control for Expressive Speech Synthesis | This paper introduces a multi-scale speech style modeling method for end-to-end expressive speech synthesis. The proposed method employs a multi-scale reference encoder to extract both the global-scale utterance-level and the local-scale quasi-phoneme-level style features of the target speech, which are then fed into the speech synthesis model as an extension to the input phoneme sequence. During training time, the multi-scale style model could be jointly trained with the speech synthesis model in an end-to-end fashion. By applying the proposed method to style transfer task, experimental results indicate that the controllability of the multi-scale speech style model and the expressiveness of the synthesized speech are greatly improved. Moreover, by assigning different reference speeches to extraction of style on each scale, the flexibility of the proposed method is further revealed. | ['Helen Meng', 'Jia Jia', 'Zhiyong Wu', 'Jingbei Li', 'Changhe Song', 'Xiang Li'] | 2021-04-08 | null | null | null | null | ['expressive-speech-synthesis'] | ['speech'] | [ 3.89582366e-01 2.62040645e-01 -1.14616398e-02 -6.09539211e-01
-7.94735193e-01 -6.20557308e-01 4.80218709e-01 -6.47491693e-01
-6.56452924e-02 6.14190638e-01 5.53288579e-01 -1.62898511e-01
2.70324737e-01 -5.95142305e-01 -4.22725469e-01 -5.55632651e-01
6.16765440e-01 1.15222618e-01 -1.51945695e-01 -5.37943900e-01
-1.79028079e-01 5.68201005e-01 -1.24773836e+00 5.20552516e-01
7.67475843e-01 1.08642042e+00 7.32406676e-01 1.01834488e+00
-1.59141898e-01 6.47123575e-01 -8.04320574e-01 -1.52222142e-01
6.25083819e-02 -7.35896111e-01 -5.78645706e-01 5.88275731e-01
9.93162096e-02 -3.80470186e-01 -5.64896584e-01 1.05143774e+00
5.78450143e-01 5.24373829e-01 3.58199358e-01 -7.57322192e-01
-4.53980476e-01 4.61544812e-01 1.90966919e-01 -9.15178210e-02
2.78133810e-01 1.29401088e-01 9.12784874e-01 -1.01090991e+00
6.29841745e-01 1.62886500e+00 3.52483504e-02 7.67034888e-01
-1.07260883e+00 -7.07009673e-01 1.34295687e-01 -2.46444613e-01
-1.13921034e+00 -9.08076227e-01 1.29205763e+00 -9.84137878e-02
9.20181334e-01 3.44723284e-01 5.30803144e-01 1.27698076e+00
2.06300467e-02 7.71200716e-01 1.22749841e+00 -5.05404353e-01
7.93184564e-02 2.98317850e-01 -4.50095654e-01 7.79182315e-01
-7.36330688e-01 5.35879552e-01 -5.73539317e-01 1.71661884e-01
9.36209142e-01 -4.05146599e-01 -2.72381604e-01 6.71236441e-02
-1.08966815e+00 7.42560446e-01 2.14410171e-01 4.49906588e-01
-1.73511446e-01 -2.57911295e-01 7.79470146e-01 5.96393764e-01
6.60619199e-01 2.25038677e-01 -5.04265904e-01 -3.62938702e-01
-1.01508296e+00 -4.18296084e-02 7.18317032e-01 1.39209175e+00
5.62403560e-01 7.87468374e-01 -3.22847456e-01 1.22094905e+00
2.53528118e-01 7.00385571e-01 6.96806192e-01 -8.46164525e-01
7.91101754e-01 2.02861890e-01 -1.38252852e-02 -6.85840964e-01
-1.66349456e-01 -4.81021553e-01 -9.68544841e-01 8.04811791e-02
1.07755605e-02 -3.70041907e-01 -5.47862828e-01 1.97167969e+00
1.97256073e-01 -3.46985422e-02 3.73217434e-01 6.92322135e-01
7.09045053e-01 1.32353175e+00 -8.86864662e-02 -5.52831531e-01
1.25728941e+00 -1.31852722e+00 -1.09257686e+00 -2.53136486e-01
4.52507406e-01 -9.47517097e-01 1.39350092e+00 1.17913745e-01
-1.40833676e+00 -1.30418837e+00 -9.53501582e-01 -4.71765175e-02
-9.08349603e-02 7.09100962e-01 -1.40313402e-01 5.38577139e-01
-9.80826378e-01 4.97415543e-01 -3.19312662e-01 -1.30280331e-01
-1.75967529e-01 1.93392083e-01 -2.68394858e-01 3.28589827e-01
-1.36249292e+00 7.87052989e-01 5.55371940e-01 -3.78188826e-02
-7.24857390e-01 -4.22684520e-01 -9.29818690e-01 1.88989222e-01
9.95582864e-02 -6.47518814e-01 1.44005334e+00 -1.29214156e+00
-2.51523256e+00 4.43581522e-01 -5.35264373e-01 1.68337375e-01
3.58019561e-01 -7.92365149e-02 -7.58754075e-01 1.73912287e-01
-7.22678825e-02 6.68083251e-01 1.20142865e+00 -1.12531424e+00
-9.05311286e-01 3.59106846e-02 -1.35624260e-01 5.58846116e-01
-3.97626013e-01 2.53396690e-01 -1.61341771e-01 -1.12374854e+00
-2.75249958e-01 -8.53681445e-01 4.64542806e-02 -2.14507788e-01
-1.57097980e-01 -1.67550087e-01 1.02108169e+00 -9.09956217e-01
1.29947972e+00 -2.42094898e+00 6.53915882e-01 -1.41091928e-01
-3.27144086e-01 3.89084488e-01 -3.81260127e-01 4.06314909e-01
7.58327544e-02 -1.64484411e-01 -5.24927452e-02 -5.90833426e-01
-5.72603457e-02 2.29534641e-01 -4.93403435e-01 -3.54812182e-02
2.78181881e-01 7.65639365e-01 -9.69906926e-01 -2.64789879e-01
4.93877113e-01 3.20070267e-01 -5.22352874e-01 9.16630685e-01
-1.01108551e-01 8.91434848e-01 -3.02188933e-01 2.16371551e-01
2.46667475e-01 4.13963228e-01 9.47718918e-02 -2.14649260e-01
-7.80316293e-02 5.75722158e-01 -9.29695427e-01 1.93629968e+00
-1.22114372e+00 4.22075808e-01 4.03338641e-01 -6.28844440e-01
1.32264888e+00 8.59194934e-01 8.18827674e-02 -5.64831376e-01
8.69710967e-02 2.38276929e-01 1.02954209e-01 -3.34968776e-01
6.32257640e-01 -5.94688475e-01 -3.13604385e-01 8.59740525e-02
3.69567841e-01 -6.73003793e-01 -1.42509595e-01 -4.21251893e-01
4.28958833e-01 1.55291259e-01 4.11205173e-01 -3.40101242e-01
1.00569463e+00 -6.85101867e-01 5.47721803e-01 1.69662625e-01
-4.80072796e-02 2.56094873e-01 2.60659847e-02 -2.70780288e-02
-1.40634227e+00 -1.03771245e+00 2.97447443e-01 1.27171087e+00
-3.30203772e-01 -3.17847729e-01 -9.97103155e-01 -6.03987157e-01
-4.89890665e-01 8.41133177e-01 -1.88153714e-01 -2.44590312e-01
-8.01192284e-01 2.78439105e-01 7.46031940e-01 5.30580521e-01
6.24026358e-01 -1.26684415e+00 -2.29454692e-02 5.76090872e-01
-4.09753263e-01 -1.36000156e+00 -1.14438164e+00 -1.04366980e-01
-8.77434611e-01 -1.95288926e-01 -8.10636401e-01 -1.14085829e+00
3.13595146e-01 -1.74388885e-01 5.71258187e-01 -4.65652645e-01
3.97744775e-01 -9.66373831e-02 -3.08590055e-01 -3.14003862e-02
-1.27859747e+00 1.86733156e-01 3.53548467e-01 4.23171461e-01
-4.40414786e-01 -6.58623159e-01 -2.24521175e-01 2.23010600e-01
-7.64742315e-01 3.74587834e-01 4.54478264e-01 1.09985173e+00
4.29136932e-01 -8.43515769e-02 1.26208162e+00 -4.69829589e-01
7.85532653e-01 2.12338567e-02 -5.91557264e-01 9.73627642e-02
-8.71198773e-02 1.50050089e-01 1.36651397e+00 -5.96669555e-01
-1.63816106e+00 7.98319578e-02 -7.12771714e-01 -5.58657229e-01
-3.71905446e-01 1.83410242e-01 -8.77769172e-01 2.07230285e-01
1.51785761e-01 6.05835199e-01 7.37642571e-02 -5.54620206e-01
7.73592651e-01 1.11687219e+00 5.38422823e-01 -7.01412559e-01
7.50045538e-01 -2.18164191e-01 -1.48124844e-01 -1.23498869e+00
-5.96481562e-01 -2.22228378e-01 -6.24216318e-01 -8.20419341e-02
7.80934453e-01 -1.02230322e+00 -3.39565277e-01 6.13556325e-01
-1.46231914e+00 -3.90467733e-01 -4.34464753e-01 3.62473100e-01
-1.08770740e+00 2.77212322e-01 -8.15326154e-01 -7.84528613e-01
-4.25310820e-01 -1.47808754e+00 1.31313932e+00 -2.94917598e-02
-3.61832231e-01 -1.16826630e+00 -1.08098269e-01 2.22909868e-01
4.52370793e-01 -2.70716846e-01 1.11300111e+00 -5.69530904e-01
-1.23001456e-01 7.98850581e-02 5.64247696e-03 7.90107071e-01
6.65335894e-01 -2.44486153e-01 -1.09409368e+00 -3.43430400e-01
2.56641358e-01 -3.14721227e-01 9.43782181e-02 3.65552306e-02
9.29653227e-01 -7.36938059e-01 2.34863415e-01 7.45915771e-01
9.79523063e-01 6.12493336e-01 3.60829860e-01 -2.62905061e-01
7.49661624e-01 6.52631819e-01 6.44919336e-01 2.92011768e-01
1.20025948e-01 1.03522825e+00 -2.33913049e-01 -1.97757706e-01
-4.94601518e-01 -5.41872919e-01 7.81757891e-01 1.73815060e+00
1.55478030e-01 -2.75227875e-01 -6.60590008e-02 3.01798880e-01
-1.44137657e+00 -1.05621493e+00 5.00317752e-01 1.84226751e+00
1.15086639e+00 1.47251738e-02 1.42274559e-01 1.15882993e-01
7.29676306e-01 4.88833427e-01 -3.28882903e-01 -9.53264832e-01
-3.29569615e-02 2.07572535e-01 -2.07093060e-01 7.73554862e-01
-8.65187466e-01 1.34686291e+00 5.85284328e+00 1.37760592e+00
-1.42547870e+00 1.13766663e-01 4.42381114e-01 9.15754810e-02
-2.66020715e-01 -2.46529385e-01 -8.19865823e-01 5.32594562e-01
1.28112042e+00 -3.07466507e-01 9.22273695e-01 8.15163672e-01
7.35607505e-01 6.31566942e-01 -1.19302034e+00 8.67042780e-01
-1.31348863e-01 -1.01878846e+00 2.32986987e-01 -1.97153211e-01
8.02568018e-01 -4.61038142e-01 3.12860347e-02 4.67526495e-01
-1.99893042e-01 -6.24959111e-01 1.05588877e+00 3.04774851e-01
1.44064772e+00 -9.33428049e-01 2.32266739e-01 5.65355778e-01
-1.54708397e+00 -2.41451263e-02 -9.24155191e-02 -5.89782149e-02
4.20038640e-01 9.47870389e-02 -8.95162046e-01 6.96131229e-01
-7.72217289e-02 5.18437684e-01 -1.90602124e-01 2.25246191e-01
-3.76393110e-01 7.94843972e-01 3.38271223e-02 -1.67847037e-01
3.32659841e-01 -3.08094859e-01 8.76651585e-01 1.46998715e+00
6.01180315e-01 2.32510753e-02 2.57928401e-01 7.56083429e-01
-1.70298010e-01 2.41619930e-01 -7.47660637e-01 -3.17104399e-01
5.47407210e-01 1.14125299e+00 -1.98776230e-01 -5.78231275e-01
-3.35617840e-01 1.30894220e+00 1.83047980e-01 5.22842050e-01
-5.32031119e-01 -6.08125806e-01 8.01489592e-01 -3.51049304e-01
3.44348758e-01 -2.22637087e-01 -3.36149156e-01 -1.16354585e+00
-9.89418700e-02 -1.22674274e+00 -3.21137279e-01 -9.20985103e-01
-1.02699792e+00 1.09323978e+00 -2.30379611e-01 -1.26246488e+00
-7.54361987e-01 -5.82010567e-01 -6.98608279e-01 1.24169362e+00
-1.12016881e+00 -1.40428424e+00 2.63090491e-01 5.84097862e-01
1.39156961e+00 -7.25623250e-01 1.09747231e+00 8.40740725e-02
-2.76633799e-01 9.22480881e-01 1.91411644e-01 1.53555244e-01
4.57260847e-01 -1.00750983e+00 4.56732273e-01 7.81922638e-01
-5.25412383e-03 3.99192840e-01 6.06709003e-01 -5.48520803e-01
-1.00976539e+00 -1.40636182e+00 1.01819420e+00 8.78425315e-02
6.80244863e-01 -5.35630345e-01 -6.53449059e-01 4.54060286e-01
4.23306018e-01 -2.86380649e-01 4.33116436e-01 -2.68624187e-01
-5.16702980e-02 -3.46771628e-01 -1.08109605e+00 8.52008998e-01
1.02917373e+00 -1.07210839e+00 -7.68614650e-01 -8.09983984e-02
1.26172888e+00 -5.12911499e-01 -1.14195359e+00 2.96622723e-01
5.12053311e-01 -5.04387140e-01 7.00511396e-01 -5.81469059e-01
5.00350893e-01 -8.32241550e-02 -5.64361811e-01 -1.87424386e+00
-3.52858990e-01 -8.47417295e-01 -1.53841404e-02 1.62246859e+00
2.70562500e-01 -4.62711841e-01 1.77007437e-01 -2.49368232e-02
-6.17195964e-01 -6.32041097e-01 -1.14156246e+00 -9.47194219e-01
1.95386648e-01 -1.26290008e-01 7.81083465e-01 5.38116813e-01
4.05629985e-02 7.96016574e-01 -5.71125984e-01 2.23004967e-01
1.54326513e-01 8.23742896e-02 6.52486145e-01 -6.02716863e-01
-6.63442373e-01 -5.40082812e-01 -8.24858993e-02 -1.38632786e+00
6.17349982e-01 -8.53052974e-01 2.93593615e-01 -1.05897820e+00
-4.51291442e-01 -9.43284929e-02 -1.13862537e-01 5.70254587e-03
-2.33940855e-01 -3.05782378e-01 4.77983713e-01 -2.34619007e-01
1.08975507e-01 1.00586915e+00 1.77443588e+00 -9.06934440e-02
-3.13958287e-01 4.35791731e-01 -2.79241025e-01 6.27073824e-01
8.01116705e-01 -1.29430443e-01 -7.97259212e-01 -1.35264769e-01
-5.95285654e-01 8.61618280e-01 -8.86745155e-02 -7.65254855e-01
-2.50715196e-01 -3.36852819e-01 6.08741231e-02 -3.06223780e-01
7.56670117e-01 -8.00735235e-01 -1.29709780e-01 2.04605728e-01
-7.35198200e-01 -4.06408459e-02 3.07216942e-01 3.41498226e-01
-5.62220991e-01 -1.97851464e-01 9.95814800e-01 4.09248136e-02
-4.42410469e-01 1.55035257e-01 -3.88802409e-01 -1.52824819e-01
5.85311174e-01 -3.04248482e-02 1.29126385e-01 -5.11253417e-01
-9.26599979e-01 -3.83298278e-01 1.31000683e-01 6.70180321e-01
6.01285160e-01 -1.77893746e+00 -6.51456892e-01 6.35010898e-01
6.98487461e-02 -1.84870064e-01 3.02214026e-01 2.75437504e-01
4.62985858e-02 5.34585357e-01 -2.98055589e-01 -3.03322583e-01
-1.20912337e+00 6.00376308e-01 3.13442349e-01 -3.26112658e-01
-4.06545371e-01 6.94900334e-01 5.04807055e-01 -6.63020611e-01
1.40859291e-01 -4.96770233e-01 1.90734863e-01 -1.63523704e-01
4.03085500e-01 3.63242835e-01 -6.21557683e-02 -9.38538551e-01
1.00056466e-03 4.33386236e-01 1.96600527e-01 -6.76880658e-01
1.05616033e+00 -5.92070401e-01 1.34737030e-01 8.34422231e-01
1.41100252e+00 2.93543398e-01 -1.45299816e+00 -2.29059964e-01
-3.54020804e-01 -2.21610531e-01 -6.76676333e-02 -8.28051209e-01
-8.11910510e-01 1.16456759e+00 3.21088135e-01 -9.31648314e-02
1.36258090e+00 -3.81830037e-01 1.05498362e+00 3.00694734e-01
3.87587488e-01 -1.29868042e+00 2.15343878e-01 7.25584865e-01
1.40735078e+00 -7.90645659e-01 -7.96457767e-01 -3.58724415e-01
-8.92623305e-01 1.23560941e+00 4.52081144e-01 2.59233493e-04
4.31607246e-01 3.07474375e-01 1.72908410e-01 5.91510713e-01
-8.13841879e-01 1.00069381e-01 4.32834834e-01 5.71365237e-01
4.86994773e-01 1.98908463e-01 -2.90695354e-02 7.32362628e-01
-6.04226708e-01 1.71819068e-02 2.20395401e-01 3.94787967e-01
-5.43970585e-01 -1.30305994e+00 -2.19850332e-01 -1.10697962e-01
-9.52845812e-02 -1.76942229e-01 -1.61998078e-01 2.34162912e-01
4.37632129e-02 1.20307589e+00 -1.63457319e-01 -6.80227101e-01
6.75323129e-01 4.38921213e-01 4.71068323e-01 -6.82065487e-01
-6.09218776e-01 5.01726449e-01 5.23952067e-01 -3.53714347e-01
-1.56864319e-02 -3.04979593e-01 -1.18554246e+00 3.35720517e-02
-1.18430845e-01 1.00456238e-01 4.70697314e-01 1.12801957e+00
2.42822170e-01 9.17827606e-01 1.30385518e+00 -1.02465892e+00
-8.76207590e-01 -1.33062482e+00 -5.86199462e-01 4.51484382e-01
5.82972229e-01 -2.56290138e-01 -4.59623672e-02 3.62867892e-01] | [14.939289093017578, 6.550870895385742] |
e34c16a9-7eb3-4d7b-9c9b-857be11a0a21 | an-algorithm-and-heuristic-based-on | 2210.13456 | null | https://arxiv.org/abs/2210.13456v1 | https://arxiv.org/pdf/2210.13456v1.pdf | An Algorithm and Heuristic based on Normalized Mutual Information for Dimensionality Reduction and Classification of Hyperspectral images | In the feature classification domain, the choice of data affects widely the results. The Hyperspectral image (HSI), is a set of more than a hundred bidirectional measures (called bands), of the same region (called ground truth map: GT). The HSI is modelized at a set of N vectors. So we have N features (or attributes) expressing N vectors of measures for C substances (called classes). The problematic is that it's pratically impossible to investgate all possible subsets. So we must find K vectors among N, such as relevant and no redundant ones; in order to classify substances. Here we introduce an algorithm based on Normalized Mutual Information to select relevant and no redundant bands, necessary to increase classification accuracy of HSI. Keywords: Feature Selection, Normalized Mutual information, Hyperspectral images, Classification, Redundancy. | ['Driss Aboutajdine', 'Ahmed Hammouch', 'Elkebir Sarhrouni'] | 2022-10-22 | null | null | null | null | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 8.85945916e-01 -4.95658010e-01 -3.57362151e-01 -4.68955010e-01
-3.36369723e-01 -6.75655305e-01 2.57482260e-01 -5.34044988e-02
-7.69237205e-02 9.05570686e-01 1.15655944e-01 9.19033065e-02
-1.03566206e+00 -1.22109437e+00 -1.84671767e-02 -9.55081522e-01
-3.02941591e-01 1.93516184e-02 -3.20482880e-01 -1.78699240e-01
6.86069369e-01 6.99222863e-01 -2.26778150e+00 4.37293291e-01
1.11154974e+00 1.30522871e+00 2.97428727e-01 3.74745637e-01
-3.76040608e-01 5.21520078e-01 -3.45933318e-01 1.04074590e-01
6.23549223e-01 -6.34330392e-01 -1.14187193e+00 4.58790869e-01
8.57745949e-03 1.64887071e-01 1.02026232e-01 1.67394853e+00
1.03004593e-02 2.66905487e-01 1.26520836e+00 -1.32358074e+00
-6.12394691e-01 7.50800788e-01 -5.90961397e-01 -6.23659566e-02
2.28242666e-01 -2.54953235e-01 1.08457708e+00 -6.39582813e-01
5.71696758e-01 8.12724590e-01 4.19961452e-01 -8.00846051e-03
-1.26584327e+00 -5.61079681e-01 -4.44081575e-01 3.28591645e-01
-1.77562118e+00 -2.39175394e-01 7.14453101e-01 -6.23832226e-01
6.14417195e-01 1.07855463e+00 1.04463506e+00 3.40385675e-01
2.44875878e-01 2.14741871e-01 1.21281338e+00 -6.74665987e-01
2.79342592e-01 3.76862198e-01 3.83033395e-01 2.48146415e-01
5.04257083e-01 -4.48712222e-02 -2.65088528e-01 -3.41575086e-01
3.38975042e-01 3.00956577e-01 -6.18473470e-01 -3.04554939e-01
-1.03034556e+00 1.00056767e+00 4.53347415e-01 8.46956849e-01
-5.78278840e-01 -6.89625561e-01 1.36656880e-01 6.67524636e-01
2.35774234e-01 8.57379317e-01 -3.19620371e-01 5.62142134e-01
-7.21813798e-01 -1.70735240e-01 6.06763303e-01 9.22073483e-01
1.44262230e+00 -1.35272086e-01 4.02410142e-02 8.41583788e-01
1.11763887e-01 6.65961385e-01 6.08534336e-01 -7.06588447e-01
3.09427321e-01 1.15217984e+00 -2.75569111e-01 -1.38756001e+00
-5.58952153e-01 -3.10790896e-01 -9.85006511e-01 1.84404880e-01
-3.96141917e-01 -4.77523394e-02 -8.58822167e-01 1.29317629e+00
-5.72962686e-02 -3.66451323e-01 2.54026383e-01 9.49321270e-01
8.90690446e-01 1.00883758e+00 -5.02924845e-02 -5.70382774e-01
9.63157535e-01 -2.91693985e-01 -7.23415494e-01 1.27706870e-01
4.38549459e-01 -8.60739112e-01 6.57072067e-01 6.79814041e-01
-5.49015820e-01 -3.16144347e-01 -1.25300801e+00 4.80649173e-01
-1.11176217e+00 1.56522736e-01 1.00898767e+00 8.15376759e-01
-8.11704457e-01 8.02964330e-01 5.17358258e-02 -2.89768279e-01
1.20707273e-01 5.09373784e-01 -8.33325565e-01 5.26265129e-02
-9.97667134e-01 8.81578684e-01 9.43494380e-01 -5.09446152e-02
-1.57930866e-01 -3.41093898e-01 -6.26347423e-01 -4.27300707e-02
1.78305954e-01 -3.19951504e-01 2.44195983e-01 -1.30628836e+00
-8.79502714e-01 9.61243153e-01 8.02169144e-02 2.82538682e-01
-2.08775982e-01 4.21668380e-01 -8.19503605e-01 1.66892484e-01
8.01092535e-02 2.85584152e-01 4.67338145e-01 -1.42336452e+00
-7.92550147e-01 -7.46832907e-01 -2.72346914e-01 1.77238747e-01
-4.52823043e-01 1.10766608e-02 4.41588014e-01 -3.13980728e-01
1.03158915e+00 -6.81950748e-01 -2.65515029e-01 -3.67874384e-01
-7.66199827e-01 1.83432057e-01 7.68373907e-01 -5.09038329e-01
9.42068636e-01 -2.11238575e+00 -2.89890207e-02 1.03005660e+00
-5.20612895e-02 5.13028577e-02 -1.58286944e-01 3.75463933e-01
-6.23955846e-01 5.06682277e-01 -7.17006564e-01 6.10118330e-01
-2.25637689e-01 7.40323439e-02 5.80777274e-03 7.01156259e-01
1.43985059e-02 2.03406662e-01 -6.30283773e-01 -3.92548054e-01
3.70325029e-01 2.93838441e-01 1.27485767e-01 -2.08269238e-01
2.28679612e-01 -9.29628685e-03 -6.89154685e-01 9.05104220e-01
1.22492588e+00 -7.17039257e-02 5.99407367e-02 -5.50905168e-01
-5.70189714e-01 -1.48508117e-01 -1.43953872e+00 1.40057194e+00
1.76107481e-01 5.17358780e-01 -4.33236599e-01 -1.20031071e+00
1.21280122e+00 3.87029231e-01 1.04294384e+00 -4.20292944e-01
3.99383187e-01 1.39676034e-01 2.83551943e-02 -5.69835067e-01
3.57228220e-01 -6.56053647e-02 3.78884673e-01 3.88681352e-01
6.14325739e-02 -3.00218612e-01 3.07708025e-01 -1.98955759e-01
6.12420678e-01 -5.15132129e-01 7.92368293e-01 -6.83870018e-01
7.00551629e-01 3.67014289e-01 5.46315432e-01 4.94142503e-01
9.71308127e-02 4.77418154e-01 3.38587433e-01 -3.27550918e-01
-1.02441847e+00 -7.59149373e-01 -6.31543756e-01 5.70787609e-01
4.20422629e-02 3.59144341e-03 -3.43942702e-01 -1.37275204e-01
1.36559635e-01 5.09613872e-01 -5.87308109e-01 -2.85626143e-01
4.02641624e-01 -1.12098277e+00 1.34019196e-01 -3.43476027e-01
7.04754472e-01 -7.44715810e-01 -3.98210377e-01 1.51649997e-01
-6.40987931e-03 -3.32336694e-01 1.36200666e-01 5.06746173e-01
-8.73967767e-01 -1.17735672e+00 -7.82316104e-02 -5.60110807e-01
5.66424251e-01 7.90160120e-01 9.99673486e-01 -2.14820698e-01
-4.98102963e-01 8.54223371e-02 -7.37305820e-01 -5.89277923e-01
-7.71875679e-03 -3.70599121e-01 -5.79702146e-02 2.33527467e-01
4.06688631e-01 -8.87991965e-01 -3.41911435e-01 2.01210871e-01
-1.11231506e+00 -1.96509376e-01 7.23554075e-01 6.16285861e-01
8.66900921e-01 8.97873461e-01 2.77760565e-01 -7.99500227e-01
3.65546584e-01 -5.29249907e-01 -5.02070010e-01 6.01840794e-01
-6.50979280e-01 -8.75796676e-02 4.01506215e-01 9.01891440e-02
-6.67105794e-01 9.82111096e-02 6.08547688e-01 -3.14626703e-03
-1.84290722e-01 8.00539613e-01 -5.38630068e-01 -6.39304996e-01
1.00276577e+00 4.13865209e-01 6.81695342e-02 -3.02852690e-01
2.92235762e-01 9.30012047e-01 1.69466212e-01 -2.13127702e-01
5.62808335e-01 3.06205511e-01 4.68630672e-01 -1.16282666e+00
-5.52333713e-01 -7.40533769e-01 -9.26003158e-01 -3.26995313e-01
7.43597925e-01 -5.41179597e-01 -6.40211046e-01 1.45227164e-02
-8.27952683e-01 5.34122348e-01 -3.68378401e-01 8.10525537e-01
-5.20028532e-01 1.06303439e-01 -8.17080513e-02 -1.01305699e+00
-2.69884169e-01 -9.58663046e-01 5.92721760e-01 2.42559969e-01
2.54147023e-01 -6.13926947e-01 -5.02935685e-02 4.33567613e-02
1.92133218e-01 4.79905397e-01 9.41814721e-01 -7.04003751e-01
-2.87995011e-01 -6.43070415e-02 -3.81621987e-01 5.47959566e-01
8.64685774e-01 4.28597659e-01 -1.03140414e+00 2.76820548e-02
3.10612321e-01 9.37458947e-02 9.10017014e-01 4.41952825e-01
1.30284929e+00 -4.34221596e-01 -3.67774934e-01 6.99873149e-01
1.86357784e+00 8.72774005e-01 8.26178133e-01 2.80748397e-01
3.63466948e-01 7.59496868e-01 8.34056377e-01 6.75462902e-01
-3.58458966e-01 -5.66818789e-02 3.99174243e-01 -1.67957917e-01
4.68418211e-01 5.14764488e-01 -4.15419370e-01 8.54326844e-01
-5.47085345e-01 -3.19106281e-01 -7.88410962e-01 1.06823213e-01
-1.41022229e+00 -1.27290106e+00 -3.61989230e-01 2.38078713e+00
5.57962239e-01 -5.40306747e-01 -1.14989325e-01 8.57187688e-01
9.58141744e-01 7.26026297e-02 -2.85955548e-01 -1.19890235e-01
-6.62705719e-01 1.71112373e-01 8.46833527e-01 3.09080511e-01
-1.46976805e+00 3.97123218e-01 6.82073879e+00 6.16844714e-01
-1.08452547e+00 -3.33020300e-01 6.33888543e-01 3.48192453e-01
-5.25203347e-01 -2.83292998e-02 -1.84167415e-01 2.60768652e-01
7.05875278e-01 -6.06791437e-01 5.27166903e-01 6.32362843e-01
1.59838393e-01 -2.29522884e-01 -6.94151521e-01 1.13160574e+00
2.14238420e-01 -1.01929641e+00 4.85056341e-01 2.79912978e-01
1.06175828e+00 -2.54651278e-01 8.52297917e-02 -3.72022837e-01
2.87861764e-01 -9.56917644e-01 3.25842053e-01 9.49844658e-01
5.84513724e-01 -1.23846734e+00 8.31512928e-01 1.84838608e-01
-1.31998599e+00 -1.23384804e-01 -9.52361703e-01 1.15231559e-01
-6.36183321e-01 9.88327920e-01 -5.19023240e-01 1.13893032e+00
5.24925113e-01 7.89910793e-01 -5.01332998e-01 1.29480731e+00
3.02849978e-01 5.00345509e-03 -9.79991481e-02 -2.56599873e-01
3.86605620e-01 -1.02467275e+00 4.14502352e-01 9.24757838e-01
6.89678490e-01 6.65428042e-01 2.62960464e-01 7.46648371e-01
3.57000768e-01 6.99323237e-01 -1.04986489e+00 -1.54779866e-01
6.29551470e-01 1.18831146e+00 -6.55616879e-01 -2.89191246e-01
-2.66211689e-01 7.34475911e-01 -4.42797750e-01 2.72530288e-01
-3.89896154e-01 -6.60934329e-01 6.79502189e-01 -3.91788244e-01
-2.88924575e-01 1.99075297e-01 -3.66724432e-01 -8.15566719e-01
-1.18512243e-01 -7.83040583e-01 6.90956354e-01 -8.57180238e-01
-1.55444121e+00 5.41885078e-01 -3.40289995e-02 -1.89427650e+00
3.09611857e-02 -8.50730419e-01 -1.03529200e-01 1.19275820e+00
-1.18264413e+00 -8.09647679e-01 -3.28303158e-01 8.99004400e-01
-9.51810852e-02 -4.42286372e-01 1.08299422e+00 1.79119229e-01
-2.48228818e-01 -1.82334334e-01 6.52383268e-01 -2.31615901e-01
2.11386248e-01 -1.17477942e+00 -6.46181881e-01 5.16915023e-01
-1.20546721e-01 4.68511969e-01 6.13433182e-01 -7.08466589e-01
-1.32310379e+00 -1.02179122e+00 6.85278714e-01 3.32423270e-01
5.10935903e-01 3.16306114e-01 -6.95969760e-01 3.65812987e-01
-1.19625740e-02 -2.12195635e-01 1.39019382e+00 4.61678579e-02
-1.67628407e-01 -3.51523817e-01 -1.49941862e+00 4.43802118e-01
9.32584167e-01 -4.45799738e-01 -3.22827309e-01 9.43595052e-01
4.04832065e-01 4.17538464e-01 -1.10688400e+00 3.85100961e-01
5.55888653e-01 -1.14547563e+00 9.65602934e-01 -5.68987489e-01
1.30167589e-01 -4.81682986e-01 -7.03722239e-01 -1.36534023e+00
-8.42294574e-01 1.69656664e-01 1.08895707e+00 9.22385275e-01
6.03565156e-01 -9.25957084e-01 6.90202892e-01 4.65328217e-01
-8.69700611e-02 -8.05599317e-02 -9.36429441e-01 -9.78830159e-01
-5.32889664e-01 -1.01434350e-01 1.12215436e+00 1.69841921e+00
3.13654810e-01 1.42177427e-02 -2.89872676e-01 2.07480490e-01
7.35301673e-01 4.49673653e-01 1.13340892e-01 -1.99381483e+00
2.10007995e-01 -7.46209025e-01 -1.05003846e+00 7.30343685e-02
-6.92292973e-02 -1.01603186e+00 -1.26862824e-01 -1.42338014e+00
5.75512409e-01 -4.61004645e-01 -5.44167995e-01 5.28686583e-01
2.52942353e-01 1.97178856e-01 4.95673120e-02 3.93720925e-01
7.29300752e-02 2.37514317e-01 1.20368493e+00 -5.36812901e-01
-2.83503085e-01 3.92833445e-03 -6.86727524e-01 5.55953145e-01
9.40211236e-01 -4.33444291e-01 -5.51222384e-01 1.79146469e-01
5.88445216e-02 9.14720073e-02 -1.91198617e-01 -1.17432034e+00
1.22719727e-01 -7.98025489e-01 7.55158722e-01 -9.21855986e-01
3.06941539e-01 -1.21725941e+00 1.08396471e+00 4.36476171e-01
-2.17478603e-01 -2.96454728e-01 -3.33691150e-01 -7.89813846e-02
-3.99231851e-01 -8.93036067e-01 7.22994030e-01 -5.19069254e-01
-8.32199991e-01 4.79326367e-01 -3.16206127e-01 -9.56679583e-01
9.62537348e-01 -7.04106033e-01 -3.37094128e-01 -1.60388038e-01
-7.37112105e-01 -2.17330575e-01 3.40983719e-01 -1.84519276e-01
6.47403002e-01 -1.55697286e+00 -6.97654963e-01 2.00744182e-01
4.00602400e-01 -5.45135975e-01 2.79915839e-01 3.39234889e-01
-5.96118987e-01 4.23257619e-01 -7.18434930e-01 -4.17074323e-01
-1.25216067e+00 5.21665692e-01 3.68858844e-01 1.79489255e-01
-1.27218723e-01 5.86332858e-01 -1.05334811e-01 -3.34516913e-01
-4.45118487e-01 -7.85381272e-02 -9.09033000e-01 4.32417214e-01
6.03394032e-01 4.60909039e-01 3.05683404e-01 -9.66815710e-01
-2.38390371e-01 8.16180706e-01 4.91981864e-01 -1.02445088e-01
1.44263148e+00 3.66059393e-02 -9.07827914e-01 5.30318141e-01
1.35709846e+00 -1.86194256e-01 -2.33269498e-01 -4.53131609e-02
-3.11574657e-02 -1.01190567e+00 1.87388897e-01 -6.58478618e-01
-1.14079082e+00 1.85622692e-01 1.04747295e+00 7.85867155e-01
1.56996715e+00 -4.82415408e-01 -2.91401535e-01 7.37057686e-01
5.93166828e-01 -1.10938597e+00 -6.20862067e-01 2.72440642e-01
8.84108305e-01 -1.24868345e+00 2.38588080e-01 -8.89452577e-01
-3.99682254e-01 1.40203977e+00 9.34751257e-02 1.93499222e-01
8.32101464e-01 -1.25171438e-01 -2.41931826e-01 -3.99837852e-01
-2.26055071e-01 -5.88750899e-01 4.96090889e-01 7.37249553e-01
5.27163982e-01 5.87935328e-01 -6.25902712e-01 2.33211398e-01
-4.86318558e-01 -4.11103100e-01 3.75776768e-01 9.34678197e-01
-1.20598149e+00 -8.89634192e-01 -8.54425251e-01 1.07535064e+00
1.90565661e-02 -8.56658965e-02 -6.57694042e-01 7.56465971e-01
5.19947171e-01 1.27795446e+00 -4.25359979e-03 -7.76375830e-01
3.47673804e-01 -5.81046231e-02 3.43605489e-01 -3.52269441e-01
-1.47480547e-01 -2.55183131e-01 -3.30193862e-02 -4.16404963e-01
-7.66137600e-01 -7.28429556e-01 -9.24553394e-01 -5.55100381e-01
-5.64181626e-01 4.70065832e-01 9.96422112e-01 7.54805982e-01
-2.45505184e-01 2.48038217e-01 1.36279905e+00 -6.98300838e-01
-2.14560628e-01 -9.42193985e-01 -1.60360658e+00 3.01482201e-01
1.05602063e-01 -7.17495382e-01 -5.97568929e-01 1.64801225e-01] | [9.747203826904297, -1.8299728631973267] |
02025425-ed21-435c-8b78-71e13f9ee769 | textdragon-an-end-to-end-framework-for | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Feng_TextDragon_An_End-to-End_Framework_for_Arbitrary_Shaped_Text_Spotting_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Feng_TextDragon_An_End-to-End_Framework_for_Arbitrary_Shaped_Text_Spotting_ICCV_2019_paper.pdf | TextDragon: An End-to-End Framework for Arbitrary Shaped Text Spotting | Most existing text spotting methods either focus on horizontal/oriented texts or perform arbitrary shaped text spotting with character-level annotations. In this paper, we propose a novel text spotting framework to detect and recognize text of arbitrary shapes in an end-to-end manner, using only word/line-level annotations for training. Motivated from the name of TextSnake, which is only a detection model, we call the proposed text spotting framework TextDragon. In TextDragon, a text detector is designed to describe the shape of text with a series of quadrangles, which can handle text of arbitrary shapes. To extract arbitrary text regions from feature maps, we propose a new differentiable operator named RoISlide, which is the key to connect arbitrary shaped text detection and recognition. Based on the extracted features through RoISlide, a CNN and CTC based text recognizer is introduced to make the framework free from labeling the location of characters. The proposed method achieves state-of-the-art performance on two curved text benchmarks CTW1500 and Total-Text, and competitive results on the ICDAR 2015 Dataset.
| [' Cheng-Lin Liu', ' Xu-Yao Zhang', ' Fei Yin', ' Wenhao He', 'Wei Feng'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['text-spotting'] | ['computer-vision'] | [ 0.3557838 -0.3645509 0.05119157 -0.24534364 -0.7214719 -0.7466567
0.7308 -0.10035124 -0.37480918 -0.12115435 0.09267446 -0.42694512
0.3718867 -0.54760784 -0.47344688 -0.65645045 0.9243107 0.6613372
0.32344556 0.0173509 0.5271839 0.46774352 -0.914036 0.5538726
0.7814977 0.989852 0.29604912 0.70912886 -0.5591438 0.3940118
-0.5481156 -0.5065592 0.25485134 -0.2413036 -0.28084263 0.59178233
0.47625044 -0.46417454 -0.65441954 0.8813234 0.5685152 -0.24763939
1.0052478 -0.7043916 -0.50506383 0.817945 -0.9688425 -0.15070151
0.21076505 0.07209083 1.0766454 -1.5456476 0.4258033 1.1067725
0.73183924 0.35853723 -0.7750407 -0.56249666 0.15674916 -0.33622244
-1.6001387 -0.22481959 0.6848646 -0.5790747 0.78957504 0.47731212
0.2606264 0.8457939 0.22124012 1.5534332 0.5461023 -0.6635173
-0.12436823 -0.06808651 0.05165637 0.82683814 0.16474435 -0.3872211
-0.4609567 0.11753722 0.67054504 0.38149875 -0.18898073 -0.3333316
-1.6006916 0.72057664 0.05726494 0.13374297 0.27244207 -0.13753085
0.75168884 -0.0831022 0.54911923 0.00798716 -0.0803391 -0.02008254
-1.245615 -0.1411366 0.48560384 1.1754279 0.27996498 -0.03890258
-0.6916834 0.97273535 0.38915637 1.0912217 0.5509325 0.53783333
1.0987821 1.1133691 -0.05073463 -0.7871236 -0.6130762 -0.39039648
-1.0302509 -0.2193264 0.33043128 -0.21414721 -1.257902 0.52025115
0.2072632 -0.30126226 -0.22296703 0.9096785 0.74105555 0.8449811
-0.59843916 0.34510273 1.5501143 -1.202413 -0.6496795 -0.22834186
1.074419 -1.1010118 1.2333939 0.51965463 -0.57228756 -0.28660735
-1.1195545 -0.45791453 -0.4916947 1.3640981 -0.12870647 0.7803816
-0.5649082 0.02721771 -0.92282635 -0.51021934 0.42690486 0.1912733
0.09366015 0.24418941 -0.6264698 0.5588532 0.45298544 0.30793068
-0.39126027 -0.09650934 -0.74332666 0.14207228 0.49812126 -0.22037113
0.8723152 -0.6769694 -1.5480595 0.9406652 -0.06972075 -0.27201813
1.1529778 -0.3495503 -0.48610044 0.16768797 0.09380191 0.40499628
1.3653815 -0.7430932 -0.86173975 -0.39004338 -0.8757635 0.37716323
-0.6490049 0.10362972 -0.8852087 -1.0634271 0.27625173 -0.6524155
0.27163735 0.3190398 -1.2025651 -0.32443345 1.5048066 -0.63456833
1.381083 -2.274693 -0.1555572 0.22211592 0.37871826 0.24377549
0.059964 0.47258827 0.29433054 0.12118137 -0.12001432 -0.6625168
0.34019 -0.30821362 -0.7433714 0.67058885 0.1797202 1.0318013
-0.42885894 -0.77008307 0.72328687 0.26558593 -0.07741242 -0.12052788
-0.28660658 -0.06660563 -0.7535623 0.6376031 0.9551641 -0.09556375
-0.02993645 -0.10279993 -0.41817245 -0.06163255 -1.0995196 1.2518601
-0.20363125 1.0506269 -0.2202938 -0.5775581 1.2750626 0.09517099
0.15571521 -0.4702744 0.4611558 0.24280822 -0.392057 -0.44409126
0.7378067 0.4953129 -0.17408976 0.48134732 -0.3933892 -0.20722689
0.10977986 0.08859073 0.94506717 0.10006998 0.01563299 -0.22852379
0.60447115 -0.07949255 -0.03436912 0.79738367 0.18254013 0.99667335
0.5196176 -0.57540375 -1.2904972 -0.7125 -0.38784328 1.3077991
0.07677031 -0.4744069 -0.75134706 -0.9287673 -0.05457065 0.51591426
-0.71773523 0.32356563 -0.71742177 -0.5605879 0.9124224 0.47400308
0.7287034 -1.0523523 -0.688722 -0.01540538 -0.08560269 -1.357771
-1.4233104 0.34483778 -0.6632839 -0.8331095 -1.2532924 -1.1934358
0.9896035 0.2753031 0.47951612 -0.14649424 -0.42880282 0.02725181
-0.688008 -0.23966564 -0.14170657 0.35481876 -0.46613488 0.5427081
0.18723527 0.30269486 -0.6846263 0.7121196 -1.0909822 0.5507609
0.8199811 0.9768583 0.4144515 -0.09222312 -0.03801748 -0.70465237
0.5759304 0.1946637 -0.5727447 0.4417778 -0.33441055 0.12545708
1.0556158 -0.54935884 -0.6783877 0.5470599 -0.04995199 -0.44706
-0.0456173 0.28166917 -0.09965998 0.01306538 0.61882406 1.1359928
-0.44316667 -0.42328903 0.43749106 1.1955997 0.414258 -0.2654702
0.8431964 0.7816087 -0.26839557 -1.1570219 -0.65326333 -0.8284818
-1.0358243 -0.01420139 0.7841773 -0.5590387 -0.6245852 0.8758665
-1.0173352 -0.32317272 0.08790109 0.0426225 -0.32603922 0.76709116
-0.4767511 -0.8087341 -0.9039122 -1.0322269 1.9300206 0.05013769
0.24526548 -0.81815106 -0.30592552 0.10106249 0.10492975 -0.0382098
0.710191 -0.8900577 -0.15309289 -0.64374727 -0.53466356 -0.0969474
0.04401968 0.1962869 -0.67274487 -0.32895225 -0.3505701 -0.15794863
1.1188909 0.10565843 1.2077104 -0.2141491 -0.70826423 0.7858088
1.1667075 0.19408417 0.61203223 0.2293481 1.0410551 0.06229568
0.507411 0.83010703 0.05131088 0.6802547 0.07346027 -0.27355766
-0.0224831 -0.43063632 0.3047554 0.51182514 0.7158314 -0.6328491
-1.0651363 0.3319403 -1.7742137 -0.53499466 -0.456093 1.9619607
0.5586737 0.4332406 0.12202672 0.21890935 1.2273139 0.15926026
-0.7792317 -0.12480583 -0.2330964 -0.0257153 0.6804501 0.15082072
-1.4104435 1.2111635 4.941246 1.4121703 -1.4419373 -0.39492482
0.5453297 0.12976156 0.24722761 -0.38268277 -1.1467605 0.52722037
0.07590242 0.03634748 0.18479678 0.6552445 0.45758647 0.15613447
-1.0640918 1.2727997 0.40362743 -1.160009 0.31771284 -0.25598997
0.58702993 -0.10394051 0.24968554 0.23575096 -0.08529866 -0.9551878
1.0371655 0.34822276 1.2526282 -0.48262855 0.48445755 0.45790952
-1.6177235 0.06985287 -0.39118752 0.6095095 -0.14182918 0.5303921
-1.2084135 0.42182237 0.21200037 0.81943065 -0.78892636 1.2455477
-0.11813416 0.75254697 -0.46041003 -0.7071158 0.32019016 -0.09453082
0.4103413 1.7218026 0.49384692 -0.25465742 0.27847344 0.8844133
-0.28526902 0.522086 -0.3280715 -0.17424048 0.3087622 1.5019777
-1.4747163 -0.34701747 -0.30477533 1.3923246 -0.06056947 0.07727937
-0.8046175 -1.0440102 -0.44535297 -0.00610822 0.7505201 -0.26151228
-0.70565903 -1.3507938 0.33686936 -0.7876875 0.2835131 -0.90282714
-1.0054185 0.40886587 -0.695165 -1.5099496 0.4306574 -0.93144494
-0.8110412 0.60127723 -1.0312824 -1.579665 -0.4429686 0.57067764
1.0598265 -0.03946369 0.30007634 0.03238503 -0.7900483 0.9655254
0.79117435 0.98200953 0.9018465 -1.1975029 0.9985201 0.77684534
0.3335541 0.23009588 0.3976125 -0.8558496 -1.5326643 -1.3724756
0.4717587 -0.37068522 0.67774826 -1.0947618 -0.7027145 0.692691
-0.05630853 -0.03794416 -0.06955661 -0.39288056 -0.31936812 0.21371475
-0.8303192 0.8319333 0.7346182 -0.33909267 -0.38681462 0.66955745
0.42600054 -0.7518041 -0.12566428 0.04341893 0.5780773 -0.5922882
0.5317708 -0.01918035 0.33064556 -0.35901624 0.24548563 -0.7809194
-0.01845759 -0.70987976 0.17648835 1.070158 0.5872564 -0.5974331
1.0984138 -0.03543326 -0.38401365 -0.5744625 -0.9471452 -0.657487
0.22753054 -0.3521547 0.32142976 0.75396556 0.24172822 0.34936783
-0.49309337 -0.01488825 0.29697448 0.22243327 0.7703342 -0.9814747
0.09776497 -0.866697 -0.2736634 -1.8232222 -0.16550589 -1.0187999
0.25372237 -1.4284513 0.31752947 -0.12910315 0.34914392 0.4877425
-0.07228242 0.13138096 0.18757014 0.27314243 -0.88068736 0.7147363
1.2846895 -0.47379956 -0.2267226 0.11108348 -0.185932 0.65134585
0.5603702 -0.28761458 0.07335442 -0.53155303 0.23494425 0.03940143
0.14193328 -0.86117196 0.62620366 0.1706695 0.5923426 -1.5509157
-0.01177724 -0.7811066 -0.736719 0.2284811 -0.56294906 -0.13269363
0.14874668 0.39999995 0.27335528 -0.43671286 0.773237 0.526239
-0.23267661 0.24644354 -0.4835951 0.07495376 0.95687115 -0.3310348
-0.4630854 -0.08188368 -0.43417138 0.43055493 0.36336228 0.40595636
0.8628437 -1.1217406 -0.792878 0.32717907 0.38770416 0.16662478
0.04878056 0.7575017 -0.9076333 0.83558875 0.4068339 -0.8751556
-1.2811332 0.52305675 0.5841431 -0.38093683 -1.2805929 0.49543428
0.35109362 -0.50109875 0.42061722 -0.7334435 -0.07958315 -0.14063764
0.43507764 0.16325563 0.14372759 -0.4267265 -0.12304109 1.0169921
-0.42281246 -0.20616566 0.6868637 -0.13470033 0.26295006 0.31505638
0.9226016 0.27903736 -1.3813678 -0.37852502 0.08301559 -0.33246955
-0.03885124 -0.964921 -0.90331507 1.1918746 0.582625 0.19280125
0.78342235 -0.19571832 0.89901054 0.56246895 -0.09979127 -1.2756486
0.27658302 0.7743572 0.922043 -1.1026298 -0.0531326 -0.3692271
-0.7007346 1.7917453 0.501501 0.0339506 0.33440858 0.61469895
-0.11602052 -0.1355677 -0.34485266 -0.16576527 0.42330706 0.19412981
0.27979016 0.10308598 -0.0217753 0.40957075 -0.13584748 -0.3197157
0.49217665 0.72561634 -0.7981564 -0.6665648 -0.6794178 0.8781938
-0.43636584 -0.39485887 -0.7350988 0.7805787 -0.34347814 0.8374872
0.24747281 -0.43429857 0.29287466 0.07605083 0.00997508 -0.5875127
-0.4528815 0.57609403 -0.3230455 0.26843157 0.49887422 -0.67561495
-1.5406463 -0.12030479 -0.8255564 -0.102721 0.59099454 0.97019553
0.20938127 0.3778502 0.8127479 -0.6610267 -0.55588585 -1.19431
-0.7217358 0.20859937 0.22634403 -0.15945429 -0.25180614 0.21519896] | [11.999177932739258, 2.266062021255493] |
bec562d2-6b3c-483b-ae86-da77b2744343 | application-of-adnn-for-background | 2301.00264 | null | https://arxiv.org/abs/2301.00264v1 | https://arxiv.org/pdf/2301.00264v1.pdf | Application Of ADNN For Background Subtraction In Smart Surveillance System | Object movement identification is one of the most researched problems in the field of computer vision. In this task, we try to classify a pixel as foreground or background. Even though numerous traditional machine learning and deep learning methods already exist for this problem, the two major issues with most of them are the need for large amounts of ground truth data and their inferior performance on unseen videos. Since every pixel of every frame has to be labeled, acquiring large amounts of data for these techniques gets rather expensive. Recently, Zhao et al. [1] proposed one of a kind Arithmetic Distribution Neural Network (ADNN) for universal background subtraction which utilizes probability information from the histogram of temporal pixels and achieves promising results. Building onto this work, we developed an intelligent video surveillance system that uses ADNN architecture for motion detection, trims the video with parts only containing motion, and performs anomaly detection on the trimmed video. | ['Neeraj Goyal', 'Gagan Raj Singh', 'Piyush Batra'] | 2022-12-31 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 4.32206362e-01 -4.11341488e-01 -3.45332436e-02 -2.65043676e-01
-1.37882277e-01 -1.94716185e-01 3.67621690e-01 -8.44803900e-02
-6.35454476e-01 6.29094124e-01 -3.72876167e-01 -3.63542497e-01
3.53529751e-01 -8.59552979e-01 -6.34294689e-01 -1.07218206e+00
7.82484636e-02 3.42402495e-02 8.56921852e-01 1.20392081e-03
1.27126023e-01 4.07771111e-01 -1.58743215e+00 2.69065529e-01
6.83109462e-01 1.38717902e+00 6.99944571e-02 7.55083084e-01
-2.48444274e-01 1.08575845e+00 -7.56139815e-01 -2.25813434e-01
4.95371342e-01 -5.65419137e-01 -3.94463271e-01 2.31039137e-01
6.62154675e-01 -7.67506421e-01 -5.34217596e-01 1.58938968e+00
3.28294128e-01 1.51049197e-01 4.38722372e-01 -1.40234077e+00
-4.38857615e-01 2.55705148e-01 -7.31988490e-01 9.90556896e-01
-1.92066357e-01 1.45804912e-01 1.87800884e-01 -5.43223441e-01
2.68051982e-01 1.06571019e+00 6.81478798e-01 6.38317227e-01
-5.91672897e-01 -5.18678784e-01 1.90151468e-01 7.66679227e-01
-1.19916058e+00 -2.40675539e-01 7.32665658e-01 -5.10626256e-01
4.69052166e-01 1.01393774e-01 7.03054726e-01 1.07386029e+00
2.17239827e-01 9.68092620e-01 5.81203759e-01 -3.16450864e-01
3.09663862e-01 -2.29385331e-01 3.30974489e-01 5.27498841e-01
5.23904085e-01 -1.32090569e-01 5.01708463e-02 1.23661883e-01
7.05180049e-01 4.69071150e-01 -4.19800937e-01 -3.35433930e-01
-9.75021124e-01 5.32921910e-01 1.96104631e-01 3.65786970e-01
-4.78612751e-01 -3.05742938e-02 5.25320590e-01 8.31729844e-02
2.97110707e-01 -2.86929250e-01 -4.95557845e-01 -2.37941407e-02
-1.04791057e+00 5.90660796e-02 3.71758848e-01 6.10268891e-01
5.63576400e-01 4.38619941e-01 -1.30456954e-01 3.22821617e-01
1.64495349e-01 5.96037805e-01 7.78617620e-01 -8.23964119e-01
1.51701853e-01 8.70235682e-01 1.49123639e-01 -1.54402649e+00
-2.78530359e-01 -1.47363394e-01 -1.16699505e+00 4.33003783e-01
7.52715886e-01 -3.66435409e-01 -1.13281620e+00 1.46512043e+00
4.24790442e-01 4.97797668e-01 -2.92786565e-02 1.09699321e+00
7.05641866e-01 8.97431910e-01 5.78611456e-02 -4.05634165e-01
9.93928015e-01 -9.52015817e-01 -9.41598475e-01 -2.94946432e-01
2.97010064e-01 -4.90321398e-01 4.95039165e-01 6.52194858e-01
-8.12091470e-01 -8.31818283e-01 -9.29199398e-01 1.25173986e-01
-4.65571612e-01 2.45321631e-01 5.38390815e-01 6.66127384e-01
-8.67654502e-01 5.10078728e-01 -9.12905872e-01 -3.89812499e-01
4.76674020e-01 3.90616953e-01 -1.85089767e-01 -6.93266913e-02
-1.06858242e+00 6.68360174e-01 8.10626864e-01 4.90322948e-01
-1.08762491e+00 2.03333553e-02 -6.90438569e-01 -4.46485430e-02
5.33584237e-01 -3.79876912e-01 1.09786439e+00 -1.72953618e+00
-1.10567987e+00 6.75981402e-01 6.88203648e-02 -6.43666208e-01
6.00268960e-01 -4.52713549e-01 -7.00495601e-01 1.23957291e-01
-1.15002267e-01 4.33754385e-01 1.20902383e+00 -7.32214272e-01
-1.05976057e+00 -5.30868888e-01 -1.93046778e-01 -2.07203284e-01
-4.72976863e-01 1.05525762e-01 -4.78606999e-01 -7.41538107e-01
1.09543562e-01 -5.81666291e-01 -1.49294347e-01 2.51499321e-02
-6.83560669e-02 -2.78087854e-01 1.47265542e+00 -1.14656591e+00
1.32921398e+00 -2.28640413e+00 1.75493732e-02 -2.02291235e-01
1.06510334e-01 8.58472705e-01 4.33732122e-02 -4.35296357e-01
-3.52368131e-02 -3.57308775e-01 -4.25732791e-01 4.15738940e-01
-4.31884766e-01 1.34591267e-01 -1.14413418e-01 6.07021093e-01
3.04105103e-01 4.85012650e-01 -7.73244083e-01 -5.87239563e-01
4.51548785e-01 3.16927642e-01 -3.06985587e-01 1.37464628e-01
-2.81274825e-01 4.21719968e-01 -3.66925240e-01 7.00779676e-01
9.77565467e-01 1.43489912e-02 -1.07841045e-01 -2.84387469e-01
-1.32879168e-01 -4.92843390e-01 -1.30541492e+00 1.20827091e+00
3.85620117e-01 8.74482870e-01 1.67543679e-01 -1.39646757e+00
6.46007538e-01 2.59074569e-01 5.88820577e-01 -4.58244562e-01
5.99543333e-01 1.90058071e-02 1.63514614e-01 -1.02443683e+00
4.22563404e-01 1.68216541e-01 3.68183672e-01 4.98381071e-03
-9.59329307e-02 5.13372719e-01 3.29037458e-01 -2.99509078e-01
1.16055810e+00 2.60059029e-01 2.21984193e-01 5.81442006e-03
8.58717680e-01 3.04679900e-01 1.11276495e+00 6.38897061e-01
-7.14950144e-01 5.12858152e-01 9.61262435e-02 -9.19869721e-01
-9.51974869e-01 -7.36423373e-01 1.49435177e-01 9.73269284e-01
3.47864360e-01 3.67667049e-01 -1.05277836e+00 -6.34841859e-01
-3.80009830e-01 4.22785670e-01 -4.86787140e-01 -1.57343879e-01
-8.68290246e-01 -1.16224444e+00 4.57421362e-01 6.70150101e-01
1.06784797e+00 -1.04917622e+00 -8.27654839e-01 1.22989245e-01
-2.50700146e-01 -1.12401032e+00 -1.59194067e-01 -1.50044844e-01
-9.61449325e-01 -1.36272073e+00 -7.78459489e-01 -9.36690271e-01
7.03235447e-01 5.32181144e-01 7.94063032e-01 7.60148047e-03
-4.54168379e-01 6.39138976e-04 -3.28308344e-01 -5.16961157e-01
-4.53498244e-01 -3.59284967e-01 1.36951536e-01 3.83404225e-01
8.24019551e-01 -1.61723018e-01 -7.04008520e-01 2.79549897e-01
-1.14920926e+00 -2.16690272e-01 8.19283783e-01 5.45117915e-01
3.47534239e-01 5.61274409e-01 2.89138764e-01 -3.79274726e-01
4.56638038e-02 -3.72908294e-01 -9.01280999e-01 2.44273260e-01
2.17809305e-02 -3.38301808e-01 5.84124327e-01 -5.59923232e-01
-1.02012134e+00 1.54670656e-01 2.13720240e-02 -5.75465620e-01
-4.87417161e-01 -2.43755020e-02 -4.36100453e-01 1.70703176e-02
3.66482913e-01 4.21478271e-01 -7.82319461e-04 -3.24868977e-01
1.01809774e-03 6.81560814e-01 8.39082658e-01 1.19473994e-01
5.95967829e-01 6.88658834e-01 -1.86164498e-01 -1.18003631e+00
-6.69912279e-01 -4.46780354e-01 -6.58129454e-01 -4.01128232e-01
1.41181898e+00 -6.80874884e-01 -4.75904197e-01 1.11324489e+00
-1.17996931e+00 -1.37494490e-01 1.19269211e-02 5.69163382e-01
-1.03031307e-01 8.27857912e-01 -5.06394804e-01 -1.02995455e+00
-2.19159216e-01 -1.00931358e+00 6.92176342e-01 5.72086990e-01
1.62428483e-01 -7.37262845e-01 -1.29086211e-01 2.68286973e-01
4.58291441e-01 4.20003861e-01 7.14528203e-01 -6.03637338e-01
-8.77345562e-01 -4.80622351e-01 -2.84361660e-01 7.55745530e-01
3.01445633e-01 2.90187508e-01 -8.07898700e-01 -1.60196602e-01
3.95204008e-01 1.22348443e-01 9.87128854e-01 8.33407044e-01
1.32029045e+00 -2.03965873e-01 -3.66804868e-01 5.58219135e-01
1.22693336e+00 8.11502337e-01 7.10883617e-01 6.24594569e-01
7.68171728e-01 2.97506928e-01 6.13976419e-01 2.63920426e-01
-9.44445580e-02 3.64836484e-01 6.70902133e-01 -2.83166282e-02
1.41546294e-01 3.55294198e-01 5.58311284e-01 5.93086660e-01
-2.49739453e-01 -3.56340021e-01 -8.15981686e-01 5.86219966e-01
-1.87853014e+00 -1.33904743e+00 -2.48392001e-01 2.12059546e+00
4.22463924e-01 2.44063675e-01 1.46042556e-01 3.58031899e-01
1.19654226e+00 3.41418870e-02 -6.50068521e-01 1.91728637e-01
-3.52676928e-01 -1.93204731e-01 4.46711659e-01 3.75569835e-02
-1.73982477e+00 7.34488010e-01 5.81049442e+00 6.89630032e-01
-1.24451852e+00 -2.37545334e-02 8.38737965e-01 9.56617519e-02
6.18812680e-01 -3.95073175e-01 -7.07639933e-01 8.01847339e-01
6.57942355e-01 2.05980033e-01 1.33730039e-01 1.01267874e+00
2.96943128e-01 -4.30190951e-01 -7.30712831e-01 1.09985530e+00
2.84109145e-01 -9.54407215e-01 2.28194565e-01 -3.67018491e-01
7.62280166e-01 -1.20957546e-01 -1.14881031e-01 2.77952850e-01
-5.71487360e-02 -7.99023986e-01 4.58092362e-01 5.34776330e-01
4.20557335e-03 -6.33620977e-01 1.24586082e+00 4.84343737e-01
-9.63408172e-01 -3.50417346e-01 -5.33595741e-01 -2.34464198e-01
8.52892101e-02 4.79291022e-01 -3.34891766e-01 3.02014560e-01
1.05522823e+00 4.80805188e-01 -5.81798196e-01 1.29311717e+00
9.91440415e-02 6.39638364e-01 -2.62316942e-01 4.77385409e-02
2.52420187e-01 -3.62105906e-01 5.48525512e-01 1.02196276e+00
4.07439858e-01 9.01429504e-02 2.82035798e-01 4.67751890e-01
1.18209794e-01 -2.13473350e-01 -6.47798955e-01 -1.59166623e-02
1.57987643e-02 1.17459428e+00 -1.00551116e+00 -6.31756186e-01
-5.58109939e-01 1.08640754e+00 -3.70620728e-01 2.53471941e-01
-1.05312479e+00 -2.91991383e-01 4.57844585e-01 4.02204366e-03
4.17837620e-01 -2.40055844e-01 1.56753697e-02 -1.04459763e+00
1.44046731e-02 -9.14795339e-01 5.03524125e-01 -5.90399086e-01
-1.08738029e+00 5.33290386e-01 -1.24057829e-01 -1.30378056e+00
-4.18195575e-02 -1.01887596e+00 -7.58670866e-01 3.37165117e-01
-1.20404530e+00 -7.74658024e-01 -8.89942706e-01 6.67679429e-01
8.22815835e-01 -4.27923679e-01 1.98016614e-01 7.70982504e-01
-1.05578065e+00 1.85855746e-01 1.96134090e-01 7.87083983e-01
3.93033952e-01 -9.01703298e-01 3.28720391e-01 1.48397076e+00
-1.10560007e-01 1.23748221e-01 7.15903759e-01 -6.95281386e-01
-1.22293961e+00 -1.36967206e+00 2.78728515e-01 -3.69903684e-01
4.43535060e-01 1.58248872e-01 -1.11965203e+00 5.87329149e-01
2.03907415e-01 1.74857125e-01 2.66059667e-01 -7.48778105e-01
4.25922275e-01 -1.57988176e-01 -1.04678273e+00 5.08046687e-01
7.97421813e-01 1.25715524e-01 -6.57197118e-01 6.05583489e-02
4.44291592e-01 -3.40777040e-01 -1.83085173e-01 6.55707777e-01
1.73635870e-01 -1.16980743e+00 7.57980525e-01 -5.95438421e-01
1.47604242e-01 -8.62626791e-01 -1.96650282e-01 -7.76512384e-01
-1.46855295e-01 -3.41472983e-01 -2.86184371e-01 1.08914006e+00
-2.39949316e-01 -2.62338609e-01 1.03010178e+00 4.11530465e-01
-9.72318873e-02 -2.02577904e-01 -7.70408690e-01 -7.41686702e-01
-2.89098591e-01 -2.70969719e-01 2.51167327e-01 9.58653033e-01
-7.88475871e-01 6.89088032e-02 -5.39175153e-01 5.30651748e-01
7.18644202e-01 -1.42806992e-01 8.03575277e-01 -1.26018596e+00
7.72002526e-03 -3.88819098e-01 -9.63680804e-01 -8.52637589e-01
-1.22436546e-01 -3.34232569e-01 9.81003270e-02 -1.42643118e+00
1.47860005e-01 1.89000577e-01 -3.96440446e-01 2.67772287e-01
-3.70684445e-01 3.39108020e-01 -1.52544111e-01 -4.13549356e-02
-7.58716464e-01 4.57765520e-01 8.07532370e-01 -4.31339473e-01
-2.08937321e-02 1.47469148e-01 -4.67649028e-02 1.31753337e+00
8.90745878e-01 -2.25091577e-01 -1.21167518e-01 -5.86892664e-01
-3.23290974e-01 -2.23260477e-01 6.65696204e-01 -1.53838110e+00
3.23023707e-01 -3.03324997e-01 8.55313122e-01 -9.99380410e-01
8.43204781e-02 -1.04459715e+00 4.61278744e-02 5.69992304e-01
1.76485941e-01 3.27526718e-01 4.06175971e-01 7.30603576e-01
-4.57602292e-01 -2.12409556e-01 9.72528577e-01 -2.77363896e-01
-1.48059165e+00 3.06209117e-01 -6.29630566e-01 -3.12402118e-02
1.43601131e+00 -4.35878366e-01 -3.59873086e-01 -1.44942865e-01
-5.35004139e-01 -5.91006950e-02 3.10351819e-01 3.82615656e-01
5.61144292e-01 -1.13030410e+00 -4.77532595e-01 2.53557235e-01
-3.56726050e-01 2.04574838e-01 4.04564917e-01 8.49915445e-01
-1.02268672e+00 2.60632306e-01 -5.92331171e-01 -7.43223250e-01
-1.35882342e+00 9.42413807e-01 5.45717895e-01 2.51350671e-01
-5.55529237e-01 7.11858332e-01 2.49146640e-01 2.69568294e-01
4.32540387e-01 -3.20244968e-01 -3.82950425e-01 -5.05418554e-02
9.26555991e-01 6.73536003e-01 -9.26916748e-02 -6.78342283e-01
-3.05450201e-01 4.94665712e-01 -4.82298657e-02 3.76004726e-01
1.08287549e+00 5.29456921e-02 -2.90524989e-01 1.88487336e-01
8.71480525e-01 -3.42351884e-01 -1.38383269e+00 3.05188298e-02
2.04074278e-01 -5.51163077e-01 8.30638260e-02 -2.97492415e-01
-1.34676349e+00 9.78178918e-01 1.20365214e+00 3.76779199e-01
1.47200918e+00 -5.54134607e-01 9.06094730e-01 6.64257944e-01
1.67535692e-01 -1.28753984e+00 6.75777271e-02 3.72103542e-01
2.83164442e-01 -1.56748104e+00 -1.34126693e-01 -8.83518457e-02
-3.53894323e-01 1.23029351e+00 9.06009734e-01 -1.77128881e-01
5.06544590e-01 2.39494577e-01 2.25705236e-01 1.43780723e-01
-1.42258868e-01 -2.54813462e-01 4.49630506e-02 7.11047173e-01
1.13575026e-01 -3.94371778e-01 -1.22560985e-01 3.01934421e-01
3.99215937e-01 2.46471509e-01 4.18884486e-01 1.04364884e+00
-8.73260319e-01 -6.90476954e-01 -7.44290173e-01 4.12905335e-01
-7.56466210e-01 2.10523158e-01 -2.91124880e-01 7.03241825e-01
5.95518947e-01 8.61576855e-01 2.44123250e-01 -2.37714410e-01
5.57369813e-02 7.01633915e-02 1.61103040e-01 -1.27698943e-01
-1.65923491e-01 -1.97004721e-01 -3.63188446e-01 -5.08828878e-01
-7.30095506e-01 -5.65454066e-01 -1.06715131e+00 -2.41703659e-01
-2.85679400e-01 -4.80596311e-02 4.14799362e-01 8.85527372e-01
1.09138906e-01 6.45147204e-01 1.88420773e-01 -8.37382793e-01
-2.71805763e-01 -9.78872776e-01 -4.21471328e-01 5.73484838e-01
5.55574536e-01 -6.38600469e-01 -3.00861478e-01 4.67572182e-01] | [8.822426795959473, -0.7407134175300598] |
24de91cd-5d42-42a2-bf27-1fbf7dbe5283 | sequential-recommendation-with-bidirectional | 2112.0646 | null | https://arxiv.org/abs/2112.06460v4 | https://arxiv.org/pdf/2112.06460v4.pdf | Sequential Recommendation with Bidirectional Chronological Augmentation of Transformer | Sequential recommendation can capture user chronological preferences from their historical behaviors, yet the learning of short sequences (cold-start problem) in many benchmark datasets is still an open challenge. Recently, data augmentation with pseudo-prior items generated by Transformers has drawn considerable attention. These methods can generate pseudo-prior items sequentially in reverse chronological order to extend the original sequences. Nevertheless, the performance may still dramatically degrade in very short sequences; most notably, the generation of pseudo-prior items does not take into account the forward direction (from the past to the future), and so the underlying temporal correlations are not preserved in terms of conditional probabilities. Motivated by this, we propose a Bidirectional Chronological Augmentation of Transformer (BiCAT) that uses a forward learning constraint in the reverse generative process to capture contextual information more effectively. Then, self-knowledge distillation is adopted in augmented and original sequences to bridge the gap between data augmentation and model representation, which enhances the robustness of sequence encoder. Moreover, an informative positive and negative sampling strategy is proposed to accelerate optimization and prevent overfitting. Extensive experiments on two popular real-world datasets demonstrate the efficacy of our method on very short sequences (L < 3) and long sequences (20 < L < 50) as well, our approach outperforms state-of-the-art models by an average of 35.04% and 8.76% respectively, in terms of Recall@5. Source code is available at https://github.com/juyongjiang/BiCAT. | ['Jae Boum Kim', 'Yingtao Luo', 'Sunghun Kim', 'Kai Zhang', 'Juyong Jiang'] | 2021-12-13 | null | null | null | null | ['self-knowledge-distillation'] | ['computer-vision'] | [ 3.29825461e-01 -3.83922756e-01 -6.58497512e-01 -2.87333548e-01
-4.22143877e-01 -5.47680736e-01 5.55547297e-01 5.59313186e-02
-4.60087240e-01 1.00004399e+00 5.74084640e-01 -2.08176836e-01
-2.49355566e-02 -7.49984443e-01 -7.80541241e-01 -6.93131447e-01
9.37998965e-02 2.78380573e-01 -1.22710718e-02 -3.82234544e-01
8.03119838e-02 -2.48636425e-01 -1.28794277e+00 2.61095345e-01
1.11995602e+00 8.63689959e-01 4.77899104e-01 2.96883494e-01
-9.86453295e-02 7.72334278e-01 -8.58707950e-02 -4.54157203e-01
1.55365705e-01 -4.98664498e-01 -4.72573519e-01 -3.22139636e-02
-5.08947968e-02 -2.83664078e-01 -4.35320050e-01 8.31607997e-01
3.21251690e-01 4.00130093e-01 3.45792472e-01 -9.03070509e-01
-9.93727803e-01 9.76272821e-01 -6.90290570e-01 2.12252408e-01
3.75582993e-01 2.44719848e-01 1.22281206e+00 -8.06804478e-01
2.66650677e-01 1.01509869e+00 6.45368278e-01 4.86943245e-01
-1.06273997e+00 -7.86120951e-01 7.92075872e-01 3.93879503e-01
-1.21755636e+00 -7.97198191e-02 7.51626968e-01 -3.62734169e-01
5.57258725e-01 3.68743747e-01 8.01815331e-01 1.59004605e+00
-2.28460997e-01 1.09215045e+00 7.92093039e-01 -1.67142361e-01
5.68279251e-02 3.83750908e-02 2.45701492e-01 2.63886541e-01
1.36564136e-01 2.08185822e-01 -5.94371140e-01 3.11940666e-02
7.45581090e-01 4.24870223e-01 -3.04395169e-01 2.71876603e-02
-1.21724534e+00 6.59674466e-01 1.42267540e-01 2.45001078e-01
-5.60021758e-01 -8.13025832e-02 5.00163972e-01 1.92122087e-01
4.08391327e-01 2.63229102e-01 -6.34808421e-01 -4.21020657e-01
-6.98841035e-01 2.24942178e-01 4.08132762e-01 1.21510005e+00
5.20744801e-01 7.68572688e-02 -4.28311616e-01 9.32982981e-01
1.55583084e-01 3.44848156e-01 7.71218538e-01 -3.67748737e-01
6.88548923e-01 3.59936625e-01 4.11675304e-01 -8.15784514e-01
-1.91405527e-02 -9.21752155e-01 -1.08445752e+00 -8.77278149e-01
2.01085076e-01 -1.66207299e-01 -7.78839290e-01 1.89017284e+00
2.46866599e-01 6.87116444e-01 -1.91662192e-01 8.67845356e-01
3.84518892e-01 9.66705024e-01 1.48919567e-01 -5.63451827e-01
1.40123904e+00 -1.02001011e+00 -7.94983685e-01 -4.95912910e-01
4.53737497e-01 -7.30586827e-01 1.41376591e+00 4.38894480e-01
-8.76912415e-01 -8.32840025e-01 -8.75278473e-01 2.66433537e-01
7.67289102e-02 2.46224597e-01 5.55622041e-01 4.98432577e-01
-3.66820365e-01 4.44595188e-01 -8.09001982e-01 -9.47537050e-02
2.82053620e-01 1.19221531e-01 6.39933944e-02 -2.15201035e-01
-1.69418561e+00 4.61027741e-01 8.73461366e-01 8.59597623e-02
-7.75292277e-01 -9.14145529e-01 -6.23622119e-01 -2.07224637e-02
6.98836148e-01 -5.97372949e-01 1.28773916e+00 -8.42281818e-01
-1.49437058e+00 1.92226306e-01 -2.07611606e-01 -5.34289718e-01
6.07397079e-01 -7.53518224e-01 -7.18394101e-01 -3.82154465e-01
-1.35096163e-01 1.95981413e-01 6.25703931e-01 -9.45624769e-01
-8.25091183e-01 -9.01494548e-02 -2.67783161e-02 4.21498090e-01
-8.49215686e-01 -3.60497713e-01 -8.69852245e-01 -1.23669517e+00
-2.92658091e-01 -1.09782743e+00 -4.44377482e-01 -5.90249479e-01
-3.49698573e-01 -5.72810099e-02 4.57235903e-01 -8.37204337e-01
1.85240459e+00 -2.11295891e+00 -5.07249497e-02 9.53372419e-02
-3.00680369e-01 3.76003563e-01 -2.02764302e-01 5.63479722e-01
-5.09262644e-02 1.05802871e-01 -2.75127649e-01 -2.90723711e-01
-1.62745029e-01 3.93334836e-01 -4.69815463e-01 1.35508299e-01
-1.80099174e-01 8.02577257e-01 -1.24054396e+00 -2.30105609e-01
1.19545095e-01 3.49375993e-01 -6.40634000e-01 2.18541577e-01
-4.93526995e-01 7.34045684e-01 -4.49536234e-01 2.62420058e-01
7.11348951e-01 -4.51541811e-01 4.17229801e-01 -5.69365099e-02
-9.26426426e-02 5.85253239e-01 -1.11144638e+00 1.90983796e+00
-5.49944520e-01 5.71046863e-03 -7.03896284e-01 -8.10199142e-01
9.13691401e-01 4.47602689e-01 5.90345263e-01 -8.12018037e-01
-1.57460630e-01 3.67678329e-03 -7.98003525e-02 -4.36920375e-01
6.49825931e-01 -1.51997104e-01 -2.46086344e-02 4.68373030e-01
-2.78161049e-01 6.32087708e-01 4.15415734e-01 1.50943995e-01
6.64123714e-01 4.01742280e-01 2.58832216e-01 5.60669601e-02
6.21002614e-01 -3.87849033e-01 1.03974044e+00 5.62947035e-01
1.36533558e-01 4.96816665e-01 1.76885232e-01 -3.09653044e-01
-1.17068136e+00 -8.49188507e-01 2.84951270e-01 1.10405982e+00
2.77901977e-01 -7.06733644e-01 -3.17821920e-01 -7.18275726e-01
-1.38539851e-01 9.55361247e-01 -6.69829726e-01 -3.66873503e-01
-7.95828104e-01 -1.11506259e+00 2.06230134e-01 7.40791321e-01
3.31061661e-01 -8.95589054e-01 -1.44948483e-01 4.73469436e-01
-6.76825285e-01 -1.06992221e+00 -9.34341073e-01 -1.35828957e-01
-1.01507175e+00 -6.44843519e-01 -7.40827918e-01 -5.50308347e-01
4.52269912e-01 2.54760116e-01 9.66050982e-01 -4.90487330e-02
2.96382010e-01 -2.06516474e-01 -7.73145556e-01 -9.33992043e-02
-2.32230827e-01 3.46377075e-01 7.92526752e-02 2.37432003e-01
3.76491129e-01 -8.93322349e-01 -9.08624113e-01 5.03035367e-01
-8.76824379e-01 4.23810989e-01 6.50851250e-01 1.03665674e+00
5.22294343e-01 -7.61146620e-02 7.33486831e-01 -1.02123272e+00
4.43717331e-01 -8.40504706e-01 -3.27920824e-01 2.84875810e-01
-6.93315387e-01 1.39871657e-01 9.76967096e-01 -7.72624433e-01
-1.37213969e+00 -1.75008148e-01 -3.20459247e-01 -3.81283343e-01
6.87663704e-02 7.07804501e-01 -1.76090360e-01 8.32996726e-01
3.54834408e-01 7.76657641e-01 -2.16492832e-01 -7.60589600e-01
4.23754632e-01 4.22540396e-01 4.03724045e-01 -5.90431809e-01
6.35616422e-01 3.01336110e-01 -4.79351968e-01 -4.37552780e-01
-9.08656299e-01 -4.63149816e-01 -3.57596248e-01 1.12650782e-01
2.26844683e-01 -1.08722043e+00 -5.76581538e-01 2.65714467e-01
-7.35128701e-01 -2.14818671e-01 -2.22128406e-01 6.22893035e-01
-3.81271690e-01 6.20957851e-01 -7.13967741e-01 -8.72728944e-01
-5.38785398e-01 -7.89456606e-01 4.74431574e-01 2.07316846e-01
-2.66602844e-01 -8.82994652e-01 -3.18607246e-03 2.18514368e-01
2.24862143e-01 -8.16225410e-02 9.31853175e-01 -6.84452832e-01
-5.12071311e-01 -6.31500110e-02 1.30403891e-01 4.29064840e-01
4.04753864e-01 -2.92162448e-01 -5.99038303e-01 -4.05027002e-01
-1.03219502e-01 -9.90197528e-03 7.65557468e-01 7.60098249e-02
1.26607335e+00 -6.12549484e-01 -3.29126686e-01 3.06842864e-01
1.36449862e+00 5.75525343e-01 6.93515241e-01 1.80808306e-01
8.02607477e-01 3.24970305e-01 1.03297651e+00 9.84929621e-01
4.15948004e-01 6.35503709e-01 3.34731013e-01 3.36379200e-01
7.61490017e-02 -6.92201793e-01 4.84236926e-01 1.17488873e+00
-1.40933111e-01 -5.01795232e-01 -5.63908398e-01 6.98822618e-01
-2.12950635e+00 -1.11037397e+00 -5.41573055e-02 2.53735971e+00
1.18085277e+00 3.10094118e-01 3.18398565e-01 1.53505087e-01
7.24909961e-01 1.90376893e-01 -5.64267755e-01 2.16561720e-01
1.11331660e-02 -1.05919205e-01 3.97044986e-01 2.46864289e-01
-9.45906341e-01 7.57975280e-01 4.64912081e+00 1.13678682e+00
-9.54276443e-01 1.79237127e-01 7.61999428e-01 -2.54826367e-01
-6.02677882e-01 1.24221653e-01 -8.88409078e-01 1.19953215e+00
8.22143495e-01 -1.26927793e-01 5.27789235e-01 6.34680569e-01
4.45884973e-01 2.23930046e-01 -9.70668614e-01 9.33886766e-01
-1.69801831e-01 -1.14897096e+00 1.60868168e-01 -1.22230344e-01
8.67356896e-01 -2.20795810e-01 2.40108177e-01 4.51413125e-01
3.70900095e-01 -6.07420921e-01 7.09865928e-01 7.08370686e-01
6.67392254e-01 -1.02342975e+00 5.47760129e-01 6.81305110e-01
-1.40862286e+00 -2.61173904e-01 -1.40843123e-01 -1.11291520e-01
4.91853267e-01 7.79733419e-01 -8.27644169e-01 9.92909610e-01
5.50003827e-01 1.14416468e+00 -3.38096410e-01 9.54446554e-01
-2.94497758e-01 1.04286492e+00 -1.91451997e-01 -5.29066883e-02
2.19651550e-01 -3.34043682e-01 4.07630116e-01 1.19984090e+00
4.31361169e-01 7.47190043e-02 2.99584717e-01 3.66918147e-01
-3.94658707e-02 1.53474480e-01 -1.28633067e-01 -6.92817196e-02
7.59482503e-01 9.08327341e-01 -3.51184964e-01 -4.46137786e-01
-4.66692865e-01 8.55409861e-01 1.31086200e-01 4.25792634e-01
-1.12968957e+00 -6.12769723e-02 6.09080791e-01 8.58562887e-02
4.61280555e-01 -7.68404901e-02 -3.05217415e-01 -1.41128731e+00
1.73077017e-01 -1.10345948e+00 5.59698880e-01 -3.70715767e-01
-1.36919248e+00 4.56610620e-01 -1.42148718e-01 -1.60384345e+00
-2.44327947e-01 -1.47931688e-02 -3.60193759e-01 7.79735863e-01
-1.32964051e+00 -9.76792037e-01 -1.79141670e-01 3.66643310e-01
8.08300614e-01 9.05218944e-02 4.99263942e-01 6.90273166e-01
-7.12349594e-01 7.96043992e-01 3.76199484e-01 -4.95735668e-02
6.23821497e-01 -9.43802536e-01 6.27678990e-01 1.00469029e+00
7.66198039e-02 9.89722908e-01 6.93303049e-01 -8.40395033e-01
-1.27341616e+00 -1.28856421e+00 7.24313974e-01 -2.15135992e-01
4.33875263e-01 -4.21136945e-01 -1.07602942e+00 7.13421702e-01
2.47426197e-01 -2.74080068e-01 8.69298518e-01 2.06388712e-01
-3.42222840e-01 -2.64111727e-01 -5.61137676e-01 9.05663610e-01
1.37532008e+00 -2.68444687e-01 -3.15887928e-01 2.02413574e-01
8.67830932e-01 -2.31732845e-01 -8.50804150e-01 5.17700076e-01
7.65443683e-01 -6.95680916e-01 1.04887545e+00 -5.65802813e-01
4.41098422e-01 -3.63112211e-01 -1.06385127e-01 -1.26088488e+00
-5.74385881e-01 -8.32689881e-01 -6.01774037e-01 1.37540901e+00
2.67059952e-01 -4.86393929e-01 7.82457352e-01 1.38436586e-01
-1.14398994e-01 -1.01925552e+00 -3.45210135e-01 -9.84947741e-01
-2.76706994e-01 -5.25145233e-01 9.42335486e-01 9.26795304e-01
-1.58730187e-02 4.66746151e-01 -1.14635599e+00 -4.00903225e-02
3.27581823e-01 2.93725342e-01 6.83294058e-01 -8.77212167e-01
-5.82188129e-01 -2.23918706e-01 2.89806843e-01 -1.66161084e+00
-2.62611568e-01 -7.09928274e-01 -1.27160354e-02 -1.23672652e+00
2.11811170e-01 -6.31831825e-01 -6.98684335e-01 3.50604236e-01
-6.25010312e-01 9.82734933e-03 1.56135872e-01 3.33238363e-01
-4.77662325e-01 9.81183112e-01 1.33647859e+00 1.43355072e-01
-2.13525504e-01 2.34364405e-01 -8.46446335e-01 4.83009368e-01
9.60582852e-01 -3.63047600e-01 -8.86121333e-01 -3.16751003e-01
3.34720463e-01 1.10996969e-01 -6.05586804e-02 -7.42291689e-01
2.53400620e-04 -4.72912461e-01 1.86921641e-01 -7.46001303e-01
3.22208673e-01 -7.28009164e-01 5.12051046e-01 4.43438411e-01
-4.73404467e-01 1.71012431e-01 5.20829335e-02 9.75298345e-01
-9.05019641e-02 -1.68778375e-01 5.18065631e-01 -1.60089627e-01
-7.57761896e-01 5.84898651e-01 -1.91034019e-01 1.02026179e-01
8.64205003e-01 -3.02508101e-02 -2.20019259e-02 -3.52035046e-01
-5.31786025e-01 4.18556690e-01 2.42491841e-01 7.48198211e-01
5.22902429e-01 -1.53136122e+00 -6.71412706e-01 3.62364948e-02
1.74543157e-01 -9.39316154e-02 6.93539321e-01 9.06440198e-01
-4.80323806e-02 3.83856267e-01 -2.10353378e-02 -3.71803343e-01
-1.00936747e+00 9.14005280e-01 -2.01615036e-01 -6.91688895e-01
-4.66611177e-01 7.96830237e-01 2.50670105e-01 -1.33383363e-01
2.28542164e-01 -2.39598945e-01 -3.26228291e-01 1.29704505e-01
6.20602369e-01 3.82119477e-01 -5.83574101e-02 -3.09765428e-01
-2.16433316e-01 1.94876090e-01 -6.45927727e-01 2.19064206e-02
1.28241146e+00 -3.17163885e-01 2.54187346e-01 4.80149299e-01
1.01275873e+00 2.09773537e-02 -1.46953046e+00 -6.74996555e-01
8.52598473e-02 -7.25331545e-01 -4.27434802e-01 -7.96501219e-01
-9.77824509e-01 6.75389647e-01 3.19598317e-01 -6.37769373e-03
1.33518326e+00 -4.60257828e-01 1.20809686e+00 2.82323975e-02
2.73762017e-01 -9.74761009e-01 2.33187765e-01 6.51625037e-01
7.18825161e-01 -9.74922895e-01 -7.07984269e-02 -4.30668235e-01
-9.22283590e-01 6.55435324e-01 7.10065365e-01 1.91649497e-02
5.10937810e-01 -1.32101119e-01 -2.87978917e-01 4.72660184e-01
-1.00609612e+00 -3.01483646e-02 3.01011056e-01 1.17274709e-01
5.98827004e-01 2.00002983e-01 -6.14433050e-01 8.57090473e-01
-9.67518762e-02 1.85145542e-01 2.88415730e-01 9.54151034e-01
-7.56116807e-02 -1.39153576e+00 -1.45829711e-02 4.58628088e-01
-4.19911087e-01 -3.87771100e-01 3.67165595e-01 3.84638637e-01
1.87907428e-01 7.78580666e-01 -1.15526818e-01 -6.65552378e-01
1.77804634e-01 5.91289736e-02 1.81705043e-01 -4.37423080e-01
-5.12682319e-01 3.73558551e-01 1.13583654e-01 -1.70376822e-01
-2.47660130e-01 -9.33143914e-01 -1.02222931e+00 -2.39119947e-01
-2.65799284e-01 3.78954470e-01 1.00787096e-01 8.41781557e-01
5.67396045e-01 6.05915010e-01 8.16133380e-01 -3.42826903e-01
-6.65206194e-01 -1.17276657e+00 -2.93275654e-01 5.25802076e-01
2.47526288e-01 -5.70256114e-01 1.34471610e-01 2.30597928e-01] | [10.186452865600586, 5.484277248382568] |
696e4658-cd23-49d9-8210-8660d9c438ba | unet-a-nested-u-net-architecture-for-medical | 1807.10165 | null | http://arxiv.org/abs/1807.10165v1 | http://arxiv.org/pdf/1807.10165v1.pdf | UNet++: A Nested U-Net Architecture for Medical Image Segmentation | Implementation of different kinds of Unet Models for Image Segmentation - Unet , RCNN-Unet, Attention Unet, RCNN-Attention Unet, Nested Unet | ['Jianming Liang', 'Nima Tajbakhsh', 'Md Mahfuzur Rahman Siddiquee', 'Zongwei Zhou'] | 2018-07-18 | null | null | null | null | ['thermal-image-segmentation', 'video-polyp-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.14402346e-01 3.46072555e-01 5.28311551e-01 -5.40884793e-01
-1.24646239e-01 -6.87594116e-01 3.93433958e-01 -4.63457674e-01
-6.52572215e-01 4.52866465e-01 3.73579681e-01 -7.81422257e-01
-8.68354514e-02 -6.54622138e-01 -8.92265737e-01 -3.93539578e-01
-2.07485661e-01 1.15865517e+00 7.43402243e-01 -4.08142768e-02
5.85447073e-01 3.21704596e-01 -1.37676430e+00 9.53258693e-01
9.03082550e-01 1.20889270e+00 2.37993941e-01 1.81950498e+00
-4.46079344e-01 1.22671926e+00 -7.86623716e-01 -4.04368132e-01
2.94336766e-01 -5.99864244e-01 -1.23215294e+00 -4.49403554e-01
7.91229129e-01 2.02323273e-01 2.75221700e-03 9.57733691e-01
5.88331878e-01 7.00695276e-01 1.56705928e+00 -1.23251331e+00
-1.07130921e+00 1.22357595e+00 -5.35678923e-01 7.47447670e-01
1.59389496e-01 1.99994996e-01 3.92159373e-01 -8.81855547e-01
8.21268320e-01 1.86202919e+00 1.13833117e+00 6.86724424e-01
-1.17600894e+00 -4.54808474e-01 5.73061593e-02 3.28213364e-01
-7.30444551e-01 2.85393056e-02 -5.55985332e-01 -2.12620478e-02
1.92795420e+00 8.41956079e-01 1.32372510e+00 9.68326569e-01
7.11713314e-01 1.56512964e+00 1.39517641e+00 1.00073107e-01
-4.69592437e-02 -2.48244077e-01 8.45191181e-01 5.06493568e-01
-2.68051058e-01 -1.55669935e-02 1.68303728e-01 4.69164759e-01
1.04230845e+00 -5.66248834e-01 6.50083482e-01 4.40489471e-01
-1.23562837e+00 9.94724870e-01 1.15960407e+00 1.99527934e-01
-6.63632154e-01 2.80654401e-01 8.36834908e-01 3.93542111e-01
5.08159339e-01 5.83326101e-01 -7.21461773e-01 -2.46256012e-02
-9.85550344e-01 2.24239171e-01 5.55816710e-01 1.82316566e+00
4.29740131e-01 5.43900609e-01 -1.01289332e+00 4.23630208e-01
1.63406171e-02 4.80291218e-01 6.81467652e-01 -1.36755514e+00
1.39529213e-01 4.94729429e-01 -9.66852158e-02 -4.80491847e-01
-8.40603471e-01 -2.47481942e-01 -1.05119324e+00 -6.46772683e-02
2.35131726e-01 -5.40089607e-01 -2.02063656e+00 1.24344707e+00
-1.64317563e-01 2.47470900e-01 4.83674347e-01 1.18188715e+00
1.52835953e+00 1.25197363e+00 7.32919395e-01 9.25749540e-02
1.43897414e+00 -1.57895994e+00 -8.85104716e-01 -1.26871020e-01
1.84325680e-01 -9.79467869e-01 6.46071911e-01 4.29917723e-01
-1.39436257e+00 -1.10531044e+00 -5.08982062e-01 -7.82110155e-01
-1.17162669e+00 -2.77279973e-01 1.17175221e+00 5.43829560e-01
-1.59185708e+00 9.54308286e-02 -3.45098078e-01 -3.61374557e-01
5.32216072e-01 8.33848000e-01 -1.16095029e-01 2.32480958e-01
-5.90955496e-01 1.15400708e+00 8.82295132e-01 5.92027843e-01
-1.41459799e+00 -1.10725492e-01 -6.62941754e-01 -3.37531209e-01
3.74145597e-01 -6.41580105e-01 1.56946218e+00 -1.28791678e+00
-1.29533243e+00 6.14567399e-01 -2.47922271e-01 -9.17487979e-01
5.02065420e-01 -6.10699117e-01 -1.17209949e-01 -1.24652475e-01
3.20410281e-01 1.53699577e+00 7.48398304e-01 -9.81972575e-01
-4.82158601e-01 -2.92155445e-01 -3.06901224e-02 2.08788529e-01
6.32461965e-01 3.32594126e-01 -1.03843085e-01 -5.75535357e-01
-2.45408751e-02 -6.77452087e-01 -4.98088121e-01 -7.32425392e-01
-1.03767622e+00 -1.04438923e-01 1.66424036e+00 -4.25730288e-01
7.23079801e-01 -1.74322200e+00 2.22253844e-01 -3.88158448e-02
6.04617260e-02 8.85963559e-01 -4.27687198e-01 -1.18009023e-01
-5.69742978e-01 8.57253075e-01 -3.33646238e-01 -4.88627627e-02
-4.37444687e-01 1.12239313e+00 5.99962592e-01 -1.02230571e-01
1.69029966e-01 1.70120633e+00 -7.78342068e-01 -9.49503899e-01
4.25675303e-01 2.85817198e-02 -2.61277854e-01 4.98464108e-01
-6.51042044e-01 4.12954628e-01 -9.18074608e-01 3.09060872e-01
5.26451886e-01 -4.53821383e-02 -1.05900359e+00 -5.96494302e-02
-6.86219573e-01 -3.95468473e-01 -8.23484063e-01 1.33270311e+00
-2.52931923e-01 7.30698586e-01 2.66373940e-02 -8.41823995e-01
2.40543038e-01 8.07271719e-01 -1.64097011e-01 -5.42992234e-01
9.13654447e-01 5.04421651e-01 6.49442852e-01 -1.14804196e+00
6.07047856e-01 2.84343362e-01 6.19736314e-02 4.78136003e-01
4.73671108e-01 -3.15197080e-01 4.80944037e-01 1.52845517e-01
8.66184354e-01 3.48495036e-01 5.99397309e-02 -4.64701891e-01
8.10801208e-01 4.77827132e-01 4.97803718e-01 1.02443802e+00
-3.26042861e-01 1.20757246e+00 7.70640075e-01 -6.64280832e-01
-1.54843247e+00 -4.42435712e-01 8.46447721e-02 1.34865129e+00
-4.31850217e-02 6.09347820e-01 -1.65522099e+00 -1.16794705e+00
-5.01020312e-01 1.26211548e+00 -1.28502154e+00 3.16361576e-01
-1.24842942e+00 -9.22680199e-01 5.61144710e-01 1.07197070e+00
1.18759894e+00 -2.32589841e+00 -1.10143352e+00 3.23170066e-01
2.37225574e-02 -1.07564259e+00 -9.02960300e-01 6.77322388e-01
-1.03060710e+00 -8.38431180e-01 -9.66643870e-01 -9.54933167e-01
5.51577747e-01 -1.70989677e-01 1.51715589e+00 -3.21777761e-01
-7.18781531e-01 1.77281424e-01 -9.61678028e-02 -7.81808257e-01
-4.18582350e-01 2.10050941e-01 -5.63365877e-01 -1.61092088e-01
3.25015008e-01 -2.38913387e-01 -6.57217860e-01 4.73543078e-01
-8.06981862e-01 1.20869882e-01 9.05313134e-01 4.14156318e-01
3.85298699e-01 -2.55259901e-01 3.40379253e-02 -1.30144894e+00
9.23690677e-01 -2.81408876e-01 -2.43249476e-01 3.79714876e-01
1.13803461e-01 -1.45254984e-01 4.62596804e-01 -8.42689693e-01
-1.44342029e+00 -3.35680723e-01 -6.35905504e-01 -8.02384436e-01
-9.74419773e-01 7.68129481e-03 2.40549549e-01 3.07084192e-02
5.29364944e-01 2.66530484e-01 -2.81141490e-01 -5.66024520e-02
8.65708232e-01 5.27425170e-01 5.02656996e-01 -3.66985530e-01
2.72216201e-01 -1.28841370e-01 -4.60978746e-01 -6.74542010e-01
-1.07970262e+00 -3.25953633e-01 -9.66823995e-01 -4.85827059e-01
2.15998530e+00 -3.37247133e-01 -5.38121819e-01 8.80706966e-01
-1.22552228e+00 -7.44277060e-01 -7.14724600e-01 2.85522163e-01
-5.50804794e-01 -2.86472678e-01 -1.19754708e+00 -1.04829514e+00
-1.01138961e+00 -1.43534124e+00 9.55018759e-01 1.02265441e+00
-1.97888732e-01 -1.03982663e+00 -2.27037176e-01 1.50917917e-01
6.54418111e-01 5.44881880e-01 1.10205925e+00 -1.09064746e+00
-4.13213342e-01 1.52594298e-01 -6.19628906e-01 6.27384245e-01
-3.31901699e-01 3.89999866e-01 -7.49491513e-01 3.47556382e-01
2.93544494e-03 -2.79318422e-01 9.44679439e-01 1.32939446e+00
1.15569913e+00 -1.36844277e-01 7.04311132e-02 6.71338141e-01
1.38170588e+00 7.89409041e-01 6.76378369e-01 3.39651585e-01
5.90346396e-01 2.46140093e-01 7.77914584e-01 -7.20786870e-01
1.14462495e-01 -8.34645405e-02 1.01831388e+00 -3.62263978e-01
-4.22387362e-01 3.33173305e-01 1.58164382e-01 1.13437212e+00
-1.11368358e+00 -8.82010162e-01 -6.71052158e-01 3.87891799e-01
-2.00364947e+00 -9.14827287e-01 -1.19964039e+00 1.18120551e+00
2.74903208e-01 4.43448037e-01 2.55847842e-01 5.18739223e-03
7.97753811e-01 -1.22998677e-01 -7.28239238e-01 -1.61823940e+00
-3.09752136e-01 2.16098219e-01 4.45424616e-01 2.82152325e-01
-1.25230968e+00 1.55048203e+00 7.79278708e+00 9.42093194e-01
-6.69108868e-01 4.62461531e-01 1.28440213e+00 7.64644682e-01
-6.72392845e-02 -3.20042223e-01 -7.31222272e-01 8.96038264e-02
1.53158677e+00 5.55148482e-01 1.64847374e-01 7.60698497e-01
3.86048481e-03 -1.94294021e-01 -9.75670874e-01 4.57712799e-01
5.55404127e-02 -1.06461358e+00 3.84348631e-01 -4.57448140e-02
6.88387156e-01 6.38952374e-01 2.04669639e-01 7.49625325e-01
7.25423574e-01 -1.23143613e+00 1.20808744e+00 4.21045959e-01
1.63927197e-01 -1.01674259e+00 1.39898670e+00 2.00430349e-01
-8.66050124e-01 -2.53195673e-01 -7.89888322e-01 5.22780418e-01
-3.72105539e-02 -1.08517580e-01 -7.99584150e-01 7.00830162e-01
8.04998100e-01 6.52752042e-01 -1.07382905e+00 1.45579410e+00
-4.18826997e-01 4.98482406e-01 -2.79718399e-01 -4.07288581e-01
1.53950763e+00 -3.93534631e-01 8.66944611e-01 1.67126417e+00
-2.19048932e-02 3.17577064e-01 6.51129931e-02 9.50788140e-01
4.56545293e-01 8.93452391e-02 -3.42790961e-01 3.78764689e-01
-4.92139459e-01 1.41842973e+00 -1.58517075e+00 -5.55839837e-01
2.03215659e-01 1.06156433e+00 3.64106782e-02 3.13375443e-01
-1.19115484e+00 -3.70107144e-01 9.81058329e-02 -2.22321689e-01
4.68631089e-01 3.02880257e-01 -2.11121589e-01 -6.60039127e-01
-9.43467319e-01 -1.12476873e+00 4.27029371e-01 -1.71377134e+00
-1.24185967e+00 1.53739130e+00 2.32929945e-01 -4.88642603e-01
-1.19579867e-01 -1.06852388e+00 -6.75310194e-01 8.14746320e-01
-1.09476507e+00 -1.22273874e+00 3.97855416e-02 1.19485974e+00
1.14087260e+00 -1.35718256e-01 8.07863295e-01 -3.48789036e-01
-9.81838942e-01 7.91361034e-02 -3.33767682e-01 3.60487044e-01
-3.60016018e-01 -2.07389331e+00 1.10191405e+00 5.21628737e-01
-6.67764386e-03 3.77457887e-01 7.92617857e-01 -6.26962781e-01
-6.17546201e-01 -1.29017341e+00 7.39636198e-02 -7.38140285e-01
1.75565705e-01 -3.01532954e-01 -9.87788558e-01 1.27773190e+00
1.26023614e+00 -4.59087687e-03 4.06150550e-01 -4.20699924e-01
-1.33318594e-02 6.67396843e-01 -1.63337421e+00 6.14574611e-01
8.98493588e-01 2.18114883e-01 -6.38996840e-01 4.77426171e-01
9.54846561e-01 -9.31905568e-01 -1.31863320e+00 2.58228958e-01
4.47120577e-01 -1.39334869e+00 9.35827076e-01 -8.92210603e-01
1.69781819e-01 -1.46452829e-01 2.49920562e-01 -9.58551586e-01
-3.28386039e-01 -6.49135709e-01 -9.91034731e-02 9.99034345e-01
1.01221180e+00 -8.79493892e-01 7.22159564e-01 4.85258371e-01
-9.08186913e-01 -4.82301980e-01 -1.19890022e+00 -6.22986019e-01
2.18085676e-01 -7.18237996e-01 5.17919362e-01 2.11343437e-01
-1.17472625e+00 6.46644890e-01 -3.15359116e-01 1.20791845e-01
5.66409826e-01 -1.23056816e-02 4.53526199e-01 -1.11581898e+00
2.63425350e-01 -3.28501701e-01 -2.53639519e-01 -1.10415721e+00
-3.66541147e-01 -7.89978430e-02 -3.76347341e-02 -2.01557827e+00
2.21590891e-01 -4.96209264e-01 -1.94144115e-01 7.11578846e-01
-2.28734594e-02 7.95321286e-01 1.20746255e-01 -2.45000664e-02
-9.94372606e-01 -9.53709036e-02 1.95581901e+00 1.55157611e-01
5.11270538e-02 3.36209655e-01 -4.32163104e-02 9.41164911e-01
5.47643423e-01 -8.20528388e-01 -7.21539378e-01 -8.52880418e-01
-8.48679990e-03 1.37458280e-01 1.88404113e-01 -1.19564199e+00
3.57149601e-01 4.16331232e-01 6.58262491e-01 -1.50078142e+00
2.75643647e-01 -7.42924511e-01 -2.61466712e-01 1.46101236e-01
-4.80631590e-01 8.67878258e-01 6.60608828e-01 3.30466062e-01
-3.55534852e-01 -9.05630052e-01 6.22934997e-01 -1.25371289e+00
-8.56368244e-01 2.08473876e-01 -1.12596929e+00 3.51946324e-01
9.87223268e-01 -6.03545725e-01 -3.83686304e-01 -3.08975875e-01
-1.51337242e+00 3.74854416e-01 -4.49932814e-01 7.61002839e-01
7.68310726e-01 -5.72240651e-01 -6.03267789e-01 2.06184223e-01
-8.09694707e-01 5.37148237e-01 2.20744178e-01 9.82358158e-01
-1.01784146e+00 7.26452470e-01 -5.85842311e-01 -7.46875107e-01
-1.13247204e+00 6.65930390e-01 7.55054295e-01 -1.69001088e-01
-8.43292892e-01 1.66509426e+00 1.65519595e-01 -9.30589497e-01
-6.17703684e-02 -8.86089325e-01 -1.22759926e+00 -4.43545021e-02
2.98500031e-01 5.55343390e-01 -3.01129348e-03 -1.22769821e+00
3.11292142e-01 9.84422445e-01 -7.73974359e-01 3.91720757e-02
1.44286907e+00 1.36698574e-01 -5.35533547e-01 4.65759218e-01
4.94644254e-01 -1.07008529e+00 -6.89065456e-01 8.60772133e-01
3.44327331e-01 1.54148132e-01 1.92427784e-01 -1.23547578e+00
-1.28502524e+00 7.97078848e-01 8.69313836e-01 7.77090311e-01
1.36354744e+00 -7.40688145e-01 6.23761773e-01 7.47594833e-02
3.34929347e-01 -8.98618340e-01 -2.92668611e-01 5.43567955e-01
1.03881419e+00 -1.07640195e+00 -5.26605129e-01 -4.38731998e-01
-8.71341109e-01 1.38274014e+00 9.77563083e-01 -2.42430434e-01
8.60081136e-01 2.54787505e-01 2.66930521e-01 -5.95341027e-01
-1.04725909e+00 -3.77865463e-01 2.98176438e-01 9.50823069e-01
1.73105702e-01 -3.33205342e-01 4.65992063e-01 -5.18287122e-02
-6.14490986e-01 -1.85799241e-01 4.50097978e-01 7.60870159e-01
8.50173309e-02 -4.40180451e-01 -1.34221405e-01 6.22538567e-01
-8.14761639e-01 -4.42018092e-01 -5.52441120e-01 1.42932093e+00
6.85385883e-01 9.46482301e-01 1.24143481e-01 -6.29961342e-02
9.93002594e-01 -6.89258147e-03 -1.08792573e-01 -9.10416484e-01
-1.98621535e+00 8.53808105e-01 1.51025355e-01 -4.61597800e-01
-8.26311290e-01 -8.21205258e-01 -1.09722018e+00 -3.66837606e-02
-5.57088077e-01 5.31133175e-01 3.63326430e-01 9.65632081e-01
-1.99000150e-01 1.11954141e+00 -3.60809326e-01 -1.12092924e+00
5.65276384e-01 -1.54386616e+00 -1.71861500e-01 -2.82097101e-01
2.73667723e-01 -7.97195211e-02 -3.60069543e-01 -3.81868750e-01] | [9.649759292602539, 0.3686968684196472] |
5daeb89e-bf98-43d4-9137-5b0ec53459f6 | multi-scale-cross-contrastive-learning-for | 2306.14293 | null | https://arxiv.org/abs/2306.14293v1 | https://arxiv.org/pdf/2306.14293v1.pdf | Multi-Scale Cross Contrastive Learning for Semi-Supervised Medical Image Segmentation | Semi-supervised learning has demonstrated great potential in medical image segmentation by utilizing knowledge from unlabeled data. However, most existing approaches do not explicitly capture high-level semantic relations between distant regions, which limits their performance. In this paper, we focus on representation learning for semi-supervised learning, by developing a novel Multi-Scale Cross Supervised Contrastive Learning (MCSC) framework, to segment structures in medical images. We jointly train CNN and Transformer models, regularising their features to be semantically consistent across different scales. Our approach contrasts multi-scale features based on ground-truth and cross-predicted labels, in order to extract robust feature representations that reflect intra- and inter-slice relationships across the whole dataset. To tackle class imbalance, we take into account the prevalence of each class to guide contrastive learning and ensure that features adequately capture infrequent classes. Extensive experiments on two multi-structure medical segmentation datasets demonstrate the effectiveness of MCSC. It not only outperforms state-of-the-art semi-supervised methods by more than 3.0% in Dice, but also greatly reduces the performance gap with fully supervised methods. | ['Fani Deligianni', 'Paul Henderson', 'Xiao Gu', 'Qianying Liu'] | 2023-06-25 | null | null | null | null | ['contrastive-learning', 'semi-supervised-medical-image-segmentation', 'medical-image-segmentation', 'contrastive-learning'] | ['computer-vision', 'computer-vision', 'medical', 'methodology'] | [ 5.19393206e-01 2.84058213e-01 -5.08626103e-01 -6.70513451e-01
-1.10840333e+00 -4.35952991e-01 2.20525384e-01 4.37565297e-01
-3.61859471e-01 5.21413982e-01 2.33147606e-01 -5.70057742e-02
-1.96887463e-01 -7.17597067e-01 -4.91111428e-01 -6.20023310e-01
6.89398497e-02 4.76879478e-01 3.85119289e-01 3.48808169e-02
-8.61067250e-02 4.71767932e-01 -1.23708105e+00 6.36468589e-01
1.05341768e+00 9.97710884e-01 2.14046061e-01 1.71702951e-01
-2.45671764e-01 9.56927836e-01 -3.84444326e-01 -2.94973940e-01
-4.54315450e-03 -6.67884469e-01 -1.15984666e+00 3.85984898e-01
3.75685126e-01 8.00780803e-02 1.38643697e-01 1.08554041e+00
3.45670521e-01 -3.51272255e-01 7.90002465e-01 -8.96802306e-01
-4.17659461e-01 5.80864668e-01 -7.23834395e-01 2.30623364e-01
-1.35554269e-01 -9.00438055e-02 1.05211425e+00 -4.59933668e-01
7.06918478e-01 7.94027686e-01 8.60613108e-01 4.95143235e-01
-1.29607821e+00 -5.88560581e-01 1.17552243e-01 -7.33879358e-02
-1.25291908e+00 -2.37236954e-02 9.90742385e-01 -5.95051885e-01
5.40131807e-01 2.61970252e-01 5.83407402e-01 6.45881593e-01
1.79733276e-01 1.07514858e+00 1.42204189e+00 -4.29023534e-01
1.50984600e-01 -1.52411517e-02 2.53904313e-02 9.90425050e-01
6.26718625e-02 -6.64000809e-02 -2.13165358e-01 1.79057790e-03
7.44497478e-01 1.10387534e-01 -2.39602953e-01 -7.10222185e-01
-1.34013629e+00 9.20974433e-01 8.64057004e-01 7.04798579e-01
-3.38863343e-01 -2.11148053e-01 4.92632270e-01 -1.14403203e-01
6.56922519e-01 3.67611378e-01 -4.92234975e-01 3.31560314e-01
-1.27833593e+00 -2.70197839e-01 2.37319171e-01 6.59861147e-01
7.29722142e-01 -2.84365654e-01 -2.74223417e-01 9.94919658e-01
3.32586616e-01 1.87409058e-01 7.91547656e-01 -6.76688135e-01
2.93784320e-01 1.08166230e+00 -4.76325452e-01 -9.40702260e-01
-7.09049284e-01 -7.83215940e-01 -1.07107615e+00 2.41337139e-02
3.33431095e-01 1.62857458e-01 -1.10621357e+00 1.50823545e+00
5.23613513e-01 8.37620944e-02 -5.58750369e-02 8.15748155e-01
1.13119340e+00 2.38840327e-01 9.65919122e-02 -2.87614524e-01
1.17500412e+00 -1.09345937e+00 -6.79814577e-01 -1.47922128e-01
8.93434227e-01 -6.93213761e-01 8.55059087e-01 1.94808438e-01
-9.24475610e-01 -5.22243977e-01 -1.11154926e+00 1.51984483e-01
-2.25846514e-01 3.13796788e-01 5.57222128e-01 5.39272666e-01
-8.31100583e-01 7.15289116e-01 -1.05283475e+00 7.21891150e-02
1.06665039e+00 3.28106761e-01 -3.87984276e-01 -1.84401318e-01
-9.39959884e-01 6.98185086e-01 4.01791066e-01 1.43443480e-01
-6.62264109e-01 -7.19716847e-01 -1.02495050e+00 -2.11390659e-01
2.68264741e-01 -3.63496482e-01 8.37335765e-01 -1.11929762e+00
-1.09107685e+00 1.27044952e+00 1.37935758e-01 -3.23902786e-01
5.71407378e-01 1.13757424e-01 -2.75994331e-01 5.11135459e-01
4.30735379e-01 9.06496227e-01 5.43530703e-01 -1.48815525e+00
-5.12702286e-01 -5.94707906e-01 -2.47945696e-01 1.72674224e-01
-1.65018380e-01 -1.08151041e-01 -5.31976283e-01 -8.41786087e-01
3.81134301e-01 -8.11782420e-01 -4.68899697e-01 2.71025836e-01
-6.34403706e-01 -2.06138231e-02 5.75834572e-01 -5.18337905e-01
9.84466434e-01 -2.04841971e+00 -2.52138004e-02 2.85499841e-01
3.45500529e-01 2.89839000e-01 9.33359377e-03 -2.15169817e-01
7.80580426e-03 -1.24550827e-01 -8.24687481e-01 -3.51081282e-01
-4.11236316e-01 5.12429595e-01 1.97076291e-01 6.31673694e-01
4.16365623e-01 1.09669471e+00 -8.72927785e-01 -1.13432348e+00
2.55788654e-01 4.36921388e-01 -4.88261402e-01 6.29854724e-02
8.28537792e-02 6.61513746e-01 -4.74948853e-01 7.19674647e-01
6.57755375e-01 -6.50148809e-01 3.05140615e-01 -2.55162120e-01
2.32114464e-01 4.89822142e-02 -9.31772172e-01 1.87435365e+00
-4.72681910e-01 2.91749209e-01 -3.17408256e-02 -1.56069219e+00
8.98675501e-01 1.47548780e-01 8.87755156e-01 -9.64858472e-01
6.88759387e-02 3.90273511e-01 -1.74222112e-01 -3.36348385e-01
-7.54326656e-02 -3.13711762e-01 9.59053822e-03 2.07837284e-01
8.10385346e-02 -1.68787390e-01 1.75030544e-01 9.06604826e-02
6.07738614e-01 8.68856385e-02 3.54865074e-01 -5.04214942e-01
5.77080488e-01 1.37702227e-01 8.64318788e-01 4.94608372e-01
-3.35093766e-01 1.12992787e+00 4.19199705e-01 -4.21016216e-01
-5.42741239e-01 -9.85889852e-01 -4.42983389e-01 7.09536254e-01
3.45878601e-01 -1.74319878e-01 -8.86171341e-01 -1.28813756e+00
-1.63376018e-01 2.11951777e-01 -1.00271416e+00 -1.22273877e-01
-6.27824247e-01 -9.73830283e-01 3.75389308e-01 7.83727527e-01
4.53977138e-01 -1.08733952e+00 -7.84914613e-01 2.67263979e-01
-4.50363100e-01 -1.15467167e+00 -3.78941089e-01 3.98528010e-01
-1.10540545e+00 -1.37188256e+00 -9.69808340e-01 -1.11107492e+00
1.01715934e+00 9.57367793e-02 1.36340618e+00 2.73666590e-01
-6.62932277e-01 4.30248976e-02 -4.07970190e-01 -1.34302780e-01
-3.75950933e-01 1.27806813e-01 -5.80075443e-01 1.08035684e-01
6.25569746e-02 -2.27824807e-01 -6.85646117e-01 5.37398398e-01
-9.02120471e-01 2.57457912e-01 7.65050173e-01 1.10208130e+00
1.13041854e+00 5.49592115e-02 4.81604308e-01 -1.32422781e+00
4.15530428e-02 -2.05444545e-01 -2.29565054e-01 4.85895306e-01
-6.83448911e-01 -5.04997782e-02 3.78422558e-01 -2.00968370e-01
-1.06318545e+00 2.60772884e-01 -9.87096429e-02 -1.10241495e-01
-2.52643734e-01 4.98503029e-01 1.65973920e-02 2.82825828e-02
5.27979493e-01 2.44718134e-01 1.31554753e-01 -2.01345533e-01
2.35676959e-01 6.17545366e-01 4.60895777e-01 -4.14003491e-01
3.96203309e-01 8.68287742e-01 -6.42419457e-02 -3.64410460e-01
-1.54833293e+00 -6.11895025e-01 -1.14740324e+00 -1.38017371e-01
1.09058249e+00 -9.38014090e-01 -2.86453553e-02 4.42247927e-01
-7.20048606e-01 -5.43239117e-01 -4.98529494e-01 5.29838502e-01
-5.49293816e-01 3.66513222e-01 -7.08766997e-01 -1.22972600e-01
-3.47633809e-01 -1.38464570e+00 1.19561422e+00 1.76878572e-01
-1.43281817e-01 -1.14363742e+00 9.61566493e-02 7.49173582e-01
2.93814242e-01 5.28987825e-01 7.61737406e-01 -5.86607993e-01
-2.12606266e-01 1.00086518e-02 -4.07425463e-01 4.84636962e-01
3.99526507e-01 -1.89359561e-01 -7.42692232e-01 -2.94208825e-01
-1.68868229e-01 -5.16165555e-01 1.09723723e+00 6.84869766e-01
1.46622384e+00 1.28304273e-01 -3.88273090e-01 6.92098975e-01
1.37950873e+00 -1.42297342e-01 4.49241787e-01 2.23396182e-01
8.41444194e-01 7.81535625e-01 7.26080418e-01 1.61141217e-01
3.90405744e-01 5.02429724e-01 4.82083619e-01 -8.76558900e-01
-3.66515696e-01 -4.53915223e-02 -2.94461071e-01 8.84920895e-01
2.27674261e-01 2.60858536e-01 -9.09052312e-01 6.96977079e-01
-1.61168754e+00 -4.40163583e-01 -1.69050679e-01 1.77601528e+00
1.14516211e+00 3.39960963e-01 4.16454114e-02 1.99493557e-01
7.65317380e-01 5.14076650e-02 -5.08737266e-01 5.00023961e-02
-1.28381476e-01 3.94259751e-01 4.59959984e-01 1.91552728e-01
-1.44254947e+00 7.21573830e-01 6.06137609e+00 9.32941258e-01
-1.21867299e+00 2.12323591e-01 1.11826158e+00 1.80923030e-01
-2.31689423e-01 -2.99956948e-01 -4.93359864e-01 3.58837187e-01
4.33909327e-01 5.00107408e-01 -1.51010290e-01 6.27378047e-01
-1.30377024e-01 -2.82782912e-02 -8.66683662e-01 8.56119573e-01
3.03072691e-01 -1.37592089e+00 -5.83160156e-03 -2.06891373e-02
1.16299498e+00 3.50717902e-02 -1.77197307e-02 5.13003767e-02
2.08027557e-01 -1.14109921e+00 5.13108015e-01 2.51390249e-01
8.42422485e-01 -7.90150583e-01 9.81566370e-01 1.75188869e-01
-1.15126538e+00 1.59399629e-01 -1.91835895e-01 4.73449022e-01
4.49407138e-02 8.61660063e-01 -4.70077068e-01 7.52788126e-01
6.95628226e-01 9.55843151e-01 -8.17724407e-01 9.34271753e-01
-1.53552502e-01 6.95195377e-01 -7.58026838e-02 2.90631950e-01
4.27348107e-01 -7.60179386e-02 -1.63498148e-01 1.32970762e+00
-2.47086082e-02 -1.11552127e-01 4.45305049e-01 7.46483564e-01
-6.01918530e-03 4.03003484e-01 -1.48159146e-01 2.60076076e-01
2.30569756e-04 1.32531953e+00 -1.37347138e+00 -4.98291224e-01
-4.41666365e-01 9.11502957e-01 3.67211878e-01 1.49391231e-03
-8.52675200e-01 -1.17848985e-01 5.69902733e-02 7.92768374e-02
4.25076962e-01 2.30950177e-01 -5.29715061e-01 -1.03513706e+00
2.85083205e-02 -6.39421880e-01 7.83705056e-01 -3.56724977e-01
-1.35421574e+00 7.20349193e-01 -1.26232669e-01 -1.38750613e+00
7.39167407e-02 -3.82892787e-01 -4.08560842e-01 4.10539240e-01
-1.87776446e+00 -1.52179742e+00 -3.36349636e-01 5.75598121e-01
4.83333230e-01 2.68913787e-02 8.00007164e-01 3.35791349e-01
-5.01805842e-01 6.08198464e-01 1.08732395e-01 3.70477289e-01
7.43598223e-01 -1.40973985e+00 -5.46479896e-02 5.45293152e-01
3.49875540e-01 4.06650096e-01 1.34495735e-01 -5.19136906e-01
-5.58594942e-01 -1.31368387e+00 6.07398391e-01 -9.74265561e-02
2.95423388e-01 -1.35261387e-01 -1.10713506e+00 3.93668801e-01
-1.46682918e-01 8.63033056e-01 1.01807904e+00 -4.28169891e-02
-3.76542717e-01 -1.27326414e-01 -1.37963963e+00 7.41539001e-02
8.79948974e-01 -4.85955030e-01 -5.09628952e-01 3.49521130e-01
5.07375896e-01 -4.81473565e-01 -1.07839906e+00 7.72463381e-01
3.24586630e-01 -1.10633874e+00 9.86884415e-01 -4.27785635e-01
7.14353144e-01 -2.69053161e-01 6.70326352e-02 -1.08590662e+00
-1.90930769e-01 4.37864996e-02 1.72414258e-01 1.18693018e+00
4.15595472e-01 -4.60393310e-01 9.46009874e-01 9.17298570e-02
-2.76681155e-01 -1.10803390e+00 -9.43596423e-01 -5.35097480e-01
4.19652581e-01 -2.72949874e-01 3.34129155e-01 1.24741793e+00
-1.06677920e-01 1.14018209e-01 -6.04343005e-02 2.76349746e-02
7.24791825e-01 6.21230423e-01 1.47185937e-01 -1.32558262e+00
-2.68673122e-01 -4.63597804e-01 -5.02015710e-01 -7.58939683e-01
2.40702465e-01 -1.19944084e+00 8.70755389e-02 -1.78448427e+00
6.31769478e-01 -6.82639182e-01 -5.89848816e-01 6.20782256e-01
-3.40254128e-01 8.93732905e-01 -2.31691644e-01 2.67492712e-01
-9.08920467e-01 5.46528280e-01 1.62619078e+00 -3.81798863e-01
1.53297158e-02 -3.94944921e-02 -8.04554820e-01 7.93594897e-01
8.39208424e-01 -5.72322905e-01 -3.74117255e-01 -2.95351684e-01
-2.75740564e-01 -7.87900090e-02 3.45898986e-01 -9.25844312e-01
-2.60042213e-02 -1.82667412e-02 6.50018394e-01 -7.28480518e-01
-1.73725352e-01 -6.92923903e-01 -3.40823084e-01 5.93439281e-01
-7.08422959e-01 -3.22679222e-01 1.65630057e-02 4.53001320e-01
-5.71989596e-01 -1.74579501e-01 1.10160208e+00 -1.94406927e-01
-4.44176823e-01 2.59984136e-01 -9.51299444e-02 3.22751880e-01
1.04322314e+00 -1.19580902e-01 2.90307272e-02 6.77917525e-02
-9.33684349e-01 2.93905169e-01 2.61126459e-01 1.47495776e-01
6.15539074e-01 -1.20766461e+00 -6.44176900e-01 1.83939874e-01
4.08926487e-01 5.02793849e-01 4.67793524e-01 9.50917482e-01
-6.01897478e-01 2.66550303e-01 -2.42808789e-01 -9.89469290e-01
-1.29536581e+00 3.71565789e-01 4.86933559e-01 -5.99014282e-01
-7.14301646e-01 8.48511577e-01 3.89133304e-01 -7.37778366e-01
6.83443919e-02 -4.72875625e-01 -3.94750059e-01 7.21651539e-02
3.43483925e-01 -9.19784606e-02 2.61473417e-01 -8.67283106e-01
-4.66242015e-01 8.33101213e-01 -2.97736526e-01 2.90484339e-01
1.33689487e+00 -2.56000366e-02 -5.50628081e-02 2.99022973e-01
1.53267324e+00 -2.35593140e-01 -1.37688577e+00 -6.17819726e-01
4.95759398e-02 -3.01200390e-01 3.60009670e-01 -9.25361514e-01
-1.55101657e+00 8.39007616e-01 9.21268046e-01 -1.56477958e-01
1.14646649e+00 3.16223323e-01 8.59466136e-01 -2.18543380e-01
1.56429842e-01 -1.07740796e+00 3.14643830e-01 5.22186384e-02
4.71505553e-01 -1.69802260e+00 2.14790419e-01 -7.63469040e-01
-8.36387575e-01 1.01109397e+00 5.28949797e-01 -1.65477037e-01
6.36432767e-01 2.44196936e-01 2.90842116e-01 -5.47429025e-01
-2.04278529e-01 -3.91904235e-01 6.19022489e-01 6.38744056e-01
6.42230213e-01 2.39032269e-01 -1.85891926e-01 3.89695019e-01
2.65816394e-02 -1.56597793e-02 1.43592700e-01 1.01155186e+00
-1.29845560e-01 -1.01813579e+00 -1.22874491e-01 6.00449502e-01
-7.63932765e-01 1.36773199e-01 -1.98481634e-01 7.78928578e-01
3.29521656e-01 7.79124260e-01 2.01415330e-01 -1.31378975e-03
1.86564416e-01 -4.87243831e-01 5.25503218e-01 -6.72429740e-01
-5.81347585e-01 3.56593251e-01 -2.11292475e-01 -5.13985813e-01
-8.81854594e-01 -6.58828199e-01 -1.52024639e+00 5.05098045e-01
-4.39148962e-01 5.98327406e-02 3.62716436e-01 9.83896017e-01
1.05857037e-01 7.80530393e-01 7.24890769e-01 -5.28549075e-01
-3.59907597e-01 -7.50560105e-01 -4.81828809e-01 7.48676777e-01
2.30488092e-01 -7.56134033e-01 -1.34130508e-01 1.02321856e-01] | [14.744087219238281, -2.120853900909424] |
99567c90-5421-4625-b538-633b549310e3 | isles-2022-a-multi-center-magnetic-resonance | 2206.06694 | null | https://arxiv.org/abs/2206.06694v1 | https://arxiv.org/pdf/2206.06694v1.pdf | ISLES 2022: A multi-center magnetic resonance imaging stroke lesion segmentation dataset | Magnetic resonance imaging (MRI) is a central modality for stroke imaging. It is used upon patient admission to make treatment decisions such as selecting patients for intravenous thrombolysis or endovascular therapy. MRI is later used in the duration of hospital stay to predict outcome by visualizing infarct core size and location. Furthermore, it may be used to characterize stroke etiology, e.g. differentiation between (cardio)-embolic and non-embolic stroke. Computer based automated medical image processing is increasingly finding its way into clinical routine. Previous iterations of the Ischemic Stroke Lesion Segmentation (ISLES) challenge have aided in the generation of identifying benchmark methods for acute and sub-acute ischemic stroke lesion segmentation. Here we introduce an expert-annotated, multicenter MRI dataset for segmentation of acute to subacute stroke lesions. This dataset comprises 400 multi-vendor MRI cases with high variability in stroke lesion size, quantity and location. It is split into a training dataset of n=250 and a test dataset of n=150. All training data will be made publicly available. The test dataset will be used for model validation only and will not be released to the public. This dataset serves as the foundation of the ISLES 2022 challenge with the goal of finding algorithmic methods to enable the development and benchmarking of robust and accurate segmentation algorithms for ischemic stroke. | ['Jan S. Kirschke', 'Benedikt Wiestler', 'Benno Ikenberg', 'Maria Berndt', 'Tobias Boeckh-Behrens', 'Claus Zimmer', 'Lukas Meyer', 'Gabriel Broocks', 'Bjoern Menze', 'The Anh Baran', 'Johannes Bürkle', 'Teresa Zarth', 'Tassilo Friedrich', 'Alexander Hutton', 'David Robben', 'Ivan Ezhov', 'Florian Kofler', 'Sook-Lei Liew', 'Maria Ines Meyer', 'Mauricio Reyes', 'Waldo Enrique Valenzuela Pinilla', 'Roland Wiest', 'Uta Hanning', 'Ezequiel de la Rosa', 'Moritz Roman Hernandez Petzsche'] | 2022-06-14 | null | null | null | null | ['ischemic-stroke-lesion-segmentation'] | ['medical'] | [ 1.52246594e-01 -4.63682204e-01 -3.22378308e-01 -2.85544842e-01
-1.09044647e+00 -9.76312578e-01 5.13373554e-01 4.98684019e-01
-7.58110821e-01 6.44908369e-01 6.66562259e-01 -9.71045494e-01
-7.19212443e-02 -5.61416507e-01 2.15784591e-02 -4.51403439e-01
-3.90719861e-01 1.04120445e+00 4.40882683e-01 3.21610123e-01
1.87186360e-01 1.21112061e+00 -8.27913702e-01 6.48809493e-01
5.97985327e-01 5.97143829e-01 4.35333282e-01 6.98243260e-01
-9.73096937e-02 5.76110244e-01 -4.74879265e-01 3.41133326e-01
4.45649147e-01 -6.66938543e-01 -9.39113736e-01 -3.83277863e-01
-4.49301079e-02 -5.93023956e-01 -4.53701794e-01 3.78081411e-01
1.21339571e+00 6.21292703e-02 8.59530568e-01 -8.16385865e-01
4.40290272e-01 3.82018417e-01 -1.77339882e-01 1.20373094e+00
-5.45060448e-02 2.32703343e-01 1.61991417e-01 -5.56665838e-01
1.08171844e+00 4.89706069e-01 4.23268020e-01 2.89354324e-01
-7.53723860e-01 -3.27671826e-01 -2.39887461e-01 6.89034939e-01
-9.17078197e-01 -3.61482382e-01 1.68256894e-01 -8.29024851e-01
8.13315749e-01 4.24004555e-01 1.03378117e+00 7.97661126e-01
2.08421946e-01 8.34691286e-01 1.19847107e+00 -1.46609515e-01
4.26100045e-01 -5.09640992e-01 4.17243123e-01 1.16691485e-01
3.20284605e-01 2.16569588e-01 3.29665914e-02 -3.70710135e-01
7.73966134e-01 8.25178903e-03 -5.95168650e-01 -3.81816983e-01
-1.23441339e+00 7.79304862e-01 2.88716555e-01 3.92200381e-01
-7.07763553e-01 -3.93714309e-01 7.18374252e-01 9.97739807e-02
3.82785760e-02 2.33978972e-01 -2.56771326e-01 -5.68811595e-01
-1.07383847e+00 1.91240281e-01 3.77177328e-01 2.00267106e-01
-2.27338344e-01 -2.40362719e-01 -3.55244756e-01 8.99039805e-01
-2.85531521e-01 3.09826493e-01 9.72006440e-01 -8.48527014e-01
4.26071256e-01 3.52958351e-01 1.12048328e-01 -1.91337407e-01
-9.39835489e-01 -4.43479180e-01 -7.11674452e-01 5.18179595e-01
6.67000651e-01 -3.57455760e-01 -1.08928025e+00 1.14177811e+00
-8.31843987e-02 2.37498134e-01 -1.17800020e-01 1.21749866e+00
7.90072978e-01 6.57411441e-02 2.90318429e-01 -1.68720514e-01
1.40462601e+00 -5.48044324e-01 -1.14111099e-02 -7.66537264e-02
9.54352140e-01 -6.54247463e-01 7.00192034e-01 3.00819963e-01
-1.14965725e+00 1.91993192e-01 -5.33521116e-01 3.25743169e-01
-4.00578618e-01 1.99057773e-01 2.55287617e-01 8.40486228e-01
-8.46572340e-01 2.55915225e-01 -1.10270107e+00 -4.25077766e-01
9.64756370e-01 2.24626571e-01 -5.44143736e-01 -2.56286189e-02
-8.04364264e-01 1.68275201e+00 3.05117875e-01 -1.08965166e-01
-6.60913825e-01 -1.10182512e+00 -3.53038639e-01 -2.26335362e-01
-1.65565789e-01 -6.71811938e-01 1.00136709e+00 -4.55771536e-01
-1.06366110e+00 1.06025350e+00 -3.62150580e-01 -7.14389324e-01
9.09808040e-01 6.52088523e-02 -1.42254084e-01 6.57051444e-01
2.77275294e-01 3.65849018e-01 2.99291521e-01 -9.37325001e-01
-7.69871950e-01 -6.30043626e-01 -2.90531158e-01 2.91391969e-01
5.31646013e-01 5.27704060e-01 8.36984143e-02 -7.39437103e-01
1.81313172e-01 -7.67768443e-01 -4.50654536e-01 -3.59134853e-01
3.16059254e-02 3.22103649e-01 8.71814191e-01 -1.15332770e+00
7.55568802e-01 -1.53651881e+00 -1.84901386e-01 3.50453645e-01
5.09848595e-01 5.94192982e-01 8.98362026e-02 1.00661926e-01
-3.87084961e-01 9.81632695e-02 -3.90300393e-01 3.45623106e-01
-5.22360206e-01 -1.67126000e-01 -1.99928895e-01 5.34123182e-01
-2.25904226e-01 1.22078049e+00 -7.55974412e-01 -3.06494534e-01
6.02179825e-01 2.77618170e-01 1.23366015e-02 2.24055782e-01
5.55882096e-01 9.33068812e-01 -4.19683576e-01 2.24931315e-01
3.78553033e-01 1.45301178e-01 -2.16200035e-02 -1.55217135e-02
-3.61959547e-01 2.16512140e-02 -7.72324383e-01 1.29415238e+00
-2.70972013e-01 8.19954157e-01 4.10296023e-02 -1.29509270e+00
4.53590244e-01 4.35884356e-01 1.00241256e+00 -6.38414502e-01
4.80629146e-01 2.71594733e-01 4.04543608e-01 -5.42777181e-01
-3.07263851e-01 -1.52718961e-01 5.19055724e-01 7.96871185e-01
-5.55439115e-01 -2.01942511e-02 5.07014394e-01 1.67100921e-01
1.41719055e+00 -3.13390315e-01 1.06249243e-01 -3.00016344e-01
3.25870305e-01 5.21826744e-01 2.00857297e-01 1.07116926e+00
-5.92033148e-01 9.86061692e-01 4.10355359e-01 -6.42897785e-01
-1.09631395e+00 -1.43740320e+00 -5.81256449e-01 4.36709791e-01
-8.17352012e-02 -1.18550786e-03 -7.43625045e-01 -4.27740812e-01
-3.52579027e-01 4.90275174e-01 -3.42531979e-01 9.20247883e-02
-9.76093590e-01 -1.10636055e+00 4.60432887e-01 7.85114825e-01
5.82547307e-01 -1.14733934e+00 -1.40419936e+00 4.48946238e-01
-5.37722170e-01 -8.01605761e-01 -5.04786782e-02 -1.64587833e-02
-9.86973405e-01 -1.58455634e+00 -1.39512265e+00 -6.35627449e-01
3.65385860e-01 1.67903081e-01 7.81941116e-01 9.61090252e-02
-8.10685277e-01 5.34232795e-01 -6.29428864e-01 -5.50897896e-01
-3.36687893e-01 1.54299289e-01 -2.91188359e-01 -3.65364194e-01
1.52397498e-01 -4.86581504e-01 -1.25440824e+00 1.80879861e-01
-7.53408372e-01 4.38247807e-02 6.76793694e-01 5.23183405e-01
4.91766125e-01 -5.95802128e-01 8.04162383e-01 -5.67115545e-01
9.37593281e-01 -6.16434634e-01 -2.16043130e-01 3.43856335e-01
-3.23062450e-01 -4.48120594e-01 3.96191359e-01 -2.56156951e-01
-7.22570419e-01 -7.48757496e-02 -2.81415111e-03 3.98759656e-02
-5.73345780e-01 7.27443159e-01 2.71651655e-01 -5.50034456e-02
6.68672144e-01 8.73215422e-02 2.50328153e-01 -4.63358015e-01
8.77081901e-02 7.75412917e-01 8.07743073e-01 -3.21391404e-01
2.06682280e-01 6.01488113e-01 2.01718181e-01 -8.44706178e-01
-6.22432046e-02 -8.81177068e-01 -8.83849084e-01 -3.91865909e-01
7.75849342e-01 -4.16695952e-01 2.20885128e-02 4.62967902e-01
-7.76545644e-01 -7.15984762e-01 -2.94653118e-01 7.43463278e-01
-5.38151145e-01 3.81577373e-01 -3.31667602e-01 -2.39502713e-01
-5.72994053e-01 -1.57300019e+00 6.30049169e-01 -1.99859552e-02
-3.57957214e-01 -9.11530018e-01 2.16825619e-01 2.60000944e-01
6.76343560e-01 5.21036804e-01 1.05559003e+00 -6.98088467e-01
-9.38390791e-02 -7.03600824e-01 -2.93465674e-01 -1.58440500e-01
7.10796043e-02 -3.42436045e-01 -3.14402640e-01 -1.84449941e-01
-2.37097219e-01 8.53197500e-02 7.63509631e-01 1.02576470e+00
9.05719340e-01 5.79880953e-01 -2.47394979e-01 3.94045830e-01
9.74433899e-01 5.34875393e-01 8.97655427e-01 7.97254205e-01
2.50722110e-01 4.89023417e-01 -2.24130638e-02 1.84511453e-01
2.11183980e-01 6.91499293e-01 -1.06961064e-01 -4.86756638e-02
-6.09684467e-01 5.66528559e-01 -2.17395753e-01 1.50663465e-01
-3.25507730e-01 1.76334023e-01 -1.57338989e+00 5.71582735e-01
-1.65732646e+00 -1.09081781e+00 -5.46692491e-01 2.28530192e+00
4.22383606e-01 -1.97894722e-01 4.61171538e-01 8.27357545e-03
7.21932709e-01 -3.41821700e-01 -4.85307187e-01 -1.36991426e-01
-2.22264349e-01 5.69710970e-01 6.10358298e-01 4.24387306e-01
-1.04034424e+00 6.66229129e-01 6.34747314e+00 5.45679450e-01
-1.42517698e+00 4.64690775e-01 7.77397871e-01 -1.31897122e-01
2.65346646e-01 7.39013450e-03 -9.75005701e-02 5.27486563e-01
9.04546380e-01 -3.49692523e-01 2.13166744e-01 3.29308987e-01
8.45109522e-01 -6.06162310e-01 -3.84781599e-01 8.32588434e-01
-3.15340638e-01 -1.73222804e+00 -1.17923751e-01 -3.18225980e-01
4.18426573e-01 7.08391905e-01 -4.72539544e-01 -5.82062081e-02
-1.11084394e-01 -9.55103636e-01 4.93851572e-01 6.48329139e-01
1.03338015e+00 -4.83024836e-01 9.34008241e-01 1.98131844e-01
-9.65098083e-01 3.61539237e-02 7.50690699e-02 1.55568019e-01
6.97409511e-01 2.26883262e-01 -6.66141987e-01 3.65374178e-01
6.73728824e-01 3.41537714e-01 -5.64091146e-01 1.84572423e+00
6.43625334e-02 8.21165502e-01 -4.12319571e-01 5.18257618e-01
2.19885096e-01 -2.42991984e-01 7.45630085e-01 1.01529384e+00
5.77359200e-02 6.41873956e-01 2.17827544e-01 2.18541801e-01
4.88316536e-01 3.93694162e-01 -3.70535702e-01 4.36065137e-01
2.68051684e-01 9.85786378e-01 -1.47201359e+00 -6.05552971e-01
-1.94585443e-01 5.70748508e-01 -4.62573953e-03 4.78875726e-01
-2.72335798e-01 -3.16505939e-01 6.21500835e-02 5.75479031e-01
-6.25062957e-02 -2.54442751e-01 -8.46523941e-01 -1.03033257e+00
-1.67630151e-01 -6.68544114e-01 6.16578281e-01 -8.11899722e-01
-7.73616552e-01 6.50085270e-01 3.57794434e-01 -1.01781213e+00
-4.08863962e-01 -6.64193332e-01 -1.13950872e+00 1.29399943e+00
-1.13672686e+00 -7.15865433e-01 -5.38761854e-01 3.64151210e-01
3.16471994e-01 -5.37281334e-01 8.70537341e-01 1.35474429e-01
-4.04815793e-01 7.79845342e-02 2.95143574e-01 3.34575921e-01
4.78983551e-01 -1.00912833e+00 -6.50084466e-02 8.14909935e-01
-4.19208080e-01 3.81736130e-01 4.17768329e-01 -8.00307810e-01
-6.02003992e-01 -9.14803028e-01 5.74263632e-01 -5.57355165e-01
6.37160182e-01 4.29987401e-01 -6.19745553e-01 3.44434261e-01
-1.97335452e-01 -2.07876638e-01 7.45566785e-01 -7.17340052e-01
1.93353266e-01 4.29307759e-01 -1.25714803e+00 6.16825581e-01
7.74497747e-01 -1.68256178e-01 -7.14397848e-01 3.67898971e-01
-4.52933669e-01 -5.48997164e-01 -8.01690578e-01 6.80746973e-01
8.15776050e-01 -6.86363161e-01 9.48690891e-01 -1.02828193e+00
2.99300522e-01 8.53428841e-02 5.11821091e-01 -1.25674653e+00
-8.38764310e-02 -4.65742081e-01 4.01857138e-01 2.33208075e-01
1.87903151e-01 -8.06209922e-01 1.05091596e+00 9.10667717e-01
-3.26804906e-01 -8.33577693e-01 -1.07113779e+00 -6.19315863e-01
6.83061600e-01 -4.69326049e-01 3.89241338e-01 5.42147577e-01
-4.45519090e-02 -3.55100840e-01 5.35145044e-01 -8.32380727e-02
5.13958216e-01 2.46821180e-01 3.41816396e-01 -1.07089353e+00
2.84975916e-01 -1.25650084e+00 -5.63543141e-01 -4.10020649e-01
6.84006438e-02 -1.37324405e+00 -3.71492833e-01 -2.12082744e+00
2.94248194e-01 -6.59214437e-01 -1.31584138e-01 3.58709127e-01
-1.23142518e-01 3.14564407e-01 1.15758575e-01 5.19641340e-01
2.49978825e-01 -1.17434748e-01 9.82416630e-01 1.55894473e-01
-4.33144748e-01 2.31170401e-01 -1.95335567e-01 5.31653523e-01
1.45989799e+00 -5.07191002e-01 -4.40101415e-01 -8.88304114e-02
-6.39090538e-01 2.00532705e-01 6.21666014e-01 -9.25792992e-01
-4.87113111e-02 -8.03412348e-02 4.23871279e-01 -6.41820729e-01
-2.51874089e-01 -5.18281341e-01 -8.47120397e-03 7.69997060e-01
-3.32881391e-01 2.41806641e-01 7.60456324e-02 -3.48517336e-02
6.21532686e-02 -5.76121330e-01 7.66151547e-01 -2.11249247e-01
-5.82183599e-01 4.98186231e-01 -1.10991323e+00 5.27479291e-01
1.07643819e+00 -4.47892070e-01 -4.22083795e-01 -4.35936242e-01
-1.30597806e+00 9.13108140e-02 1.23415351e-01 2.66182542e-01
5.66166997e-01 -9.17359889e-01 -1.09165859e+00 -6.00126721e-02
-1.52679995e-01 -2.66554117e-01 3.42661798e-01 1.47602952e+00
-1.23939955e+00 5.87411225e-01 -6.59931123e-01 -3.98959488e-01
-1.30442178e+00 -1.02724694e-01 5.32588482e-01 -3.27913731e-01
-1.37568998e+00 2.87666589e-01 -3.07960123e-01 6.48913160e-02
1.21319108e-01 7.62845427e-02 -3.45888138e-01 -8.22920632e-03
8.75304937e-01 6.15966618e-01 3.52339596e-01 -9.18458164e-01
-3.12771767e-01 8.69525224e-02 -2.81972829e-02 -5.64465821e-01
1.35631692e+00 1.02039233e-01 -1.42432064e-01 -3.18010263e-02
1.06776309e+00 -4.72666740e-01 -7.28787243e-01 2.71342874e-01
5.81315011e-02 -2.18516454e-01 4.22362059e-01 -1.23285782e+00
-1.15098798e+00 1.00698769e+00 1.15899873e+00 -1.63945183e-01
8.92483532e-01 3.54902744e-02 8.94328892e-01 3.96491997e-02
3.89404327e-01 -7.26712048e-01 -5.57329893e-01 2.38548458e-01
9.04777586e-01 -9.20014620e-01 -1.65547177e-01 -5.86750247e-02
-6.81060135e-01 1.42700660e+00 -6.84283078e-02 -1.08592011e-01
8.75335574e-01 4.08213913e-01 4.79566693e-01 -2.98331082e-01
-7.50843016e-03 -3.41207117e-01 2.07918644e-01 8.43741298e-01
3.09549063e-01 3.97262752e-01 -9.36642945e-01 6.68823898e-01
-2.50477016e-01 2.37248808e-01 4.22210157e-01 1.10204732e+00
-4.39521343e-01 -1.39063859e+00 -1.71789423e-01 1.31227982e+00
-3.81399482e-01 -1.37200326e-01 -4.07528251e-01 6.00675166e-01
-1.83653086e-02 6.91169679e-01 -1.76874280e-01 3.88356805e-01
4.66726393e-01 2.05214798e-01 6.62104249e-01 -2.13504866e-01
-7.63587952e-01 -1.76004127e-01 2.47973710e-01 -2.59227157e-01
-2.10482001e-01 -1.11462843e+00 -1.45958388e+00 1.49853498e-01
2.06227869e-01 2.22572967e-01 7.71823645e-01 1.14515603e+00
4.17907149e-01 3.85321796e-01 4.58394848e-02 -8.09696555e-01
-1.11024790e-01 -8.06073308e-01 -3.34063053e-01 4.37530756e-01
-1.85064748e-02 -6.22167528e-01 -1.57863513e-01 1.66620836e-01] | [14.214051246643066, -2.0397300720214844] |
17507b0b-d8ec-40e7-8c13-0ee5d2d67f7d | earthquake-prediction-with-artificial-neural | 1907.02209 | null | https://arxiv.org/abs/1907.02209v1 | https://arxiv.org/pdf/1907.02209v1.pdf | Earthquake Prediction With Artificial Neural Network Method: The Application Of West Anatolian Fault In Turkey | A method that exactly knows the earthquakes beforehand and can generalize them cannot still been developed. However, earthquakes are tried to be predicted through numerous methods. One of these methods, artificial neural networks give appropriate outputs to different patterns by learning the relationship between the determined inputs and outputs. In this study, a feedforward back propagation artificial neural network that is connected to Gutenberg-Richter relationship and that bases on b value used in earthquake predictions was developed. The artificial neural network was trained employing earthquake data belonging to four different regions which have intensive seismic activity in the west of Turkey. After the training process, the earthquake data belonging to later dates of the same regions were used for testing and the performance of the network was put forward. When the prediction results of the developed network are examined, the prediction results that the network predicts that an earthquake is not going to occur are quite high in all regions. Furthermore, the earthquake prediction results that the network predicts that an earthquake is going to occur are different to some extent for the studied regions. | ['Osman Duman', 'Handan Cam'] | 2019-05-26 | null | null | null | null | ['earthquake-prediction'] | ['computer-vision'] | [-1.77713156e-01 3.17544267e-02 2.60589659e-01 -4.81544137e-01
3.49243656e-02 -1.89960033e-01 1.96555853e-01 8.99718329e-02
-4.25716758e-01 9.92537916e-01 4.81700040e-02 -4.29318607e-01
-3.20858657e-01 -1.15981257e+00 -5.81274629e-01 -6.79420769e-01
-2.11916342e-01 2.40111887e-01 4.07715499e-01 -4.96500462e-01
5.28766572e-01 7.75658965e-01 -1.27508903e+00 5.24052441e-01
2.83016652e-01 5.46486974e-01 3.13568264e-01 6.33061171e-01
3.49416614e-01 1.05877125e+00 -7.76715994e-01 2.54096210e-01
1.88325956e-01 -3.70192081e-01 -6.73067689e-01 -2.67808795e-01
-4.95830685e-01 -2.99339831e-01 -3.71242076e-01 5.23454130e-01
3.48743588e-01 3.02499920e-01 1.08830404e+00 -5.76526046e-01
-6.39760137e-01 7.53275335e-01 -1.09409936e-01 2.88645178e-01
1.59818277e-01 -3.83332022e-03 3.50380898e-01 -8.94753993e-01
2.10672855e-01 7.54845858e-01 9.75126207e-01 8.49243253e-02
-8.60365033e-01 -4.84803885e-01 -3.54178727e-01 2.18763247e-01
-1.51814413e+00 1.55573696e-01 8.83385003e-01 -7.75459588e-01
1.02343309e+00 1.04926102e-01 5.21607816e-01 5.39690375e-01
5.23126543e-01 -8.29532295e-02 8.76086175e-01 -5.13032079e-01
4.06625092e-01 1.74433127e-01 -1.05561599e-01 1.81593388e-01
5.35518527e-01 3.81289691e-01 -1.49444371e-01 1.84134245e-01
6.16538286e-01 -1.26805706e-02 -1.62431419e-01 6.83298230e-01
-8.54685128e-01 8.16719115e-01 7.04660952e-01 1.06585491e+00
-7.86453962e-01 -2.18600780e-01 3.69368106e-01 2.62322605e-01
2.41881460e-01 6.15612626e-01 -7.51956880e-01 3.98740113e-01
-1.11126542e+00 3.53571773e-02 8.80642533e-01 8.38060305e-02
7.58323967e-01 5.39196134e-01 4.75468576e-01 5.36711454e-01
1.06269799e-01 5.09021699e-01 7.13126719e-01 -4.20672715e-01
4.59130436e-01 5.98198950e-01 7.98980296e-02 -1.73401177e+00
-7.97798395e-01 -2.81960398e-01 -1.01988494e+00 3.81184369e-01
5.74344575e-01 -8.77238333e-01 -7.37639368e-01 1.36317146e+00
-1.72502846e-01 1.62838310e-01 6.46066427e-01 7.91837811e-01
7.89230645e-01 1.16221368e+00 -3.51173729e-02 -2.30104700e-02
6.60569072e-01 -3.48233819e-01 -3.73037338e-01 -2.91494519e-01
6.11524701e-01 -5.04088044e-01 4.19804245e-01 5.13224006e-01
-5.29930651e-01 -8.67883980e-01 -1.06985247e+00 6.62212014e-01
-6.34876013e-01 2.67614245e-01 3.97222459e-01 1.93995073e-01
-8.21957886e-01 8.47492218e-01 -6.48205161e-01 -3.51631820e-01
-4.84439105e-01 4.04454410e-01 -4.06596720e-01 4.84341979e-01
-1.73454344e+00 1.35459411e+00 1.19722533e+00 6.26020551e-01
-3.47745478e-01 -9.51185748e-02 -5.90510726e-01 2.98753083e-01
-1.44294590e-01 -1.70364380e-01 7.30779946e-01 -1.04660928e+00
-1.18637550e+00 2.19872788e-01 4.60779458e-01 -5.21438897e-01
2.29424506e-01 2.30273172e-01 -8.49296451e-01 1.04433425e-01
5.70551381e-02 1.76459998e-01 1.98329479e-01 -1.10118198e+00
-6.69114769e-01 -1.07485019e-02 -4.00963128e-01 -2.11390350e-02
-2.59757370e-01 -9.54067707e-02 3.05086404e-01 -7.22347200e-01
3.69000584e-01 -6.38945460e-01 -3.88775527e-01 -8.05496812e-01
-3.66819978e-01 -1.04534388e-01 4.37895715e-01 -9.67262030e-01
1.40778279e+00 -2.10559297e+00 -2.48207539e-01 8.11816454e-01
-4.86579001e-01 2.30100974e-02 4.18746948e-01 7.85749972e-01
-3.28175992e-01 4.73718829e-02 -3.30247432e-01 8.32372248e-01
-4.35923278e-01 3.84862453e-01 -1.91058323e-01 3.66389751e-02
1.84826553e-01 -9.43661258e-02 -5.86326003e-01 -1.87365621e-01
1.90498576e-01 3.09512347e-01 -2.24272072e-01 4.25888211e-01
1.66442409e-01 4.03461397e-01 -5.22170365e-01 2.03513250e-01
3.75227451e-01 2.54089326e-01 1.43420011e-01 -2.02156007e-02
-3.20956737e-01 -3.09035718e-01 -1.08997941e+00 4.60379064e-01
-2.89568305e-01 7.34390140e-01 -7.39686370e-01 -1.40350103e+00
1.77279341e+00 9.12750542e-01 1.80756062e-01 -2.65225500e-01
1.81224570e-01 7.57439733e-01 1.93112701e-01 -1.13316715e+00
4.98302013e-01 -2.90756792e-01 -2.91450582e-02 1.00927673e-01
-2.43482485e-01 -3.01683217e-01 3.11170161e-01 -6.00899398e-01
6.19880736e-01 1.43527025e-02 2.73418486e-01 5.08118607e-02
5.93602598e-01 3.07344168e-01 5.41486025e-01 6.76974535e-01
5.69444001e-01 4.51116830e-01 4.17529255e-01 -9.93792951e-01
-1.22187531e+00 -5.04055381e-01 -2.51179278e-01 6.41227007e-01
-4.03354287e-01 5.49878240e-01 -5.04710317e-01 -4.38545384e-02
-2.15482563e-01 9.96829689e-01 -5.16121268e-01 -4.54651378e-02
-5.88853717e-01 -8.65654409e-01 5.91469884e-01 5.08486688e-01
4.98094022e-01 -1.40178561e+00 -7.85103202e-01 4.89179760e-01
4.59028557e-02 -6.22529566e-01 5.65474093e-01 4.81486827e-01
-1.14083588e+00 -9.08217967e-01 -8.16806078e-01 -9.47244525e-01
7.04861403e-01 -4.84504342e-01 6.22461617e-01 1.82130903e-01
4.12100941e-01 -4.89446431e-01 -3.19991052e-01 -7.20809340e-01
-6.70153260e-01 2.25299209e-01 4.28840742e-02 -1.40419632e-01
3.51551205e-01 -5.84959388e-01 -1.54456720e-01 3.78004223e-01
-9.37294543e-01 -2.28004053e-01 6.99890733e-01 6.88497484e-01
1.66619241e-01 7.01824248e-01 1.20863199e+00 -7.46469080e-01
5.80820322e-01 -1.04695201e+00 -4.43043083e-01 1.80811882e-01
-4.65265483e-01 1.20256105e-02 8.89494836e-01 -4.75092053e-01
-1.12451661e+00 1.67657822e-01 -3.69933009e-01 2.26099998e-01
-6.23954475e-01 1.23883903e+00 3.84121418e-01 1.55562267e-01
9.93982971e-01 1.70187399e-01 -4.37470734e-01 -3.85446340e-01
-3.55147660e-01 9.32060182e-01 7.78510749e-01 -2.93529063e-01
5.63922167e-01 -2.52547562e-01 6.55513331e-02 -7.63771713e-01
-3.95343423e-01 -5.49526587e-02 -5.84398746e-01 -7.43230820e-01
7.10651755e-01 -4.94498044e-01 -1.76728085e-01 6.86705232e-01
-1.16308272e+00 -1.20361924e-01 -1.56175159e-02 1.14548051e+00
-2.68962413e-01 -6.68566599e-02 -5.06155550e-01 -1.02182245e+00
-1.73862681e-01 -8.46599340e-01 -1.47681385e-01 4.18209493e-01
-4.26444829e-01 -1.15768349e+00 2.51235247e-01 -3.30750883e-01
5.22501826e-01 5.93887269e-01 9.73654509e-01 -1.11767590e+00
1.55459091e-01 -6.62043929e-01 2.56373286e-01 5.04014552e-01
2.08845288e-01 3.94765854e-01 -7.07685411e-01 1.30438507e-01
2.49458000e-01 5.78349978e-02 6.22137368e-01 4.47480530e-01
6.70180261e-01 -4.15538937e-01 -2.92739682e-02 2.22047627e-01
1.66280985e+00 9.13439870e-01 7.28610098e-01 8.96598577e-01
7.47005492e-02 7.14140654e-01 1.96058169e-01 2.88176954e-01
3.85457277e-02 3.22211713e-01 2.76666045e-01 -3.07835072e-01
4.35831010e-01 -4.59014736e-02 2.32717216e-01 9.03804183e-01
-5.85753798e-01 -6.43462688e-02 -1.16414499e+00 6.91716909e-01
-1.52216065e+00 -1.25050712e+00 -4.66916144e-01 2.00325990e+00
7.09936440e-01 5.46909273e-01 -5.00555709e-02 7.70521462e-01
7.31294930e-01 -1.49240181e-01 -1.87808514e-01 -9.82786298e-01
-2.65911639e-01 2.27199897e-01 4.08252239e-01 3.62749398e-01
-7.77354777e-01 3.88845146e-01 6.29882288e+00 3.10754031e-01
-1.50335550e+00 -5.41276872e-01 7.51175642e-01 5.13896823e-01
1.07117452e-01 -2.44599264e-02 -2.44320944e-01 5.06447375e-01
1.23824835e+00 -2.10186150e-02 1.11563735e-01 8.48457456e-01
5.17941236e-01 -4.68326271e-01 -5.75977445e-01 3.91543388e-01
-1.39571086e-01 -1.13406122e+00 -4.30565998e-02 -5.16838789e-01
8.81097913e-01 -3.63761455e-01 -3.30897510e-01 2.07081214e-01
1.80067196e-02 -8.76175642e-01 5.80365658e-01 1.27727151e+00
1.17160313e-01 -9.16333616e-01 1.56737566e+00 5.44082940e-01
-8.20571899e-01 -2.96020448e-01 -5.11459351e-01 -7.18338609e-01
5.65323383e-02 7.14938998e-01 -1.26065445e+00 6.46623015e-01
7.50012934e-01 3.41826141e-01 -3.87580812e-01 1.03316426e+00
-4.26415145e-01 1.17804575e+00 -4.44654912e-01 -2.67454177e-01
2.62479812e-01 -2.24829778e-01 2.10888311e-01 1.16749871e+00
8.67058516e-01 1.30203262e-01 -2.16658816e-01 5.84032178e-01
5.46706319e-01 2.78849095e-01 -9.33171809e-01 3.33558440e-01
4.84330148e-01 4.77912158e-01 -5.97893119e-01 -3.87894213e-01
-1.72661141e-01 3.90780747e-01 6.01826515e-03 7.03970253e-01
-5.14043510e-01 -8.98803115e-01 -4.09886360e-01 6.60748929e-02
1.34159178e-01 1.18269371e-02 -6.59753740e-01 -4.64265168e-01
-2.98682332e-01 -3.26571792e-01 3.32323581e-01 -1.35440636e+00
-1.18498147e+00 1.01431108e+00 3.68910909e-01 -1.14118540e+00
-8.42396736e-01 -6.52904212e-01 -1.04058981e+00 1.18186963e+00
-8.30924869e-01 -5.49001276e-01 5.22635952e-02 2.95807183e-01
2.12769032e-01 -5.23902357e-01 8.37384641e-01 1.66785747e-01
-5.11295438e-01 1.07272111e-01 3.04772675e-01 5.38705587e-01
2.58329600e-01 -7.27944434e-01 -1.49731785e-01 8.59150350e-01
-3.74966413e-01 1.73125520e-01 8.56804192e-01 -7.82321632e-01
-1.83943346e-01 -8.10601532e-01 1.54319835e+00 2.90646940e-01
5.96965432e-01 5.62156200e-01 -1.28186905e+00 7.83328295e-01
1.06117114e-01 -4.09187824e-01 4.76200253e-01 -3.57268840e-01
3.18715364e-01 -6.38589710e-02 -1.14316070e+00 3.71742845e-01
-2.84544736e-01 -1.79983616e-01 -1.09898078e+00 2.00641707e-01
1.67800084e-01 -1.67128742e-01 -1.10604668e+00 4.12659377e-01
4.60497260e-01 -1.05057061e+00 8.61333609e-01 -6.35204911e-01
7.17537284e-01 -3.75594676e-01 -1.13149479e-01 -1.34000158e+00
-4.87228096e-01 1.99816808e-01 5.18764615e-01 9.49329019e-01
9.39207196e-01 -9.19683576e-01 6.22757852e-01 6.06585026e-01
-1.45422325e-01 -8.89713705e-01 -7.33594060e-01 -5.91326296e-01
2.13846132e-01 -2.70115346e-01 5.84508717e-01 9.71572578e-01
1.16182238e-01 6.08966090e-02 -3.21537793e-01 4.38260376e-01
1.16599232e-01 -2.21779525e-01 3.20363313e-01 -1.17948246e+00
-1.57700360e-01 -2.44377911e-01 -5.99375069e-01 -1.76287904e-01
-1.77387163e-01 -5.24855614e-01 -4.87963930e-02 -1.62864983e+00
-3.97442788e-01 -4.21653807e-01 -3.57809693e-01 7.96392083e-01
-1.06038079e-01 1.69831112e-01 -1.80192059e-03 3.72630149e-01
7.92430460e-01 -3.24639119e-02 7.39427507e-01 1.16956502e-01
-4.86895740e-01 4.93707240e-01 -3.57681066e-02 1.03002346e+00
1.32086003e+00 -6.48933053e-01 -1.85918733e-01 -3.38393986e-01
4.63343471e-01 6.03817701e-01 1.33895338e-01 -1.49927580e+00
2.63475001e-01 -4.27586973e-01 7.99129128e-01 -5.48907697e-01
-1.17159113e-01 -1.04466999e+00 8.33528638e-01 7.67168999e-01
-3.83180439e-01 2.57226259e-01 2.03064173e-01 7.50101134e-02
-5.24353445e-01 -9.10249710e-01 4.65888977e-01 -1.41911462e-01
-7.15930879e-01 -3.69924217e-01 -9.34947848e-01 -4.38884944e-01
9.42041099e-01 -5.71349859e-01 2.69156814e-01 -4.11831230e-01
-1.10973406e+00 -2.76162446e-01 -4.11173292e-02 -1.13558382e-01
6.05323672e-01 -1.17672527e+00 -9.41313505e-01 7.86781684e-02
-2.86735833e-01 -2.44434208e-01 2.00351208e-01 6.24173045e-01
-1.04342389e+00 3.52040827e-01 -5.59974790e-01 -1.06406258e-02
-6.24081910e-01 3.54402035e-01 6.49674535e-01 -9.80118215e-02
-3.62396240e-01 5.73628128e-01 -3.87505829e-01 -4.51073766e-01
-1.99865192e-01 -4.47514683e-01 -9.05972242e-01 1.07628796e-02
2.19861805e-01 4.23673630e-01 1.53166220e-01 -6.74137115e-01
-4.57345396e-02 4.98624116e-01 5.04421949e-01 -1.82614163e-01
1.62687004e+00 3.54376644e-01 -1.65983941e-02 7.82813311e-01
7.83625066e-01 -1.97227702e-01 -7.19582200e-01 2.50496063e-02
2.45011315e-01 -1.01699680e-01 -2.15173319e-01 -9.53960478e-01
-1.16631961e+00 5.75695574e-01 4.96413022e-01 5.37977755e-01
1.35677254e+00 -5.08694291e-01 5.07930458e-01 8.06635559e-01
1.38990045e-01 -1.25550890e+00 -5.34595788e-01 6.21746957e-01
1.01405787e+00 -7.76012361e-01 -1.20884523e-01 1.44831151e-01
-5.69924295e-01 1.90711379e+00 3.88098925e-01 -3.09979588e-01
7.93413460e-01 3.21698189e-01 1.43872807e-02 4.50116135e-02
-3.28419328e-01 2.87983626e-01 1.05938025e-01 5.44110358e-01
4.33005840e-01 1.88751921e-01 -7.63079882e-01 7.96560824e-01
-3.43684405e-01 3.71766925e-01 7.39434242e-01 7.02742636e-01
-8.21143687e-01 -5.67741632e-01 -1.03836250e+00 4.92806643e-01
-4.15059835e-01 -5.15849963e-02 -2.22840607e-01 9.89670515e-01
2.47062430e-01 9.72927630e-01 1.00011937e-01 -6.92575872e-01
4.57453430e-01 -1.70333795e-02 -2.87689388e-01 -2.54045695e-01
-7.87923813e-01 -5.15498877e-01 1.85481802e-01 2.71790028e-01
-4.24724042e-01 -3.87540817e-01 -1.82807839e+00 -2.58761883e-01
-3.77945006e-01 6.37771904e-01 6.94797933e-01 1.12450743e+00
-2.98369855e-01 4.43819553e-01 9.60483670e-01 -9.67439234e-01
-4.64483052e-01 -1.37713075e+00 -6.48654997e-01 3.06913465e-01
-2.56531332e-02 -3.55653763e-01 -6.45214319e-01 2.97783881e-01] | [6.329278469085693, 3.094984531402588] |
3fdd9498-e2bc-499b-987d-e111b6e77ad7 | m-3-video-masked-motion-modeling-for-self | 2210.06096 | null | https://arxiv.org/abs/2210.06096v2 | https://arxiv.org/pdf/2210.06096v2.pdf | Masked Motion Encoding for Self-Supervised Video Representation Learning | How to learn discriminative video representation from unlabeled videos is challenging but crucial for video analysis. The latest attempts seek to learn a representation model by predicting the appearance contents in the masked regions. However, simply masking and recovering appearance contents may not be sufficient to model temporal clues as the appearance contents can be easily reconstructed from a single frame. To overcome this limitation, we present Masked Motion Encoding (MME), a new pre-training paradigm that reconstructs both appearance and motion information to explore temporal clues. In MME, we focus on addressing two critical challenges to improve the representation performance: 1) how to well represent the possible long-term motion across multiple frames; and 2) how to obtain fine-grained temporal clues from sparsely sampled videos. Motivated by the fact that human is able to recognize an action by tracking objects' position changes and shape changes, we propose to reconstruct a motion trajectory that represents these two kinds of change in the masked regions. Besides, given the sparse video input, we enforce the model to reconstruct dense motion trajectories in both spatial and temporal dimensions. Pre-trained with our MME paradigm, the model is able to anticipate long-term and fine-grained motion details. Code is available at https://github.com/XinyuSun/MME. | ['Chuang Gan', 'Mingkui Tan', 'Changhao Li', 'Thomas H. Li', 'LiangWei Chen', 'Peihao Chen', 'Xinyu Sun'] | 2022-10-12 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Sun_Masked_Motion_Encoding_for_Self-Supervised_Video_Representation_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Sun_Masked_Motion_Encoding_for_Self-Supervised_Video_Representation_Learning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['self-supervised-action-recognition'] | ['computer-vision'] | [ 1.30309686e-01 -3.98287117e-01 -3.34612668e-01 -1.53242096e-01
-5.78722894e-01 -6.11769497e-01 4.49330240e-01 -4.16628629e-01
2.03073062e-02 5.94435871e-01 3.73461694e-01 2.05111474e-01
2.37814978e-01 -4.40741628e-01 -8.54135454e-01 -7.68502772e-01
-2.88644403e-01 -2.47908279e-01 3.05452824e-01 2.48067137e-02
1.85711741e-01 5.38341701e-01 -1.79248619e+00 7.78341651e-01
3.58349770e-01 9.03662860e-01 5.53026736e-01 9.22398627e-01
1.29761577e-01 1.18706882e+00 -1.51502699e-01 2.41110712e-01
1.90081939e-01 -7.38214910e-01 -7.23882139e-01 4.57401484e-01
6.35159910e-01 -7.72084534e-01 -9.51901317e-01 9.22856331e-01
4.31888700e-02 3.81483078e-01 5.45707941e-01 -9.32028592e-01
-5.91945589e-01 5.88970967e-02 -6.43507421e-01 7.21718550e-01
7.21528947e-01 4.49146479e-01 6.80776477e-01 -1.04370391e+00
9.50243473e-01 1.05566776e+00 3.10203135e-01 9.43561554e-01
-1.01390088e+00 -2.81792462e-01 6.05507016e-01 6.31826520e-01
-1.40347934e+00 -7.66758680e-01 9.12155151e-01 -7.18163669e-01
6.42707705e-01 3.21700066e-01 6.78277016e-01 1.39727294e+00
-6.59287348e-03 1.10049736e+00 7.09596097e-01 -2.27368593e-01
3.49284746e-02 -3.62269372e-01 -6.29408732e-02 8.03852439e-01
-8.84479880e-02 2.59160250e-01 -7.24638641e-01 1.06046207e-01
1.15309405e+00 3.53866100e-01 -7.55005658e-01 -1.87344089e-01
-1.38663530e+00 4.24262315e-01 1.95886955e-01 5.21502137e-01
-5.96949577e-01 3.88363689e-01 1.05008312e-01 1.86555266e-01
3.76457691e-01 1.67156160e-01 -4.12374020e-01 -2.76853442e-01
-1.10598409e+00 6.53803349e-02 2.78333098e-01 7.15183139e-01
8.75631034e-01 3.42362016e-01 -2.39848450e-01 5.28806508e-01
1.45349503e-01 3.57293278e-01 6.26119196e-01 -1.11856937e+00
3.20242286e-01 1.78777352e-01 2.81526625e-01 -1.15566194e+00
-7.73320273e-02 5.82304932e-02 -8.91961515e-01 1.75466388e-02
5.71659088e-01 1.02671005e-01 -9.67927814e-01 1.89945817e+00
2.59349406e-01 7.82739818e-01 -1.93241864e-01 1.21637380e+00
7.99466431e-01 9.73940790e-01 -1.73820212e-01 -4.81467903e-01
1.11257839e+00 -9.72226322e-01 -8.11056197e-01 -2.34362498e-01
2.99007684e-01 -5.95148921e-01 8.46701324e-01 2.68887490e-01
-1.22337592e+00 -9.38954890e-01 -8.08929622e-01 -1.24843858e-01
5.95062710e-02 2.39869595e-01 4.95906025e-01 -1.16113998e-01
-1.07193947e+00 6.89908206e-01 -1.20138645e+00 -3.11411351e-01
2.52248943e-01 1.23692833e-01 -4.90597278e-01 -5.27089000e-01
-1.01230145e+00 4.91550148e-01 1.27181634e-02 3.51844311e-01
-1.27817142e+00 -4.69663501e-01 -1.05864429e+00 -8.87390152e-02
2.91764170e-01 -7.06187606e-01 9.92409468e-01 -1.20675886e+00
-1.20733857e+00 7.04513907e-01 -8.28206360e-01 -3.09314616e-02
3.53253216e-01 -1.30867422e-01 -4.50498581e-01 6.11512542e-01
4.05974984e-02 6.61705732e-01 1.17137885e+00 -1.20990586e+00
-5.04013896e-01 -2.42598355e-01 2.31976602e-02 9.07383412e-02
-1.20895356e-01 -1.23715140e-01 -7.98558116e-01 -1.12503791e+00
2.28153467e-01 -8.71558666e-01 -3.02006990e-01 2.67660737e-01
-6.81890994e-02 2.31925651e-01 9.40924287e-01 -9.76627588e-01
1.28387499e+00 -2.28034210e+00 3.46942186e-01 -2.32340097e-01
1.49496317e-01 2.18303561e-01 -3.28889728e-01 1.96781248e-01
-6.59314692e-02 -2.51352072e-01 -1.83959723e-01 -3.69567811e-01
-3.90480727e-01 3.09256762e-01 -4.83422965e-01 5.95712423e-01
3.68980646e-01 1.01981032e+00 -1.08676124e+00 -2.26965845e-01
3.57657135e-01 6.44272745e-01 -6.15589738e-01 4.66306269e-01
-2.88470060e-01 1.03264821e+00 -5.44912755e-01 8.31757665e-01
5.63032150e-01 -4.47191894e-01 1.96059078e-01 -2.73512751e-01
5.53004481e-02 8.72863363e-03 -1.15300083e+00 1.92914593e+00
-1.24166258e-01 8.82493258e-01 5.66604771e-02 -1.05661535e+00
4.38264728e-01 3.88030469e-01 8.08776915e-01 -6.13047540e-01
-1.24225214e-01 3.53493616e-02 -4.36808348e-01 -9.38005984e-01
5.37650228e-01 1.59447209e-03 2.35527650e-01 3.10039520e-01
-8.25603399e-03 3.61055851e-01 1.11019731e-01 1.16845556e-01
1.09366596e+00 4.89793420e-01 2.82679290e-01 1.96005240e-01
6.10243380e-01 -1.82209611e-01 8.75504434e-01 3.73478115e-01
-2.98845917e-01 1.13554800e+00 1.33216381e-01 -8.20764184e-01
-8.17487299e-01 -8.84972215e-01 2.77706832e-01 9.86274481e-01
4.03466105e-01 -4.05654550e-01 -3.38625699e-01 -6.14752352e-01
-1.78616211e-01 1.81944057e-01 -8.32788765e-01 -8.64713266e-02
-9.60195780e-01 -4.38338667e-01 -1.06165046e-02 5.46366274e-01
1.50638819e-01 -1.07699418e+00 -5.72631299e-01 1.80943131e-01
-6.91080511e-01 -1.33179569e+00 -9.29160714e-01 -1.26674280e-01
-9.40224290e-01 -1.09651923e+00 -8.44103515e-01 -7.96014249e-01
7.90215552e-01 6.82637990e-01 9.58986521e-01 2.51507491e-01
-4.69031811e-01 5.09752691e-01 -4.80606228e-01 5.73714733e-01
-1.68078527e-01 -6.81504548e-01 1.20612077e-01 4.64969933e-01
7.52124190e-02 -7.76996493e-01 -9.76245761e-01 3.53253394e-01
-1.02336538e+00 1.41713858e-01 4.16869551e-01 6.22416496e-01
8.13158691e-01 -1.01529263e-01 1.83377177e-01 -6.00440502e-01
-1.60389006e-01 -6.80050492e-01 -2.16252059e-01 2.60704458e-01
-4.10934165e-02 1.82820410e-02 6.02223814e-01 -7.80274272e-01
-8.81565213e-01 3.85733157e-01 -7.78965056e-02 -1.12900937e+00
-3.38538080e-01 2.66979039e-01 -7.19563961e-02 1.02434680e-01
3.99937153e-01 7.05107927e-01 -1.06612988e-01 -5.89947402e-01
4.18834805e-01 7.22766295e-02 7.91150093e-01 -5.91309547e-01
5.64883232e-01 8.11335206e-01 -1.51128054e-01 -8.29375803e-01
-7.30248630e-01 -7.02722609e-01 -8.88438165e-01 -3.96966517e-01
9.41060305e-01 -1.05119872e+00 -3.78928661e-01 2.60053575e-01
-1.15812111e+00 -5.56244254e-01 -2.39214107e-01 4.93008286e-01
-8.38964403e-01 6.64493084e-01 -8.33531380e-01 -5.39101481e-01
1.08739480e-01 -1.06720960e+00 1.02744734e+00 -7.74042606e-02
-2.76924372e-01 -9.21815336e-01 1.29578918e-01 7.79632181e-02
1.45648509e-01 4.17113394e-01 4.90842521e-01 1.88865140e-01
-9.94058073e-01 1.76617224e-02 9.01319310e-02 1.43858835e-01
3.79819393e-01 1.22387283e-01 -9.92873430e-01 -4.58173096e-01
2.37275377e-01 -6.78420514e-02 1.10458934e+00 7.81705856e-01
1.28661990e+00 -4.95697141e-01 -1.96863085e-01 9.26358402e-01
1.23641157e+00 1.11503646e-01 6.90496624e-01 1.15120225e-01
9.93564367e-01 6.03837430e-01 6.89899504e-01 5.69357395e-01
2.74837732e-01 9.20575798e-01 3.19568634e-01 2.38501668e-01
-6.32517278e-01 -4.29356396e-01 7.73736358e-01 9.06268001e-01
-3.96295607e-01 -7.76464045e-02 -4.53906506e-01 6.11993253e-01
-2.04701853e+00 -1.46680093e+00 3.22182328e-02 1.92895341e+00
5.72938561e-01 -2.93967336e-01 1.48226216e-01 -6.51571080e-02
6.90565050e-01 5.60462773e-01 -6.86880648e-01 1.99962378e-01
-2.71428049e-01 -2.51856506e-01 -1.37407565e-02 5.67095220e-01
-1.04208398e+00 8.48855913e-01 6.10100317e+00 5.97605526e-01
-1.35066843e+00 9.94070843e-02 6.60647631e-01 -2.98600614e-01
-3.84080052e-01 -1.05507504e-02 -5.93278646e-01 6.92310393e-01
7.62276113e-01 1.59734726e-01 5.18753767e-01 5.52203298e-01
4.71185774e-01 2.80620158e-02 -1.23759222e+00 1.30037642e+00
1.60468102e-01 -1.58963120e+00 2.89213419e-01 -9.91470590e-02
8.92420828e-01 -1.92035928e-01 1.65755942e-01 2.80338377e-02
-1.81397289e-01 -1.00742781e+00 8.78018439e-01 8.08398187e-01
9.64926481e-01 -3.39621812e-01 1.99739516e-01 4.39514160e-01
-1.68749571e+00 -1.19871296e-01 -2.75765955e-01 -1.76794291e-01
3.52688193e-01 3.91769677e-01 -3.60455036e-01 4.66005027e-01
6.20009959e-01 1.40167165e+00 -3.92042100e-01 7.99785733e-01
-2.36600012e-01 4.19851273e-01 2.84298360e-02 6.39320135e-01
9.59173664e-02 -2.02682108e-01 4.70867157e-01 1.19600582e+00
5.95915675e-01 4.32984799e-01 2.68338799e-01 7.27206349e-01
2.79923171e-01 -4.17632073e-01 -6.68992817e-01 -1.44920424e-01
2.07439184e-01 9.44951057e-01 -6.26328647e-01 -3.54060173e-01
-5.76154053e-01 1.45599544e+00 2.94191569e-01 7.80856490e-01
-7.49039829e-01 4.04659986e-01 9.67556417e-01 3.77817214e-01
6.25046790e-01 -5.33654213e-01 2.48843625e-01 -1.69428372e+00
1.54759809e-01 -9.30723846e-01 5.81217587e-01 -7.86019504e-01
-1.05411899e+00 6.93929136e-01 -1.93608046e-01 -1.70524800e+00
-6.33042097e-01 -4.67629850e-01 -5.61424017e-01 5.26246488e-01
-1.55872238e+00 -1.07427764e+00 -4.58474785e-01 1.00280809e+00
1.08198977e+00 1.17811456e-01 5.72229981e-01 4.19857383e-01
-5.20365775e-01 3.70002180e-01 -3.28345895e-02 6.30705357e-02
5.95136344e-01 -8.00899565e-01 2.47827649e-01 1.20626664e+00
4.93723482e-01 5.33925414e-01 6.53450668e-01 -4.98682529e-01
-1.71597087e+00 -1.20923960e+00 5.98356068e-01 -6.38921559e-01
4.38367665e-01 -9.23530310e-02 -1.09821856e+00 8.08094621e-01
-1.14584967e-01 7.54625082e-01 4.26096767e-01 -4.22718287e-01
-3.78131986e-01 1.75236955e-01 -7.88057983e-01 4.62279856e-01
1.18506324e+00 -8.57358873e-01 -4.23891515e-01 3.54710482e-02
5.03216743e-01 -3.20004195e-01 -6.16214812e-01 3.90302926e-01
5.58490336e-01 -9.88563061e-01 1.09463108e+00 -6.20967388e-01
6.19248092e-01 -5.92384219e-01 -4.78012890e-01 -9.67479408e-01
-4.51284349e-01 -7.93357849e-01 -8.83098364e-01 8.40276420e-01
-1.22631423e-01 -2.77181771e-02 8.96310031e-01 5.33894598e-01
2.21447833e-02 -8.34481776e-01 -7.35681534e-01 -6.40722454e-01
-3.07328373e-01 -4.26188946e-01 3.03914577e-01 1.07974434e+00
-1.29703552e-01 8.51929374e-03 -9.72660661e-01 4.06439245e-01
3.88691425e-01 5.18843710e-01 7.00561166e-01 -6.56686127e-01
-5.68799019e-01 -2.20619813e-01 -4.92964536e-01 -1.76596701e+00
1.88442722e-01 -5.69184840e-01 1.45671785e-01 -1.31846344e+00
4.04491186e-01 -2.51118373e-02 -4.25033212e-01 2.64983296e-01
-3.29884052e-01 4.59172726e-01 2.41928592e-01 5.91600001e-01
-6.80942774e-01 6.35034919e-01 1.50542378e+00 -9.70741734e-02
-4.15741540e-02 8.45212839e-04 -3.49379361e-01 7.75076568e-01
3.93321306e-01 -2.78512806e-01 -5.31596184e-01 -5.55461526e-01
-1.09794326e-01 6.12429082e-01 6.03919387e-01 -7.97540843e-01
2.21010908e-01 -3.96174192e-01 6.40966237e-01 -6.40140593e-01
6.44074142e-01 -5.87602258e-01 3.30620438e-01 3.30094367e-01
-2.04753295e-01 1.25419125e-01 9.77865458e-02 9.64934945e-01
-6.02809548e-01 -9.42221005e-03 6.36460483e-01 -3.81741852e-01
-1.40777135e+00 7.96590745e-01 -5.43706596e-01 -1.68735534e-01
9.50781465e-01 -4.11887854e-01 -8.94457623e-02 -5.75016141e-01
-1.05561316e+00 -3.72873124e-04 6.82749391e-01 5.49767852e-01
1.00860620e+00 -1.46182370e+00 -5.23144782e-01 3.22805166e-01
4.49575707e-02 -2.07576871e-01 8.21829796e-01 8.16129863e-01
-4.36603993e-01 3.08747947e-01 -2.94503480e-01 -6.49814308e-01
-1.25540388e+00 9.93138611e-01 1.56811953e-01 5.10583371e-02
-1.05616426e+00 8.89127076e-01 7.18662739e-01 3.48181129e-01
2.05528542e-01 -3.59094620e-01 -2.60491848e-01 -7.06270412e-02
9.85738337e-01 1.35005340e-01 -3.53357643e-01 -9.01005924e-01
-2.98225701e-01 9.09551144e-01 1.37036428e-01 3.51484492e-02
1.36602581e+00 -6.11819088e-01 9.42993611e-02 6.08846545e-01
1.43743694e+00 -2.31898366e-03 -1.97428453e+00 -2.77938724e-01
-1.07832700e-01 -9.61095273e-01 -1.32099062e-01 -1.33980572e-01
-1.29031849e+00 9.20102119e-01 3.96966934e-01 -1.08515173e-01
1.32905996e+00 1.00805226e-03 8.69910955e-01 4.36791591e-02
3.65306228e-01 -5.87205350e-01 5.16619682e-01 3.88160884e-01
9.25349712e-01 -1.34275389e+00 -7.34060556e-02 -3.32055420e-01
-7.16279924e-01 1.30775821e+00 5.89651525e-01 -3.10511366e-02
6.96242273e-01 6.47081733e-02 -6.95917681e-02 -2.40024254e-01
-9.37265038e-01 -3.36659223e-01 6.47936642e-01 5.29372871e-01
3.60306680e-01 -1.78352609e-01 1.85798258e-01 3.85522693e-01
3.72421443e-01 6.72644563e-03 4.67447609e-01 7.92127669e-01
-3.37244391e-01 -9.17931080e-01 -3.19658548e-01 -6.15470344e-03
-3.47367764e-01 1.60412773e-01 -1.61322132e-01 4.35435265e-01
6.90338686e-02 7.60281086e-01 7.00226575e-02 -4.56154108e-01
-3.61450836e-02 -1.69644415e-01 6.75707161e-01 -5.89055061e-01
1.83770433e-01 2.39436954e-01 -1.55670598e-01 -1.09709537e+00
-7.08755374e-01 -7.46252596e-01 -1.11672759e+00 -1.96503758e-01
3.20765942e-01 -7.30829015e-02 6.37966022e-02 7.36931801e-01
3.55864376e-01 4.64745671e-01 7.40226805e-01 -1.28063488e+00
-5.46565205e-02 -6.80043638e-01 -6.36316121e-01 7.84402251e-01
9.16397870e-01 -7.27618039e-01 -4.59035933e-01 6.04041994e-01] | [8.816534042358398, 0.48715218901634216] |
882d5cac-d396-4505-8ea3-cb4f70fe325c | neural-code-search-revisited-enhancing-code | 2008.12193 | null | https://arxiv.org/abs/2008.12193v1 | https://arxiv.org/pdf/2008.12193v1.pdf | Neural Code Search Revisited: Enhancing Code Snippet Retrieval through Natural Language Intent | In this work, we propose and study annotated code search: the retrieval of code snippets paired with brief descriptions of their intent using natural language queries. On three benchmark datasets, we investigate how code retrieval systems can be improved by leveraging descriptions to better capture the intents of code snippets. Building on recent progress in transfer learning and natural language processing, we create a domain-specific retrieval model for code annotated with a natural language description. We find that our model yields significantly more relevant search results (with absolute gains up to 20.6% in mean reciprocal rank) compared to state-of-the-art code retrieval methods that do not use descriptions but attempt to compute the intent of snippets solely from unannotated code. | ['Geert Heyman', 'Tom Van Cutsem'] | 2020-08-27 | null | null | null | null | ['code-search', 'annotated-code-search', 'code-search'] | ['computer-code', 'computer-code', 'computer-vision'] | [-3.47764641e-02 -2.77442753e-01 -6.49187803e-01 -2.30342597e-01
-1.38047707e+00 -9.24742877e-01 6.88940287e-01 6.45477474e-01
-4.17037398e-01 2.89976716e-01 4.15325135e-01 -6.07957125e-01
-1.36214063e-01 -4.91967559e-01 -5.44566274e-01 3.49264629e-02
-2.32376322e-01 3.99218053e-01 2.54069179e-01 -2.88737357e-01
9.07719851e-01 -8.19159076e-02 -1.56399822e+00 6.78095222e-01
9.60706353e-01 6.31235480e-01 3.48887891e-01 5.58846295e-01
-4.35798645e-01 1.26778924e+00 -4.49674010e-01 -3.87764812e-01
-9.59663168e-02 9.17969365e-03 -1.24797928e+00 -5.70982337e-01
7.08623350e-01 -4.03282225e-01 -3.63178372e-01 1.08824039e+00
1.53096482e-01 -1.04385532e-01 7.45002270e-01 -9.51102555e-01
-1.16466582e+00 6.66711450e-01 -4.92997527e-01 4.40000296e-01
1.13181221e+00 -1.14816479e-01 1.54220140e+00 -1.04615045e+00
7.48497784e-01 8.63440514e-01 6.61761880e-01 5.95690548e-01
-1.28141057e+00 -7.55968571e-01 -3.13828468e-01 2.70205792e-02
-1.59190679e+00 -4.70062762e-01 3.74816954e-01 -9.01744246e-01
1.74634314e+00 6.15941472e-02 1.52803332e-01 6.75575733e-01
4.07477498e-01 9.19961393e-01 1.92917496e-01 -7.04260826e-01
7.12284539e-03 5.48985489e-02 5.64840972e-01 9.72739995e-01
3.27827007e-01 -6.10995255e-02 -2.61673033e-01 -7.88165569e-01
8.93216878e-02 2.54388899e-01 -3.26907672e-02 -5.33501029e-01
-1.12872612e+00 1.16119409e+00 5.53852260e-01 3.60376239e-01
-1.23430550e-01 5.36013544e-01 8.69684637e-01 5.31285286e-01
3.00474882e-01 1.15340912e+00 -6.28716648e-01 -3.38651180e-01
-1.22250950e+00 4.71362978e-01 9.59597111e-01 1.55121922e+00
1.21822846e+00 -5.13735414e-01 -4.32391107e-01 9.62984383e-01
5.53993046e-01 5.13983548e-01 8.09839308e-01 -8.32609534e-01
6.25061870e-01 9.26660895e-01 4.29055616e-02 -7.42389917e-01
5.74496649e-02 -1.82816416e-01 6.88423812e-02 -3.31002384e-01
-9.92109403e-02 1.69464588e-01 -6.38927519e-01 1.35108101e+00
-7.02260852e-01 -3.85634035e-01 1.51354283e-01 3.19536626e-01
7.59050012e-01 5.29610634e-01 2.37299994e-01 2.89260089e-01
1.30732739e+00 -1.00684893e+00 -2.53118634e-01 -5.15064478e-01
1.14367938e+00 -9.22283709e-01 1.08654690e+00 -1.09187134e-01
-6.82389259e-01 -3.39291990e-01 -8.82387042e-01 -2.37192467e-01
-4.40568209e-01 4.09857929e-01 6.21982574e-01 4.49903011e-01
-1.55996430e+00 1.70463219e-01 -6.72013640e-01 -4.29452926e-01
2.37511054e-01 2.16696545e-01 -1.75724506e-01 -3.32077473e-01
-8.74836028e-01 6.74635470e-01 3.12004179e-01 -9.00350213e-01
-1.10310674e+00 -9.64906693e-01 -1.12773275e+00 3.93771797e-01
3.30991209e-01 -3.92394364e-01 1.80332756e+00 -6.57361269e-01
-4.98425096e-01 1.06881261e+00 -2.10424230e-01 -3.57093662e-01
-1.97613314e-01 -3.69164616e-01 -2.39441693e-01 1.12587549e-01
7.55649149e-01 5.71175933e-01 5.12597322e-01 -1.06761873e+00
-4.76104438e-01 2.02424049e-01 4.65276033e-01 -2.81350374e-01
-6.41240954e-01 5.13897359e-01 -5.94039619e-01 -5.46526134e-01
-4.38676387e-01 -1.07792437e+00 2.97443010e-02 1.07156046e-01
-6.63563237e-02 -5.95014155e-01 6.14528418e-01 -4.79576170e-01
1.71451938e+00 -2.30681825e+00 -1.02822356e-01 -5.31549053e-03
5.15837371e-01 3.56859744e-01 -4.02497262e-01 9.42720294e-01
-4.29996774e-02 5.38011730e-01 -3.05755466e-01 -1.43287942e-01
3.51193696e-01 -1.11612096e-01 -7.59566486e-01 6.30893856e-02
2.05846623e-01 1.04773068e+00 -1.33724427e+00 -7.53603697e-01
-3.61098707e-01 6.29273951e-02 -1.04852116e+00 2.57872075e-01
-3.80724043e-01 -4.81608570e-01 -9.52641428e-01 7.10969567e-01
9.52790231e-02 -5.47890663e-01 -2.85593778e-01 3.30879271e-01
6.87402338e-02 6.45318806e-01 -1.42173275e-01 1.87784553e+00
-8.10391128e-01 9.32147443e-01 -2.35810280e-01 -5.03255010e-01
7.90604174e-01 3.55757266e-01 3.30571264e-01 -8.02753329e-01
-4.72149312e-01 5.36673248e-01 -4.13575500e-01 -8.44497800e-01
6.20219529e-01 1.46117926e-01 -6.10065639e-01 8.63654733e-01
1.07982449e-01 -2.32219130e-01 4.49433178e-01 5.47049105e-01
1.88635194e+00 -5.26289828e-02 5.21991849e-01 -5.23586512e-01
6.40955210e-01 3.90065998e-01 -1.03614718e-01 1.17823076e+00
4.77973707e-02 3.52478355e-01 3.74873132e-01 -4.16689903e-01
-1.04177105e+00 -9.35868621e-01 -3.73221897e-02 1.70142341e+00
5.15075400e-02 -8.90131712e-01 -5.20771623e-01 -9.78205144e-01
3.17453235e-01 6.55972123e-01 -5.03483415e-01 -4.83552873e-01
-5.60968935e-01 -1.11395665e-01 9.86280084e-01 5.95732152e-01
2.43240505e-01 -1.08431292e+00 -7.81402647e-01 5.34784887e-03
-6.72651529e-02 -7.23027050e-01 -9.86391544e-01 1.71713114e-01
-5.97651362e-01 -1.13047135e+00 -6.79673374e-01 -1.04145348e+00
6.51368678e-01 3.02052200e-01 1.81461430e+00 8.93975556e-01
-3.31203431e-01 7.72465467e-01 -8.32470238e-01 2.39025280e-02
-8.21761310e-01 4.91445303e-01 -4.32095498e-01 -1.02056253e+00
8.78935754e-01 -1.60383388e-01 -3.23083431e-01 -2.47252248e-02
-1.30633938e+00 -3.78884733e-01 7.88711131e-01 9.54882681e-01
-2.41146460e-01 -4.19000953e-01 4.79689091e-01 -8.60051394e-01
1.21917105e+00 -8.91006768e-01 -6.82347357e-01 5.98137796e-01
-1.07127154e+00 7.33597219e-01 4.61602807e-01 -2.43579537e-01
-9.60763156e-01 1.37845337e-01 3.54104936e-02 -2.59794444e-01
7.62738213e-02 1.02873862e+00 8.32928896e-01 -3.31670910e-01
1.23175740e+00 3.82671386e-01 -3.05648565e-01 -3.85068953e-01
3.08236063e-01 9.55745101e-01 2.28800178e-01 -1.11361885e+00
7.46987045e-01 -6.84606358e-02 -7.94474006e-01 -3.02846879e-01
-4.22843874e-01 -1.17361617e+00 -5.89670658e-01 1.87809184e-01
6.59708083e-01 -9.82777834e-01 -2.06216678e-01 -2.55171448e-01
-1.36686933e+00 -1.24768326e-02 2.47969091e-01 2.44594455e-01
-6.46198452e-01 5.35361707e-01 -8.15105855e-01 -3.73406678e-01
-5.67241371e-01 -1.38761175e+00 1.70799470e+00 -2.73000181e-01
-5.29495239e-01 -7.64407873e-01 6.52962267e-01 1.31635368e-01
7.96527863e-01 -4.72236395e-01 1.44512737e+00 -1.14197111e+00
-7.60747850e-01 -5.74718654e-01 -5.03789008e-01 -1.00488730e-01
9.33555737e-02 -4.40776311e-02 -7.61357009e-01 -4.41810966e-01
-4.88128066e-01 -6.24528944e-01 9.10131335e-01 -3.54768634e-01
7.89437711e-01 -4.21902925e-01 -7.30197072e-01 2.29393482e-01
1.87253141e+00 3.35784525e-01 3.66262794e-01 3.75999689e-01
3.04810166e-01 2.19911307e-01 4.55153227e-01 4.03838485e-01
1.73819736e-01 5.81009984e-01 3.15060198e-01 3.94292295e-01
-7.26292003e-03 -2.44509727e-01 2.92132109e-01 7.98068702e-01
5.16903520e-01 -6.44832663e-03 -1.67290246e+00 9.93345559e-01
-1.66825664e+00 -7.71712780e-01 3.39704871e-01 1.89939511e+00
1.15648282e+00 -7.01726377e-02 -2.69646466e-01 -5.25454283e-01
4.50212419e-01 -4.43393625e-02 -4.00840014e-01 -2.83672810e-01
6.07494295e-01 2.56496191e-01 4.29088771e-01 3.99543405e-01
-1.02775908e+00 8.99109602e-01 7.60584593e+00 7.73729086e-01
-7.73400187e-01 -1.37634631e-02 -1.25942051e-01 2.82433420e-01
-5.80790102e-01 1.31722257e-01 -5.67445517e-01 2.66422689e-01
8.81224215e-01 -7.36303031e-01 6.80938721e-01 1.31842160e+00
-4.86938089e-01 9.69910622e-02 -1.65517867e+00 7.44196177e-01
2.99419165e-01 -1.34656703e+00 1.76424354e-01 -1.81944013e-01
9.18197334e-01 4.54719782e-01 -1.40411705e-01 9.47147548e-01
6.85367584e-01 -7.90803790e-01 5.99864185e-01 4.60381389e-01
7.01973319e-01 -1.72254518e-01 7.12482512e-01 2.68039316e-01
-1.39261985e+00 -5.91340601e-01 -3.45136553e-01 1.09823346e-02
-6.75998688e-01 1.74381211e-01 -1.04891133e+00 1.38573378e-01
5.21379828e-01 9.90610361e-01 -1.17267799e+00 1.42942405e+00
-1.07945804e-03 4.00232345e-01 2.25684419e-01 -5.35556972e-01
3.96518886e-01 4.85907465e-01 2.53006369e-01 1.70211291e+00
2.98711151e-01 -4.36405152e-01 2.38768533e-01 1.46211588e+00
-4.24852312e-01 1.49723798e-01 -1.00000095e+00 -4.05801356e-01
6.15474045e-01 9.40852761e-01 -4.63325948e-01 -6.84143603e-01
-7.74064183e-01 7.30652332e-01 4.31693137e-01 5.86110353e-01
-6.16779327e-01 -8.58352482e-01 4.62935716e-01 -1.40671119e-01
3.47091556e-01 -1.11650929e-01 2.99533725e-01 -1.26269746e+00
2.55553782e-01 -9.18398201e-01 4.29118633e-01 -1.08133698e+00
-1.52721941e+00 5.29657423e-01 4.33214575e-01 -1.19834352e+00
-7.77803302e-01 -5.35464883e-01 -4.84151065e-01 6.80952966e-01
-1.42614996e+00 -7.12990999e-01 -1.02978684e-01 2.22129405e-01
7.39061296e-01 -4.33165312e-01 1.05314648e+00 6.46490812e-01
2.14483052e-01 5.23901224e-01 3.75444651e-01 5.51363885e-01
7.99891233e-01 -1.32183456e+00 6.30211353e-01 5.17214000e-01
3.30100775e-01 1.39017057e+00 5.20667493e-01 -6.81372523e-01
-1.47231662e+00 -8.41509342e-01 1.00074697e+00 -7.54841685e-01
1.05270052e+00 -3.01050037e-01 -8.66513312e-01 8.06055069e-01
4.42951769e-01 4.33409698e-02 6.54042840e-01 1.08198330e-01
-1.04502261e+00 1.86406344e-01 -7.98815012e-01 3.32217485e-01
7.23022044e-01 -1.52563882e+00 -1.08122063e+00 4.43588167e-01
1.02518642e+00 3.01099923e-02 -6.55554891e-01 1.85194314e-01
5.40577114e-01 -4.87815559e-01 1.08931136e+00 -6.28897071e-01
8.19042146e-01 -1.49756208e-01 -1.85002416e-01 -9.04832423e-01
-2.37487301e-01 -3.12223166e-01 2.33871713e-01 8.30605924e-01
6.88649774e-01 -2.56914407e-01 4.60348576e-01 7.83356488e-01
-1.59919173e-01 -3.40566695e-01 -5.81931889e-01 -6.85826778e-01
2.79748350e-01 -1.17076576e-01 3.74493867e-01 9.19725418e-01
8.09237301e-01 2.04247236e-01 2.53380060e-01 -3.04816842e-01
2.35653996e-01 2.68619716e-01 3.34307015e-01 -1.24361503e+00
-1.70043319e-01 -7.30725050e-01 -3.70463163e-01 -1.17012966e+00
7.27845728e-01 -1.27786648e+00 4.53910381e-01 -1.44328797e+00
8.05793285e-01 -2.68360585e-01 -1.92769095e-01 8.65036130e-01
-1.53985649e-01 3.68865505e-02 -8.91184509e-02 6.09135866e-01
-1.24069905e+00 2.00170964e-01 4.34313655e-01 -7.15114713e-01
2.02578947e-01 -2.14431658e-01 -6.42884195e-01 2.55295336e-01
4.12367821e-01 -7.91845083e-01 -5.87446868e-01 -7.81461716e-01
7.51614571e-01 2.89870560e-01 1.73572421e-01 -9.70713317e-01
6.80045307e-01 1.98553041e-01 -3.26237440e-01 -1.70268282e-01
-1.76992580e-01 -7.83694506e-01 -3.08345228e-01 5.35776556e-01
-1.13318348e+00 4.96150494e-01 4.72890824e-01 5.25224149e-01
-3.97139579e-01 -1.01061368e+00 1.30710483e-01 -4.65379953e-01
-9.19033468e-01 -1.44067198e-01 -9.05993104e-01 3.00799280e-01
6.78254962e-01 3.32988650e-01 -6.73007309e-01 -3.14969242e-01
-5.83970249e-02 2.18234643e-01 7.82032967e-01 6.60684586e-01
6.96252465e-01 -1.21459997e+00 -5.42082131e-01 7.62091763e-03
1.07293820e+00 -7.59530187e-01 -5.28652310e-01 1.85179248e-01
-5.47264934e-01 1.02881801e+00 1.30835781e-02 -3.71615678e-01
-1.15639305e+00 8.96922588e-01 2.02212200e-01 -3.00047398e-01
-9.52215791e-02 5.45645595e-01 -7.17942119e-02 -6.97077632e-01
1.94472462e-01 -4.68168855e-01 -7.56523162e-02 -2.08323032e-01
4.94091511e-01 2.60559488e-02 1.79095149e-01 -3.54069769e-01
-6.40028000e-01 6.11167669e-01 -3.55804652e-01 1.91819131e-01
1.06890512e+00 2.10981786e-01 -4.78759706e-01 1.44094884e-01
1.75925040e+00 2.88137525e-01 -5.04650235e-01 -4.79889452e-01
7.29586244e-01 -5.42815983e-01 -1.04948595e-01 -7.84649491e-01
-7.83929586e-01 6.89635038e-01 3.07556003e-01 2.75644243e-01
8.66127372e-01 4.56677020e-01 5.40369153e-01 1.32152569e+00
6.56767726e-01 -5.39651394e-01 4.63671118e-01 7.43178546e-01
6.81276619e-01 -1.25918937e+00 3.47924158e-02 3.85462828e-02
-6.18315004e-02 1.25392640e+00 4.62982863e-01 -2.16180682e-01
6.87006712e-01 3.60603929e-01 -7.92974606e-02 -6.36531830e-01
-9.85820293e-01 -1.01598300e-01 4.95501190e-01 4.20147955e-01
9.26071167e-01 -3.92099321e-01 -1.80210441e-01 1.73811853e-01
1.99456528e-01 9.06906556e-03 4.13906962e-01 1.30734730e+00
-6.37242138e-01 -1.16116118e+00 -9.32391658e-02 8.02799404e-01
-4.32871312e-01 -8.32854152e-01 -4.95933533e-01 5.66608131e-01
-4.36932147e-01 9.21665370e-01 -1.73528239e-01 -5.34444034e-01
-3.64819379e-03 3.22454572e-01 -2.79441066e-02 -1.31664777e+00
-8.41828465e-01 -4.02611762e-01 -2.84204856e-02 -6.58470392e-01
-2.78783202e-01 -4.94812995e-01 -1.25424135e+00 1.20833650e-01
-4.91228789e-01 6.20761931e-01 4.44307625e-01 8.30896199e-01
5.48669875e-01 2.34240294e-01 1.91408113e-01 -4.38106269e-01
-6.76921546e-01 -8.36536705e-01 6.67095333e-02 4.86836344e-01
8.00763488e-01 -5.36896646e-01 -4.27234173e-01 3.13423246e-01] | [7.524329662322998, 8.074219703674316] |
d72aa180-9e7d-4ff1-959e-16ffa3741501 | data-integration-for-toxic-comment | null | null | https://aclanthology.org/2021.woah-1.17 | https://aclanthology.org/2021.woah-1.17.pdf | Data Integration for Toxic Comment Classification: Making More Than 40 Datasets Easily Accessible in One Unified Format | With the rise of research on toxic comment classification, more and more annotated datasets have been released. The wide variety of the task (different languages, different labeling processes and schemes) has led to a large amount of heterogeneous datasets that can be used for training and testing very specific settings. Despite recent efforts to create web pages that provide an overview, most publications still use only a single dataset. They are not stored in one central database, they come in many different data formats and it is difficult to interpret their class labels and how to reuse these labels in other projects. To overcome these issues, we present a collection of more than thirty datasets in the form of a software tool that automatizes downloading and processing of the data and presents them in a unified data format that also offers a mapping of compatible class labels. Another advantage of that tool is that it gives an overview of properties of available datasets, such as different languages, platforms, and class labels to make it easier to select suitable training and test data. | ['Ralf Krestel', 'Philipp Schmidt', 'Julian Risch'] | null | null | null | null | acl-woah-2021-8 | ['data-integration', 'toxic-comment-classification'] | ['knowledge-base', 'natural-language-processing'] | [-5.28585017e-02 -2.90714025e-01 -4.36524123e-01 -5.70225716e-01
-6.27256930e-01 -1.03317821e+00 7.71615386e-01 8.71863544e-01
-4.88093197e-01 6.76645219e-01 2.49583274e-01 -3.62130404e-01
-1.19860105e-01 -6.21379733e-01 7.06683993e-02 -3.94303203e-01
5.17800391e-01 6.82458580e-01 5.86749494e-01 -7.56211020e-03
3.74452978e-01 3.89693648e-01 -1.87451351e+00 7.17512310e-01
5.05121589e-01 6.72490001e-01 1.90642715e-01 1.91372618e-01
-5.43641269e-01 6.04819596e-01 -8.45693290e-01 -6.50296688e-01
2.03655869e-01 -3.45280081e-01 -1.06530190e+00 -1.86856277e-02
2.65517324e-01 3.40836972e-01 2.08666116e-01 1.00412369e+00
5.52890897e-01 -4.12462473e-01 7.60029614e-01 -1.41876268e+00
-5.17021954e-01 7.61846244e-01 4.92917523e-02 3.79316360e-02
5.08982539e-01 -2.07571700e-01 9.52389419e-01 -4.44392532e-01
1.00687122e+00 1.02948785e+00 5.25062382e-01 4.75611329e-01
-1.13150954e+00 -6.92786217e-01 -2.01220915e-01 1.79104283e-01
-1.07421219e+00 -3.95233892e-02 3.77809495e-01 -9.75952983e-01
7.60709584e-01 6.87533081e-01 5.90060174e-01 1.35177171e+00
-2.73539990e-01 1.78555444e-01 1.56191230e+00 -7.74649680e-01
1.01748161e-01 1.04469562e+00 5.34957290e-01 3.74470532e-01
5.29019892e-01 -6.00574434e-01 -3.20203424e-01 -3.77338052e-01
1.66241173e-03 -3.46856900e-02 -7.97929242e-02 -2.98188001e-01
-1.02360237e+00 7.75098503e-01 2.12159171e-03 8.70815814e-01
2.51854062e-01 -5.50258338e-01 9.13256109e-01 4.99457359e-01
3.74132574e-01 4.49662447e-01 -6.44375980e-01 -1.22076459e-01
-6.82698846e-01 3.87268543e-01 1.30656278e+00 9.20497358e-01
8.35134804e-01 -5.90746224e-01 6.77221641e-03 9.37270880e-01
3.39721292e-01 1.26414701e-01 6.94623172e-01 -6.58557713e-01
6.49263263e-01 1.03374052e+00 1.33347809e-01 -9.25740063e-01
-6.90163553e-01 -2.15897694e-01 -3.29791546e-01 3.34998906e-01
6.99739516e-01 6.62584603e-02 -3.85923654e-01 1.25170588e+00
2.15383872e-01 -6.65809572e-01 -4.95242886e-02 5.05353868e-01
1.14073288e+00 2.26418123e-01 3.93854111e-01 4.91360240e-02
1.44340599e+00 -6.53613746e-01 -7.72403181e-01 1.04925390e-02
1.25483143e+00 -9.64110553e-01 1.27731562e+00 4.38972026e-01
-5.95977783e-01 -2.22436801e-01 -9.25406635e-01 -2.03857124e-01
-1.29891038e+00 -2.76140831e-02 4.69438940e-01 1.12484038e+00
-7.87643492e-01 5.49017549e-01 -3.27013373e-01 -7.78842628e-01
2.40826324e-01 3.18319201e-02 -4.79486763e-01 1.29439101e-01
-1.16371679e+00 1.14164519e+00 5.64530373e-01 -5.04321456e-01
-3.02966893e-01 -6.12816751e-01 -5.55891871e-01 -2.32766613e-01
1.87443778e-01 -2.36256495e-01 1.10868216e+00 -9.77087855e-01
-9.91771519e-01 1.19395173e+00 3.16623181e-01 5.50795831e-02
6.27208591e-01 3.26471776e-01 -5.35752416e-01 -3.65526259e-01
3.09365958e-01 2.72319824e-01 2.63457298e-01 -1.12110543e+00
-9.46927607e-01 -4.46037203e-01 1.30641147e-01 6.58470988e-02
-5.38391411e-01 7.63977945e-01 -2.91876912e-01 -5.20251751e-01
-3.02642375e-01 -6.80904806e-01 1.16405487e-02 -3.08865488e-01
-2.26144090e-01 -4.30394024e-01 7.84216166e-01 -5.01743555e-01
1.55472326e+00 -2.05335617e+00 -5.95084019e-02 -4.25045677e-02
6.92531690e-02 3.09169322e-01 2.47980803e-01 7.99000800e-01
-2.28872702e-01 7.73863673e-01 -1.63612321e-01 -3.31961691e-01
2.73756593e-01 2.70501554e-01 6.12927563e-02 4.68598127e-01
-3.91053855e-01 2.90526450e-01 -7.83445835e-01 -5.70122421e-01
1.10606082e-01 3.05233330e-01 -1.72641605e-01 -1.13581993e-01
-1.89218253e-01 4.52630252e-01 -4.69285876e-01 5.43752849e-01
5.87945938e-01 -1.52081149e-02 3.82643014e-01 2.03785300e-01
-4.50952441e-01 6.07910514e-01 -1.46775997e+00 1.35621536e+00
-3.96783173e-01 6.82938457e-01 -1.22397549e-01 -7.57265151e-01
8.98880363e-01 6.13037407e-01 4.15303230e-01 -3.58342528e-01
2.70786017e-01 5.16772211e-01 -2.87866145e-01 -7.18580246e-01
4.06214297e-01 4.30221222e-02 -3.73828948e-01 8.65236044e-01
3.65798995e-02 -9.55901518e-02 7.95651734e-01 -1.05678057e-02
7.51789689e-01 4.10167985e-02 1.37981787e-01 -1.91421166e-01
6.48592055e-01 1.51238041e-02 4.16257352e-01 4.59251851e-01
1.97766528e-01 6.13437831e-01 7.56238282e-01 -6.76006436e-01
-1.22738719e+00 -3.82486284e-01 -5.76616049e-01 1.12199295e+00
-3.95485908e-01 -9.38623011e-01 -7.20906615e-01 -6.81444168e-01
-1.83557466e-01 5.83021998e-01 -5.91563880e-01 3.34736943e-01
-1.39953956e-01 -7.32876956e-01 5.05009890e-01 1.24891855e-01
2.83082247e-01 -1.21685266e+00 -6.73092723e-01 -8.95064622e-02
-9.36888605e-02 -8.82932603e-01 1.74315169e-01 2.97872394e-01
-7.21242845e-01 -1.25226736e+00 -1.88289061e-01 -7.38863051e-01
3.82610559e-01 5.27304038e-03 1.31208467e+00 2.82127202e-01
5.56979515e-03 2.16432917e-03 -7.48582065e-01 -7.72277355e-01
-7.90825963e-01 6.18380249e-01 -3.80795479e-01 -2.98435837e-01
4.96032119e-01 -2.00468749e-01 9.57963318e-02 2.08827049e-01
-1.21704280e+00 -2.14574844e-01 3.79024968e-02 2.86973804e-01
1.95731997e-01 1.40476346e-01 4.96550947e-01 -1.62693083e+00
9.23094630e-01 -6.23100162e-01 -6.69635177e-01 5.12754798e-01
-9.75695729e-01 -1.72476858e-01 5.65130711e-01 -3.47222760e-02
-7.00493276e-01 -1.84002947e-02 -1.79298326e-01 4.19775814e-01
-6.33066714e-01 7.58112848e-01 -2.48110637e-01 2.66922086e-01
8.31086159e-01 -4.25448060e-01 -1.51548937e-01 -1.01761675e+00
2.20147759e-01 1.35264957e+00 -1.62917793e-01 -2.80393392e-01
5.59308350e-01 4.51620594e-02 -4.29982841e-01 -5.57438910e-01
-9.89629149e-01 -6.69847310e-01 -1.08701289e+00 -2.41055682e-01
7.92529166e-01 -5.86796105e-01 -3.81727278e-01 4.81042385e-01
-8.06282282e-01 -3.20801914e-01 -2.11700097e-01 2.34890789e-01
-9.70845893e-02 3.64800483e-01 -3.01048934e-01 -5.47777236e-01
-2.06329063e-01 -1.30332172e+00 4.07893747e-01 9.82874259e-02
-6.15625501e-01 -1.25036788e+00 1.32054031e-01 6.28074467e-01
3.28724086e-01 2.22648323e-01 1.15075231e+00 -9.42507863e-01
-1.25069559e-01 -5.52417219e-01 -1.18065156e-01 2.10436627e-01
6.11081608e-02 3.93760532e-01 -9.21083033e-01 -1.54193237e-01
-1.30658016e-01 -4.69458967e-01 3.89823735e-01 -3.41160655e-01
1.13994300e+00 -1.80901855e-01 -3.71115953e-01 1.50212541e-01
1.38096189e+00 8.02828893e-02 5.73558509e-01 8.62122953e-01
4.47720617e-01 8.94671440e-01 3.95669937e-01 2.46324673e-01
3.98377657e-01 9.41479743e-01 3.69993538e-01 1.93755120e-01
-4.73306067e-02 1.40515834e-01 1.85038507e-01 7.92149603e-01
5.81203699e-02 -2.59370953e-01 -9.58323061e-01 1.89042643e-01
-1.69940972e+00 -9.71873164e-01 -6.73966408e-01 2.38325047e+00
9.85628307e-01 2.28723828e-02 4.66223359e-01 3.94167155e-01
5.86950243e-01 2.71249842e-02 5.40093184e-02 -6.15114272e-01
-4.50870730e-02 1.18067063e-01 4.22511429e-01 4.32562321e-01
-1.00716817e+00 6.65243745e-01 7.24416161e+00 6.73667252e-01
-1.21634531e+00 4.27763551e-01 1.35639578e-01 4.31254119e-01
-4.18064505e-01 1.58683851e-01 -9.52473521e-01 6.05288744e-01
1.29319656e+00 -2.67921418e-01 4.25620005e-02 7.99468458e-01
1.67160854e-01 -8.57794136e-02 -1.00266361e+00 8.36477518e-01
1.11735210e-01 -1.37409735e+00 -4.24174778e-02 5.83130158e-02
3.31686467e-01 2.12461069e-01 -1.87811643e-01 1.54450178e-01
3.20601523e-01 -9.00592208e-01 1.11085093e+00 2.59255290e-01
7.11379647e-01 -3.89134586e-01 9.86763358e-01 4.25017834e-01
-5.82927287e-01 -1.49380788e-01 -3.34189206e-01 -3.54806840e-01
-2.76396155e-01 3.77691627e-01 -6.38800383e-01 7.53376842e-01
9.16030824e-01 8.12595069e-01 -1.20398891e+00 1.30631077e+00
-2.21645549e-01 4.21516001e-01 -1.55869827e-01 -3.86128962e-01
-1.31242812e-01 -3.18277299e-01 3.16363350e-02 1.55937278e+00
1.11089826e-01 -5.26118517e-01 3.80348772e-01 3.94633263e-01
1.58338174e-01 8.21632922e-01 -4.51968938e-01 -3.94839048e-02
4.28328872e-01 1.56742251e+00 -9.48153019e-01 -1.68699890e-01
-7.06560731e-01 4.33623374e-01 3.60718071e-01 -7.72684962e-02
-6.90974176e-01 -3.77968669e-01 3.44416529e-01 4.04050559e-01
-2.37871125e-01 -2.15339571e-01 -2.46887922e-01 -1.08734512e+00
-1.64189383e-01 -1.08565843e+00 7.90435016e-01 -8.28204393e-01
-1.21719265e+00 8.34862173e-01 3.53849411e-01 -1.28793740e+00
1.98844150e-02 -7.08163619e-01 -1.91359669e-02 8.82013321e-01
-1.05334771e+00 -9.04172719e-01 -3.97422403e-01 3.51936072e-01
2.68673539e-01 -2.05226466e-01 1.25915051e+00 8.59241962e-01
-6.90235317e-01 2.88615227e-01 2.65505761e-01 3.08309406e-01
1.09885693e+00 -1.34139705e+00 -1.09803930e-01 2.63636947e-01
3.46959800e-01 4.07460034e-01 6.39183581e-01 -4.00735945e-01
-3.77204210e-01 -8.09068799e-01 1.27217543e+00 -1.00652540e+00
9.18067217e-01 -6.73619509e-01 -8.69854331e-01 6.04403138e-01
4.64400411e-01 -3.29944700e-01 1.20898736e+00 1.98569983e-01
-7.47410357e-01 -2.60285586e-01 -9.51537192e-01 2.83956736e-01
5.22834957e-01 -3.72300088e-01 -5.27128220e-01 7.95887291e-01
5.90475872e-02 -1.78419441e-01 -8.20021987e-01 -1.79169491e-01
3.16057593e-01 -1.09964120e+00 2.91858852e-01 -6.78941190e-01
1.23464651e-01 -3.99003625e-01 7.01956451e-03 -1.01350331e+00
-1.47322446e-01 -2.37645447e-01 7.56613910e-01 1.57029390e+00
1.07954109e+00 -8.10327709e-01 4.85807598e-01 7.76561141e-01
-1.10199653e-01 -3.68669868e-01 -6.49910450e-01 -5.08547306e-01
2.18762025e-01 -4.21703249e-01 5.07366836e-01 1.16090214e+00
3.11297804e-01 5.93149900e-01 5.68013303e-02 -3.51491332e-01
1.05026066e-01 1.78247184e-01 7.10210323e-01 -1.95193446e+00
1.83377951e-01 -4.70474005e-01 -4.86958236e-01 -1.26679003e-01
6.38033673e-02 -1.50216675e+00 -5.27964652e-01 -1.93995869e+00
4.20612723e-01 -1.05983651e+00 -4.15322855e-02 8.97969961e-01
2.75801420e-01 3.75906199e-01 1.13615528e-01 6.60104096e-01
-4.59651351e-01 -3.32027048e-01 6.01632416e-01 5.93613796e-02
-1.33170322e-01 -5.36213890e-02 -6.52316868e-01 7.79632866e-01
9.03084457e-01 -1.05060923e+00 -2.04998717e-01 -2.90461838e-01
6.77134991e-01 -8.50626588e-01 -5.13560846e-02 -1.25094187e+00
-1.26178354e-01 -9.77080017e-02 1.40394881e-01 -4.66930151e-01
-2.92579643e-02 -1.01880050e+00 6.07793033e-01 3.24061811e-01
-5.73564529e-01 6.18356802e-02 -1.35642484e-01 2.07558095e-01
-1.00576207e-01 -1.04416668e+00 8.40769827e-01 -3.88989240e-01
-4.17434692e-01 -2.40093388e-04 -6.37704015e-01 1.43698514e-01
1.28711557e+00 -1.65386483e-01 -4.21889514e-01 -1.39949962e-01
-8.44871044e-01 -1.43468916e-01 9.15650189e-01 6.81098044e-01
-3.89041543e-01 -9.49496269e-01 -5.13316393e-01 -4.30119745e-02
3.82742554e-01 -4.07781899e-01 -6.98654726e-02 5.84136367e-01
-8.34422052e-01 4.24848765e-01 -5.22088885e-01 -2.67048657e-01
-1.59651518e+00 7.08791375e-01 3.36718321e-01 -4.50990617e-01
-5.51376283e-01 1.46647274e-01 -5.26189029e-01 -7.08047807e-01
8.95954743e-02 -7.09816068e-02 -8.77303779e-01 8.08863580e-01
5.70013165e-01 2.23777264e-01 6.06007934e-01 -9.34279859e-01
-2.22570300e-01 2.87185907e-01 5.28585203e-02 -1.25183448e-01
1.55380976e+00 -3.79165038e-02 -4.92007315e-01 8.87468636e-01
1.04455507e+00 3.74165624e-01 -5.60061097e-01 -2.38756649e-02
4.94226694e-01 -4.10088718e-01 -3.19908053e-01 -8.52197111e-01
-9.59591866e-01 5.90250015e-01 4.34721798e-01 9.61755335e-01
6.33330643e-01 2.35635370e-01 -1.77475344e-02 2.13823780e-01
2.56567061e-01 -1.22939944e+00 -3.13650936e-01 5.73169947e-01
8.09689820e-01 -9.22762096e-01 2.79725581e-01 -7.36407518e-01
-6.76413119e-01 1.25179744e+00 9.08820033e-02 5.28184116e-01
7.77472079e-01 3.22354198e-01 6.55871928e-01 -2.53491789e-01
-5.31134248e-01 -1.86764762e-01 -4.08206321e-02 7.79435158e-01
8.60343277e-01 -1.01182386e-01 -9.82427299e-01 6.44805908e-01
-3.59959871e-01 5.21853659e-03 9.59253490e-01 9.45488214e-01
-3.31763774e-01 -1.98481727e+00 -3.03826839e-01 6.42960131e-01
-8.45919669e-01 2.17158198e-01 -7.61177540e-01 9.42298770e-01
4.48823720e-01 1.09364188e+00 -2.17437461e-01 -8.18996653e-02
5.10914028e-01 3.99257123e-01 2.15172485e-01 -1.23643684e+00
-1.12789547e+00 -2.86353678e-01 4.36206013e-01 2.28910595e-02
-7.11554408e-01 -8.97791922e-01 -9.94338036e-01 -1.99712783e-01
-3.43206763e-01 6.10238254e-01 1.00268328e+00 8.24737251e-01
2.94371426e-01 7.99956843e-02 2.06982315e-01 -4.65961337e-01
-2.26065353e-01 -1.16820490e+00 -7.30387449e-01 7.64670432e-01
-3.25060546e-01 -6.66960776e-01 -4.64538157e-01 2.65891314e-01] | [9.861549377441406, 9.619660377502441] |
e265ff38-824f-400f-b23f-2178fe02a2d8 | panoptic-feature-pyramid-networks | 1901.02446 | null | http://arxiv.org/abs/1901.02446v2 | http://arxiv.org/pdf/1901.02446v2.pdf | Panoptic Feature Pyramid Networks | The recently introduced panoptic segmentation task has renewed our
community's interest in unifying the tasks of instance segmentation (for thing
classes) and semantic segmentation (for stuff classes). However, current
state-of-the-art methods for this joint task use separate and dissimilar
networks for instance and semantic segmentation, without performing any shared
computation. In this work, we aim to unify these methods at the architectural
level, designing a single network for both tasks. Our approach is to endow Mask
R-CNN, a popular instance segmentation method, with a semantic segmentation
branch using a shared Feature Pyramid Network (FPN) backbone. Surprisingly,
this simple baseline not only remains effective for instance segmentation, but
also yields a lightweight, top-performing method for semantic segmentation. In
this work, we perform a detailed study of this minimally extended version of
Mask R-CNN with FPN, which we refer to as Panoptic FPN, and show it is a robust
and accurate baseline for both tasks. Given its effectiveness and conceptual
simplicity, we hope our method can serve as a strong baseline and aid future
research in panoptic segmentation. | ['Piotr Dollár', 'Kaiming He', 'Ross Girshick', 'Alexander Kirillov'] | 2019-01-08 | panoptic-feature-pyramid-networks-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Kirillov_Panoptic_Feature_Pyramid_Networks_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Kirillov_Panoptic_Feature_Pyramid_Networks_CVPR_2019_paper.pdf | cvpr-2019-6 | ['thermal-image-segmentation'] | ['computer-vision'] | [ 4.52878982e-01 2.01739147e-01 -3.33236158e-01 -4.23319399e-01
-5.43824792e-01 -8.65080714e-01 6.21741951e-01 -2.54937381e-01
-2.64649510e-01 1.52532026e-01 8.35049972e-02 -2.95610458e-01
8.16378519e-02 -1.02909064e+00 -5.25105059e-01 -5.77869654e-01
1.07893862e-01 4.80882436e-01 5.87774158e-01 -8.07003379e-02
1.38698965e-01 6.59790218e-01 -1.54073739e+00 2.49074161e-01
8.67699981e-01 1.07923853e+00 1.45317838e-01 5.87557852e-01
-6.90881982e-02 5.05685329e-01 -4.91095454e-01 -5.42795718e-01
6.35820150e-01 -3.74305397e-01 -1.30995619e+00 1.26090452e-01
7.05931306e-01 -8.03537518e-02 -8.24203640e-02 9.91770089e-01
1.75389975e-01 2.62500674e-01 4.89323169e-01 -1.00895369e+00
-5.08473516e-01 9.32197392e-01 -6.41864240e-01 -1.35364115e-01
-3.88070881e-01 1.39109403e-01 1.55335510e+00 -4.78254557e-01
5.01364291e-01 1.18723774e+00 1.17199492e+00 4.93407577e-01
-1.41946507e+00 -4.92315322e-01 4.36898768e-01 -4.17873532e-01
-1.32984269e+00 -1.56021580e-01 5.49461722e-01 -3.17070127e-01
8.44011545e-01 2.74186820e-01 8.73302698e-01 5.86649597e-01
-1.81464285e-01 1.25874078e+00 1.13829899e+00 -3.69013399e-02
8.19199607e-02 -1.71558499e-01 3.79023641e-01 5.41209996e-01
2.41368487e-01 -1.30428687e-01 7.26338625e-02 2.44131058e-01
1.00577617e+00 6.49097785e-02 -5.22902124e-02 -3.19054425e-01
-1.06021142e+00 9.03577626e-01 7.88426578e-01 6.01409376e-01
-9.73836184e-02 4.24396962e-01 2.16573626e-01 1.02064893e-01
9.08685088e-01 6.78560376e-01 -7.99740613e-01 1.79001257e-01
-1.65264452e+00 3.67193550e-01 1.11207712e+00 6.81656063e-01
1.15329969e+00 -1.13140076e-01 -2.50911981e-01 9.26859438e-01
1.34220347e-01 1.64653793e-01 1.85771048e-01 -1.16910839e+00
-1.19131850e-02 6.90096498e-01 -1.79722711e-01 -8.48344088e-01
-6.74338520e-01 -5.59199691e-01 -7.50490904e-01 -6.64712256e-03
4.87872630e-01 -1.79106683e-01 -1.39317703e+00 1.72393537e+00
2.52327055e-01 3.28190953e-01 -4.51925285e-02 7.48259962e-01
8.92381966e-01 5.63985229e-01 1.04093745e-01 5.44981956e-01
1.60436893e+00 -1.55629039e+00 -1.85700014e-01 -3.26168865e-01
5.22188008e-01 -7.58973181e-01 9.28929329e-01 2.60788351e-01
-1.11963880e+00 -6.48494124e-01 -9.46438313e-01 -1.55773476e-01
-6.90396369e-01 9.57530066e-02 1.08889544e+00 7.65827537e-01
-1.45055783e+00 1.07751346e+00 -8.75616193e-01 -6.20467603e-01
7.10171163e-01 4.05468792e-01 3.44546214e-02 2.03267217e-01
-1.00850916e+00 6.80363655e-01 4.42111284e-01 2.39093546e-02
-6.78451300e-01 -8.43689620e-01 -7.38768816e-01 -6.71782345e-03
5.87117016e-01 -7.14118183e-01 1.42486656e+00 -1.08018649e+00
-1.58743703e+00 1.26718092e+00 3.00841421e-01 -7.60834813e-01
2.80620217e-01 -4.25715744e-01 -2.42360011e-02 2.82745868e-01
2.62767434e-01 1.33199668e+00 7.51909435e-01 -1.39617765e+00
-7.57668316e-01 -2.59427220e-01 2.43254870e-01 5.87820187e-02
2.74806377e-02 -3.49479839e-02 -6.59226358e-01 -8.02541912e-01
2.96519518e-01 -8.03922951e-01 -6.22053266e-01 -6.76125959e-02
-5.64872801e-01 -2.94200599e-01 7.50798106e-01 -5.55268586e-01
1.12594104e+00 -2.10438013e+00 2.05371380e-01 2.00346857e-01
3.36684406e-01 1.96942911e-01 -2.40394726e-01 1.61822420e-02
4.64029573e-02 4.47550148e-01 -1.10053146e+00 -5.67931890e-01
4.12378088e-02 5.11614740e-01 -2.88883388e-01 4.39733833e-01
2.31541902e-01 1.28110826e+00 -7.27723777e-01 -4.48698342e-01
2.23315075e-01 2.57832110e-01 -5.60216665e-01 1.22299992e-01
-5.34087002e-01 4.25437450e-01 -2.54860431e-01 7.50460625e-01
8.91235411e-01 -3.25378418e-01 -9.66192037e-02 -2.00690985e-01
-4.52553093e-01 6.56073466e-02 -1.11696231e+00 1.96007931e+00
-3.46896946e-01 3.40360969e-01 4.81383950e-01 -1.15611124e+00
8.63085330e-01 -9.82839316e-02 5.79034150e-01 -5.82296729e-01
5.67732528e-02 4.24631029e-01 -1.21636756e-01 -2.51077669e-04
6.92877889e-01 -6.16831183e-02 -3.09742093e-01 5.37918746e-01
1.76308006e-01 -9.10573363e-01 2.41982043e-01 1.83312744e-02
8.37466300e-01 4.75210518e-01 8.39311332e-02 -6.97053850e-01
3.68661672e-01 9.53694284e-02 5.35204113e-01 9.99340117e-01
-3.20218027e-01 1.05011809e+00 3.95108789e-01 -5.18537879e-01
-9.07817483e-01 -1.06294608e+00 -2.72222221e-01 1.29531074e+00
2.73384392e-01 -2.42753476e-01 -1.08370304e+00 -7.44360089e-01
8.32125247e-02 2.73651332e-01 -6.58467829e-01 3.81136000e-01
-8.36599767e-01 -9.73571777e-01 8.81344736e-01 7.50334799e-01
9.95365977e-01 -1.22192466e+00 -5.93549609e-01 2.17263505e-01
-9.14903544e-03 -1.15076387e+00 -3.07737559e-01 3.41894329e-01
-8.36593509e-01 -1.06961644e+00 -1.12596977e+00 -8.56555939e-01
2.14952424e-01 4.54423130e-01 1.48856282e+00 1.65230334e-01
-3.34350377e-01 3.28817844e-01 -2.58806139e-01 -8.38901401e-02
-6.62507713e-02 5.91606975e-01 -6.64000511e-01 1.03924163e-01
1.64375052e-01 -7.67711639e-01 -8.65505040e-01 3.23333085e-01
-1.27444816e+00 1.57029763e-01 3.85714322e-01 2.63211131e-01
5.64719379e-01 -2.70439893e-01 2.70755976e-01 -1.08823323e+00
3.27571213e-01 -4.13803309e-01 -6.96250677e-01 1.64335757e-01
-2.45652050e-01 -2.44365260e-01 4.28745687e-01 7.65580162e-02
-9.80400920e-01 -1.76357981e-02 -5.03813744e-01 3.59616950e-02
-3.91681373e-01 3.45936984e-01 7.23671243e-02 -6.08415492e-02
3.58100444e-01 -4.30086069e-02 -1.74519241e-01 -7.73125410e-01
9.10556555e-01 5.05305588e-01 6.86279297e-01 -5.88844001e-01
7.19331980e-01 8.24874163e-01 3.34415436e-02 -7.49418974e-01
-1.40876126e+00 -7.68140674e-01 -9.13229883e-01 2.24542364e-01
1.35126853e+00 -9.94970798e-01 -2.82316685e-01 7.82350183e-01
-1.01656795e+00 -7.94888854e-01 -5.78406870e-01 -1.49495393e-01
-7.20804513e-01 4.98159021e-01 -7.05145538e-01 -4.24876660e-01
-3.23394179e-01 -1.16544175e+00 1.35309017e+00 3.07982206e-01
-8.49029347e-02 -1.22699749e+00 1.07592218e-01 4.11270201e-01
6.67983651e-01 3.36264074e-01 7.35401511e-01 -8.41084063e-01
-7.24559724e-01 1.44248098e-01 -7.21198976e-01 6.53308868e-01
1.05560571e-01 1.31955430e-01 -1.12126756e+00 1.18812546e-02
-8.41242000e-02 -2.54126310e-01 1.62241518e+00 6.39258623e-01
1.31977689e+00 8.38218108e-02 -3.23868394e-01 9.69906092e-01
1.51662207e+00 -2.34741345e-01 6.85813963e-01 2.92077482e-01
9.34846818e-01 8.51076841e-01 1.22639999e-01 2.40025464e-02
3.50942075e-01 5.29622793e-01 3.36569250e-01 -6.56033099e-01
-4.98563886e-01 1.10895507e-01 -4.76688445e-02 6.65139973e-01
-4.46894281e-02 -2.41895080e-01 -1.05443347e+00 7.54306436e-01
-2.07302284e+00 -6.01354420e-01 -1.03142552e-01 1.79542696e+00
8.29601467e-01 -1.59168988e-01 1.45535842e-01 -1.08831547e-01
7.39978313e-01 5.93202174e-01 -3.19511652e-01 -3.95677745e-01
-2.40537420e-01 8.03049266e-01 5.54920316e-01 4.20908421e-01
-1.68777430e+00 1.44662678e+00 6.62093687e+00 1.00594175e+00
-1.00583398e+00 1.73265800e-01 8.38595748e-01 3.57352287e-01
-1.52069792e-01 1.72019377e-01 -8.59577298e-01 1.14960201e-01
8.35975230e-01 3.42136294e-01 3.57041478e-01 7.54065216e-01
-2.45650291e-01 -2.11909503e-01 -8.95035386e-01 5.35636783e-01
-1.50556102e-01 -1.41856658e+00 1.07316054e-01 -2.42463723e-01
9.90385711e-01 4.60706472e-01 2.19549965e-02 2.79525250e-01
6.32800400e-01 -1.25783467e+00 5.85456252e-01 2.14701489e-01
4.24338251e-01 -5.12436152e-01 6.38439238e-01 -3.94142233e-02
-1.53360820e+00 1.68437093e-01 -1.55726209e-01 -4.69699651e-02
2.23203704e-01 6.87733233e-01 -6.26884252e-02 7.07201004e-01
7.81061351e-01 1.08449912e+00 -4.95735615e-01 1.04803705e+00
-2.93192208e-01 7.18862474e-01 -5.13632178e-01 5.18371105e-01
9.35574889e-01 -3.44844222e-01 2.32620209e-01 1.54894066e+00
-1.19171731e-01 -2.68091351e-01 6.19033337e-01 1.17814136e+00
-3.43671620e-01 7.22448304e-02 -4.71489429e-01 -3.78309451e-02
-1.01262957e-01 1.53232014e+00 -1.61861205e+00 -5.68379402e-01
-5.05802870e-01 9.36801732e-01 1.36591330e-01 2.17631638e-01
-7.22931445e-01 -3.79832327e-01 7.32234418e-01 -3.31560493e-01
7.12283611e-01 -1.43519983e-01 -6.65183663e-01 -1.10829890e+00
-3.03774446e-01 -5.54296315e-01 2.80447155e-01 -3.61457258e-01
-1.31861281e+00 5.03957033e-01 5.98593391e-02 -7.43117392e-01
3.36772591e-01 -6.59415841e-01 -6.50912404e-01 6.44305885e-01
-1.87599909e+00 -1.51528811e+00 -2.09604919e-01 3.95812184e-01
7.12123573e-01 3.90667558e-01 6.06379747e-01 1.15273304e-01
-5.11262834e-01 2.25110888e-01 -5.81115894e-02 2.69538134e-01
4.49499071e-01 -1.49004006e+00 8.90218019e-01 8.19685102e-01
2.19289988e-01 3.95551026e-01 3.43530923e-01 -5.88209212e-01
-7.04125881e-01 -1.41337645e+00 6.73756242e-01 -4.18382198e-01
8.06788146e-01 -4.16385591e-01 -8.06654990e-01 7.41498649e-01
2.51350790e-01 -8.69151298e-03 5.52759349e-01 1.26121521e-01
-3.95182908e-01 -2.56418553e-03 -1.13538122e+00 6.51781023e-01
1.10113287e+00 -2.21724734e-01 -6.79865956e-01 3.83134067e-01
1.15765405e+00 -2.09760979e-01 -8.38379025e-01 7.16010630e-01
5.23359716e-01 -1.16851807e+00 1.03610635e+00 -9.48006064e-02
4.80064958e-01 -1.67440355e-01 -2.06365123e-01 -9.52478051e-01
-2.12973699e-01 -7.48287797e-01 2.60493785e-01 1.09187567e+00
4.34921890e-01 -6.55965209e-01 9.05850232e-01 2.28893533e-01
-4.91950542e-01 -8.48622024e-01 -8.07374239e-01 -8.02039862e-01
4.72292602e-01 -5.35715401e-01 7.34278381e-01 6.90961957e-01
-7.06933439e-01 1.18337683e-01 -1.50738344e-01 -5.20594902e-02
4.44588870e-01 6.00185871e-01 6.88529611e-01 -1.56053567e+00
-1.30322620e-01 -8.75787735e-01 -4.70285714e-02 -1.61252773e+00
1.82814047e-01 -1.02095354e+00 2.74385184e-01 -1.62263119e+00
2.06440881e-01 -5.23398995e-01 -2.64275253e-01 5.65469265e-01
2.87176762e-02 8.59856069e-01 3.60850215e-01 3.54076803e-01
-6.95888042e-01 3.41311425e-01 1.47065830e+00 -1.34463906e-01
-3.38164806e-01 2.46177658e-01 -9.53724384e-01 9.90876853e-01
8.88687551e-01 -3.21111232e-01 -1.37067184e-01 -4.68379796e-01
1.44194521e-03 -4.57327485e-01 4.81972873e-01 -1.09217763e+00
4.05673571e-02 -2.19532885e-02 -5.29387705e-02 -6.88865900e-01
2.46546611e-01 -6.34377301e-01 -2.86305934e-01 3.55290622e-01
-5.49889430e-02 -2.13299215e-01 2.78714359e-01 2.12315172e-01
-3.69463712e-01 -3.71248662e-01 8.96531165e-01 -7.14446783e-01
-8.33343506e-01 5.09665787e-01 -2.17423707e-01 2.42313430e-01
7.74306536e-01 -5.22446036e-01 -1.27503678e-01 5.76200858e-02
-8.29133689e-01 3.34120780e-01 5.19151211e-01 2.84234285e-01
1.23646699e-01 -7.21928835e-01 -5.34323752e-01 2.28881813e-03
-2.04248279e-01 4.22325999e-01 6.59347996e-02 7.97280133e-01
-7.30969012e-01 4.69181478e-01 -3.37684341e-02 -7.03511059e-01
-8.08035195e-01 1.63780749e-01 4.51082915e-01 -6.10278726e-01
-7.56333530e-01 1.05196595e+00 5.88445544e-01 -8.48741651e-01
1.26802444e-01 -6.42047524e-01 -2.71518320e-01 3.16249251e-01
1.80080563e-01 2.80788630e-01 -1.04571395e-01 -4.05614465e-01
-1.32247642e-01 8.75581980e-01 -1.23084644e-02 7.88756981e-02
1.58329225e+00 -4.93764877e-03 -6.24294281e-01 3.68989050e-01
1.03568029e+00 -3.20131421e-01 -1.29872632e+00 -2.25921825e-01
8.53933170e-02 1.36904325e-02 1.15905907e-02 -8.34004879e-01
-1.39968383e+00 8.99887681e-01 2.27551654e-01 5.68513215e-01
1.20120788e+00 2.19067067e-01 1.12791288e+00 1.96253181e-01
2.01494575e-01 -1.04486859e+00 -1.02097914e-01 8.28136384e-01
4.99585301e-01 -1.03424478e+00 9.00718719e-02 -7.65072227e-01
-3.09521228e-01 1.07338262e+00 5.40215909e-01 -4.59548891e-01
8.26738656e-01 1.54060423e-01 2.31164675e-02 -4.26790029e-01
-9.14279670e-02 -7.55237937e-01 2.75352448e-01 4.10262167e-01
3.82423460e-01 2.53666371e-01 -1.40122399e-01 4.78575528e-01
-2.14743629e-01 -1.99702933e-01 1.29166797e-01 9.29866731e-01
-6.78457320e-01 -1.18978083e+00 -8.91681314e-02 3.81085485e-01
-6.15118921e-01 -3.59932572e-01 -4.69099730e-01 1.06807733e+00
4.13961172e-01 7.34864235e-01 3.27812970e-01 -8.16637948e-02
1.02599606e-01 -3.85542698e-02 3.92203808e-01 -8.95117283e-01
-1.20483291e+00 6.10969253e-02 -1.35842413e-01 -6.78896725e-01
-9.14653420e-01 -2.92035103e-01 -1.00100422e+00 -6.55330867e-02
-2.15728372e-01 -2.18338780e-02 5.34588218e-01 9.42615747e-01
1.46329790e-01 3.28916848e-01 3.87828171e-01 -1.20680034e+00
-2.59477198e-01 -8.36919546e-01 -5.90481937e-01 3.62780750e-01
1.63806632e-01 -3.52311105e-01 -3.93776983e-01 -6.66148141e-02] | [9.579144477844238, 0.37181025743484497] |
7f71fced-cf44-4738-8d0e-7137d97e080e | procedural-editing-of-3d-building-point | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Demir_Procedural_Editing_of_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Demir_Procedural_Editing_of_ICCV_2015_paper.pdf | Procedural Editing of 3D Building Point Clouds | Thanks to the recent advances in computational photography and remote sensing, point clouds of buildings are becoming increasingly available, yet their processing poses various challenges. In our work, we tackle the problem of point cloud completion and editing and we approach it via inverse procedural modeling. Contrary to the previous work, our approach operates directly on the point cloud without an intermediate triangulation. Our approach consists of 1) semi-automatic segmentation of the input point cloud with segment comparison and template matching to detect repeating structures, 2) a consensus-based voting schema and a pattern extraction algorithm to discover completed terminal geometry and their patterns of usage, all encoded into a context-free grammar, and 3) an interactive editing tool where the user can create new point clouds by using procedural copy and paste operations, and smart resizing. We demonstrate our approach on editing of building models with up to 1.8M points. In our implementation, preprocessing takes up to several minutes and a single editing operation needs from one second to one minute depending on the model size and the operation type. | ['Daniel G. Aliaga', 'Bedrich Benes', 'Ilke Demir'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['point-cloud-completion'] | ['computer-vision'] | [ 7.51219153e-01 -7.50520602e-02 5.11125565e-01 -2.97063440e-01
-6.25567615e-01 -8.20419371e-01 6.31757379e-01 5.78687847e-01
-2.91630507e-01 2.84197897e-01 -5.96886039e-01 -5.42230487e-01
-2.10420221e-01 -1.18687546e+00 -5.83015740e-01 -1.75782412e-01
2.09398400e-02 9.09085810e-01 6.49814010e-01 -1.74854830e-01
5.41319013e-01 9.68116999e-01 -1.88523853e+00 -8.10954943e-02
1.27158642e+00 8.23758245e-01 4.54208314e-01 8.80540133e-01
-5.64995110e-01 4.69981097e-02 -2.01894954e-01 -2.15580046e-01
5.52738070e-01 -8.12209994e-02 -8.48071992e-01 4.71138000e-01
2.46120781e-01 -8.52483511e-02 6.08933091e-01 9.37097251e-01
2.01274648e-01 -8.52715150e-02 5.05323887e-01 -1.03493464e+00
8.72855708e-02 2.19479442e-01 -5.84039450e-01 -5.00724494e-01
3.44141304e-01 -4.00116369e-02 6.59671009e-01 -1.17028034e+00
6.68396592e-01 7.11471796e-01 8.24151993e-01 -6.50366619e-02
-1.34688449e+00 -4.54852670e-01 -1.58707649e-01 -1.79945361e-02
-1.84000945e+00 -3.30240160e-01 6.95994198e-01 -5.89864433e-01
7.63998270e-01 7.34380305e-01 1.03536475e+00 -1.43029943e-01
-3.38840157e-01 -4.70976233e-02 1.05352962e+00 -7.74414003e-01
5.97276688e-01 9.55119822e-03 -7.88335726e-02 5.37178159e-01
-1.85520668e-02 -2.59919763e-01 2.24708989e-01 -4.90882844e-01
9.54375982e-01 2.62973644e-02 -7.89876580e-02 -5.07810950e-01
-1.04245698e+00 6.06154025e-01 2.62472510e-01 1.49683461e-01
-5.16333818e-01 1.53229073e-01 -6.58848882e-02 9.04308930e-02
5.18059850e-01 1.80382386e-01 -4.37116235e-01 3.28465812e-02
-1.41555107e+00 2.43815452e-01 7.21348763e-01 1.24097800e+00
1.34612548e+00 -4.61748809e-01 6.11933947e-01 7.11388886e-01
3.68533909e-01 4.12546009e-01 -2.37797618e-01 -1.07635450e+00
4.64792609e-01 9.95747566e-01 2.77370036e-01 -1.11610460e+00
-3.40951711e-01 7.11537478e-03 -5.72177291e-01 5.46907902e-01
2.10704356e-01 1.38990059e-01 -9.93261158e-01 9.16909933e-01
7.72095144e-01 1.69095751e-02 -4.82217401e-01 3.55932385e-01
3.32179010e-01 4.71805185e-01 -1.02063477e-01 -2.19219357e-01
1.34986091e+00 -3.86037469e-01 -2.68002987e-01 -1.32003864e-02
7.63011098e-01 -1.13496792e+00 7.73653388e-01 3.56415749e-01
-1.01182759e+00 -2.92888105e-01 -9.16664660e-01 -7.45673850e-02
-3.22005630e-01 1.78868577e-01 2.05840811e-01 6.50801182e-01
-1.12288809e+00 7.46496439e-01 -7.79132247e-01 -5.23874402e-01
3.25803399e-01 6.20962083e-01 -1.33083388e-01 8.30486491e-02
-4.01899070e-01 7.10568547e-01 3.25902194e-01 3.52712542e-01
-1.27626806e-02 -7.73027778e-01 -5.19234538e-01 3.90291400e-02
5.19661486e-01 -8.75749230e-01 1.02751899e+00 -8.83274615e-01
-1.60266781e+00 1.05455959e+00 -1.68386072e-01 -7.85936601e-04
6.40553236e-01 -6.20226283e-03 -4.94156964e-02 -2.21490450e-02
1.65283635e-01 3.03389579e-01 6.11027837e-01 -1.54453468e+00
-8.03607702e-01 -6.58315599e-01 3.90069224e-02 4.53864411e-02
3.98503780e-01 2.04027265e-01 -6.92908466e-01 -4.74056184e-01
7.31833041e-01 -1.05868030e+00 -6.54130638e-01 3.39148432e-01
-1.94412649e-01 7.66123012e-02 8.41603577e-01 -7.87658334e-01
1.16036999e+00 -2.01440859e+00 -8.05870106e-04 9.85359967e-01
1.62409768e-02 4.77151610e-02 2.71534622e-01 6.21549845e-01
4.54898700e-02 2.10645974e-01 -9.88198280e-01 -3.36082816e-01
-1.22518376e-01 3.70168835e-01 -2.74021327e-01 3.74746084e-01
-3.96593697e-02 3.98989767e-01 -7.69063115e-01 -6.63312256e-01
6.38914585e-01 2.73286670e-01 -4.12188768e-01 6.60381373e-03
-3.05609018e-01 3.79828095e-01 -5.17516971e-01 6.93271935e-01
1.03892303e+00 8.72102845e-03 3.42201501e-01 1.01198688e-01
-8.17839384e-01 2.36595888e-02 -1.74043155e+00 1.73186135e+00
-5.72295427e-01 5.88109195e-02 4.40235972e-01 -6.45169854e-01
1.14996648e+00 1.67017654e-01 3.99560452e-01 -2.03593612e-01
-4.75927889e-02 5.10637105e-01 -4.94377583e-01 -1.60406038e-01
7.96553671e-01 1.27430692e-01 2.03234315e-01 6.00599170e-01
-5.36345840e-01 -9.01278675e-01 6.05955496e-02 -1.63976047e-02
7.87588358e-01 4.07270461e-01 5.01270175e-01 -3.22124511e-01
5.50333321e-01 5.05296171e-01 4.07518744e-01 4.05340165e-01
6.65697217e-01 6.52895331e-01 9.36497822e-02 -5.60649991e-01
-1.12100542e+00 -7.36728191e-01 -1.20935261e-01 7.18147814e-01
1.41622514e-01 -5.81058264e-01 -8.88013184e-01 -1.13574015e-02
-1.49823248e-01 7.08250999e-01 -2.22069442e-01 6.03918672e-01
-8.08418214e-01 -5.29082239e-01 6.75042570e-02 3.03001702e-01
3.49198401e-01 -8.19051862e-01 -1.06928813e+00 3.98779452e-01
2.12009624e-01 -8.00963044e-01 -1.56406432e-01 -1.33142188e-01
-1.36004400e+00 -1.08863592e+00 -3.95429671e-01 -5.75005412e-01
1.16274512e+00 2.63833135e-01 1.13716853e+00 6.25427842e-01
-3.12893391e-01 4.02856857e-01 -3.79238099e-01 -5.43595850e-01
-4.89617288e-01 8.07419568e-02 -4.76809978e-01 2.44532581e-02
-1.01163663e-01 -1.09477949e+00 -4.86733377e-01 3.51679981e-01
-9.86983180e-01 4.18533117e-01 4.35905099e-01 3.57172042e-02
9.63886797e-01 4.33792844e-02 -1.93167970e-01 -9.65697825e-01
2.17255041e-01 -1.83513448e-01 -1.21721995e+00 3.72961730e-01
-5.97459257e-01 -2.55696148e-01 3.83269995e-01 2.49677792e-01
-8.39377820e-01 7.84472644e-01 -7.40248710e-02 -2.60974318e-01
-2.92577296e-01 5.44428527e-01 -8.53552297e-02 -3.28671843e-01
4.98218179e-01 1.40296698e-01 -7.44276196e-02 -6.04687333e-01
6.62525415e-01 5.86574018e-01 4.98512268e-01 -4.95030940e-01
1.25703645e+00 7.20958889e-01 2.87781078e-02 -1.13652432e+00
-3.11233457e-02 -4.46967065e-01 -1.49339402e+00 -3.71592432e-01
8.52692008e-01 -4.77001667e-01 -4.34384674e-01 4.55816537e-01
-1.40136731e+00 -1.72164410e-01 -3.28476548e-01 7.07740337e-02
-4.90741402e-01 6.25460625e-01 2.40773335e-02 -7.97753274e-01
-3.02574515e-01 -1.03580439e+00 1.16595733e+00 -1.53173387e-01
-1.75074860e-01 -4.95178372e-01 1.75109223e-01 2.01555774e-01
2.44144961e-01 5.28824091e-01 9.13494587e-01 2.66812975e-03
-1.12607396e+00 -1.41348869e-01 -1.09913997e-01 -1.16212316e-01
1.00242019e-01 6.11443341e-01 -6.92781568e-01 4.96062972e-02
-2.43121669e-01 5.17216682e-01 1.12701863e-01 1.70842558e-01
1.06840789e+00 2.64057610e-02 -5.86109102e-01 6.36453390e-01
1.58846831e+00 2.43719041e-01 8.88734639e-01 3.83570731e-01
7.39816308e-01 7.08141506e-01 5.41859388e-01 4.76700336e-01
4.96356815e-01 8.75202715e-01 5.94439209e-01 -5.48629947e-02
3.44569802e-01 1.67190339e-02 -3.16786557e-01 9.24096107e-01
-7.28579760e-01 3.38432461e-01 -1.35301912e+00 4.10720170e-01
-1.71162748e+00 -8.57876599e-01 -8.34278524e-01 2.62050676e+00
4.81451124e-01 -2.40248084e-01 -8.76877829e-02 3.80445063e-01
8.10465991e-01 -3.30852568e-02 -4.02719602e-02 -4.53027695e-01
4.34926063e-01 5.80083787e-01 6.78411424e-01 7.11861968e-01
-8.56496513e-01 8.43430579e-01 5.87282610e+00 4.28194612e-01
-1.13681960e+00 -9.79355630e-03 -2.84893508e-03 2.68854171e-01
-3.31504017e-01 5.59245646e-01 -3.59305888e-01 2.59937078e-01
4.46505845e-01 -3.42487060e-02 4.61546868e-01 9.33565259e-01
2.66013473e-01 -4.06743139e-01 -6.90416753e-01 9.94371831e-01
-2.35717699e-01 -1.35653973e+00 -7.86622018e-02 4.03716207e-01
5.86638093e-01 1.54667851e-02 -6.51063025e-01 -2.64962226e-01
2.26300701e-01 -7.55850315e-01 1.00374603e+00 6.67825401e-01
8.72823834e-01 -7.65732825e-01 2.17338815e-01 5.24039984e-01
-1.52370763e+00 2.64003575e-01 -2.34676942e-01 -2.02558741e-01
3.19311172e-01 7.85287917e-01 -8.19377363e-01 9.56965566e-01
5.51551402e-01 2.36555532e-01 -4.83302534e-01 1.03485382e+00
-8.78182575e-02 2.92533487e-01 -8.78676951e-01 2.10691750e-01
-9.06871632e-02 -8.38667154e-01 4.00100976e-01 1.00796425e+00
6.49443984e-01 4.29978073e-01 3.03076804e-01 9.22507167e-01
2.63733923e-01 4.14242685e-01 -3.90918881e-01 5.72101831e-01
6.75117254e-01 1.22407746e+00 -1.20988321e+00 -3.78497601e-01
-8.80718306e-02 8.29644442e-01 2.84581631e-02 -6.26980066e-02
-5.90696692e-01 -5.21674573e-01 2.50516504e-01 4.20283079e-01
5.56007743e-01 -7.14130282e-01 -6.61455274e-01 -7.99745917e-01
3.84440601e-01 -4.40728396e-01 -3.95129658e-02 -8.39966536e-01
-6.39103472e-01 5.14347613e-01 1.04914509e-01 -1.37923324e+00
-1.92203507e-01 -1.74200043e-01 -8.13108563e-01 9.16809857e-01
-1.22261870e+00 -1.17858911e+00 -5.81174314e-01 3.46563965e-01
3.24818194e-01 5.27990997e-01 9.33098495e-01 3.18741858e-01
-1.79071158e-01 -1.69031814e-01 2.39511132e-01 -2.16201484e-01
9.69884619e-02 -1.00794399e+00 7.43400574e-01 9.05760169e-01
4.45574410e-02 4.39550936e-01 6.78578675e-01 -7.57988095e-01
-1.28446341e+00 -1.14555550e+00 1.04883885e+00 -3.06982428e-01
2.88770527e-01 -4.42386597e-01 -1.02901268e+00 6.78145766e-01
-1.11168995e-01 -4.62664291e-02 3.72491151e-01 -1.72135234e-01
9.22371149e-02 -1.25397304e-02 -1.17138791e+00 3.85532409e-01
1.22920489e+00 -3.31378639e-01 -5.05758941e-01 3.71705711e-01
1.70444518e-01 -5.50404370e-01 -9.05099630e-01 3.60109597e-01
4.34692979e-01 -9.79147077e-01 7.84433186e-01 1.00602299e-01
1.71515644e-02 -9.66554821e-01 -2.57743839e-02 -9.81851697e-01
-1.98547289e-01 -6.51371181e-01 6.45738423e-01 1.22733390e+00
3.57243627e-01 -5.12272298e-01 6.06710434e-01 8.25960875e-01
-3.01492751e-01 -5.02721906e-01 -1.06708765e+00 -3.64230156e-01
-3.93391788e-01 -7.46779799e-01 8.42861354e-01 9.87302005e-01
-5.46368599e-01 -2.90262252e-01 9.32707489e-02 6.40366495e-01
5.77655852e-01 5.26795864e-01 1.21269429e+00 -1.89301753e+00
-2.74871867e-02 -3.02244276e-01 -3.01251352e-01 -8.53849888e-01
-2.86183238e-01 -7.31508553e-01 -2.54416782e-02 -1.77178049e+00
-3.36242020e-01 -1.07935238e+00 6.57435119e-01 3.70588094e-01
2.78299361e-01 1.69954523e-01 1.47759125e-01 4.17539746e-01
-8.66117552e-02 1.96561053e-01 5.58263302e-01 1.35989234e-01
-4.82040048e-01 1.28538370e-01 -1.75768822e-01 9.47886884e-01
6.26618743e-01 -7.11064756e-01 -1.92620993e-01 -7.57743835e-01
4.08746123e-01 -9.49991569e-02 4.76428717e-01 -1.00778091e+00
3.00526649e-01 -2.63354599e-01 7.67398849e-02 -9.35624123e-01
4.00173157e-01 -1.22196388e+00 9.60272074e-01 4.08374012e-01
3.76387924e-01 2.29008287e-01 6.86228797e-02 3.99150580e-01
1.26846313e-01 -6.02929354e-01 6.16353869e-01 -4.19040799e-01
-5.36449134e-01 2.04020068e-01 -3.68967623e-01 -5.67006648e-01
1.25486338e+00 -7.28414476e-01 2.31780022e-01 7.28705004e-02
-8.71838033e-01 8.95535573e-02 1.18768597e+00 -3.65618914e-02
4.69070733e-01 -9.99955654e-01 -5.48987925e-01 3.09575647e-01
-7.61543065e-02 5.64669192e-01 1.47178382e-01 6.07455969e-01
-1.27238643e+00 2.13590581e-02 -3.03226355e-02 -8.84964705e-01
-1.57752872e+00 2.23094210e-01 2.67286271e-01 -7.48586236e-03
-6.70221806e-01 3.23713273e-01 -3.21838886e-01 -6.28820717e-01
-4.68273938e-01 -6.56569898e-01 -1.91263203e-02 7.14993626e-02
2.17189461e-01 6.29589856e-01 5.06323457e-01 -8.06112468e-01
-4.78198677e-01 1.30768347e+00 4.26790267e-01 -2.14645416e-01
1.49270511e+00 -9.03871879e-02 -6.06999516e-01 2.06638992e-01
5.85251927e-01 3.52369100e-01 -7.54621387e-01 -6.29174784e-02
3.74966502e-01 -6.09774053e-01 -5.91168217e-02 -2.62875885e-01
-6.57418191e-01 6.37224495e-01 3.85021865e-01 3.42723370e-01
1.08521867e+00 -5.69526777e-02 4.87066686e-01 2.56655633e-01
9.10246193e-01 -1.01927543e+00 -9.54554081e-01 4.31825846e-01
9.82637286e-01 -6.54444516e-01 4.01957154e-01 -9.28791821e-01
-3.17555629e-02 1.34819961e+00 -1.16912521e-01 -1.57833010e-01
8.63772273e-01 2.17307061e-01 4.65588234e-02 -3.64553809e-01
-2.54388422e-01 -2.09653631e-01 -5.87589853e-03 4.30736721e-01
-2.92044226e-02 1.49080321e-01 -4.69947636e-01 -6.86011985e-02
-5.31191528e-01 2.38847882e-01 4.94438171e-01 1.35103703e+00
-6.49513602e-01 -1.33908057e+00 -7.88888931e-01 4.43738341e-01
8.00045431e-02 4.05088067e-02 -3.87797773e-01 7.70084679e-01
1.98848888e-01 7.54004776e-01 2.91122705e-01 -2.40875095e-01
7.24184752e-01 7.30660409e-02 3.24914783e-01 -8.13468874e-01
-6.41823351e-01 -6.03045635e-02 -1.01042710e-01 -4.21627492e-01
-5.59685528e-01 -8.35365295e-01 -1.27682698e+00 -1.68660402e-01
-5.00303805e-01 6.63266778e-02 1.16325510e+00 8.48252356e-01
6.01495981e-01 -8.08892176e-02 6.50244296e-01 -1.18864799e+00
-7.74008185e-02 -6.48677349e-01 -4.48955506e-01 6.03650697e-02
-2.51390755e-01 -4.57841784e-01 -1.18912958e-01 3.35336298e-01] | [8.311504364013672, -2.723179578781128] |
cd54cc05-dfb7-4e57-a021-458aa73f8dca | statistically-consistent-k-mer-methods-for | 1511.01956 | null | http://arxiv.org/abs/1511.01956v2 | http://arxiv.org/pdf/1511.01956v2.pdf | Statistically-Consistent k-mer Methods for Phylogenetic Tree Reconstruction | Frequencies of $k$-mers in sequences are sometimes used as a basis for
inferring phylogenetic trees without first obtaining a multiple sequence
alignment. We show that a standard approach of using the squared-Euclidean
distance between $k$-mer vectors to approximate a tree metric can be
statistically inconsistent. To remedy this, we derive model-based distance
corrections for orthologous sequences without gaps, which lead to consistent
tree inference. The identifiability of model parameters from $k$-mer
frequencies is also studied. Finally, we report simulations showing the
corrected distance out-performs many other $k$-mer methods, even when sequences
are generated with an insertion and deletion process. These results have
implications for multiple sequence alignment as well, since $k$-mer methods are
usually the first step in constructing a guide tree for such algorithms. | [] | 2016-01-14 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 7.02633619e-01 -4.26124722e-01 -1.89492553e-01 -6.09955490e-01
-7.70703733e-01 -9.74630237e-01 1.10530213e-01 3.68714392e-01
-7.18192160e-01 1.00712132e+00 -3.93839777e-01 -7.69572198e-01
-1.53066859e-01 -5.04011333e-01 -5.91491520e-01 -8.89010131e-01
-2.91795939e-01 6.01601243e-01 1.62553087e-01 -6.80284435e-03
7.39216030e-01 5.62395811e-01 -1.29634833e+00 -1.49076924e-01
8.53164375e-01 6.19349107e-02 5.54699361e-01 1.01829278e+00
-2.95750678e-01 -1.55531839e-01 -4.97792870e-01 -6.00166976e-01
2.11263284e-01 -9.02487278e-01 -9.04810846e-01 -1.94981053e-01
3.67983073e-01 -3.03195179e-01 1.40582338e-01 9.10881937e-01
3.10269803e-01 -1.74547821e-01 9.35887158e-01 -9.91993964e-01
-2.19488859e-01 6.72955215e-01 -7.31600821e-01 1.54218093e-01
2.21659586e-01 2.95940161e-01 1.32684004e+00 -5.69591820e-01
9.08279121e-01 1.40114295e+00 1.11030984e+00 8.46346915e-02
-1.70470190e+00 -3.53736192e-01 -4.43134785e-01 2.80389398e-01
-1.60672200e+00 -9.82512087e-02 3.73667508e-01 -6.51699185e-01
1.19093609e+00 3.73747647e-01 5.16001284e-01 5.88267148e-01
4.64263827e-01 2.24379286e-01 8.35379064e-01 -4.25990820e-01
1.56765625e-01 -3.44144702e-01 2.06232488e-01 5.29514790e-01
5.87227106e-01 -1.41558751e-01 -2.38162860e-01 -8.74559700e-01
6.56172991e-01 -3.40618193e-01 -7.01939911e-02 -2.48143151e-01
-1.17931998e+00 1.01891220e+00 3.11047793e-03 2.89101571e-01
-2.60493070e-01 3.34398538e-01 3.73313546e-01 5.99684119e-02
2.88002249e-02 5.68842530e-01 -4.60638642e-01 -4.20088083e-01
-7.72845984e-01 4.29793864e-01 6.39975965e-01 9.63598490e-01
8.43039334e-01 -9.30967182e-02 6.72764421e-01 9.41732883e-01
1.83518216e-01 1.99369669e-01 1.79095671e-01 -1.17738473e+00
-8.09873268e-02 1.30420610e-01 3.27560186e-01 -7.56606400e-01
-1.76935554e-01 1.92963451e-01 -5.27721286e-01 -1.25398800e-01
8.42445076e-01 2.37719238e-01 -5.40513754e-01 1.85897446e+00
5.70289969e-01 1.46438375e-01 -1.34708360e-01 3.59875858e-01
-1.59926265e-01 5.13755143e-01 1.31057233e-01 -7.66509831e-01
1.34826958e+00 -2.65146822e-01 -3.40097427e-01 6.08345605e-02
9.81878579e-01 -9.63723898e-01 8.63530040e-01 3.26735526e-01
-7.64293015e-01 -1.97242007e-01 -9.64541852e-01 3.84919383e-02
1.24191726e-02 -5.69500864e-01 5.24964929e-01 7.36660242e-01
-9.90517259e-01 1.02779353e+00 -7.97823131e-01 -7.04748809e-01
-1.85952663e-01 2.05314502e-01 -3.38908166e-01 1.06672324e-01
-6.86937213e-01 1.09159827e+00 5.88455260e-01 -1.59480050e-01
-3.77324671e-01 -3.98582876e-01 -3.37059647e-01 -2.77399957e-01
-8.07977691e-02 -4.80110765e-01 1.11115694e+00 -4.88838881e-01
-8.92758608e-01 1.02860439e+00 -5.33196092e-01 -3.42764974e-01
2.11199760e-01 2.71388382e-01 1.15789562e-01 2.06359997e-01
5.73807545e-02 5.43613851e-01 3.40008974e-01 -8.60696018e-01
-4.13073331e-01 -4.52752858e-01 -6.27554953e-01 -7.55705014e-02
4.32079375e-01 2.19183818e-01 1.49160743e-01 -5.25778115e-01
4.22672719e-01 -1.06468260e+00 -5.56152642e-01 1.54672757e-01
-1.91655144e-01 -2.47483298e-01 3.00100833e-01 -6.62676036e-01
1.11796546e+00 -1.80607998e+00 2.92911291e-01 3.08928221e-01
5.23685701e-02 1.19518496e-01 -4.01453018e-01 1.01264548e+00
-4.54008192e-01 4.74886030e-01 -7.80623555e-01 2.68964231e-01
-2.05452442e-01 4.63510573e-01 -1.45364285e-01 7.30097950e-01
4.28279229e-02 3.82171184e-01 -8.57121289e-01 -4.99189079e-01
1.49604515e-03 2.36908168e-01 -4.42506433e-01 6.83660358e-02
-1.62249997e-01 1.80694610e-01 -1.59278065e-01 3.29265505e-01
7.74828315e-01 -2.99716741e-01 8.78512442e-01 1.29835397e-01
-2.13593438e-01 3.25334162e-01 -8.52550924e-01 1.50791395e+00
8.95003304e-02 4.17755723e-01 -5.08532114e-02 -1.14107490e+00
1.10832846e+00 -7.92286098e-02 5.03977299e-01 1.96740359e-01
2.65823919e-02 5.53245902e-01 6.73325539e-01 -4.72716205e-02
3.59507203e-01 -5.10743678e-01 2.10503951e-01 3.32812637e-01
-2.20820874e-01 -4.14374083e-01 2.86167711e-01 1.48282334e-01
1.13280571e+00 8.76517594e-02 5.95158696e-01 -5.70740819e-01
3.91539514e-01 3.40060204e-01 8.72033477e-01 7.72896409e-01
-3.08013678e-01 3.55799228e-01 5.01354277e-01 -3.48996878e-01
-1.77427733e+00 -9.77330625e-01 -6.42609358e-01 6.80951357e-01
6.36675060e-02 -5.20665824e-01 -8.55354548e-01 -1.42801210e-01
2.65984267e-01 9.01185095e-01 -2.98422396e-01 -2.58998293e-02
-5.64614475e-01 -1.12650108e+00 1.06922615e+00 1.94260746e-01
-2.52532929e-01 -5.67277074e-01 -6.10060871e-01 7.30405450e-01
-4.57196355e-01 -5.84367692e-01 -4.78468418e-01 6.71538115e-01
-1.04343581e+00 -1.25276589e+00 -6.38251126e-01 -6.26228273e-01
4.48457450e-01 4.65024650e-01 8.87811542e-01 3.69143546e-01
-6.48168802e-01 -1.66109100e-01 -3.78869116e-01 9.28697139e-02
-7.12903678e-01 -1.69584557e-01 3.28315496e-01 -7.59913564e-01
7.75185764e-01 -1.03877234e+00 -6.34077489e-01 5.76937914e-01
-8.58776689e-01 -2.43400440e-01 2.76047379e-01 1.07163119e+00
7.41198599e-01 -3.17850441e-01 4.49369967e-01 -7.07875788e-01
4.79961842e-01 -4.90837187e-01 -7.58958280e-01 4.20160025e-01
-6.12984419e-01 1.64788157e-01 7.38388538e-01 -3.60275358e-01
-4.48755592e-01 -3.26793827e-02 -5.82823038e-01 1.06505170e-01
-8.24529454e-02 3.70864213e-01 4.09439690e-02 1.13624623e-02
1.00366318e+00 6.08747482e-01 2.55111426e-01 -5.65164506e-01
2.95685798e-01 8.78638148e-01 3.71655494e-01 -8.05433154e-01
5.07763147e-01 3.00012738e-01 1.99035645e-01 -1.35408759e+00
6.95044547e-02 -6.38501287e-01 -8.10378313e-01 3.02486122e-01
5.54817438e-01 -4.84930038e-01 -9.91575897e-01 4.21691209e-01
-1.12979734e+00 -1.65980026e-01 2.51738757e-01 5.06140351e-01
-1.06577086e+00 1.21589613e+00 -5.89236856e-01 -8.84659827e-01
-4.25760821e-02 -1.16872561e+00 8.46706808e-01 -2.00016230e-01
-5.44622719e-01 -6.23171806e-01 5.51944554e-01 9.95631516e-02
-7.33539313e-02 2.43645906e-01 1.41851509e+00 -5.82897246e-01
-2.81895190e-01 7.36539960e-02 1.28814960e-02 1.56565178e-02
3.65965813e-01 7.06314504e-01 -4.52444792e-01 -3.44944835e-01
-1.63307056e-01 -1.71010509e-01 5.94774187e-01 5.02804160e-01
9.29759860e-01 -1.44473389e-01 -5.19212484e-01 6.45473659e-01
1.56425941e+00 6.42170668e-01 6.10722542e-01 2.54139274e-01
2.43227839e-01 8.00476074e-01 8.18907797e-01 4.81404275e-01
1.18426204e-01 6.15349472e-01 -2.94591542e-02 6.27275229e-01
5.62997878e-01 -3.02280903e-01 4.26665545e-02 9.52715635e-01
2.06862077e-01 -3.15334767e-01 -1.17420435e+00 6.26872361e-01
-1.50021231e+00 -1.20533073e+00 -4.60219234e-01 2.44302225e+00
1.17725563e+00 -2.94680983e-01 5.56699932e-01 1.34569407e-01
8.66059899e-01 -5.02435490e-02 -9.60023463e-01 -6.58642411e-01
-1.83664814e-01 1.98321059e-01 8.27101648e-01 7.58643448e-01
-4.37270939e-01 9.06141102e-01 8.34425545e+00 9.38475072e-01
-7.85653472e-01 -5.25565386e-01 4.02800381e-01 2.96623170e-01
-6.89896882e-01 4.79028225e-01 -7.82609880e-01 5.75588226e-01
1.22367215e+00 -4.52533245e-01 3.27631712e-01 8.46887469e-01
3.13254446e-01 -3.64500880e-01 -1.14692223e+00 1.00653493e+00
-5.30313909e-01 -1.37806761e+00 -1.03501678e-01 3.08560252e-01
6.80630878e-02 2.05077790e-02 -3.71746123e-01 -5.76679826e-01
8.69857311e-01 -9.91017938e-01 2.98545714e-02 -5.63567728e-02
8.71076226e-01 -1.00432146e+00 2.27262393e-01 6.12398505e-01
-1.21761262e+00 5.54538727e-01 -1.00009847e+00 -3.92410718e-03
4.58632320e-01 5.67415237e-01 -1.25629783e+00 3.50147337e-01
4.42035049e-01 4.00518447e-01 -1.09837309e-01 9.97978270e-01
1.77143291e-01 7.45093524e-01 -7.05478430e-01 -2.58537084e-01
1.46622121e-01 -8.47152829e-01 4.31600958e-01 1.32139957e+00
5.65079570e-01 2.91710347e-01 -2.57196367e-01 6.39169276e-01
5.65556943e-01 2.11189002e-01 -7.64294744e-01 -5.15297353e-01
9.10560191e-01 7.18233943e-01 -9.16498303e-01 -3.44664901e-01
-1.21397123e-01 1.15358579e+00 1.97327569e-01 6.85230047e-02
-6.58530951e-01 -7.68787146e-01 1.40289271e+00 -1.88078284e-01
4.23722327e-01 -5.59669793e-01 -9.38831568e-02 -9.43377078e-01
-2.70591706e-01 -1.10366452e+00 2.70244211e-01 -6.82949066e-01
-1.34167349e+00 2.03773722e-01 3.58213447e-02 -8.21511567e-01
-7.98592508e-01 -6.35764599e-01 -1.52280971e-01 1.10394144e+00
-8.18125486e-01 -4.14744496e-01 3.52728009e-01 -3.10937539e-02
3.55397969e-01 3.37831318e-01 1.02110076e+00 -1.95659474e-01
-3.36534977e-01 6.80017889e-01 8.83198082e-01 -3.34778577e-01
5.57886183e-01 -1.24285412e+00 9.55200195e-01 6.65308535e-01
4.73638102e-02 1.26367915e+00 1.21993005e+00 -8.77052188e-01
-1.42906713e+00 -5.44007480e-01 7.75485992e-01 -1.83543518e-01
7.36758709e-01 -2.14969903e-01 -1.24216270e+00 7.30146468e-01
-1.66953281e-01 -6.79213822e-01 1.33315313e+00 1.25471586e-02
-7.70510375e-01 3.44882250e-01 -1.32104635e+00 6.81408167e-01
1.30444324e+00 -5.56529701e-01 -3.79428983e-01 2.45011717e-01
5.37038743e-01 3.11087221e-01 -1.16717017e+00 1.87240139e-01
1.04089773e+00 -9.99367356e-01 1.12873662e+00 -8.29552948e-01
1.66853473e-01 -6.07152760e-01 -4.99603808e-01 -1.12616146e+00
-4.81212318e-01 -6.30833447e-01 7.79435575e-01 7.35258341e-01
1.97816059e-01 -5.94973564e-01 7.80141890e-01 2.04663593e-02
1.37581512e-01 -2.08517224e-01 -1.17019951e+00 -1.01316929e+00
5.13813913e-01 -9.75918770e-02 3.67382377e-01 1.04582286e+00
2.49218300e-01 2.25967869e-01 -4.98388648e-01 -1.38396785e-01
9.04991210e-01 -8.80221725e-02 9.96458232e-01 -1.05974305e+00
-6.22425795e-01 -3.65053475e-01 -6.04453564e-01 -1.25739992e+00
1.27193881e-02 -7.12021232e-01 2.51761466e-01 -8.92038405e-01
3.14043760e-01 -5.76430380e-01 2.58868545e-01 4.07133251e-03
-4.22507316e-01 -8.92820954e-02 -1.29238546e-01 1.53665379e-01
1.90224111e-01 1.69335350e-01 6.42129481e-01 3.37361395e-01
2.76940793e-01 -4.44608033e-01 -5.08921027e-01 7.97580838e-01
7.06857145e-01 -6.83144212e-01 -2.80504614e-01 -4.52782549e-02
1.94303498e-01 3.61133993e-01 -2.48407602e-01 -2.28337735e-01
-2.38193423e-01 -7.53494561e-01 2.23475154e-02 -9.53398943e-01
3.38413566e-01 -6.32365286e-01 8.65304172e-01 8.64549637e-01
-2.05399796e-01 5.15970588e-01 1.32316630e-02 5.41858554e-01
1.55948222e-01 -6.64351702e-01 1.11278653e+00 -2.99837679e-01
-3.87858510e-01 -8.41063075e-03 -9.09409225e-01 -2.77584642e-01
9.37958658e-01 -6.58286631e-01 -2.77149439e-01 -1.70635447e-01
-3.26787442e-01 -6.73281327e-02 1.02784050e+00 -1.95657700e-01
4.16147828e-01 -8.31752360e-01 -7.92822719e-01 -1.67494625e-01
2.62172341e-01 -4.48797405e-01 8.26962739e-02 4.96009499e-01
-1.31031728e+00 1.81848884e-01 -3.28750730e-01 -8.81838500e-01
-1.73828566e+00 7.19191253e-01 2.13508338e-01 2.61266410e-01
-3.02968502e-01 9.78795409e-01 8.20040926e-02 -5.77858627e-01
-1.95033416e-01 8.88871998e-02 5.41060746e-01 -4.86004725e-02
5.21592319e-01 3.14049155e-01 -3.72267902e-01 -7.83464611e-01
-6.98162079e-01 8.34414184e-01 -2.60302663e-01 -2.45573670e-01
1.19507515e+00 -1.99640587e-01 -4.08183515e-01 6.08195484e-01
1.27118170e+00 -8.66857320e-02 -7.42921174e-01 4.07394469e-02
3.96529317e-01 -9.26976800e-01 -9.82894421e-01 -1.81830317e-01
-1.34409145e-01 8.65898788e-01 3.28284800e-01 -9.01648104e-02
6.19750142e-01 -1.79723591e-01 5.22396147e-01 6.48315430e-01
5.22501349e-01 -7.64524221e-01 -3.86381596e-01 2.63909817e-01
4.45127845e-01 -8.43469560e-01 -9.45880115e-02 -5.49212217e-01
-6.50659800e-02 1.06983101e+00 1.47854939e-01 -4.62898500e-02
1.68651581e-01 1.93397447e-01 1.18856199e-01 2.84421027e-01
-8.11703920e-01 -1.86776798e-02 -5.51708996e-01 7.10548997e-01
8.28324735e-01 8.24868008e-02 -1.33541656e+00 -2.14803621e-01
-5.33775985e-01 -3.00787330e-01 7.64277160e-01 9.54163134e-01
-8.99092257e-01 -1.76501644e+00 -3.80998492e-01 3.79742801e-01
-4.90568459e-01 -2.21315071e-01 -7.00724185e-01 5.87735415e-01
-3.71074259e-01 9.45679605e-01 8.88426900e-02 -5.04636288e-01
-2.23612934e-01 4.99146283e-01 6.02501273e-01 -3.00259382e-01
-3.10362220e-01 3.67294461e-01 2.17920244e-01 -2.12365821e-01
-1.33760452e-01 -8.58521998e-01 -1.03722727e+00 -1.01322281e+00
-6.06442273e-01 7.69044995e-01 7.44000375e-01 4.97384071e-01
2.74877459e-01 -2.05713883e-01 6.87109470e-01 -2.68371075e-01
-9.99248087e-01 -8.93275440e-01 -7.48180747e-01 2.41662815e-01
-4.72747795e-02 -4.15194899e-01 -5.65465629e-01 1.27009958e-01] | [4.820062160491943, 5.137710094451904] |
2111d4c4-be48-4310-a3bd-7737a9034c65 | phrase-detectives-corpus-10-crowdsourced | null | null | https://aclanthology.org/L16-1323 | https://aclanthology.org/L16-1323.pdf | Phrase Detectives Corpus 1.0 Crowdsourced Anaphoric Coreference. | Natural Language Engineering tasks require large and complex annotated datasets to build more advanced models of language. Corpora are typically annotated by several experts to create a gold standard; however, there are now compelling reasons to use a non-expert crowd to annotate text, driven by cost, speed and scalability. Phrase Detectives Corpus 1.0 is an anaphorically-annotated corpus of encyclopedic and narrative text that contains a gold standard created by multiple experts, as well as a set of annotations created by a large non-expert crowd. Analysis shows very good inter-expert agreement (kappa=.88-.93) but a more variable baseline crowd agreement (kappa=.52-.96). Encyclopedic texts show less agreement (and by implication are harder to annotate) than narrative texts. The release of this corpus is intended to encourage research into the use of crowds for text annotation and the development of more advanced, probabilistic language models, in particular for anaphoric coreference. | ['Udo Kruschwitz', 'Jon Chamberlain', 'Massimo Poesio'] | 2016-05-01 | phrase-detectives-corpus-10-crowdsourced-1 | https://aclanthology.org/L16-1323 | https://aclanthology.org/L16-1323.pdf | lrec-2016-5 | ['text-annotation'] | ['natural-language-processing'] | [-3.03886712e-01 7.13879049e-01 -1.85680583e-01 -5.41826129e-01
-1.14068151e+00 -1.00374413e+00 8.62995028e-01 6.67318523e-01
-9.74332750e-01 9.34080064e-01 1.06337333e+00 1.04493238e-01
1.67113051e-01 -3.23169023e-01 -2.12487072e-01 -2.21686974e-01
7.40978837e-01 1.44909573e+00 7.93198228e-01 -5.06147087e-01
3.73458773e-01 1.26159087e-01 -1.19228363e+00 6.05703592e-01
4.48717177e-01 8.20659250e-02 2.76847512e-01 4.54072356e-01
-5.52691936e-01 9.65133667e-01 -8.04968417e-01 -9.58018005e-01
-1.32734090e-01 -1.58928819e-02 -1.43197596e+00 -2.68579274e-01
2.17738464e-01 1.65818945e-01 1.19244337e-01 9.92540538e-01
7.85352170e-01 1.88572556e-01 4.02484179e-01 -1.10638893e+00
-4.84477907e-01 1.06958735e+00 -4.53719705e-01 3.44861060e-01
8.88303101e-01 -6.07427545e-02 1.35211217e+00 -8.08763683e-01
1.28783453e+00 1.54997432e+00 7.81298041e-01 5.07098258e-01
-1.06903076e+00 -5.22488236e-01 -3.16384256e-01 -2.73329094e-02
-1.20567703e+00 -5.11462569e-01 3.63993049e-01 -7.98459709e-01
1.24639022e+00 1.51538715e-01 1.96655750e-01 1.15632594e+00
-9.54644457e-02 5.34125388e-01 1.00022876e+00 -7.26783216e-01
2.14004949e-01 4.87215787e-01 1.02405772e-01 1.29761606e-01
3.72699827e-01 -5.26955605e-01 -8.75340223e-01 -8.27081442e-01
3.03961098e-01 -8.19799304e-01 -2.64852792e-01 -1.54343307e-01
-9.91501391e-01 9.63843346e-01 -1.89791620e-01 8.92718017e-01
-3.95985365e-01 -2.96448171e-01 8.51364195e-01 1.42450377e-01
4.71938699e-01 9.92018878e-01 -2.93742299e-01 -5.52523673e-01
-7.07844496e-01 8.10082912e-01 1.44025826e+00 9.26259935e-01
3.92227173e-01 -7.56204128e-01 -1.43416256e-01 1.16752136e+00
5.04172683e-01 2.17521966e-01 4.64784533e-01 -1.40821648e+00
8.13303888e-01 7.00728297e-01 8.23879898e-01 -1.20755732e+00
-4.96471375e-01 3.74421299e-01 1.33668989e-01 -3.98489274e-02
8.37137341e-01 -8.81816968e-02 1.63975373e-01 1.45369136e+00
2.71921545e-01 -5.78763425e-01 3.79943579e-01 1.11924231e+00
8.69248927e-01 4.24345911e-01 5.63164711e-01 -3.27216148e-01
1.74422586e+00 -6.93042815e-01 -1.02668226e+00 -5.56986511e-01
9.74043071e-01 -1.22397029e+00 1.08625102e+00 3.15969363e-02
-1.41236103e+00 3.14175710e-02 -6.22637749e-01 -3.74969333e-01
-4.91851158e-02 -2.92191416e-01 3.41082811e-01 4.90837783e-01
-7.47158110e-01 2.60139793e-01 -5.86159706e-01 -9.01917398e-01
1.94164261e-01 2.47440755e-01 -6.63917005e-01 -9.00033191e-02
-1.55633080e+00 1.40381551e+00 5.86118996e-01 -4.67623860e-01
-8.43539685e-02 -2.85985202e-01 -7.90003717e-01 -2.48881191e-01
4.19879109e-01 -1.17949285e-01 1.60718191e+00 -7.95707047e-01
-1.03792489e+00 1.71157598e+00 -1.68305740e-01 -3.41564268e-01
5.74006498e-01 -4.12666231e-01 -2.39519283e-01 2.75916725e-01
8.99235487e-01 5.30618668e-01 3.26642841e-02 -1.26203287e+00
-9.34714615e-01 -3.93627346e-01 1.21371157e-01 4.89035010e-01
-2.49806136e-01 1.11858487e+00 -2.25918427e-01 -4.09055591e-01
-2.04018071e-01 -9.24772978e-01 4.25905958e-02 -4.03563589e-01
-1.22721188e-01 -1.02931452e+00 4.63715017e-01 -7.44384706e-01
1.07111967e+00 -1.72270441e+00 -2.58673906e-01 -2.77052164e-01
4.12362128e-01 1.95745319e-01 1.39203057e-01 6.73403978e-01
1.87231973e-01 1.94614634e-01 4.02316358e-03 -1.50498971e-01
2.08231971e-01 3.48240882e-01 -1.69008881e-01 4.08667147e-01
1.72539175e-01 3.32834601e-01 -1.46537292e+00 -1.01046133e+00
-3.07097703e-01 4.57586423e-02 -2.75500447e-01 2.12238207e-02
-2.15626419e-01 2.12601751e-01 -6.39262140e-01 3.59595835e-01
7.37857744e-02 -9.68334824e-02 3.76894295e-01 2.10509747e-02
-3.99747789e-01 3.98849726e-01 -1.14126003e+00 1.40855253e+00
-1.53831556e-01 1.00670862e+00 4.27635103e-01 -4.99434590e-01
1.07724464e+00 8.69235575e-01 2.86999136e-01 -3.65082115e-01
1.98434338e-01 5.46217740e-01 1.15206735e-02 -8.94600749e-01
9.00958955e-01 -3.06914061e-01 -5.28527677e-01 1.01666284e+00
-7.51430988e-02 -5.25620639e-01 3.58337492e-01 4.25734639e-01
1.24736488e+00 -2.01850552e-02 4.60596114e-01 -4.93577033e-01
4.42692280e-01 9.71244752e-01 7.41054773e-01 7.16996729e-01
-4.98080105e-01 2.20298678e-01 5.10239303e-01 -6.10610962e-01
-1.39569652e+00 -4.05502409e-01 -2.13303253e-01 1.54386044e+00
8.71442482e-02 -4.56628710e-01 -8.60371351e-01 -5.52527308e-01
-3.65066975e-01 9.07087326e-01 9.32355318e-03 3.49646300e-01
-7.39163399e-01 -5.75527251e-01 8.91775012e-01 6.09357297e-01
1.56712607e-01 -1.25511098e+00 -6.50109172e-01 5.94367266e-01
-8.81173849e-01 -1.59989488e+00 -2.30652690e-01 -1.63253561e-01
9.29002929e-03 -1.02370226e+00 -4.11359042e-01 -8.84543061e-01
1.65559471e-01 -1.37590066e-01 1.42901158e+00 1.43357664e-02
4.10774313e-02 5.62120020e-01 -4.44160700e-01 -8.22034299e-01
-8.46755683e-01 5.06656766e-02 2.95454800e-01 -6.87788725e-01
1.18904400e+00 -1.35935143e-01 1.50726279e-02 5.46774983e-01
-6.35783553e-01 -4.17167276e-01 7.74165392e-02 7.70354688e-01
8.54331851e-02 -2.47445449e-01 4.32762384e-01 -8.03064466e-01
1.11230230e+00 -4.51549679e-01 -3.21152717e-01 1.39771372e-01
-1.93520576e-01 -2.21346036e-01 1.86586589e-01 -4.91538584e-01
-1.57530618e+00 -7.55581930e-02 2.06811920e-01 3.32050890e-01
-3.08872938e-01 3.79139930e-01 5.75793199e-02 2.59586394e-01
1.41697133e+00 -8.66107583e-01 -2.35803276e-01 -2.66348124e-01
1.39968634e-01 1.12046266e+00 7.70210922e-01 -1.02388287e+00
5.52824676e-01 3.60959530e-01 -6.68787003e-01 -7.37740755e-01
-1.08375013e+00 -9.79023218e-01 -7.07464397e-01 -1.54843733e-01
9.95165467e-01 -1.16651797e+00 -3.76246661e-01 -8.41194019e-02
-1.63205862e+00 -3.64339173e-01 -1.60670117e-01 5.34671545e-01
-4.80376542e-01 3.35753083e-01 -7.86469936e-01 -7.01224208e-01
-2.94869602e-01 -7.82514393e-01 1.02652597e+00 1.50289297e-01
-1.40274823e+00 -1.01460326e+00 5.42298734e-01 9.65576410e-01
-2.91409399e-02 1.71792343e-01 5.01542211e-01 -1.50929451e+00
3.07959378e-01 -6.06734455e-01 -2.90620834e-01 -2.08807290e-01
-2.81807244e-01 -9.16547403e-02 -8.59734237e-01 1.56914651e-01
-3.33378725e-02 -8.64204466e-01 -1.44763477e-02 -2.38485351e-01
-2.75735855e-02 -3.23151439e-01 -6.65419340e-01 -8.32330227e-01
1.05908012e+00 -1.81254391e-02 3.80464375e-01 9.53420877e-01
4.04083997e-01 1.09469032e+00 8.59903514e-01 3.79027069e-01
6.56351388e-01 7.19170451e-01 -2.51567125e-01 4.82085437e-01
2.76629984e-01 1.34368092e-01 6.66775256e-02 6.85096085e-01
-2.22080573e-01 -3.71800423e-01 -1.63305902e+00 1.00123632e+00
-2.05200315e+00 -1.20393109e+00 -5.88943779e-01 1.66218126e+00
1.31470025e+00 2.56895036e-01 5.79389669e-02 1.16382711e-01
1.15929627e+00 2.48893420e-03 2.33335152e-01 -5.39456189e-01
-3.33495349e-01 6.47792369e-02 2.40286082e-01 6.38016880e-01
-9.32218850e-01 1.20431662e+00 5.90298653e+00 5.24197638e-01
-4.07059699e-01 4.43328261e-01 9.22967643e-02 6.37881383e-02
-1.34898961e-01 2.66445756e-01 -1.00498128e+00 3.45921010e-01
9.09098268e-01 -4.18639243e-01 4.25534844e-02 7.52839327e-01
1.75926253e-01 -2.87858218e-01 -9.64389443e-01 5.83475947e-01
1.81644082e-01 -1.24852526e+00 -5.50217509e-01 -2.31445104e-01
7.73291588e-01 2.37202451e-01 -1.04697895e+00 3.49841595e-01
8.22350502e-01 -6.87020183e-01 7.20242560e-01 2.06823289e-01
3.01880479e-01 -3.47554505e-01 1.41375518e+00 5.36401331e-01
-7.70440876e-01 1.14279008e-02 -3.83168519e-01 -1.84002072e-01
6.27170742e-01 3.24216872e-01 -9.26259398e-01 -1.30465701e-01
8.90412748e-01 -5.76313995e-02 -1.98127255e-01 7.45187759e-01
-3.74982446e-01 2.27041334e-01 -4.34636474e-01 -3.94055724e-01
3.38484883e-01 4.10144925e-01 8.87571931e-01 1.49063206e+00
-1.70366704e-01 6.32062912e-01 4.54843491e-01 5.86159527e-01
-8.43069628e-02 4.20853227e-01 -4.27258253e-01 -1.89037517e-01
1.10235810e+00 1.31687069e+00 -7.32483149e-01 -2.85870701e-01
-5.95083296e-01 5.84089458e-01 5.00262976e-01 -6.45849332e-02
-3.35288554e-01 -2.71308184e-01 -1.17633436e-02 4.36265439e-01
-2.38119006e-01 4.59119156e-02 -2.46280551e-01 -5.75737774e-01
-5.72679825e-02 -1.01468027e+00 5.31938910e-01 -9.95815992e-01
-1.46222651e+00 7.18062997e-01 2.45876580e-01 -5.95958531e-01
-4.88091916e-01 -5.09437799e-01 -5.27932107e-01 8.12637627e-01
-7.15023816e-01 -8.25333416e-01 -3.42017949e-01 2.72756606e-01
3.86361539e-01 -3.11546355e-01 9.40919936e-01 -4.19840496e-03
-5.59286252e-02 9.32243243e-02 -3.46685737e-01 4.77854669e-01
1.30150425e+00 -1.21260393e+00 7.47126788e-02 1.86705559e-01
-1.18005581e-01 5.10427475e-01 1.27711952e+00 -8.64918172e-01
-5.46265483e-01 -4.74367946e-01 1.62904251e+00 -1.06605446e+00
1.25301242e+00 1.18823864e-01 -1.07684386e+00 6.46913886e-01
4.23875272e-01 -2.78399199e-01 8.81618917e-01 1.62664622e-01
-2.99695522e-01 5.73183835e-01 -1.36666989e+00 4.89226907e-01
7.88324356e-01 -7.78738797e-01 -1.55589902e+00 7.08208144e-01
5.22810519e-01 -4.96687949e-01 -1.15237510e+00 1.20439135e-01
3.42415005e-01 -5.32941759e-01 4.63746339e-01 -5.15701652e-01
3.86348426e-01 -1.93196669e-01 -2.32210025e-01 -8.88408422e-01
-7.87312537e-02 -7.70352185e-01 4.80096787e-01 1.60237241e+00
5.69061220e-01 -3.55388224e-01 5.62435329e-01 1.42373097e+00
-4.28886451e-02 -1.05564795e-01 -8.79840493e-01 -4.83775288e-01
1.91213384e-01 -4.67502028e-01 1.87642485e-01 1.51371837e+00
8.70378375e-01 8.30281675e-01 3.18689018e-01 7.35832751e-02
3.13863248e-01 -4.14411336e-01 7.96379149e-01 -1.81290567e+00
1.62738428e-01 -3.14512491e-01 -3.10388654e-01 -2.66891927e-01
5.10633349e-01 -5.72834074e-01 3.77008379e-01 -1.60000646e+00
3.73303771e-01 -4.40871626e-01 5.96222878e-01 5.32173395e-01
-1.77384183e-01 2.45083347e-01 -1.10396430e-01 7.24721253e-01
-7.92871296e-01 -5.28249368e-02 8.60239148e-01 2.27993459e-01
-4.41529453e-01 -4.54401910e-01 -7.69761503e-01 1.15885830e+00
7.23746657e-01 -9.17683244e-01 1.15394250e-01 -3.63960832e-01
7.81875551e-01 -1.98512841e-02 1.41791195e-01 -6.28402889e-01
6.28901601e-01 -1.87050804e-01 -1.15009053e-02 -1.26570240e-01
2.26497322e-01 -4.04062778e-01 -1.31409794e-01 1.97345130e-02
-5.13763487e-01 1.98622197e-01 2.20701486e-01 3.50169420e-01
-2.30402187e-01 -1.02769804e+00 4.96142954e-01 -6.17329597e-01
-5.59757113e-01 -5.39017797e-01 -7.88477242e-01 7.02751219e-01
9.24613118e-01 -2.22617313e-01 -5.81576824e-01 -4.83888298e-01
-8.13314617e-01 2.48674393e-01 5.79778492e-01 3.51591587e-01
-1.44005239e-01 -1.09289587e+00 -9.67199206e-01 -7.57430792e-01
3.98478776e-01 1.79216549e-01 -3.14040661e-01 6.01842880e-01
-6.27739906e-01 3.71927857e-01 -3.22866105e-02 -3.16526055e-01
-1.30324459e+00 4.99664210e-02 -3.07330745e-03 -3.99827272e-01
-4.49295431e-01 8.14939916e-01 -4.10925031e-01 -4.97015119e-01
1.46721825e-01 4.65713918e-01 -6.19170427e-01 4.29011732e-01
7.83736706e-01 3.46514046e-01 -3.48456889e-01 -1.05984306e+00
-3.62944335e-01 1.56779036e-01 -2.93548740e-02 -6.64875448e-01
1.32312346e+00 -2.16723487e-01 -4.60713893e-01 5.56002378e-01
4.85493511e-01 4.61219311e-01 -5.79563260e-01 -4.38210487e-01
8.28623533e-01 -1.93506226e-01 -1.38279274e-01 -5.94035387e-01
-4.99586351e-02 2.49732330e-01 1.81912556e-02 4.52068806e-01
2.00808212e-01 5.38075387e-01 5.27081847e-01 7.12112427e-01
4.37015533e-01 -1.65845239e+00 1.35058269e-01 7.62468994e-01
1.07754290e+00 -1.28277946e+00 1.42860413e-01 -4.53091413e-01
-1.10423064e+00 1.04037738e+00 8.61277819e-01 8.65387470e-02
1.93042710e-01 5.44018209e-01 3.01811367e-01 -6.65316582e-01
-6.02275312e-01 8.23772997e-02 -1.22842498e-01 6.30937159e-01
8.48163843e-01 -1.68437019e-01 -8.09242845e-01 9.63873088e-01
-3.26375067e-01 -1.01466320e-01 6.06512904e-01 9.47405398e-01
-5.94144762e-01 -1.13285530e+00 -8.68143439e-01 8.39500204e-02
-8.06369126e-01 3.42952572e-02 -8.92360866e-01 9.65225995e-01
3.03787850e-02 1.29319263e+00 3.48926157e-01 2.78688878e-01
2.70508200e-01 1.89559519e-01 1.45062998e-01 -9.42700386e-01
-8.26822281e-01 -1.29213497e-01 1.12334573e+00 -2.18961731e-01
-1.02261186e+00 -8.69421005e-01 -1.48623991e+00 -2.14834452e-01
-5.98209023e-01 6.17475271e-01 4.83878851e-01 1.41826701e+00
-1.20679930e-01 -2.63775975e-01 -1.83703542e-01 -7.01626062e-01
-3.12744558e-01 -1.20251191e+00 -1.45949483e-01 6.25351489e-01
-3.79456609e-01 -5.95297396e-01 -3.61837029e-01 1.48860008e-01] | [9.515403747558594, 9.233325004577637] |
fd0e4013-bd82-4ed7-a9a4-06f78fc39417 | face-recognition-using-multi-modal-low-rank | 1703.04853 | null | http://arxiv.org/abs/1703.04853v1 | http://arxiv.org/pdf/1703.04853v1.pdf | Face Recognition using Multi-Modal Low-Rank Dictionary Learning | Face recognition has been widely studied due to its importance in different
applications; however, most of the proposed methods fail when face images are
occluded or captured under illumination and pose variations. Recently several
low-rank dictionary learning methods have been proposed and achieved promising
results for noisy observations. While these methods are mostly developed for
single-modality scenarios, recent studies demonstrated the advantages of
feature fusion from multiple inputs. We propose a multi-modal structured
low-rank dictionary learning method for robust face recognition, using raw
pixels of face images and their illumination invariant representation. The
proposed method learns robust and discriminative representations from
contaminated face images, even if there are few training samples with large
intra-class variations. Extensive experiments on different datasets validate
the superior performance and robustness of our method to severe illumination
variations and occlusion. | ['Moein Shakeri', 'Nilanjan Ray', 'Homa Foroughi', 'Hong Zhang'] | 2017-03-15 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 3.87659341e-01 -7.12357700e-01 -2.40555286e-01 -5.79400003e-01
-9.86148119e-01 -3.94087791e-01 6.52437866e-01 -4.78005767e-01
-1.47397518e-01 7.27429450e-01 2.57289886e-01 5.80009341e-01
-4.47686017e-01 -3.08810353e-01 -4.52195466e-01 -1.22698903e+00
1.07659474e-01 2.23773792e-01 -3.62317771e-01 2.22970739e-01
7.50204623e-02 7.02428639e-01 -1.98803520e+00 1.56252697e-01
4.74453062e-01 1.08420098e+00 -6.08577654e-02 1.89971030e-01
2.47141346e-01 4.00176823e-01 -5.08725107e-01 -1.35530531e-01
4.97325450e-01 -2.87426323e-01 -2.28048548e-01 6.07361197e-01
7.95797586e-01 -2.49221265e-01 -5.43576598e-01 1.26331103e+00
8.37617338e-01 8.27029496e-02 5.64263523e-01 -1.16466546e+00
-7.01654911e-01 -3.37642431e-02 -7.98811793e-01 2.62309581e-01
5.69860160e-01 -1.73767015e-01 6.27435386e-01 -1.52118790e+00
4.85013932e-01 1.49793398e+00 4.81317639e-01 4.20135587e-01
-1.43735838e+00 -7.65358925e-01 -1.77963357e-02 5.12855589e-01
-1.62787521e+00 -9.22954142e-01 1.06942427e+00 -2.79298127e-01
4.93792474e-01 1.84335753e-01 4.75262068e-02 1.30389428e+00
1.87622271e-02 6.89450443e-01 1.71735811e+00 -2.62844265e-01
-5.20898476e-02 3.98346856e-02 -6.59428835e-02 6.38476610e-01
5.39848983e-01 3.07155043e-01 -7.39126980e-01 -4.28694725e-01
6.46418989e-01 4.13996249e-01 -4.65793282e-01 -4.60385829e-01
-1.05410695e+00 7.50141263e-01 -4.33032401e-02 3.13985705e-01
-5.25921762e-01 -3.03085923e-01 2.22810194e-01 2.95095652e-01
3.14635366e-01 -3.75776827e-01 -1.46882430e-01 2.55567968e-01
-9.18424189e-01 -1.29756883e-01 5.21178305e-01 7.59127975e-01
6.15562439e-01 3.84241372e-01 -1.73494399e-01 1.23312402e+00
6.65458798e-01 1.08112800e+00 3.67293745e-01 -6.30473673e-01
1.41189039e-01 1.96916997e-01 -6.32830486e-02 -1.64338553e+00
-3.47220868e-01 -4.24073815e-01 -1.17930663e+00 -3.59777510e-02
2.78720349e-01 2.24978104e-01 -7.79619634e-01 1.51017165e+00
3.88645560e-01 5.00870645e-01 9.22685340e-02 9.82127488e-01
9.76820350e-01 4.66434032e-01 -1.08571813e-01 -7.82568216e-01
1.19372582e+00 -4.73095000e-01 -1.05090368e+00 -1.39890358e-01
-1.75221696e-01 -1.11942875e+00 6.08517230e-01 8.04787278e-01
-6.82200968e-01 -7.56457865e-01 -8.31609011e-01 2.61571079e-01
1.21492207e-01 2.59697795e-01 4.31801468e-01 8.60115528e-01
-8.06335568e-01 6.96095973e-02 -5.17242610e-01 -2.66177416e-01
6.23209953e-01 4.00810421e-01 -7.75798500e-01 -7.85546005e-01
-7.41190732e-01 5.96264958e-01 -1.03039585e-01 4.92662072e-01
-1.33121002e+00 -1.78096101e-01 -8.27774405e-01 -3.30761224e-01
3.89132679e-01 -1.85888171e-01 4.80482280e-01 -6.87085569e-01
-1.42106676e+00 6.99135780e-01 -4.62381691e-01 1.61966473e-01
1.82949156e-01 -3.38220268e-01 -7.83494294e-01 4.13134277e-01
-8.45867246e-02 4.34984788e-02 1.54887915e+00 -1.51838100e+00
-1.38053566e-01 -9.42213535e-01 -4.09050137e-01 -9.04517844e-02
-4.87113774e-01 1.13262512e-01 -4.23604280e-01 -7.96624005e-01
5.85768878e-01 -8.36661160e-01 5.18324152e-02 -3.99697050e-02
-5.97765855e-02 -2.54728366e-03 1.21124887e+00 -5.63501358e-01
7.37829328e-01 -2.14050961e+00 1.49075687e-01 2.03833744e-01
-3.74186486e-01 2.67259628e-01 -4.24913347e-01 2.72826821e-01
-7.86042288e-02 -3.97518218e-01 -1.50422841e-01 -3.78052026e-01
-5.92352487e-02 4.62229967e-01 -4.00038630e-01 1.10600877e+00
1.13498434e-01 4.26915050e-01 -6.64535403e-01 -4.85780984e-01
3.44264507e-01 1.16416216e+00 -3.16320062e-01 2.48751506e-01
2.60942608e-01 8.31092179e-01 -3.57090473e-01 1.18302882e+00
9.69193280e-01 1.19045466e-01 1.16257273e-01 -6.36613667e-01
3.87057453e-01 -4.68615562e-01 -1.73548198e+00 1.58580673e+00
-3.99795681e-01 4.04603124e-01 5.18082738e-01 -1.22945786e+00
9.65392172e-01 6.53319716e-01 6.08365476e-01 -7.44717062e-01
1.92334667e-01 2.01709896e-01 -3.75267237e-01 -7.13853359e-01
8.01308155e-02 -3.88309419e-01 4.18533891e-01 1.24442138e-01
1.60645485e-01 1.11863114e-01 4.80931066e-02 -9.50533897e-02
7.37333834e-01 1.13277555e-01 1.62263021e-01 -5.68609126e-02
8.89961600e-01 -8.03031027e-01 8.20699394e-01 5.32304645e-01
-2.14976251e-01 7.83704758e-01 -1.43341631e-01 -3.05114836e-01
-4.11980897e-01 -8.71457040e-01 -7.84634054e-01 9.64873016e-01
8.89523327e-02 -1.14568532e-01 -2.05807790e-01 -5.29585660e-01
-1.31564802e-02 2.53763311e-02 -3.89796644e-01 5.90239465e-02
-3.70253384e-01 -1.12079954e+00 4.08397526e-01 2.04400510e-01
4.65357810e-01 -6.97921395e-01 -1.26535920e-02 1.36928320e-01
-1.69305831e-01 -1.37951529e+00 -1.45253599e-01 -2.51601875e-01
-8.51354480e-01 -1.20930624e+00 -6.52721167e-01 -8.54611754e-01
9.94617522e-01 6.60632014e-01 7.34262288e-01 -3.39903310e-02
-6.74010575e-01 8.69284034e-01 -3.40932161e-01 -9.83295613e-04
7.11812675e-02 -8.65754545e-01 6.27664804e-01 7.76716113e-01
2.17115358e-01 -4.89070594e-01 -4.71196622e-01 6.91078484e-01
-1.10284710e+00 -7.77950644e-01 6.42010450e-01 1.30643666e+00
8.60039830e-01 2.90921122e-01 7.00419545e-01 -7.91962802e-01
2.99567103e-01 -3.95796686e-01 -5.65586388e-01 1.53546557e-01
-4.70043182e-01 -1.44737169e-01 4.57813203e-01 -3.53628606e-01
-1.17715299e+00 1.23205960e-01 3.83559540e-02 -5.82091987e-01
-3.00158083e-01 4.60080594e-01 -6.23215199e-01 -4.86692220e-01
4.21615988e-01 4.04861808e-01 -3.71568017e-02 -6.17547631e-01
2.66254336e-01 5.00971854e-01 4.50632781e-01 -7.50826120e-01
1.20808899e+00 8.68038297e-01 2.57705063e-01 -1.15244865e+00
-6.86432004e-01 -4.57357585e-01 -6.59519076e-01 -1.97747141e-01
4.38781500e-01 -1.16621470e+00 -5.63723385e-01 5.61484933e-01
-7.53968418e-01 4.76242840e-01 8.36720541e-02 6.48774564e-01
-1.64216325e-01 8.22639167e-01 -3.76951963e-01 -9.73520041e-01
-7.70945996e-02 -1.23692143e+00 1.01706517e+00 2.59682536e-01
5.31691432e-01 -8.60379279e-01 -8.99831578e-02 6.89671874e-01
4.37704921e-01 1.61191076e-01 5.63600004e-01 -3.93484563e-01
-4.08527523e-01 -4.46317077e-01 -9.94501784e-02 6.08183861e-01
6.26157463e-01 -2.54340291e-01 -1.27792382e+00 -8.85253310e-01
2.36214802e-01 -6.28455520e-01 7.33653426e-01 3.12424272e-01
1.14268374e+00 -4.18945462e-01 -1.50266409e-01 5.84870517e-01
1.54570007e+00 8.77150614e-03 4.46037054e-01 -2.05823243e-01
7.00859070e-01 5.57265341e-01 6.55586898e-01 6.62711680e-01
-3.68673243e-02 7.04496086e-01 4.13024515e-01 1.08495969e-02
-1.31387576e-01 1.44158065e-01 6.62670374e-01 8.72608900e-01
-1.98477060e-02 1.87177081e-02 -4.90407258e-01 4.79120702e-01
-1.59124172e+00 -1.16817713e+00 4.82233502e-02 2.27030420e+00
5.04818916e-01 -2.07436070e-01 -7.35587180e-02 3.77084762e-01
5.19452453e-01 3.86741012e-01 -4.68934268e-01 1.70420021e-01
-6.73130989e-01 3.35272908e-01 1.10302672e-01 3.22218627e-01
-1.05291295e+00 4.81346667e-01 6.65076447e+00 7.32894540e-01
-9.86148298e-01 2.05660224e-01 4.93624240e-01 -3.22867408e-02
-1.31125540e-01 -2.66764313e-01 -7.55479395e-01 1.46025717e-01
4.86428082e-01 2.38476962e-01 2.87044585e-01 5.49915731e-01
2.25633889e-01 -1.09883353e-01 -9.24555898e-01 1.60564446e+00
7.44842947e-01 -6.94274783e-01 8.40367004e-02 1.47168981e-02
1.09708416e+00 -3.07399422e-01 4.96519595e-01 1.04831733e-01
-1.53494969e-01 -1.18889892e+00 1.26459986e-01 6.67786777e-01
5.82289398e-01 -6.43489122e-01 6.36562169e-01 7.82099087e-03
-1.08841848e+00 -1.68146998e-01 -5.85546255e-01 5.26341051e-02
-4.63942438e-02 9.16437447e-01 -4.57281858e-01 5.84179699e-01
5.35255075e-01 9.12528515e-01 -5.59080303e-01 8.16672921e-01
-2.62509763e-01 7.23423719e-01 -2.91170686e-01 6.72877908e-01
-2.70775199e-01 -4.57959384e-01 6.78051531e-01 9.43106174e-01
2.23409891e-01 2.16118842e-01 4.35832828e-01 3.43026668e-01
-2.96430979e-02 2.77845979e-01 -8.35726380e-01 5.29423840e-02
3.40680033e-01 1.47122824e+00 -4.59668458e-01 -2.39632130e-02
-7.95777142e-01 8.63191366e-01 -1.81268781e-01 5.97832203e-01
-6.23058677e-01 1.11386493e-01 7.27157950e-01 -2.13805139e-01
4.74947482e-01 -4.24876630e-01 2.35126466e-02 -1.21412730e+00
2.06521243e-01 -1.22412932e+00 5.15350640e-01 -2.96237171e-01
-1.55189025e+00 4.57429320e-01 -2.98292213e-03 -1.25177932e+00
7.14077950e-02 -6.56281650e-01 -1.97019488e-01 4.41657633e-01
-1.82218802e+00 -1.24925625e+00 -2.97916859e-01 1.03016949e+00
6.18168116e-01 -6.48639739e-01 8.94702792e-01 7.40072727e-01
-8.58773708e-01 7.45022118e-01 6.04974568e-01 -3.87174264e-03
9.20376718e-01 -6.90040767e-01 -7.66736746e-01 9.80659842e-01
4.81507838e-01 8.03845465e-01 6.47277713e-01 -3.57193947e-01
-2.18617749e+00 -9.51951623e-01 3.37035805e-01 -1.93943456e-01
2.37876117e-01 -1.49736315e-01 -9.83682394e-01 3.63835722e-01
1.38458893e-01 6.78850472e-01 1.04999781e+00 8.07196051e-02
-7.03638911e-01 -5.34325540e-01 -1.29362261e+00 1.10457748e-01
8.99515927e-01 -7.19093323e-01 -4.64757740e-01 4.85098243e-01
-3.06975305e-01 -1.55365812e-02 -1.04436135e+00 6.09703362e-01
5.77084184e-01 -7.74876297e-01 1.12195969e+00 -2.93324918e-01
-1.67301729e-01 -6.85345352e-01 -7.19415128e-01 -1.06186175e+00
-2.60193825e-01 -3.99572492e-01 -9.80319381e-02 1.46814311e+00
-1.39201939e-01 -5.36664188e-01 3.34090799e-01 2.55036056e-01
2.62337357e-01 -3.09473455e-01 -1.16408813e+00 -8.58606279e-01
-5.34982800e-01 -2.62555748e-01 1.90224588e-01 9.38648462e-01
-4.77561772e-01 2.43141174e-01 -9.18901265e-01 5.67846537e-01
1.16960227e+00 4.50737238e-01 6.94279671e-01 -1.27577329e+00
-1.08309507e-01 1.81709513e-01 -6.82929158e-01 -7.37862587e-01
4.26842213e-01 -8.22929084e-01 7.76186436e-02 -1.21631324e+00
4.55398023e-01 -2.08965451e-01 -4.87394780e-01 4.58283156e-01
-1.21600419e-01 9.56363261e-01 -3.04582939e-02 3.85188669e-01
-5.73151588e-01 7.95821786e-01 9.78259921e-01 -4.46078718e-01
2.29021817e-01 -1.86261460e-01 -5.36491156e-01 6.49895072e-01
3.22648138e-01 -4.20208305e-01 -3.53633344e-01 -4.87646669e-01
-1.98976547e-01 -9.15734395e-02 3.81469548e-01 -1.06925619e+00
1.17297284e-01 -2.14007407e-01 7.56933451e-01 -4.26010519e-01
5.70948005e-01 -1.18233466e+00 3.75649840e-01 2.76950181e-01
6.01832904e-02 -2.08713621e-01 1.21380381e-01 1.00420487e+00
-5.21211088e-01 9.50821787e-02 1.11504602e+00 1.34259135e-01
-7.41882980e-01 5.31295896e-01 -6.41431957e-02 -2.34358191e-01
8.66535902e-01 -3.15408945e-01 2.18040701e-02 -3.28428000e-01
-7.56690323e-01 -2.86211759e-01 2.77870744e-01 6.44927502e-01
9.44002926e-01 -1.55104613e+00 -9.42938328e-01 6.78336084e-01
1.96815774e-01 -3.81092310e-01 3.57062995e-01 9.57276762e-01
8.09236169e-02 3.26423705e-01 -3.11630428e-01 -8.19493949e-01
-1.50102735e+00 6.28076136e-01 1.76384896e-02 2.17463002e-01
-3.84890437e-01 7.11642206e-01 1.17767140e-01 -5.68305701e-02
3.70954722e-01 2.33201519e-01 -2.78131306e-01 4.11706656e-01
8.01102281e-01 3.53738993e-01 2.89742082e-01 -1.46785986e+00
-5.70989609e-01 1.07458174e+00 1.32823214e-02 5.05955443e-02
1.49530721e+00 -3.10944766e-01 -2.18086362e-01 3.24686348e-01
1.39349759e+00 1.84612855e-01 -1.05040562e+00 -7.84227490e-01
-1.36430100e-01 -1.11557150e+00 1.87924832e-01 -4.53698426e-01
-1.55928600e+00 8.45998466e-01 1.11753631e+00 -3.50483924e-01
1.43665504e+00 -3.55155766e-01 3.92836601e-01 3.58050168e-01
7.29518533e-01 -9.40893352e-01 4.65337634e-01 2.67281890e-01
1.00222909e+00 -1.62997091e+00 4.38795447e-01 -3.75562251e-01
-3.68248105e-01 1.07649517e+00 2.93181062e-01 1.09761707e-01
8.60947251e-01 -3.08714435e-02 2.05222264e-01 -2.00843081e-01
-4.19345230e-01 -1.67480543e-01 4.16018754e-01 8.34050834e-01
3.78407210e-01 -1.77421883e-01 -2.10739493e-01 2.93871880e-01
3.94771188e-01 -2.33301312e-01 1.22682527e-01 9.64562833e-01
-2.37821743e-01 -1.16799891e+00 -9.15793478e-01 3.11860740e-01
-5.72660983e-01 3.17805856e-01 -1.09698877e-01 3.54252309e-01
1.32449910e-01 1.36534679e+00 -4.52875167e-01 -1.85963511e-01
2.91540951e-01 6.41968995e-02 6.89702809e-01 -5.25764287e-01
-8.25602338e-02 4.82600540e-01 -4.38338995e-01 -6.45597041e-01
-9.96589184e-01 -9.37996566e-01 -8.56078029e-01 1.18675627e-01
-4.04140413e-01 3.62133570e-02 5.15432835e-01 8.06628108e-01
3.20893049e-01 6.00642264e-02 1.15780604e+00 -8.99548888e-01
-5.59643686e-01 -8.53098214e-01 -9.66318429e-01 7.96737552e-01
5.53362429e-01 -1.14769447e+00 -4.39700752e-01 2.25754991e-01] | [12.653082847595215, 0.3810901343822479] |
29d378d8-79b0-4513-b0c1-84ae0f9e21e5 | an-analysis-of-universal-differential | 2306.10335 | null | https://arxiv.org/abs/2306.10335v1 | https://arxiv.org/pdf/2306.10335v1.pdf | An analysis of Universal Differential Equations for data-driven discovery of Ordinary Differential Equations | In the last decade, the scientific community has devolved its attention to the deployment of data-driven approaches in scientific research to provide accurate and reliable analysis of a plethora of phenomena. Most notably, Physics-informed Neural Networks and, more recently, Universal Differential Equations (UDEs) proved to be effective both in system integration and identification. However, there is a lack of an in-depth analysis of the proposed techniques. In this work, we make a contribution by testing the UDE framework in the context of Ordinary Differential Equations (ODEs) discovery. In our analysis, performed on two case studies, we highlight some of the issues arising when combining data-driven approaches and numerical solvers, and we investigate the importance of the data collection process. We believe that our analysis represents a significant contribution in investigating the capabilities and limitations of Physics-informed Machine Learning frameworks. | ['Michele Lombardi', 'Eleonora Misino', 'Federico Baldo', 'Mattia Silvestri'] | 2023-06-17 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 8.41694102e-02 -1.89619660e-01 1.12966329e-01 2.85474867e-01
-2.25436285e-01 -4.70089078e-01 7.85847425e-01 2.62114644e-01
-1.01486787e-01 1.03093326e+00 -3.35464984e-01 -5.96522808e-01
-7.36848831e-01 -6.15580261e-01 -4.74118829e-01 -8.01628768e-01
-2.61414081e-01 3.45917314e-01 -2.43065640e-01 -4.10035968e-01
4.52562571e-01 1.22120237e+00 -1.65001285e+00 -5.76397240e-01
1.06276560e+00 1.04490066e+00 -2.73937017e-01 4.83572483e-01
-6.25824630e-02 7.16186762e-01 -1.90144390e-01 -7.32478127e-02
1.20628528e-01 -3.39279681e-01 -7.35365927e-01 -3.12554747e-01
7.60025233e-02 2.97301561e-02 -3.36058140e-01 7.24894583e-01
5.36823928e-01 3.71253103e-01 6.61273956e-01 -1.33127046e+00
-5.09285986e-01 1.10726841e-01 -1.98488221e-01 5.58867157e-01
1.93830222e-01 4.58531231e-01 7.62076080e-01 -8.98360074e-01
5.50237298e-01 8.15111458e-01 7.50744283e-01 3.20172936e-01
-1.10703301e+00 -5.15495598e-01 -9.62674171e-02 2.37300396e-01
-1.44754970e+00 -3.78146261e-01 1.12785470e+00 -6.93780959e-01
8.76358628e-01 3.21816027e-01 6.63706601e-01 8.41534674e-01
3.85868311e-01 3.77621621e-01 1.46050584e+00 -6.58730328e-01
6.23723149e-01 3.29127520e-01 2.42215008e-01 3.87912959e-01
5.35444498e-01 4.11393464e-01 -5.29397666e-01 -2.07622290e-01
1.01074040e+00 -4.83018965e-01 -6.07804768e-02 -4.54336256e-01
-6.22032642e-01 9.88726795e-01 1.00750037e-01 6.42368317e-01
-5.82893193e-01 -7.28392228e-02 3.26321751e-01 4.88925651e-02
7.11081743e-01 8.81298244e-01 -4.71636802e-01 -3.29786956e-01
-8.99234474e-01 6.15508318e-01 1.06119859e+00 4.82175946e-01
6.58118606e-01 3.21410567e-01 2.47936115e-01 2.56758243e-01
4.91952673e-02 1.71201155e-01 2.30076641e-01 -1.15436554e+00
-1.86933622e-01 5.52302778e-01 3.25948864e-01 -1.15518546e+00
-4.60708499e-01 -5.92299759e-01 -9.40433800e-01 3.24463814e-01
4.48719740e-01 -5.47085643e-01 -2.82999098e-01 1.46203387e+00
5.50967515e-01 2.70070553e-01 6.68733865e-02 9.34301913e-01
5.92452288e-01 4.98394310e-01 1.51074126e-01 -3.01315874e-01
9.30168808e-01 -5.22367179e-01 -6.69186950e-01 2.55780190e-01
4.70640033e-01 -6.14346743e-01 6.23767674e-01 4.51004982e-01
-1.25325489e+00 -5.13832986e-01 -8.42527390e-01 1.92686677e-01
-6.85102522e-01 -2.80065745e-01 9.95237410e-01 5.17978251e-01
-9.00276601e-01 9.79316771e-01 -9.17111635e-01 -6.13441408e-01
-3.75592429e-03 3.44966471e-01 -1.81018971e-02 4.41785753e-01
-1.24759376e+00 1.08793521e+00 5.72752915e-02 1.45637140e-01
-5.63793838e-01 -1.12737894e+00 -5.71981490e-01 -9.24953371e-02
3.62226248e-01 -7.04724252e-01 9.78344619e-01 -6.89297616e-01
-1.63913000e+00 5.35648346e-01 -1.57195747e-01 -6.95659637e-01
5.24579883e-01 -2.22775601e-02 -4.08110082e-01 1.74746402e-02
-3.22164774e-01 6.60202801e-02 2.85005242e-01 -1.15286827e+00
-2.47977644e-01 -2.51548082e-01 6.71348721e-02 -2.55703509e-01
-2.87082076e-01 9.65895876e-02 5.87374382e-02 -4.09468651e-01
-7.29559138e-02 -8.48540187e-01 -3.23180407e-01 1.82489790e-02
-2.21492440e-01 -1.88034192e-01 8.20637584e-01 -5.39054215e-01
1.23372245e+00 -1.67557216e+00 3.80478173e-01 1.90737695e-01
1.49966851e-01 4.46519703e-01 4.71737146e-01 9.14498031e-01
-1.12272292e-01 6.54157698e-02 -3.37633938e-01 -2.15242356e-01
8.16497505e-02 3.37788850e-01 -4.84659672e-01 4.82122362e-01
4.63546753e-01 8.16298842e-01 -5.41487813e-01 -3.15885216e-01
5.91128111e-01 7.83051908e-01 -2.71848351e-01 5.71583845e-02
-2.28966102e-01 1.00546253e+00 -7.76807666e-01 4.58394259e-01
5.56961417e-01 -1.84611842e-01 -4.23684269e-02 -1.36945873e-01
-7.23101139e-01 7.02430829e-02 -1.31343031e+00 1.29620218e+00
-5.63554049e-01 6.07330263e-01 1.89662784e-01 -1.48872626e+00
7.11560905e-01 3.17664057e-01 9.65378225e-01 -7.69278467e-01
4.19844180e-01 3.66351068e-01 1.30192891e-01 -6.63083315e-01
3.96885127e-01 -1.24368593e-01 2.58780956e-01 3.86810273e-01
1.24777332e-02 -1.92588538e-01 2.92024076e-01 -6.94733933e-02
7.84918725e-01 3.36671591e-01 3.65099192e-01 -6.25469089e-01
9.92263734e-01 3.24876249e-01 1.90372482e-01 5.89965045e-01
-1.93906665e-01 2.56935537e-01 3.49814683e-01 -5.56013167e-01
-1.08751237e+00 -6.27688527e-01 -5.93188584e-01 4.67093170e-01
-1.38293043e-01 -1.03480138e-01 -6.61005378e-01 2.45137379e-01
2.66093969e-01 6.96139634e-01 -6.51076317e-01 -1.40378147e-01
-6.21319830e-01 -9.13287461e-01 4.77879912e-01 4.26082939e-01
2.50492960e-01 -8.93039823e-01 -8.00017357e-01 3.64567131e-01
3.85271788e-01 -1.15837646e+00 5.05189419e-01 1.36995897e-01
-1.17769134e+00 -1.07706809e+00 -6.88894212e-01 -1.16482422e-01
2.41152719e-01 -2.63360739e-01 8.72883141e-01 -9.33674648e-02
-5.91866910e-01 6.00732386e-01 -2.09587008e-01 -3.89827013e-01
-6.84601724e-01 2.60965139e-01 2.77313709e-01 -2.45668426e-01
1.54887408e-01 -1.05902863e+00 -4.29163128e-01 9.05490965e-02
-7.45779872e-01 -1.37974113e-01 3.54336649e-01 2.47902483e-01
3.02552313e-01 1.05938450e-01 7.19250858e-01 -5.48482299e-01
8.18150997e-01 -7.97004759e-01 -8.36466491e-01 -2.45092362e-02
-8.83071482e-01 1.14279635e-01 8.59165490e-01 -3.34752738e-01
-9.77877975e-01 -2.49857962e-01 -8.35037977e-02 -3.07535022e-01
-3.74766022e-01 7.80304313e-01 3.57400417e-01 -7.54782379e-01
5.81101537e-01 1.75223649e-01 1.61484644e-01 -7.16392338e-01
1.70250565e-01 4.43613857e-01 4.77052122e-01 -1.01432264e+00
7.29146659e-01 3.93677413e-01 6.28468037e-01 -1.05675244e+00
-2.70208597e-01 -3.58739972e-01 -6.17781997e-01 -2.92434484e-01
4.71741170e-01 -4.10115957e-01 -8.08617115e-01 5.99045038e-01
-9.60191727e-01 -1.20625652e-01 -4.05654818e-01 3.52236986e-01
-4.44907755e-01 2.58432537e-01 -6.03841662e-01 -1.40846539e+00
-2.85727143e-01 -1.03742933e+00 5.75153291e-01 6.15203261e-01
-3.74008477e-01 -1.38361943e+00 2.47847721e-01 -1.84018333e-02
7.59042084e-01 6.29915535e-01 8.96487534e-01 -5.66243708e-01
-4.60641652e-01 -2.31637359e-01 -1.18555844e-01 -1.26564121e-02
-7.29958266e-02 4.65997875e-01 -1.12714529e+00 2.05759276e-02
9.08118784e-02 -4.38970104e-02 2.87405908e-01 5.37568033e-01
9.01478469e-01 8.22744519e-02 -2.79323906e-01 3.82079810e-01
1.74021924e+00 2.38092780e-01 2.80676305e-01 2.94792652e-01
4.25357789e-01 6.71543419e-01 1.30142644e-01 6.93452179e-01
1.49594992e-01 6.76838815e-01 2.99408495e-01 -8.91920552e-02
7.44687095e-02 1.77427649e-01 -1.85737416e-01 6.33781016e-01
-4.21127021e-01 -2.91760312e-03 -1.11721468e+00 3.73424292e-01
-1.67695701e+00 -4.53323543e-01 -5.25194466e-01 1.96415699e+00
5.99286497e-01 -6.45738989e-02 1.61265567e-01 2.13753164e-01
5.30394137e-01 -2.96835065e-01 -6.69981539e-01 -6.22691989e-01
-9.90516692e-02 5.53596795e-01 2.88564414e-01 3.73090655e-01
-8.25401068e-01 5.33929348e-01 6.96747065e+00 4.99918401e-01
-1.42612481e+00 -5.98495826e-02 3.13921720e-01 1.70996204e-01
-8.80396292e-02 6.28899634e-02 -4.93905216e-01 3.65747929e-01
1.30969095e+00 -5.45687675e-01 4.65488732e-01 7.22436070e-01
6.89574957e-01 -3.88671517e-01 -8.98857594e-01 5.43917835e-01
-2.65433609e-01 -1.45688200e+00 -3.49578410e-01 2.24960566e-01
7.32400596e-01 -7.40838200e-02 -8.32400564e-03 1.37559071e-01
-2.03222549e-03 -1.03574479e+00 3.58260751e-01 7.69326627e-01
2.58428127e-01 -6.15442693e-01 5.21627009e-01 3.44742090e-01
-8.30318928e-01 -4.06221300e-02 3.68628465e-02 -7.42354572e-01
1.30346939e-01 9.40110147e-01 -3.22417915e-01 8.93246114e-01
5.36708176e-01 4.54977602e-01 -2.33009681e-01 1.24491501e+00
3.49185348e-01 7.32327700e-01 -4.20339167e-01 -5.72851934e-02
2.14482725e-01 -4.57100123e-01 7.10962534e-01 9.47291076e-01
2.35735342e-01 1.05989031e-01 -2.00892240e-01 1.46891475e+00
5.33176720e-01 4.93650176e-02 -6.00838840e-01 -2.61859387e-01
1.39603227e-01 1.39227641e+00 -6.08983994e-01 1.41718974e-02
-3.35505188e-01 2.36576140e-01 1.10652223e-01 4.77089614e-01
-7.80218422e-01 -5.81971779e-02 7.35139966e-01 -1.24875680e-02
7.09254816e-02 -6.68750763e-01 -6.34034157e-01 -9.31140661e-01
-1.48794398e-01 -6.87511086e-01 4.89549488e-02 -7.30428278e-01
-1.20724070e+00 1.80398464e-01 4.35078114e-01 -7.07946479e-01
-4.15758580e-01 -8.98891151e-01 -7.54952669e-01 1.16488147e+00
-1.42143822e+00 -6.55864656e-01 -1.75463453e-01 1.01468168e-01
7.31480047e-02 -4.18519303e-02 8.82706821e-01 4.44234580e-01
-9.24186110e-01 -9.26098078e-02 4.91813689e-01 -3.27817917e-01
2.65486725e-03 -9.03031707e-01 2.93229461e-01 7.30692685e-01
-2.90419340e-01 7.02698827e-01 1.34093893e+00 -4.37350810e-01
-1.82625389e+00 -4.79051441e-01 6.33273065e-01 -1.70469016e-01
1.06932366e+00 -3.43676247e-02 -1.10521984e+00 3.19281310e-01
1.11435331e-01 -1.66127846e-01 4.52561855e-01 -3.75727564e-02
3.05909783e-01 1.12053573e-01 -1.13385153e+00 3.42447996e-01
6.70579135e-01 -3.49604398e-01 -3.31186146e-01 1.57532245e-01
1.22111902e-01 -3.84013832e-01 -1.02616370e+00 4.25399095e-01
5.74409604e-01 -9.30619955e-01 1.09602940e+00 -6.48495376e-01
5.31867206e-01 8.81377012e-02 1.84514746e-01 -9.56390560e-01
-7.74737782e-05 -6.59814000e-01 -5.14901936e-01 1.16852021e+00
-4.46961522e-02 -9.97655928e-01 6.64134085e-01 1.07152987e+00
7.23333433e-02 -1.01331878e+00 -9.82903719e-01 -7.39772975e-01
7.25013077e-01 -6.45565033e-01 3.53315681e-01 9.95803654e-01
-6.12257943e-02 -3.22152048e-01 -5.57194613e-02 2.08862960e-01
3.89231145e-01 1.70817196e-01 6.34489119e-01 -1.56135046e+00
-1.32085979e-01 -5.92366934e-01 -3.04869503e-01 -2.99784094e-01
2.40904287e-01 -4.04504269e-01 -5.19749582e-01 -1.22054005e+00
-5.94330691e-02 -3.57732892e-01 -2.48963580e-01 -1.24600850e-01
2.38102719e-01 8.62035751e-02 -5.09062745e-02 2.54092455e-01
-1.31001230e-02 4.02019531e-01 1.05421662e+00 2.83756554e-01
-2.00647801e-01 -1.45625353e-01 -6.14731312e-01 6.89639270e-01
9.44562495e-01 -1.88884437e-01 -1.40119389e-01 -3.75101045e-02
2.72341698e-01 3.11715826e-02 8.36979985e-01 -1.22600293e+00
3.91550779e-01 -2.57449478e-01 1.80903539e-01 -2.61469424e-01
2.54445463e-01 -7.17787504e-01 2.60561317e-01 4.46845174e-01
-7.66917914e-02 -3.20957564e-02 6.46388650e-01 7.70295486e-02
-1.10552713e-01 -3.02328229e-01 5.87126493e-01 -7.66031668e-02
-6.98176384e-01 -2.69252826e-02 -5.83939672e-01 7.71051869e-02
9.31333542e-01 -2.68840969e-01 6.53710365e-02 -2.72537678e-01
-5.89481473e-01 5.82717173e-02 3.84092242e-01 -4.76989755e-03
2.80630607e-02 -8.27032268e-01 -3.38217080e-01 1.11899003e-01
-3.29978645e-01 -1.10726416e-01 3.45310539e-01 1.00708199e+00
-6.22717142e-01 8.77382815e-01 -2.54465461e-01 -3.61813605e-01
-7.64472485e-01 4.78328675e-01 6.97706759e-01 -2.71973014e-01
-5.66118419e-01 4.84591842e-01 -2.08369225e-01 -3.83107394e-01
6.78993389e-02 -1.50810391e-01 -2.77203415e-03 -9.14673060e-02
8.92267525e-02 8.86245549e-01 8.78473148e-02 -6.46112859e-01
-3.02782118e-01 6.57260478e-01 3.60024512e-01 -1.37392953e-01
1.33936465e+00 2.62046084e-02 -1.26404643e-01 5.98412693e-01
9.67495441e-01 -1.75736040e-01 -1.15361273e+00 8.56791660e-02
-7.77298063e-02 1.08854979e-01 4.15455073e-01 -8.50352407e-01
-9.28461015e-01 7.76913345e-01 3.66571039e-01 6.89007282e-01
1.01344764e+00 -2.42189720e-01 6.78308189e-01 3.57678711e-01
3.01646471e-01 -1.01619422e+00 -4.83754098e-01 5.15154123e-01
7.90581048e-01 -1.01115048e+00 1.62062317e-01 -4.55167770e-01
-4.60740691e-03 1.34302855e+00 4.67912078e-01 -3.05061936e-01
7.46989071e-01 3.98004144e-01 -2.28378996e-01 -2.15118885e-01
-4.32296574e-01 -7.63555989e-02 3.18075806e-01 4.57642823e-01
6.00868404e-01 -3.41744810e-01 -5.29402137e-01 5.48894107e-01
-4.03437763e-02 5.33155024e-01 3.79955262e-01 1.13046646e+00
-2.22184226e-01 -1.19606102e+00 -5.31528115e-01 4.46959585e-03
-3.47166210e-01 9.89907607e-02 -4.78110433e-01 1.36124778e+00
4.80622277e-02 7.86654413e-01 -1.83401331e-01 5.70544600e-02
1.96124524e-01 2.77142435e-01 5.30615509e-01 -1.56560652e-02
-5.47834933e-01 -4.41515505e-01 -1.92752793e-01 -2.81454295e-01
-5.95855653e-01 -7.95073688e-01 -1.25477958e+00 -6.16036415e-01
-2.02251807e-01 2.88493633e-01 9.17507231e-01 1.41952646e+00
5.64141870e-01 7.32875466e-01 2.13942602e-01 -9.63056564e-01
-4.79577810e-01 -6.62884533e-01 -6.24936521e-01 9.26741213e-02
1.72479466e-01 -9.80258882e-01 -5.89146674e-01 -2.31056511e-01] | [6.379023551940918, 3.4963746070861816] |
8f946eb0-89b1-4eab-b340-44b71ee98936 | optimization-of-the-area-under-the-roc-curve | 1901.11332 | null | http://arxiv.org/abs/1901.11332v2 | http://arxiv.org/pdf/1901.11332v2.pdf | Optimization of the Area Under the ROC Curve using Neural Network Supervectors for Text-Dependent Speaker Verification | This paper explores two techniques to improve the performance of
text-dependent speaker verification systems based on deep neural networks.
Firstly, we propose a general alignment mechanism to keep the temporal
structure of each phrase and obtain a supervector with the speaker and phrase
information, since both are relevant for a text-dependent verification. As we
show, it is possible to use different alignment techniques to replace the
global average pooling providing significant gains in performance. Moreover, we
also present a novel back-end approach to train a neural network for detection
tasks by optimizing the Area Under the Curve (AUC) as an alternative to the
usual triplet loss function, so the system is end-to-end, with a cost function
close to our desired measure of performance. As we can see in the experimental
section, this approach improves the system performance, since our triplet
neural network based on an approximation of the AUC (aAUC) learns how to
discriminate between pairs of examples from the same identity and pairs of
different identities. The different alignment techniques to produce
supervectors in addition to the new back-end approach were tested on the
RSR2015-Part I database for text-dependent speaker verification, providing
competitive results compared to similar size networks using the global average
pooling to extract supervectors and using a simple back-end or triplet loss
training. | ['Antonio Miguel', 'Alfonso Ortega', 'Victoria Mingote', 'Eduardo Lleida'] | 2019-01-31 | null | null | null | null | ['text-dependent-speaker-verification'] | ['speech'] | [ 1.01617925e-01 2.53182091e-03 3.07337791e-01 -7.74097741e-01
-1.09733045e+00 -5.44730902e-01 5.70861220e-01 6.61998838e-02
-8.30451965e-01 4.69669998e-01 -1.31812990e-01 -3.38153720e-01
-2.98141036e-02 -2.70645738e-01 -6.73630774e-01 -9.44491208e-01
-1.12735413e-01 4.14318591e-01 1.09842874e-01 -3.53631526e-01
1.08459853e-01 6.06308520e-01 -1.45420229e+00 2.42560938e-01
5.91126442e-01 1.19588017e+00 -2.97951400e-01 6.90061152e-01
2.52614886e-01 1.99198127e-01 -9.05108035e-01 -7.40569532e-01
5.60967922e-01 -4.26826030e-01 -6.64798379e-01 -1.88283518e-01
9.82957006e-01 -7.14849979e-02 -1.12469852e-01 9.81149912e-01
8.92107248e-01 1.57329082e-01 6.27799749e-01 -1.11834395e+00
-3.62602204e-01 7.66319931e-01 -3.46389651e-01 2.59989172e-01
3.13790411e-01 9.42657441e-02 1.13367987e+00 -1.01147211e+00
2.20148832e-01 1.15918469e+00 8.83626699e-01 7.18735337e-01
-1.33705330e+00 -8.21523905e-01 1.04551077e-01 3.82514089e-01
-1.49101651e+00 -8.28102827e-01 6.70540094e-01 -8.67227465e-02
9.46917832e-01 1.84579492e-01 1.34642646e-01 1.07141721e+00
-6.32192120e-02 6.25848591e-01 9.28438783e-01 -7.34043241e-01
-5.70503511e-02 5.63087404e-01 2.81566501e-01 5.96481502e-01
-1.49437502e-01 3.05579931e-01 -6.87394738e-01 -2.09387299e-02
7.11270049e-02 -4.43507552e-01 -4.55211729e-01 -3.99290979e-01
-1.08090806e+00 7.24669218e-01 5.03241122e-01 6.19380474e-01
-2.16597214e-01 -1.63290083e-01 5.92533529e-01 7.01462805e-01
4.32163268e-01 2.72068053e-01 -4.74890888e-01 3.00381631e-01
-1.25187039e+00 1.57752261e-01 9.77092326e-01 5.34123242e-01
6.95279181e-01 4.16630767e-02 -4.08966690e-01 9.18767452e-01
3.48139882e-01 3.86539459e-01 7.13264048e-01 -4.21480685e-01
7.08195150e-01 1.25792712e-01 -1.22537561e-01 -5.75763702e-01
-3.00384194e-01 -7.23953962e-01 -6.55203998e-01 3.92262220e-01
6.27796531e-01 -2.72772551e-01 -8.88191879e-01 1.99843121e+00
1.04765043e-01 -1.82119496e-02 1.16037481e-01 7.41142690e-01
6.74713731e-01 4.75705653e-01 -1.06456295e-01 -1.18904717e-01
1.47702408e+00 -9.33230579e-01 -6.54774487e-01 1.82691868e-02
6.32114172e-01 -7.34784484e-01 7.15299606e-01 2.46078640e-01
-1.13804555e+00 -7.83603609e-01 -1.48604751e+00 7.50670433e-02
-5.94766617e-01 2.02323005e-01 -1.00916862e-01 1.08039534e+00
-1.46325409e+00 6.93877697e-01 -6.69880629e-01 -2.36267924e-01
5.30405529e-02 6.70958877e-01 -6.16162419e-01 3.18638951e-01
-1.31018114e+00 1.27875268e+00 4.44153875e-01 3.98867667e-01
-6.92086399e-01 -6.24451995e-01 -9.27693069e-01 2.77253419e-01
-1.55328318e-01 -5.21761060e-01 1.31370330e+00 -1.12297285e+00
-1.84923244e+00 1.02901566e+00 -2.78996617e-01 -8.48065376e-01
6.28493190e-01 2.17693582e-01 -3.68397444e-01 -7.34745786e-02
-2.37504244e-01 5.67200661e-01 1.01280212e+00 -8.42077076e-01
-5.28095186e-01 -6.08122647e-01 -2.19890267e-01 2.72242352e-02
-5.35732925e-01 3.42940837e-01 -1.38722643e-01 -3.61652046e-01
7.30071142e-02 -8.31912398e-01 4.92343381e-02 -2.96968281e-01
-3.84634793e-01 -3.86632085e-01 8.41236532e-01 -9.69703674e-01
8.61284852e-01 -2.23718500e+00 1.12649351e-01 4.64150429e-01
-1.52364925e-01 5.48083186e-01 -2.95925468e-01 2.43593276e-01
-4.80526805e-01 -9.10723209e-02 -4.70290273e-01 -1.04655194e+00
1.63320333e-01 -2.80596763e-01 -5.31848334e-02 6.77615225e-01
2.69093663e-01 6.73536837e-01 -3.66125822e-01 -2.26729423e-01
-1.82340685e-02 7.21368015e-01 -1.26364216e-01 2.66551953e-02
2.99682915e-01 2.26852685e-01 1.40556365e-01 1.85898528e-01
9.33636904e-01 2.98825413e-01 3.78231592e-02 -5.22553660e-02
6.90772533e-02 6.32828534e-01 -1.07695627e+00 1.51969326e+00
-6.79138660e-01 8.42147946e-01 4.15729702e-01 -1.17391956e+00
1.18362415e+00 8.44499767e-01 1.81144495e-02 -6.00133061e-01
3.56815368e-01 3.75573844e-01 1.04126468e-01 -2.60053873e-02
2.73215175e-01 -4.05870914e-01 1.03351615e-01 4.84702468e-01
4.56294656e-01 1.57344371e-01 5.64821921e-02 -1.85040399e-01
6.36841416e-01 -1.16555095e-01 3.01689212e-03 -1.06352583e-01
1.16350150e+00 -6.06817544e-01 3.36134613e-01 7.17502594e-01
-4.86503363e-01 6.88911319e-01 3.40950966e-01 -1.45794481e-01
-8.22354376e-01 -8.33166897e-01 -2.26155743e-01 1.03869712e+00
-3.70020300e-01 -3.72925699e-02 -8.49395394e-01 -8.66705954e-01
-7.61898905e-02 7.27912366e-01 -6.17816329e-01 -9.49153006e-02
-7.96750128e-01 -6.03159010e-01 8.85655820e-01 4.35848176e-01
4.25779134e-01 -9.92212772e-01 3.65480855e-02 1.01409294e-01
-1.56222269e-01 -1.14571488e+00 -6.60940707e-01 6.09515846e-01
-6.36600614e-01 -5.92737496e-01 -9.79759514e-01 -9.98718977e-01
4.93504971e-01 -9.14106891e-02 9.05360222e-01 -7.35010803e-02
1.67918697e-01 -3.97223122e-02 -7.83490166e-02 -4.23956394e-01
-7.08735287e-01 2.56587654e-01 1.82955861e-01 4.69774723e-01
6.44697964e-01 -5.95431149e-01 -3.54919493e-01 3.54683369e-01
-5.29637814e-01 -6.16618812e-01 4.34138775e-01 1.04115736e+00
4.47797105e-02 -3.00352365e-01 5.97002149e-01 -2.95428783e-01
6.05027735e-01 1.93305209e-01 -8.22950780e-01 3.20640117e-01
-5.96651196e-01 2.89724946e-01 6.51729465e-01 -3.05056036e-01
-8.81699681e-01 2.67914444e-01 -4.58360225e-01 -5.29777527e-01
-2.32748672e-01 1.54937595e-01 -2.40662545e-01 -2.55924940e-01
7.54000485e-01 4.82082993e-01 3.29661340e-01 -4.71493155e-01
2.71787643e-01 9.84751761e-01 5.45999408e-01 -1.12367839e-01
9.42616463e-01 1.94315270e-01 -2.25334451e-01 -7.21697509e-01
-7.14085519e-01 -7.36836016e-01 -7.81533659e-01 4.88682240e-02
7.21176803e-01 -7.85701036e-01 -9.94618833e-01 6.29991531e-01
-1.36008573e+00 8.14081281e-02 -1.35305539e-01 5.99516749e-01
-4.29836482e-01 4.92221832e-01 -5.35820007e-01 -8.83935630e-01
-6.61597848e-01 -1.18326235e+00 9.46177185e-01 1.63481846e-01
9.09868851e-02 -9.63843763e-01 6.81781545e-02 3.24279040e-01
5.85727215e-01 -2.11179897e-01 5.63379109e-01 -1.34327638e+00
-2.39094332e-01 -6.29804730e-01 -5.02941310e-02 9.59242582e-01
-7.31406808e-02 -8.09670165e-02 -1.52367961e+00 -5.49603581e-01
3.85559618e-01 -6.49223253e-02 1.04864001e+00 2.94829786e-01
6.14733696e-01 -1.89506382e-01 -4.64632809e-02 4.85343516e-01
1.05668509e+00 4.01071608e-02 5.36943793e-01 2.56989390e-01
2.68332064e-01 7.70398319e-01 2.22826153e-01 -7.82015547e-02
7.02144355e-02 1.17839921e+00 9.74841565e-02 -6.56584278e-02
-2.18538672e-01 1.68846957e-02 7.89640129e-01 7.49248981e-01
1.03822619e-01 -6.73671439e-02 -7.33163834e-01 5.77803075e-01
-1.60932803e+00 -1.16232038e+00 4.74061333e-02 2.72206879e+00
6.73550963e-01 2.15416238e-01 3.73355120e-01 4.63318199e-01
9.97883618e-01 1.01515792e-01 -1.33284390e-01 -7.63589382e-01
-2.60524780e-01 4.30906713e-01 4.00816262e-01 7.56600678e-01
-1.19998252e+00 6.83414400e-01 6.22669172e+00 8.71636629e-01
-1.55090165e+00 3.22431386e-01 4.57901388e-01 2.55120941e-03
2.89580047e-01 -2.47153684e-01 -1.15003443e+00 2.24258736e-01
1.41706860e+00 5.09986840e-03 9.54667404e-02 5.86799443e-01
1.89468443e-01 3.58166367e-01 -1.36367369e+00 8.53136241e-01
4.26206976e-01 -8.14909458e-01 -1.67206511e-01 2.62043513e-02
4.42590386e-01 1.19726829e-01 2.30252430e-01 3.24609816e-01
-9.67645124e-02 -8.31349909e-01 8.14060688e-01 1.10656925e-01
4.34247553e-01 -7.85072684e-01 1.08282447e+00 2.61323720e-01
-1.02693391e+00 -1.15036003e-01 -1.84424207e-01 1.81527957e-01
1.46769688e-01 4.15370107e-01 -1.23592985e+00 6.90296710e-01
4.69064564e-01 2.40373418e-01 -5.30644119e-01 1.19826424e+00
-3.05265933e-01 4.03711826e-01 -3.47878397e-01 -1.17652692e-01
2.08803177e-01 2.28602402e-02 7.36290157e-01 1.33629477e+00
2.53971696e-01 -7.51003563e-01 -3.48657072e-01 7.60856628e-01
-1.49802402e-01 1.99076936e-01 -4.09677625e-01 3.02439243e-01
1.34676248e-01 1.17163658e+00 -2.54174769e-01 -2.28554443e-01
-2.26673052e-01 9.98120189e-01 2.98398167e-01 3.24900180e-01
-7.32389808e-01 -7.82384813e-01 4.58966553e-01 -1.87779188e-01
7.36803770e-01 1.64593887e-02 -2.61606246e-01 -9.29190755e-01
4.04300243e-01 -8.85937035e-01 3.06416452e-01 -2.85515815e-01
-1.26397252e+00 1.02943122e+00 -3.18570286e-01 -1.33913410e+00
-4.64248657e-01 -7.17693388e-01 -6.89743042e-01 1.39932334e+00
-1.67632127e+00 -1.06557405e+00 1.20363034e-01 5.18676281e-01
2.63605446e-01 -2.93651015e-01 9.57325041e-01 5.18839240e-01
-5.90206623e-01 1.39352274e+00 1.77397970e-02 3.83804768e-01
8.93116891e-01 -1.21211624e+00 5.63647807e-01 1.00076675e+00
3.23901862e-01 6.45947814e-01 6.81167305e-01 1.49981469e-01
-8.53132904e-01 -7.22202182e-01 1.36740160e+00 -3.78193200e-01
4.15880859e-01 -6.28540397e-01 -9.73267555e-01 5.51053107e-01
4.16787744e-01 -1.20166905e-01 5.60573876e-01 3.70139152e-01
-6.74488604e-01 -4.72974956e-01 -1.20719028e+00 2.82178670e-01
5.54208517e-01 -8.45684886e-01 -8.09699833e-01 2.64365911e-01
4.84579414e-01 -1.39212906e-01 -6.01957381e-01 3.47474158e-01
6.30851090e-01 -1.16084552e+00 8.68364096e-01 -4.45295095e-01
-6.34976551e-02 -4.05793548e-01 9.39182937e-02 -1.37452102e+00
-2.67054699e-02 -6.89439893e-01 2.09277973e-01 1.32981265e+00
1.06074679e+00 -9.13199306e-01 7.06819594e-01 3.92131329e-01
-5.20902202e-02 -4.46922719e-01 -1.46167397e+00 -1.05213487e+00
1.51056215e-01 -2.87899822e-01 4.67287481e-01 6.61806524e-01
3.02200764e-02 3.26359212e-01 -2.37356171e-01 3.07175756e-01
5.39531648e-01 -1.07029602e-01 6.24099195e-01 -1.12829125e+00
-4.00485933e-01 -7.55969822e-01 -6.95427179e-01 -1.04186594e+00
4.72519279e-01 -9.95138288e-01 2.66452283e-01 -8.78117740e-01
2.47668549e-02 -1.74281821e-01 -7.67864048e-01 4.38122898e-01
-1.05854757e-01 3.89867336e-01 2.64009833e-01 -1.17789119e-01
-1.72620043e-01 4.71828759e-01 7.57934690e-01 -5.08490503e-01
-2.55299389e-01 5.07822931e-01 -3.73236060e-01 3.20049435e-01
6.65768802e-01 -6.06520772e-01 7.31178075e-02 -2.31013164e-01
-3.59856963e-01 -3.54109071e-02 2.36412659e-01 -1.05900145e+00
4.35121268e-01 8.41893613e-01 2.15148166e-01 -5.46091735e-01
6.63743615e-01 -8.47033918e-01 -3.04808855e-01 6.21517181e-01
-4.44583684e-01 6.51957169e-02 3.46532732e-01 2.09293142e-01
-6.58695638e-01 -4.86394733e-01 1.01804280e+00 2.86865711e-01
-1.23317532e-01 3.86723951e-02 -1.10234722e-01 -5.23965597e-01
6.83074236e-01 -2.80163318e-01 -7.38705099e-02 -4.77461010e-01
-9.01327372e-01 2.79040653e-02 1.04827166e-01 4.59757805e-01
2.25046113e-01 -1.17971992e+00 -1.11788142e+00 3.64906609e-01
7.07319379e-02 -6.39383554e-01 1.81117192e-01 1.13696706e+00
-2.01015696e-01 6.57979667e-01 -2.20765114e-01 -7.92437255e-01
-1.66673577e+00 4.50407118e-01 9.20217574e-01 -3.77163738e-01
-1.87886432e-01 1.12141860e+00 -5.69458585e-03 -8.09587896e-01
5.92620969e-01 -2.27330387e-01 -1.31197616e-01 2.11286858e-01
6.83185816e-01 -8.63748044e-03 8.00128937e-01 -8.65751266e-01
-5.69485307e-01 5.39853334e-01 -2.84100771e-01 -4.28992867e-01
1.10423970e+00 5.32339662e-02 -4.46429272e-04 2.60272920e-01
1.53233039e+00 5.69870509e-02 -8.33930552e-01 -3.86421859e-01
7.40564018e-02 -8.17943290e-02 1.01349540e-02 -8.14065337e-01
-1.12323964e+00 1.17725050e+00 9.17524457e-01 3.33427578e-01
1.03976846e+00 -2.97668278e-01 4.87315953e-01 3.62272233e-01
-2.08152924e-02 -8.35168362e-01 -3.82266492e-01 5.12947202e-01
9.91503835e-01 -1.25425720e+00 -3.24041516e-01 -7.90199563e-02
-3.71580243e-01 1.14801300e+00 2.40590721e-01 1.45119935e-01
6.35985851e-01 1.14151739e-01 5.04144609e-01 1.66789934e-01
-4.57230031e-01 -1.52236328e-01 4.54543293e-01 4.99813884e-01
4.40184683e-01 -6.28748834e-02 -3.62379342e-01 2.08557248e-01
-3.50111872e-01 -4.17442828e-01 1.84863701e-01 5.36137462e-01
-1.92635208e-01 -1.38772893e+00 -5.30802548e-01 -1.68396253e-02
-5.94142139e-01 -1.15127996e-01 -4.81652498e-01 5.65410972e-01
-1.09189659e-01 1.09710312e+00 2.27194410e-02 -3.43118072e-01
4.57353204e-01 6.12020731e-01 3.94421726e-01 -3.24819863e-01
-1.10582149e+00 -2.07151234e-01 8.37093592e-02 -2.39183113e-01
-3.66141737e-01 -7.09598958e-01 -7.21146703e-01 -7.10704923e-02
-7.73178875e-01 2.98313975e-01 1.24983382e+00 1.04688406e+00
2.39429310e-01 2.33769357e-01 9.43589807e-01 -9.39012885e-01
-9.82772768e-01 -1.33553004e+00 -4.49131727e-01 3.68049771e-01
7.31514156e-01 -4.16565061e-01 -6.65598810e-01 -2.11979017e-01] | [14.306634902954102, 6.079519271850586] |
76a21fee-0637-448e-a57b-14e528b19cc0 | investigating-the-effect-of-residual-and | null | null | https://openreview.net/forum?id=rkzeXBDos7 | https://openreview.net/pdf?id=rkzeXBDos7 | Investigating the effect of residual and highway connections in speech enhancement models | Residual and skip connections play an important role in many current
generative models. Although their theoretical and numerical advantages
are understood, their role in speech enhancement systems has not been
investigated so far. When performing spectral speech enhancement,
residual connections are very similar in nature to spectral subtraction,
which is the one of the most commonly employed speech enhancement approaches.
Highway networks, on the other hand, can be seen as a combination of spectral
masking and spectral subtraction. However, when using deep neural networks, such operations would
normally happen in a transformed spectral domain, as opposed to traditional speech
enhancement where all operations are often done directly on the spectrum.
In this paper, we aim to investigate the role of residual and highway
connections in deep neural networks for speech enhancement, and verify whether
or not they operate similarly to their traditional, digital signal processing
counterparts. We visualize the outputs of such connections, projected back to
the spectral domain, in models trained for speech denoising, and show that while
skip connections do not necessarily improve performance with regards to the
number of parameters, they make speech enhancement models more interpretable. | ['Anonymous'] | 2018-10-22 | null | null | null | nips-workshop-irasl-2018 | ['speech-denoising'] | ['speech'] | [ 5.70671499e-01 1.14605926e-01 6.17946386e-01 -1.78348169e-01
-4.13446240e-02 -3.56734723e-01 7.58913636e-01 -4.17817421e-02
-5.81291318e-01 5.83014011e-01 3.34858328e-01 -4.59372491e-01
-1.68399036e-01 -7.64846325e-01 -6.26975834e-01 -1.04946947e+00
1.03685603e-01 -2.42442861e-01 1.48832694e-01 -5.93412936e-01
-2.51538187e-01 5.04164755e-01 -1.65619969e+00 -2.20531374e-02
8.15670490e-01 6.83098733e-01 1.13075450e-01 7.50765085e-01
-2.85291597e-02 5.10889709e-01 -9.67677832e-01 -3.73725504e-01
1.12705745e-01 -6.48004591e-01 -3.95635486e-01 2.31107965e-01
1.35462537e-01 -2.41657346e-01 -3.87902111e-01 1.30190992e+00
6.65572643e-01 2.90274829e-01 6.05691910e-01 -7.67193615e-01
-6.30660534e-01 7.00753272e-01 -3.64304781e-01 2.80217707e-01
1.87557846e-01 -5.75267896e-02 6.89857244e-01 -8.13024342e-01
4.33288902e-01 1.18649280e+00 9.28627849e-01 2.81200439e-01
-1.49548709e+00 -3.23467016e-01 -1.40335232e-01 6.73861504e-02
-9.40713704e-01 -6.25096500e-01 8.88024330e-01 -1.43438965e-01
8.37666571e-01 2.69769251e-01 6.37079775e-01 1.02836692e+00
-4.19478538e-03 5.47246635e-01 1.07910919e+00 -5.90561390e-01
1.64093837e-01 2.23116562e-01 -2.09725276e-02 1.59263641e-01
9.70802549e-03 3.17062348e-01 -4.90420610e-01 2.98491865e-01
5.11014283e-01 -2.08833724e-01 -7.23344386e-01 -2.57406712e-01
-9.42429304e-01 6.19495630e-01 4.56192315e-01 5.93647361e-01
-6.15894973e-01 1.75028533e-01 3.20675939e-01 5.20958185e-01
6.10202789e-01 3.12276214e-01 -1.36201099e-01 -2.45531704e-02
-1.14841759e+00 8.13052058e-02 5.17479062e-01 2.28775412e-01
5.25985062e-01 6.60422087e-01 7.81918094e-02 9.18954253e-01
1.38109133e-01 3.87430102e-01 4.84989703e-01 -8.43252897e-01
-6.13586828e-02 2.83579051e-01 -2.05870524e-01 -8.24543059e-01
-4.21074361e-01 -9.35313046e-01 -1.14252532e+00 6.69953644e-01
5.01918316e-01 -2.60592639e-01 -9.76503432e-01 1.78509521e+00
2.74557453e-02 1.91772804e-01 3.56178731e-01 9.53102946e-01
6.53328717e-01 7.47371256e-01 1.39718667e-01 -2.39105567e-01
1.00990415e+00 -5.06963909e-01 -1.00826776e+00 -2.32998922e-01
1.94228575e-01 -1.05845857e+00 9.29795086e-01 5.62268496e-01
-1.32927215e+00 -6.56016946e-01 -1.19809210e+00 -1.31104365e-01
-6.38165116e-01 1.07461661e-01 3.64816844e-01 1.10921633e+00
-1.52358627e+00 8.98208141e-01 -6.70004129e-01 -2.42931247e-01
1.77618600e-02 2.78030843e-01 -2.88422316e-01 3.20443094e-01
-1.54797709e+00 1.04592359e+00 3.05937380e-01 3.79211187e-01
-5.17718375e-01 -5.93851447e-01 -8.09160531e-01 4.73898113e-01
1.65435210e-01 -3.77964467e-01 1.23465705e+00 -1.05421734e+00
-1.48512900e+00 5.98646164e-01 -7.49738142e-02 -7.72295833e-01
5.01053751e-01 -6.11998998e-02 -6.71122730e-01 5.07336902e-03
-5.45350194e-01 5.63050866e-01 1.15043986e+00 -1.15568483e+00
-2.68810689e-01 -3.09645943e-02 1.01008952e-01 7.08994269e-02
-4.76224750e-01 1.15684178e-02 9.12408158e-02 -9.03433025e-01
2.97400683e-01 -5.76725364e-01 -8.51352215e-02 3.22555974e-02
-1.38071388e-01 2.80686855e-01 1.04988158e+00 -1.14884007e+00
1.08625281e+00 -2.48142838e+00 1.70639396e-01 2.12545007e-01
3.09912264e-02 6.71396673e-01 -6.88945353e-02 4.38809603e-01
-6.41295135e-01 1.13462500e-01 -6.36567056e-01 -5.74974000e-01
1.69154078e-01 6.69451356e-02 -2.73002595e-01 4.22446251e-01
3.54075789e-01 5.79268992e-01 -6.05799735e-01 1.13281198e-01
5.08399010e-01 1.15078259e+00 -3.56553435e-01 -3.38783413e-01
9.29271430e-02 1.88544974e-01 3.02314073e-01 -6.70778938e-03
7.47993708e-01 2.80260056e-01 1.66860640e-01 -1.10337012e-01
-1.87843099e-01 4.86288875e-01 -1.29201150e+00 1.28613627e+00
-5.93365788e-01 1.14168680e+00 4.74711984e-01 -1.20849669e+00
8.36193025e-01 7.04713821e-01 7.31815398e-02 -8.56334329e-01
5.81311509e-02 2.66171485e-01 4.39997822e-01 -1.72692668e-02
7.13965952e-01 -4.89525080e-01 4.95342404e-01 3.95783603e-01
-1.60754509e-02 -7.08688721e-02 3.32123846e-01 6.94984198e-02
8.36844444e-01 -4.89219390e-02 2.35333532e-01 -6.32926077e-02
5.28799951e-01 -4.72665727e-01 2.28138804e-01 5.04555047e-01
6.36460707e-02 8.06889713e-01 5.63267529e-01 7.82927647e-02
-1.09232843e+00 -1.10351276e+00 -1.49825230e-01 9.25361931e-01
-1.88014790e-01 -1.40130848e-01 -1.07306969e+00 -5.27883209e-02
-4.34051812e-01 7.45965660e-01 -3.34810615e-01 -4.67592001e-01
-3.63072157e-01 -7.88015544e-01 5.90981722e-01 3.82100224e-01
6.04080498e-01 -1.12767506e+00 -5.93976378e-01 4.69808757e-01
-1.69029236e-02 -7.90167689e-01 -2.31539413e-01 5.76533854e-01
-9.07793105e-01 -6.44482553e-01 -1.10459316e+00 -5.21602333e-01
3.71544689e-01 3.60729843e-01 8.49558234e-01 6.22859783e-02
-7.35399872e-02 1.72019660e-01 -3.05171549e-01 -4.83954072e-01
-8.33317339e-01 -1.67986229e-01 1.39465099e-02 9.23911408e-02
9.19717997e-02 -1.06395805e+00 -4.70030874e-01 1.37399614e-01
-1.31777382e+00 -1.53634995e-01 7.53110111e-01 7.64342010e-01
5.59067205e-02 4.73734558e-01 6.52864754e-01 -6.86491072e-01
1.11947858e+00 -1.68746840e-02 -3.84481311e-01 -1.40377292e-02
-3.89314383e-01 1.40200227e-01 6.80426002e-01 -4.25644636e-01
-1.34841263e+00 -1.18546098e-01 -6.92929626e-01 -3.20718557e-01
-4.34549809e-01 6.82461202e-01 -1.51381761e-01 -2.55750790e-02
7.73673534e-01 3.85129571e-01 6.44752532e-02 -6.53907120e-01
2.57034689e-01 5.65744817e-01 6.23585761e-01 -5.23589849e-02
9.84927058e-01 3.45911145e-01 -4.23639342e-02 -1.27421367e+00
-1.84999615e-01 -3.70056272e-01 -3.84500891e-01 -2.34108061e-01
7.78311729e-01 -5.57141185e-01 -3.22859526e-01 6.61892653e-01
-1.19789410e+00 -1.75052091e-01 -5.91722548e-01 4.14924085e-01
-3.13071519e-01 4.36239421e-01 -6.25243783e-01 -1.02921176e+00
-9.35471132e-02 -1.11155188e+00 7.40329504e-01 2.81889230e-01
-8.76576155e-02 -1.23295021e+00 -2.65081853e-01 -6.03230335e-02
7.92293370e-01 -1.59609556e-01 1.04972959e+00 -4.25019771e-01
-1.20501690e-01 -3.29625160e-02 -8.64575654e-02 8.39285195e-01
1.35693535e-01 5.42712770e-02 -1.33342969e+00 -1.48635119e-01
3.63702595e-01 2.30878815e-01 1.17475402e+00 7.22876430e-01
7.52044022e-01 4.25966047e-02 2.96041191e-01 4.45888370e-01
1.12690008e+00 4.17713314e-01 1.10964644e+00 9.68184546e-02
5.25774099e-02 8.00972581e-01 9.22123790e-02 8.37809518e-02
-4.71947491e-01 5.60541928e-01 5.23310959e-01 -6.38749540e-01
-5.86360872e-01 6.26120716e-02 4.88906920e-01 7.17274964e-01
-2.64228314e-01 -2.35695943e-01 -5.85079730e-01 3.97884279e-01
-1.33564973e+00 -1.13319409e+00 -2.45426819e-01 2.21644902e+00
6.41581059e-01 3.64157677e-01 -4.76994639e-04 9.19791639e-01
7.81494260e-01 3.64777058e-01 -2.21904740e-01 -6.15300894e-01
-5.98822594e-01 6.69025481e-01 2.93707997e-01 4.37293828e-01
-9.45440650e-01 5.28400600e-01 6.31871033e+00 7.67651320e-01
-1.31415927e+00 1.48330390e-01 5.09328008e-01 4.96304967e-02
-2.50501394e-01 -1.89448282e-01 -7.16837868e-02 1.79005995e-01
9.37071621e-01 9.55429003e-02 5.40793955e-01 4.72260416e-01
5.73738277e-01 -7.85899311e-02 -7.76765466e-01 7.85579860e-01
-2.00898111e-01 -1.10614085e+00 -2.37754360e-01 1.06065452e-01
5.27711093e-01 -4.35065240e-01 3.51111919e-01 1.86363444e-01
-7.31637999e-02 -1.12329853e+00 8.78694773e-01 3.32182944e-01
3.91069680e-01 -7.82324553e-01 7.90483952e-01 1.76390573e-01
-9.16063011e-01 9.48415548e-02 -2.78282583e-01 -1.21062733e-01
4.34321970e-01 8.41876149e-01 -6.35859132e-01 3.81326854e-01
6.65628493e-01 3.79205495e-01 -3.58762085e-01 1.04043663e+00
-3.87523741e-01 8.36878300e-01 -2.27055743e-01 2.27479309e-01
1.99747935e-01 -4.24117833e-01 7.70006180e-01 1.28707278e+00
4.50959802e-01 -2.32925445e-01 -6.56250715e-01 1.05973196e+00
-1.33086825e-02 -1.25074908e-01 -4.60602880e-01 -3.05806845e-01
6.02965951e-02 1.06574821e+00 -8.49467874e-01 -2.95070857e-02
-4.81638312e-01 8.69204640e-01 -4.18820560e-01 6.98105216e-01
-7.72344589e-01 -5.02149940e-01 8.08803678e-01 2.65546352e-01
2.54913867e-01 -1.48187429e-01 -9.39016640e-02 -6.15773737e-01
1.34183526e-01 -8.34462345e-01 -2.93812156e-01 -9.51733232e-01
-8.56385648e-01 4.46106821e-01 -3.41004223e-01 -1.10959339e+00
-2.58165538e-01 -8.04732859e-01 -6.78606868e-01 1.19486594e+00
-1.51849222e+00 -7.01590061e-01 -4.73307334e-02 5.35233557e-01
4.35462981e-01 1.36473536e-01 6.65249050e-01 5.30084848e-01
-3.72361660e-01 2.10390344e-01 3.24649483e-01 -2.83603929e-02
5.56291640e-01 -1.33587325e+00 5.56741178e-01 1.32152188e+00
2.60158479e-01 7.78032362e-01 1.21798682e+00 -3.71169537e-01
-8.04272473e-01 -7.52076447e-01 6.28812432e-01 6.47370577e-01
4.98082727e-01 -2.25349575e-01 -1.25793242e+00 2.72319406e-01
5.21010458e-01 -4.53519046e-01 2.71541148e-01 1.55099973e-01
-5.06517626e-02 -3.52327563e-02 -7.61701167e-01 7.28962600e-01
9.18100357e-01 -5.86812556e-01 -6.50820971e-01 2.54738778e-02
5.68684042e-01 -2.04286501e-01 -5.73327124e-01 1.97487712e-01
2.64021426e-01 -1.44255853e+00 1.19588852e+00 -2.63324976e-01
4.17030782e-01 -4.18707550e-01 5.85383885e-02 -1.83828008e+00
-1.18894810e-02 -5.75990856e-01 2.83440500e-01 1.30695772e+00
6.55842602e-01 -8.47300529e-01 6.93000257e-01 3.10647413e-02
-4.01953012e-01 -3.01269323e-01 -8.46765041e-01 -8.39485943e-01
3.51531454e-03 -5.78450143e-01 4.56848502e-01 7.48269022e-01
-3.68108213e-01 1.90927356e-01 -5.47333300e-01 1.22356288e-01
2.50154167e-01 -3.20418030e-01 2.81226516e-01 -1.31414342e+00
-3.37310195e-01 -7.45470464e-01 -3.48738939e-01 -8.07641447e-01
4.85009924e-02 -6.71967447e-01 9.23754871e-02 -1.73871922e+00
-3.74270231e-01 -2.77758297e-02 -9.36035216e-02 2.50937939e-01
-7.94122368e-02 3.37687850e-01 9.89462882e-02 -3.07748914e-01
3.19110036e-01 5.77196956e-01 9.35167372e-01 -1.10506117e-01
-2.79606462e-01 1.00692391e-01 -5.81449866e-01 8.10976386e-01
8.93646240e-01 -1.31696627e-01 -5.47313392e-01 -3.44318926e-01
2.69692153e-01 6.37951121e-02 5.70359468e-01 -1.19240987e+00
3.56563240e-01 3.17628384e-01 3.25109005e-01 -2.88715005e-01
7.64858186e-01 -8.18425655e-01 4.15661752e-01 4.14599329e-01
-2.37772524e-01 -1.75761700e-01 3.30453813e-01 3.22748572e-01
-6.82836890e-01 -4.66745257e-01 8.49362135e-01 -2.84911096e-02
-6.74898505e-01 -2.77051777e-01 -8.49296153e-01 -4.58825409e-01
4.76043344e-01 -6.54175460e-01 -1.45349428e-01 -8.72089148e-01
-1.09033632e+00 -4.93322402e-01 2.66639739e-01 9.92311314e-02
6.44278288e-01 -8.40842664e-01 -5.96953452e-01 1.88443556e-01
-5.37438512e-01 -2.70874590e-01 5.45969844e-01 1.09094119e+00
-4.96101886e-01 2.59285569e-01 -2.19086513e-01 -4.08751130e-01
-1.39954305e+00 4.97350574e-01 6.45082831e-01 -6.92993328e-02
-5.28157055e-01 6.39939487e-01 2.83991784e-01 -3.02091807e-01
2.62361616e-01 -2.01820523e-01 -2.51379043e-01 2.07392812e-01
4.00244236e-01 3.76685381e-01 5.48079312e-01 -5.20180941e-01
3.93876024e-02 3.10402513e-01 3.38657498e-01 -3.90838921e-01
1.43087232e+00 -6.39217347e-02 -1.55440390e-01 1.99618056e-01
9.19091523e-01 1.52910173e-01 -1.08871186e+00 -9.86919329e-02
-1.12788871e-01 -9.13258120e-02 3.02850634e-01 -7.45402515e-01
-1.35922718e+00 1.44581783e+00 7.97595799e-01 7.33117461e-01
1.67289472e+00 -4.04142022e-01 4.54850882e-01 5.26335835e-02
-2.23683417e-01 -1.08523810e+00 -9.46091786e-02 4.90938365e-01
9.01301622e-01 -8.13916445e-01 -2.87048280e-01 -5.45964658e-01
-2.49227136e-01 1.17489672e+00 7.75690749e-02 8.26199353e-03
4.65004176e-01 6.42020285e-01 3.47126313e-02 1.71996966e-01
-4.01306301e-01 -4.86707807e-01 1.28906354e-01 7.05712616e-01
7.73614585e-01 -1.97412685e-01 -3.56706828e-01 1.48801431e-01
-4.47177470e-01 -3.11949790e-01 6.31322980e-01 6.53480589e-01
-5.19059956e-01 -1.17981112e+00 -7.38261640e-01 1.12499103e-01
-3.74218315e-01 -3.65290940e-01 -4.70409811e-01 7.87644148e-01
1.59589559e-01 1.12806344e+00 -5.04753068e-02 -2.47392222e-01
4.80797678e-01 3.91055733e-01 2.21735671e-01 -2.48588666e-01
-7.86884665e-01 3.21452945e-01 1.16629004e-01 -8.52217376e-02
-5.57010174e-01 -5.70522726e-01 -1.09989202e+00 -4.07541931e-01
-3.14573079e-01 -1.91112515e-02 9.90264058e-01 9.07837808e-01
-1.02570653e-01 1.17097425e+00 1.51470333e-01 -9.14573014e-01
-3.95996541e-01 -1.07369828e+00 -7.58113444e-01 4.35553372e-01
4.56047446e-01 -4.68636513e-01 -5.78246593e-01 2.08728164e-01] | [15.086625099182129, 5.949264049530029] |
62f23506-1d54-4f54-aeb8-b51bff5a26c5 | modeling-spatial-and-temporal-cues-for-multi | 1608.00911 | null | http://arxiv.org/abs/1608.00911v1 | http://arxiv.org/pdf/1608.00911v1.pdf | Modeling Spatial and Temporal Cues for Multi-label Facial Action Unit Detection | Facial action units (AUs) are essential to decode human facial expressions.
Researchers have focused on training AU detectors with a variety of features
and classifiers. However, several issues remain. These are spatial
representation, temporal modeling, and AU correlation. Unlike most studies that
tackle these issues separately, we propose a hybrid network architecture to
jointly address them. Specifically, spatial representations are extracted by a
Convolutional Neural Network (CNN), which, as analyzed in this paper, is able
to reduce person-specific biases caused by hand-crafted features (eg, SIFT and
Gabor). To model temporal dependencies, Long Short-Term Memory (LSTMs) are
stacked on top of these representations, regardless of the lengths of input
videos. The outputs of CNNs and LSTMs are further aggregated into a fusion
network to produce per-frame predictions of 12 AUs. Our network naturally
addresses the three issues, and leads to superior performance compared to
existing methods that consider these issues independently. Extensive
experiments were conducted on two large spontaneous datasets, GFT and BP4D,
containing more than 400,000 frames coded with 12 AUs. On both datasets, we
report significant improvement over a standard multi-label CNN and
feature-based state-of-the-art. Finally, we provide visualization of the
learned AU models, which, to our best knowledge, reveal how machines see facial
AUs for the first time. | ['Fernando de la Torre', 'Wen-Sheng Chu', 'Jeffrey F. Cohn'] | 2016-08-02 | null | null | null | null | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 1.76417455e-01 -1.20203823e-01 -2.05786228e-01 -4.93706524e-01
-6.13852262e-01 -1.22430377e-01 6.07922673e-01 -3.40723068e-01
-3.94729048e-01 4.45482969e-01 3.10053259e-01 2.54900783e-01
3.47671986e-01 -6.46502793e-01 -8.09971631e-01 -6.43431127e-01
-2.21758321e-01 -1.91331670e-01 6.33109175e-03 -1.03359818e-01
-1.32298216e-01 5.39893448e-01 -1.76396811e+00 6.05545342e-01
3.19994658e-01 1.54583597e+00 -2.24482641e-01 4.12207812e-01
1.01679116e-01 1.26706362e+00 -4.47465837e-01 -4.29485142e-01
3.15062553e-02 -3.49787712e-01 -7.45236516e-01 1.24334842e-01
5.41609347e-01 -6.89009607e-01 -6.87029064e-01 7.96131313e-01
3.84654135e-01 2.03547716e-01 6.63592756e-01 -1.38718116e+00
-8.27835560e-01 9.50653851e-02 -6.74897492e-01 -2.82241199e-02
3.86525542e-01 1.62038103e-01 9.08687055e-01 -9.65999067e-01
5.55563033e-01 1.53284800e+00 5.62524021e-01 8.26685905e-01
-9.00277972e-01 -8.37202847e-01 1.75912306e-01 3.39056104e-01
-1.39782929e+00 -6.25171185e-01 6.41237855e-01 -4.21646655e-01
1.13190126e+00 -1.45613283e-01 5.35008252e-01 1.65624583e+00
1.41472891e-01 1.12472677e+00 1.03420258e+00 -4.00184542e-01
-1.86242133e-01 -8.66755322e-02 -1.64907917e-01 9.18216825e-01
-4.43900764e-01 3.30844104e-01 -6.18230283e-01 3.62873562e-02
8.92181396e-01 1.03248283e-01 -1.70790330e-01 4.68493849e-02
-1.03005791e+00 7.37294674e-01 6.74756646e-01 3.53199482e-01
-4.99297321e-01 4.11638796e-01 4.64977503e-01 1.54183775e-01
7.10812986e-01 -7.85894021e-02 -3.36154163e-01 -1.56504318e-01
-7.41932213e-01 7.85214156e-02 3.86113286e-01 8.72503519e-01
7.87577987e-01 6.07659295e-02 -4.41682488e-01 9.64974284e-01
2.68378526e-01 4.71173823e-01 3.73838782e-01 -9.11985993e-01
2.12325528e-01 4.59085375e-01 -1.23555437e-01 -1.17094946e+00
-5.10146737e-01 -1.02902181e-01 -7.73844302e-01 2.24182442e-01
3.72406274e-01 -2.65133291e-01 -9.94485855e-01 2.10172939e+00
3.06005334e-03 5.85065305e-01 -8.03799406e-02 1.02087343e+00
9.87312496e-01 5.84859550e-01 3.79528105e-01 4.22169343e-02
1.30448484e+00 -1.25499058e+00 -8.03853631e-01 -1.51847333e-01
7.45719433e-01 -5.32693684e-01 6.21741652e-01 2.19225943e-01
-9.93332624e-01 -8.35654020e-01 -7.44772673e-01 -1.34544045e-01
-4.42678005e-01 5.61337590e-01 6.29222512e-01 2.51506805e-01
-1.20717943e+00 7.11232781e-01 -7.39989877e-01 -5.30264556e-01
6.59900606e-01 3.28783512e-01 -6.45193279e-01 3.43912430e-02
-1.46066558e+00 1.01185846e+00 -6.79816678e-02 4.31744963e-01
-1.01790607e+00 -2.30193928e-01 -1.02752280e+00 -1.05936222e-01
1.15028664e-01 -3.59310687e-01 1.33848584e+00 -1.37865901e+00
-1.55933130e+00 8.63107979e-01 -5.56827366e-01 -3.47975492e-01
3.10400248e-01 -2.42235944e-01 -4.93462294e-01 2.84547955e-01
-6.16720393e-02 1.11464226e+00 9.95345771e-01 -9.57981288e-01
-5.36550522e-01 -2.75465548e-01 1.88471600e-01 2.20534578e-02
-5.73391914e-01 4.20036346e-01 -5.08136153e-01 -5.50192058e-01
-2.55566925e-01 -1.04947925e+00 -1.21642584e-02 3.15153182e-01
-1.22899516e-02 -4.61102247e-01 8.12749147e-01 -7.09438026e-01
1.09221292e+00 -2.34858775e+00 2.64011681e-01 -1.90897018e-01
9.14337486e-02 2.85285234e-01 -5.50390899e-01 1.98629081e-01
-1.65254399e-01 -2.57262178e-02 2.06533551e-01 -8.14963460e-01
-6.99860752e-02 2.88472652e-01 -2.49922186e-01 5.46987593e-01
5.84629774e-01 1.18780470e+00 -7.81882823e-01 -4.76004928e-01
2.19111815e-01 7.00837731e-01 -2.41778687e-01 1.94324180e-01
-1.57858521e-01 4.01743650e-01 -2.55984545e-01 8.96087408e-01
4.87241328e-01 -1.84704617e-01 -1.18121900e-01 -4.17375743e-01
6.81798160e-02 5.21635562e-02 -6.11614943e-01 1.75420213e+00
-5.10249674e-01 6.89120173e-01 -2.06563205e-01 -8.92847955e-01
9.59246755e-01 6.16049170e-01 5.81579089e-01 -9.47098613e-01
4.44727123e-01 9.18935463e-02 -2.29395524e-01 -5.11873424e-01
2.02446446e-01 -2.10324712e-02 1.21575542e-01 3.59220773e-01
5.08407831e-01 5.93747258e-01 3.21029425e-02 -1.11134335e-01
9.88420784e-01 5.54820418e-01 1.57343477e-01 2.09933698e-01
5.00830591e-01 -4.31275666e-01 5.96159160e-01 4.19107765e-01
-5.84118843e-01 7.10466266e-01 6.29845142e-01 -7.26102352e-01
-7.20864952e-01 -6.55978680e-01 -5.04098414e-03 1.32899046e+00
-3.44128944e-02 -5.14245510e-01 -6.29350662e-01 -7.86121964e-01
2.39134692e-02 2.57248700e-01 -1.17775524e+00 -2.11875930e-01
-5.23312688e-01 -4.74818259e-01 7.69906282e-01 8.08003485e-01
6.06857002e-01 -1.43061948e+00 -7.04266131e-01 1.79496661e-01
-1.47557512e-01 -1.46639156e+00 -2.61586130e-01 -1.55437989e-02
-4.88310367e-01 -1.07406723e+00 -7.79402971e-01 -6.18614137e-01
5.17688930e-01 3.71814787e-01 8.17627549e-01 1.19092025e-01
-2.89779603e-01 4.10519302e-01 -5.49686551e-01 -3.86391848e-01
8.78636837e-02 -2.15994418e-01 8.50229785e-02 5.27743101e-01
6.27924204e-01 -5.53174675e-01 -4.13278401e-01 2.59731084e-01
-6.74981475e-01 1.20827295e-01 8.03786874e-01 9.27053452e-01
3.94323438e-01 -4.62954313e-01 2.55985260e-01 -4.52710301e-01
3.38661224e-01 -4.15258706e-01 -2.57651806e-01 3.10951084e-01
-4.36834134e-02 -2.21231028e-01 4.13479418e-01 -5.54383516e-01
-1.06960440e+00 1.67322457e-01 -1.53107613e-01 -1.05111182e+00
-3.84423554e-01 3.20167840e-01 1.25154138e-01 -3.59440982e-01
4.25549448e-01 1.29503638e-01 2.11370483e-01 -2.80838370e-01
3.58483016e-01 5.86896062e-01 2.76318818e-01 -5.05430937e-01
2.73128986e-01 5.60769737e-01 6.64779246e-02 -8.11819553e-01
-1.03622079e+00 -1.87387347e-01 -9.05083358e-01 -4.76941973e-01
1.06171942e+00 -1.11207521e+00 -8.03141177e-01 8.27706635e-01
-1.44352138e+00 -4.00674671e-01 6.48289844e-02 4.57770884e-01
-6.07314646e-01 -2.45556366e-02 -9.58188295e-01 -9.03864801e-01
-1.70407817e-01 -1.07103097e+00 1.35111094e+00 2.70775408e-01
-2.91896403e-01 -8.88762712e-01 -1.79412767e-01 1.03691205e-01
3.85543376e-01 4.84522492e-01 3.37231338e-01 -2.51705617e-01
-2.99860299e-01 -1.45366564e-01 -5.49363554e-01 4.56431597e-01
1.78740323e-01 1.21485539e-01 -1.41865230e+00 -1.49795637e-01
-2.46704325e-01 -7.94992626e-01 1.19614220e+00 4.71749693e-01
1.28618670e+00 -2.77027428e-01 -4.31399584e-01 7.42734551e-01
9.84031379e-01 1.76835954e-01 6.76506758e-01 2.00536430e-01
7.14499176e-01 7.68083453e-01 5.82159758e-01 5.09537995e-01
3.97470951e-01 9.01981413e-01 6.36145890e-01 -2.32919395e-01
-2.71785855e-01 -2.23505452e-01 5.31856179e-01 3.01581591e-01
-4.98902023e-01 -5.05967662e-02 -6.55203640e-01 4.57158417e-01
-1.85228324e+00 -1.15006757e+00 2.24437475e-01 1.59762478e+00
4.70124543e-01 -2.15452090e-01 1.66443169e-01 -2.11277723e-01
4.52463835e-01 5.70433915e-01 -3.85080457e-01 -3.55433553e-01
-2.98511326e-01 2.48279184e-01 2.39148185e-01 2.41565019e-01
-1.48222041e+00 1.28008091e+00 6.44207573e+00 5.97084105e-01
-1.42487657e+00 1.99760795e-01 6.89230621e-01 -2.42314056e-01
1.64654717e-01 -2.97278315e-01 -7.71945953e-01 3.80192846e-01
9.99525487e-01 2.74445832e-01 3.48961771e-01 9.44715738e-01
1.44497007e-01 5.22726960e-03 -1.04202282e+00 1.22189665e+00
4.22402322e-01 -1.09262097e+00 9.97245591e-03 1.16601832e-01
6.26974046e-01 3.55066061e-02 1.98202729e-01 4.01832521e-01
4.48046438e-03 -1.33097351e+00 9.16557968e-01 7.53572226e-01
8.06513965e-01 -6.25090480e-01 8.09090316e-01 3.49535942e-02
-1.31920314e+00 -1.54118389e-01 -3.78483087e-01 -2.87980705e-01
2.18579471e-01 1.46771222e-01 -2.79123694e-01 3.86817515e-01
8.95988584e-01 1.27394080e+00 -5.70368409e-01 5.16803265e-01
-4.09414202e-01 3.53825420e-01 -6.38990477e-02 -1.64278746e-02
4.83503073e-01 -4.39379811e-02 3.07928189e-04 1.22340238e+00
3.75516802e-01 2.20880806e-01 -1.07765228e-01 7.48765826e-01
-1.13822119e-02 -4.57494408e-02 -7.35215127e-01 -1.45669222e-01
9.12644118e-02 1.41521978e+00 -2.37271264e-01 -2.12988943e-01
-8.89174581e-01 1.18403327e+00 7.30998814e-01 4.63970900e-01
-1.00994802e+00 -2.20866594e-02 9.72348273e-01 -1.47077665e-01
2.97290951e-01 -2.89562464e-01 1.37305588e-01 -1.09131980e+00
-3.37412842e-02 -7.25507081e-01 2.93229759e-01 -8.49531233e-01
-1.27738762e+00 8.75186801e-01 -1.74473494e-01 -1.26433349e+00
-4.15705711e-01 -9.04004931e-01 -4.77516323e-01 7.02388644e-01
-1.66291392e+00 -1.58193660e+00 -4.36942846e-01 7.91822672e-01
3.68395686e-01 -6.66237324e-02 1.08265567e+00 4.80363667e-01
-6.57457113e-01 7.32267141e-01 -4.25920486e-01 5.79315901e-01
7.56374598e-01 -7.02519357e-01 2.40853250e-01 6.22342110e-01
2.69847065e-01 4.92426634e-01 1.37750074e-01 -3.73766869e-01
-1.08549917e+00 -1.24757218e+00 1.00524271e+00 -3.72055978e-01
6.18856311e-01 -3.62121701e-01 -8.04730177e-01 1.00315189e+00
1.13160320e-01 6.39503539e-01 6.92183316e-01 1.42245203e-01
-5.49867868e-01 -4.15305421e-02 -7.21978784e-01 4.10742193e-01
1.30436087e+00 -6.88221514e-01 -1.99479297e-01 1.71313852e-01
3.68322939e-01 -5.74433148e-01 -7.61068046e-01 5.19692242e-01
9.37111318e-01 -1.26090884e+00 8.01816702e-01 -7.64659703e-01
6.13403618e-01 6.00989200e-02 -2.32309848e-01 -1.27399945e+00
-3.57677788e-01 -2.61851668e-01 -3.38328272e-01 1.02378917e+00
1.78864136e-01 -4.30413753e-01 5.42594373e-01 4.77458686e-01
-3.80224586e-02 -9.96494293e-01 -9.38450694e-01 -6.77157581e-01
-1.50951564e-01 -4.24089819e-01 4.70040649e-01 7.55249023e-01
-1.35443687e-01 3.08210492e-01 -9.44909275e-01 -8.89386535e-02
3.88941437e-01 -1.11450283e-02 6.50892377e-01 -1.05086684e+00
5.51921017e-02 -4.27579194e-01 -5.91531992e-01 -1.11466253e+00
6.43621504e-01 -5.69836497e-01 -2.80210599e-02 -1.28142941e+00
1.01532891e-01 -1.53926864e-01 -3.76856744e-01 1.02025044e+00
3.19689624e-02 4.71214533e-01 2.72593588e-01 1.03573926e-01
-6.51771069e-01 8.65880013e-01 1.25712943e+00 -1.89928293e-01
1.05310269e-02 -2.30990216e-01 -3.91367674e-01 7.69623458e-01
7.29227543e-01 -1.47084415e-01 -8.45038816e-02 -6.68845534e-01
-2.21875668e-01 5.68296835e-02 6.36522532e-01 -1.07580388e+00
1.86911628e-01 -1.21872522e-01 6.62880003e-01 -3.58040750e-01
7.26870418e-01 -6.56710029e-01 -1.05754390e-01 2.15154767e-01
-2.91766226e-01 -1.70330420e-01 3.07445228e-01 3.22636127e-01
-6.02939188e-01 2.74020642e-01 8.33605647e-01 -1.48097068e-01
-1.01249802e+00 7.38613129e-01 -2.44586229e-01 -5.06874144e-01
1.23338282e+00 -1.38234481e-01 -2.43629888e-01 -6.50547862e-01
-7.21928537e-01 1.10900551e-01 5.70725687e-02 7.92157590e-01
6.78959608e-01 -1.69533253e+00 -6.44456625e-01 3.58935386e-01
2.02801183e-01 -3.64905536e-01 4.73024428e-01 1.09275472e+00
-1.06156766e-01 6.60315812e-01 -4.42196131e-01 -5.52377164e-01
-1.33802867e+00 4.91945684e-01 4.35314566e-01 7.58987479e-03
-4.49999958e-01 1.06368589e+00 2.87081629e-01 -1.03702001e-01
3.17291319e-01 -1.78920031e-01 -3.89544934e-01 3.93367618e-01
7.00037837e-01 3.74049954e-02 -7.18535557e-02 -1.21507585e+00
-4.05298322e-01 7.71308243e-01 4.56796661e-02 3.32784466e-02
1.35049582e+00 -5.52822724e-02 -1.11887380e-01 2.54959762e-01
1.51730728e+00 -4.93154943e-01 -1.58658004e+00 -5.13406336e-01
-2.71317691e-01 -4.13736373e-01 -2.78691463e-02 -4.22093391e-01
-1.32391500e+00 1.22258830e+00 5.19565761e-01 -2.42687017e-01
1.35394466e+00 5.81617989e-02 7.35183418e-01 1.27302393e-01
3.62707824e-01 -8.66577268e-01 4.07702267e-01 6.69230402e-01
9.82498765e-01 -1.37540388e+00 -3.79744440e-01 -1.68585822e-01
-7.53806472e-01 1.32226682e+00 9.84509885e-01 -1.58084705e-01
5.98835170e-01 1.46066293e-01 1.66317403e-01 -1.00767456e-01
-9.15565908e-01 -5.82229316e-01 3.85949582e-01 4.86690223e-01
5.26509762e-01 -1.21761836e-01 -3.34158493e-03 6.24082804e-01
1.68102726e-01 2.85065979e-01 -4.26075347e-02 8.51687491e-01
-2.22144708e-01 -9.60628211e-01 -1.34681627e-01 2.64722139e-01
-3.64606202e-01 4.51894589e-02 -2.72940516e-01 7.37970889e-01
3.69441211e-01 8.47172201e-01 1.97589129e-01 -7.28174508e-01
2.87353367e-01 5.39531820e-02 6.90802753e-01 -3.74573886e-01
-2.92820901e-01 -4.36484590e-02 3.12761194e-03 -1.03305221e+00
-7.56461561e-01 -6.97589397e-01 -9.77210641e-01 -1.51449516e-01
3.76807638e-02 -2.63913125e-01 3.86444628e-01 1.04819906e+00
3.20369691e-01 4.57853049e-01 6.21391535e-01 -1.15014148e+00
-3.05207282e-01 -1.21100736e+00 -5.72373509e-01 5.94781816e-01
4.02159929e-01 -1.06411064e+00 -3.14186990e-01 -8.70143436e-03] | [13.558670997619629, 1.6083437204360962] |
b87c951d-d28d-44a6-ba30-f1381c829b1a | mask2cad-3d-shape-prediction-by-learning-to | 2007.13034 | null | https://arxiv.org/abs/2007.13034v1 | https://arxiv.org/pdf/2007.13034v1.pdf | Mask2CAD: 3D Shape Prediction by Learning to Segment and Retrieve | Object recognition has seen significant progress in the image domain, with focus primarily on 2D perception. We propose to leverage existing large-scale datasets of 3D models to understand the underlying 3D structure of objects seen in an image by constructing a CAD-based representation of the objects and their poses. We present Mask2CAD, which jointly detects objects in real-world images and for each detected object, optimizes for the most similar CAD model and its pose. We construct a joint embedding space between the detected regions of an image corresponding to an object and 3D CAD models, enabling retrieval of CAD models for an input RGB image. This produces a clean, lightweight representation of the objects in an image; this CAD-based representation ensures a valid, efficient shape representation for applications such as content creation or interactive scenarios, and makes a step towards understanding the transformation of real-world imagery to a synthetic domain. Experiments on real-world images from Pix3D demonstrate the advantage of our approach in comparison to state of the art. To facilitate future research, we additionally propose a new image-to-3D baseline on ScanNet which features larger shape diversity, real-world occlusions, and challenging image views. | ['Wei-cheng Kuo', 'Tsung-Yi Lin', 'Anelia Angelova', 'Angela Dai'] | 2020-07-26 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/193_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123480273.pdf | eccv-2020-8 | ['image-to-3d'] | ['computer-vision'] | [ 3.26827317e-01 4.82021645e-03 5.56370132e-02 -5.50663471e-01
-6.72740698e-01 -8.00388157e-01 7.54065990e-01 -1.70690879e-01
-2.84522027e-02 -3.13430607e-01 6.30602986e-02 -1.41085073e-01
1.56351447e-01 -8.01734805e-01 -8.70917797e-01 -2.84579039e-01
-5.39041264e-03 9.69081521e-01 5.74447215e-01 -1.22125491e-01
1.44775659e-01 1.08700776e+00 -1.81296146e+00 2.44308710e-01
1.90815516e-03 1.23558319e+00 5.01176000e-01 4.86184359e-01
-5.27619570e-02 2.46580809e-01 -3.32392693e-01 -1.45283595e-01
8.04409802e-01 5.91170155e-02 -5.08337677e-01 6.54575109e-01
1.09903455e+00 -4.34460878e-01 -4.05625761e-01 7.83639312e-01
3.34504455e-01 -5.39793596e-02 7.86939979e-01 -1.20734489e+00
-4.63197589e-01 -2.91545153e-01 -7.09821045e-01 -2.53501147e-01
5.88489652e-01 2.55976409e-01 1.00443912e+00 -1.31750917e+00
9.76557076e-01 1.63132906e+00 5.36710739e-01 1.93657413e-01
-1.57718849e+00 -3.64557892e-01 1.74899727e-01 1.70712862e-02
-1.43575573e+00 -3.43608409e-01 8.63279700e-01 -4.84272063e-01
1.04391456e+00 3.13669562e-01 8.41363013e-01 8.70709300e-01
-7.57826492e-02 8.33193541e-01 8.99133742e-01 -4.56262171e-01
1.47693217e-01 1.23721600e-01 -1.47260845e-01 7.13169456e-01
2.62932539e-01 1.31528705e-01 -5.51822722e-01 -2.53975634e-02
1.11478877e+00 3.06097697e-02 -7.64377117e-02 -1.38202679e+00
-1.28347015e+00 6.21427178e-01 6.86522365e-01 -2.23666340e-01
-4.43118244e-01 2.28896812e-01 -1.97525248e-01 -5.85131086e-02
2.64770627e-01 2.66425669e-01 -2.59686887e-01 9.37623382e-02
-6.74990296e-01 4.06811804e-01 7.32235909e-01 1.10514271e+00
8.73851359e-01 -3.48472334e-02 5.75191498e-01 5.27566016e-01
5.77698827e-01 8.43722105e-01 -1.41052961e-01 -1.27282441e+00
4.23947215e-01 9.92708802e-01 7.13864565e-02 -1.22500670e+00
-3.44447076e-01 -3.35129023e-01 -2.92331249e-01 7.27731109e-01
7.96343163e-02 6.17057502e-01 -9.98557448e-01 1.27873671e+00
6.31825745e-01 -1.15401030e-01 -1.38997808e-01 1.02920735e+00
7.45639503e-01 5.31942725e-01 -5.62604964e-01 6.12061203e-01
1.29252124e+00 -5.96309423e-01 6.67695329e-02 -7.55616248e-01
1.08723976e-01 -8.23373675e-01 8.54647279e-01 3.52985352e-01
-1.09339356e+00 -6.44689083e-01 -1.00425148e+00 -3.15149814e-01
-4.73455727e-01 2.53271699e-01 4.93802905e-01 4.75033909e-01
-9.16318536e-01 1.25398904e-01 -7.84201801e-01 -5.49452305e-01
5.41646719e-01 2.28316292e-01 -7.89774299e-01 -3.95929873e-01
-2.41327614e-01 1.08521223e+00 3.91100526e-01 -5.27860597e-03
-1.01121092e+00 -6.18800044e-01 -1.23611939e+00 -1.58793822e-01
5.37003756e-01 -6.54292703e-01 1.03835797e+00 -5.36000252e-01
-7.99837828e-01 1.33755052e+00 -4.60871346e-02 -1.06718965e-01
3.79856825e-01 -2.12946698e-01 2.60232706e-02 2.55821079e-01
9.16543454e-02 9.81146693e-01 1.01924634e+00 -1.78004158e+00
-4.12596494e-01 -7.45283544e-01 7.42132291e-02 4.40765649e-01
2.67721534e-01 -2.32448503e-01 -7.53403783e-01 -3.53566945e-01
8.74810040e-01 -9.98088896e-01 -2.29984984e-01 8.11993062e-01
-2.45779738e-01 1.12431198e-01 1.15301764e+00 -3.80250841e-01
2.72945225e-01 -2.23453522e+00 2.35753849e-01 2.86329836e-01
5.74687719e-02 -2.56558135e-02 -3.90921742e-01 5.49827456e-01
-1.11139007e-02 -8.20819065e-02 -7.17187002e-02 -5.03226161e-01
1.26178890e-01 4.32430893e-01 -4.97477978e-01 5.92022181e-01
4.93326306e-01 8.86965692e-01 -5.89365244e-01 -2.42848188e-01
6.51070535e-01 6.48544133e-01 -5.46307266e-01 3.78687680e-01
-3.92839372e-01 2.52612770e-01 -3.75360370e-01 9.99870598e-01
9.78574932e-01 -8.00776407e-02 -9.80032757e-02 -6.56572580e-01
-6.96298014e-03 1.05153762e-01 -1.57980859e+00 1.89849663e+00
-3.88014883e-01 5.50093234e-01 1.70164064e-01 -6.90111637e-01
1.05831039e+00 -2.00901404e-01 4.60803002e-01 -5.97408831e-01
8.13829303e-02 1.68875679e-01 -2.98695356e-01 -4.43063110e-01
5.29540777e-01 2.34455377e-01 -5.64502589e-02 6.61172867e-01
3.27644572e-02 -1.06519687e+00 -4.54004891e-02 1.48716718e-01
7.98173964e-01 5.10694027e-01 1.37810245e-01 6.57536760e-02
1.31072223e-01 1.28922015e-01 6.80482462e-02 7.24269807e-01
1.43005490e-01 1.08415949e+00 1.65831223e-02 -5.81738591e-01
-1.22139943e+00 -1.45260942e+00 -1.64508611e-01 5.06721437e-01
4.57170963e-01 -2.04327166e-01 -2.19252363e-01 -5.45607686e-01
4.31304365e-01 5.49111724e-01 -4.92229968e-01 -4.20418419e-02
-5.99294603e-01 -2.56411284e-01 1.13414362e-01 5.03067672e-01
2.89829403e-01 -8.07280779e-01 -1.22585320e+00 -3.26192030e-03
-1.79584976e-02 -1.26671708e+00 -2.82812387e-01 3.57141823e-01
-8.69155943e-01 -1.25467432e+00 -2.93729305e-01 -8.73841226e-01
7.71880746e-01 7.29312301e-01 1.40558302e+00 -1.78346708e-01
-9.08597052e-01 8.54638636e-01 -2.28612632e-01 -5.08994699e-01
-3.24063271e-01 -3.87213737e-01 4.43680696e-02 8.81587192e-02
2.19584420e-01 -4.46723163e-01 -6.40810013e-01 4.89412397e-01
-1.02751660e+00 1.31268516e-01 6.91450715e-01 5.39882660e-01
8.20121884e-01 -6.54771328e-02 -2.52789110e-01 -1.87320530e-01
1.52537646e-02 -1.24265783e-01 -6.76946342e-01 1.54975697e-01
-2.75770575e-01 1.80658251e-02 -1.09087110e-01 -2.72497684e-01
-9.10663128e-01 6.52771413e-01 2.21700117e-01 -5.89049935e-01
-3.35437864e-01 2.47818381e-01 -2.61485964e-01 -1.50572866e-01
6.46969378e-01 2.75178790e-01 2.01067954e-01 -7.46777833e-01
6.49299979e-01 4.71806824e-01 7.05046952e-01 -4.71217811e-01
1.18753946e+00 1.00893676e+00 1.17202677e-01 -8.78221333e-01
-7.16047883e-01 -7.26039946e-01 -1.06318891e+00 -8.80827755e-02
7.53293753e-01 -1.05434990e+00 -4.05529767e-01 2.16283411e-01
-1.30940187e+00 -1.87974364e-01 -4.49612796e-01 4.72588688e-01
-6.97941363e-01 1.83960751e-01 4.35236469e-02 -7.33765304e-01
8.63488764e-02 -1.19678915e+00 1.81803989e+00 -1.11027621e-01
-2.04219550e-01 -6.08359933e-01 -1.75172798e-02 3.84330183e-01
6.92146346e-02 3.87833297e-01 9.61406052e-01 -1.91406593e-01
-1.09001136e+00 -3.13301027e-01 -4.11463559e-01 3.48506719e-01
1.34676307e-01 1.89191438e-02 -1.02841353e+00 -2.10997611e-01
-8.03223625e-02 -4.05532151e-01 6.14774704e-01 1.28246188e-01
8.46833110e-01 1.49562851e-01 -4.21768278e-01 6.75378323e-01
1.52248454e+00 6.40929416e-02 4.22145933e-01 3.26119959e-01
6.71812594e-01 7.77143359e-01 7.30188072e-01 2.89830953e-01
3.75486016e-01 9.09488440e-01 1.02774155e+00 -4.32343274e-01
-3.22323710e-01 -3.91128331e-01 1.36462644e-01 3.28765810e-01
2.50291318e-01 -4.23963107e-02 -9.94719565e-01 4.76794630e-01
-1.44136071e+00 -6.34755075e-01 -4.78781350e-02 2.20246339e+00
3.99792492e-01 9.71048847e-02 -1.43093109e-01 -1.17849514e-01
1.89141273e-01 1.83551997e-01 -7.80108035e-01 8.22368041e-02
-2.82099903e-01 3.53917390e-01 5.71098506e-01 2.92722404e-01
-1.05884790e+00 7.68012822e-01 6.59673882e+00 2.49656111e-01
-1.04509187e+00 -3.03323776e-01 2.10913301e-01 -4.92399111e-02
-2.62477010e-01 1.19998179e-01 -6.58808291e-01 -2.65628606e-01
1.86102092e-01 1.20169325e-02 2.47511640e-01 1.06947660e+00
9.36348215e-02 -2.38710448e-01 -1.57238531e+00 1.23493850e+00
4.69743639e-01 -1.31414783e+00 3.22374940e-01 3.24978232e-01
6.62211478e-01 3.58882993e-01 1.04344152e-01 -1.10906564e-01
1.45663440e-01 -9.21293795e-01 1.25884402e+00 3.69575739e-01
7.96337128e-01 -3.32452148e-01 3.90580416e-01 4.99105960e-01
-1.20131135e+00 1.07553534e-01 -4.70999122e-01 4.04214626e-03
2.28787586e-01 2.47443482e-01 -1.34561753e+00 4.37237710e-01
8.77122045e-01 7.81217396e-01 -1.01721382e+00 1.01206207e+00
2.29363516e-02 -1.36303408e-02 -6.39541566e-01 2.76208580e-01
1.76368088e-01 -2.31387958e-01 5.92806399e-01 9.52131033e-01
2.79601425e-01 -8.74242559e-02 3.68736833e-01 1.22682261e+00
7.89047927e-02 -2.93852836e-01 -9.96743560e-01 1.04573444e-01
5.29828250e-01 1.09231937e+00 -8.79303217e-01 -8.27576220e-02
-2.77354926e-01 9.77065623e-01 3.45238931e-02 1.95656642e-01
-5.30477762e-01 6.20618276e-02 7.60812640e-01 2.11861178e-01
7.18761861e-01 -6.80837274e-01 -1.73347697e-01 -1.03524911e+00
4.18151945e-01 -9.57385659e-01 1.78348199e-02 -1.24472189e+00
-1.01179206e+00 5.39076626e-01 2.52009779e-01 -1.54037690e+00
-1.93742543e-01 -9.52057421e-01 -3.14353347e-01 8.23380530e-01
-1.33664131e+00 -1.43953121e+00 -7.17832565e-01 4.15387571e-01
6.87054813e-01 5.90849556e-02 9.08227444e-01 -1.23766154e-01
1.23312317e-01 -1.22798823e-01 -6.57205433e-02 -5.79753332e-02
4.60421711e-01 -1.09561551e+00 7.93289721e-01 7.49980152e-01
8.08875501e-01 4.46145386e-01 4.76464540e-01 -4.89268541e-01
-1.95067406e+00 -1.01884270e+00 1.95525229e-01 -1.09490418e+00
2.01847211e-01 -7.86257625e-01 -6.69266045e-01 6.87979162e-01
-2.13732883e-01 1.13697708e-01 2.07386181e-01 -2.97593504e-01
-7.10093915e-01 -1.46951243e-01 -1.01165044e+00 5.17033637e-01
1.27792525e+00 -6.31009936e-01 -6.86864257e-01 3.41359317e-01
5.69454968e-01 -7.61371136e-01 -6.55236125e-01 4.28904951e-01
7.93040395e-01 -1.14843023e+00 1.65788841e+00 -4.04376417e-01
1.27677828e-01 -5.98763287e-01 -5.86406529e-01 -1.21487260e+00
-2.05441087e-01 -1.58180788e-01 -4.72372696e-02 7.21416354e-01
-1.57101266e-02 -1.91184908e-01 9.64321315e-01 6.08669937e-01
-8.43338370e-02 -6.64076388e-01 -9.78884578e-01 -8.21227670e-01
-4.15889531e-01 -8.27296853e-01 6.44098520e-01 4.67142075e-01
-8.72641981e-01 1.06146552e-01 3.71789502e-04 4.89944428e-01
7.80344903e-01 5.32847583e-01 1.46639645e+00 -1.35021245e+00
-2.71203279e-01 -3.12187970e-01 -9.77181256e-01 -1.30035138e+00
-1.11255206e-01 -8.31558704e-01 -2.01665387e-02 -1.57458484e+00
-1.58853065e-02 -5.94976544e-01 3.25640082e-01 3.84379953e-01
3.94387126e-01 7.15549409e-01 4.51395422e-01 2.22143441e-01
-3.93638998e-01 5.38533509e-01 1.03792775e+00 -4.07020241e-01
-5.93244843e-02 -3.04297298e-01 -4.65040714e-01 8.16689372e-01
2.72814363e-01 -3.22968036e-01 -2.72691905e-01 -8.02890360e-01
4.39981259e-02 -3.78034800e-01 8.33640754e-01 -8.81145477e-01
-7.87551180e-02 -1.21774636e-01 5.61323583e-01 -1.22363222e+00
9.77026939e-01 -1.34091711e+00 4.51864302e-01 2.99832106e-01
-2.65557673e-02 4.10131291e-02 2.02184543e-01 6.41355932e-01
-6.19268864e-02 -1.44194663e-01 5.07685661e-01 -4.48072821e-01
-1.01804411e+00 3.27080101e-01 2.99572349e-01 -1.49678543e-01
1.07041264e+00 -5.85777879e-01 -4.48060594e-02 -3.79248172e-01
-5.72218299e-01 1.69172019e-01 8.60320151e-01 6.86078727e-01
1.08052397e+00 -1.37154496e+00 -7.21510470e-01 8.76140893e-01
5.60732484e-01 5.19904435e-01 -2.31684446e-02 3.98808032e-01
-8.01746249e-01 2.88694441e-01 -1.58221126e-01 -1.27311480e+00
-1.40304172e+00 2.54764020e-01 3.35905313e-01 3.46456289e-01
-8.11124086e-01 7.96298563e-01 4.86424506e-01 -7.12911725e-01
3.39224100e-01 -5.67177594e-01 3.88606697e-01 -1.26489848e-01
2.99793392e-01 1.17925838e-01 3.17668617e-01 -9.65965509e-01
-4.72608089e-01 1.01375604e+00 1.54270753e-01 -2.04981461e-01
1.54890311e+00 -5.74406274e-02 -3.80288921e-02 4.23527151e-01
1.39837611e+00 -5.44653982e-02 -1.62065482e+00 -5.27179062e-01
-1.96815297e-01 -1.03255033e+00 8.46938863e-02 -7.36920178e-01
-7.72811115e-01 9.61564660e-01 8.11986804e-01 -5.58788367e-02
7.37782657e-01 5.80711544e-01 3.79087359e-01 6.40201449e-01
7.63460159e-01 -6.41468048e-01 4.57426786e-01 3.08597773e-01
1.28817332e+00 -1.33012223e+00 2.80148208e-01 -4.54205692e-01
-4.55632716e-01 1.33808327e+00 4.85404998e-01 -1.56489372e-01
4.72781837e-01 1.42963111e-01 2.54769385e-01 -6.59735799e-01
-4.02857155e-01 -2.95381963e-01 5.93586743e-01 9.51226890e-01
-2.96754748e-01 -1.12320662e-01 7.88364649e-01 -6.24285899e-02
-2.74035871e-01 -5.40677667e-01 4.23219472e-01 1.02917576e+00
-3.32607806e-01 -1.03769565e+00 -6.94853485e-01 2.27900490e-01
3.61801952e-01 2.16595769e-01 -7.21129179e-01 9.51039612e-01
1.11397430e-01 5.33156633e-01 3.37018043e-01 -1.78238675e-01
7.42094696e-01 -5.35199754e-02 8.80760610e-01 -9.84581709e-01
6.95271045e-02 1.31088167e-01 -1.06582001e-01 -9.10564840e-01
-5.08987606e-01 -9.13721383e-01 -9.22535300e-01 3.58188957e-01
-2.51329750e-01 -6.02927506e-01 1.05269504e+00 4.69774783e-01
4.88267273e-01 -1.02797402e-02 6.43974841e-01 -1.64218283e+00
-5.57721794e-01 -5.59038579e-01 -4.97825682e-01 5.90578020e-01
3.49212706e-01 -8.97998929e-01 -2.81711519e-01 1.63018972e-01] | [7.775272369384766, -2.701007843017578] |
284244d8-635a-4379-92bc-d0507242ca6c | prompt-performance-prediction-for-generative | 2306.08915 | null | https://arxiv.org/abs/2306.08915v1 | https://arxiv.org/pdf/2306.08915v1.pdf | Prompt Performance Prediction for Generative IR | The ability to predict the performance of a query in Information Retrieval (IR) systems has been a longstanding challenge. In this paper, we introduce a novel task called "Prompt Performance Prediction" that aims to predict the performance of a query, referred to as a prompt, before obtaining the actual search results. The context of our task leverages a generative model as an IR engine to evaluate the prompts' performance on image retrieval tasks. We demonstrate the plausibility of our task by measuring the correlation coefficient between predicted and actual performance scores across three datasets containing pairs of prompts and generated images. Our results show promising performance prediction capabilities, suggesting potential applications for optimizing generative IR systems. | ['Olivier Risser-Maroix', 'Ihab Bendidi', 'Nicolas Bizzozzero'] | 2023-06-15 | null | null | null | null | ['information-retrieval'] | ['natural-language-processing'] | [ 3.87888312e-01 -1.27366081e-01 -3.91278982e-01 -6.70454562e-01
-1.65180230e+00 -6.88933969e-01 1.09815979e+00 -8.18402991e-02
-2.64910549e-01 9.39240977e-02 4.25649881e-01 -1.05811082e-01
-4.19057608e-01 -3.60356808e-01 -5.62233329e-01 -4.33209032e-01
2.76960999e-01 5.96214712e-01 -2.79678941e-01 -5.79798315e-03
7.30852544e-01 2.98813909e-01 -1.56252968e+00 7.20124543e-01
6.60070539e-01 1.15976095e+00 3.54906797e-01 1.01358712e+00
3.34264874e-01 9.98863935e-01 -9.73627567e-01 -3.09984356e-01
1.61625549e-01 -2.53046632e-01 -8.96897495e-01 -1.51100859e-01
5.05666912e-01 -5.24807811e-01 -5.17861068e-01 6.34796381e-01
5.52095652e-01 4.43282016e-02 9.52956796e-01 -1.18539476e+00
-1.12369716e+00 3.71982038e-01 -2.42002219e-01 4.59125608e-01
5.97903728e-01 3.06657225e-01 1.50691473e+00 -6.64720833e-01
6.96993113e-01 1.16045046e+00 -2.73529291e-02 4.09328759e-01
-1.56086898e+00 -3.65524083e-01 -2.07224041e-01 1.51640251e-01
-1.48571181e+00 -5.64106762e-01 6.20140135e-01 -6.23849928e-01
7.98309445e-01 4.25582290e-01 1.83478564e-01 1.02864254e+00
2.68887728e-01 1.15813982e+00 9.88040686e-01 -5.02050817e-01
-4.23894590e-03 2.38559753e-01 4.39246833e-01 4.73544717e-01
-2.73908496e-01 4.80645061e-01 -7.11182833e-01 -5.74734926e-01
3.23871344e-01 -2.96537071e-01 -2.54581869e-01 1.13706477e-01
-1.11023712e+00 1.01078379e+00 6.67426884e-01 1.54968813e-01
-5.83724558e-01 3.50981683e-01 -1.56934246e-01 3.09653550e-01
6.17366433e-01 1.14541364e+00 -1.72916710e-01 -2.41661847e-01
-7.72856236e-01 5.59663117e-01 7.38620758e-01 8.93402338e-01
6.73670888e-01 -2.70128518e-01 -1.19050455e+00 9.34309661e-01
1.63501367e-01 6.10041738e-01 5.36904514e-01 -9.63609040e-01
4.86181453e-02 5.21930695e-01 2.99708843e-01 -9.26472425e-01
-6.69325888e-02 -3.00242573e-01 2.79987343e-02 -4.34443831e-01
-1.90548256e-01 2.18573645e-01 -9.22895193e-01 1.75382578e+00
-1.17006622e-01 -8.20676237e-02 -5.95514802e-03 8.61667335e-01
7.59862185e-01 9.16215420e-01 4.26887959e-01 -2.06004247e-01
1.29003525e+00 -7.17586994e-01 -4.47475314e-01 -1.99200258e-01
6.78638995e-01 -1.13407528e+00 1.54916978e+00 1.17542297e-01
-1.08693576e+00 -5.17066240e-01 -4.78533775e-01 -1.58176199e-02
1.42838478e-01 5.19311488e-01 6.52876139e-01 3.49938273e-01
-1.46623647e+00 1.40133291e-01 -4.07003969e-01 2.98428778e-02
6.77440036e-03 2.18378648e-01 1.80022627e-01 -1.17297724e-01
-8.54967058e-01 7.22893834e-01 1.24869354e-01 -4.71022010e-01
-9.62911248e-01 -1.02352905e+00 -1.95994779e-01 5.13431430e-01
2.34679729e-01 -8.79591882e-01 1.90888119e+00 -8.27826262e-01
-1.13094664e+00 9.72219229e-01 -2.87112713e-01 -2.12255374e-01
5.80314770e-02 -3.94632608e-01 -3.65799069e-01 5.40454164e-02
4.28727753e-02 1.01096249e+00 8.51822317e-01 -1.42037463e+00
-3.23532432e-01 -4.67551529e-01 -6.17071660e-03 4.34181839e-01
-4.19693232e-01 3.31787586e-01 -6.70921028e-01 -4.65881050e-01
-1.95421129e-01 -1.18218505e+00 1.34670258e-01 -7.72140563e-01
-5.35323441e-01 -9.71779466e-01 4.13245887e-01 -5.08160770e-01
1.71200335e+00 -2.07914710e+00 -1.42511934e-01 1.87082157e-01
2.88747758e-01 1.90459061e-02 -6.47605062e-01 4.51450497e-01
1.20447017e-01 3.42565954e-01 6.39948785e-01 5.68611473e-02
1.71545938e-01 -2.48261079e-01 -7.74328947e-01 -2.63635993e-01
3.48944157e-01 1.53381932e+00 -7.48300970e-01 -6.24399424e-01
-3.22791845e-01 3.94115746e-01 -5.69606900e-01 8.95391643e-01
-6.52904451e-01 3.01610798e-01 -9.13245201e-01 5.92921436e-01
-2.17493474e-02 -9.39748287e-01 -4.64364141e-02 -2.77945340e-01
1.45494357e-01 4.46180791e-01 -2.42396474e-01 1.46559620e+00
-2.55374521e-01 5.58764637e-01 -5.39469719e-01 -5.42225540e-01
9.62958813e-01 2.84353048e-01 5.98431408e-01 -1.40649641e+00
-1.28394634e-01 -1.72987938e-01 -1.56416059e-01 -6.37107015e-01
1.01106596e+00 4.77149844e-01 -3.02643836e-01 1.00645232e+00
-1.73598856e-01 -2.81767964e-01 2.14448169e-01 5.95470250e-01
1.31052196e+00 -1.77844450e-01 -5.06000733e-03 -4.38010432e-02
-1.57420952e-02 3.81449103e-01 -1.00355167e-02 1.31493163e+00
8.47177133e-02 4.87166792e-01 4.24633771e-01 -3.20869684e-01
-1.11007130e+00 -1.12004173e+00 1.21065881e-02 1.69859660e+00
-1.32316202e-02 -3.90173137e-01 -5.55154622e-01 -6.11850977e-01
-7.23995492e-02 7.81812429e-01 -6.72909021e-01 -4.77641523e-01
-2.18630269e-01 -6.77798808e-01 4.63499218e-01 2.65443802e-01
-1.11026891e-01 -1.18251681e+00 -4.26095009e-01 -1.98279828e-01
-4.58732158e-01 -1.02880228e+00 -7.88160145e-01 -3.06385189e-01
-5.97573161e-01 -8.14417005e-01 -5.58482289e-01 -5.11640429e-01
5.17129779e-01 6.82831466e-01 1.92431641e+00 3.83028120e-01
-4.90353703e-01 1.18524563e+00 -3.82634133e-01 -4.93278950e-01
-6.22510314e-01 1.91970378e-01 -2.31152937e-01 -2.50482380e-01
4.33625698e-01 2.15360507e-01 -1.15463638e+00 4.90593523e-01
-1.18585539e+00 3.66082266e-02 9.08869803e-01 8.21139455e-01
7.43103385e-01 -4.67235267e-01 4.94667888e-01 -7.55398870e-01
1.44249809e+00 -4.81722206e-01 -5.90029299e-01 9.76433396e-01
-1.35897791e+00 4.58228499e-01 -8.68475959e-02 -5.09168506e-01
-1.14976287e+00 -1.21428691e-01 5.06864130e-01 -2.44070545e-01
-6.23298287e-02 5.03552616e-01 3.84112746e-01 2.14505389e-01
1.03107107e+00 3.37746769e-01 -2.51763105e-01 -2.84381509e-01
3.89381796e-01 7.45821059e-01 5.41877329e-01 -8.61309648e-01
4.14268494e-01 -1.30304039e-01 -2.08439097e-01 -3.22598398e-01
-1.22313559e+00 -7.72882998e-01 4.41394448e-02 -4.28836137e-01
6.83020234e-01 -1.02990735e+00 -9.40689445e-01 -2.73852706e-01
-1.17625570e+00 -3.04247051e-01 -2.89272983e-02 2.88776368e-01
-6.69478178e-01 -8.91500413e-02 -5.70643127e-01 -7.38101423e-01
-7.90204346e-01 -1.29880607e+00 1.61436248e+00 2.55349100e-01
-4.37420398e-01 -7.27373540e-01 4.87565368e-01 7.89663970e-01
5.28589904e-01 -5.49285412e-01 1.16981208e+00 -9.31111693e-01
-1.02288985e+00 -3.61355036e-01 -4.76645797e-01 2.96631102e-02
-2.24500522e-01 1.05506353e-01 -9.87129807e-01 -1.12795718e-01
-1.34943113e-01 -7.08252490e-01 6.28224492e-01 4.40356791e-01
1.44017661e+00 -5.59262097e-01 -2.57372826e-01 2.41265908e-01
1.34181941e+00 2.60754406e-01 7.36124754e-01 1.78875759e-01
1.12407319e-01 5.11479080e-01 9.61431801e-01 3.97574544e-01
2.59989589e-01 1.07750666e+00 -1.01256996e-01 1.20393991e-01
-1.74293414e-01 -4.93149430e-01 2.23386079e-01 3.46271992e-01
1.89728096e-01 -7.35616446e-01 -1.03694856e+00 5.34167811e-02
-1.65213335e+00 -8.68953884e-01 9.86672267e-02 2.19838095e+00
9.12908375e-01 -5.11738002e-01 -2.42584229e-01 -5.85432529e-01
3.95818055e-01 5.01018167e-02 -5.97234547e-01 -2.41132692e-01
1.78754061e-01 1.36543438e-01 4.47377041e-02 3.31779510e-01
-6.65240765e-01 8.88351440e-01 7.92115974e+00 8.34918082e-01
-9.28606331e-01 -1.56327114e-01 1.17861402e+00 -1.78512022e-01
-5.09871602e-01 -2.67499201e-02 -8.30086350e-01 3.42630625e-01
1.21910882e+00 -6.43159509e-01 5.37549436e-01 8.54865313e-01
2.84648836e-01 -1.61751762e-01 -1.47295940e+00 9.61635411e-01
3.03977370e-01 -1.03490496e+00 5.00859201e-01 2.10381523e-01
8.52990687e-01 -1.89237427e-02 6.33953691e-01 4.70811933e-01
5.37767470e-01 -1.06196845e+00 2.79312074e-01 8.38964880e-01
7.57430434e-01 -2.92555869e-01 2.14318335e-01 3.93931270e-01
-4.56102520e-01 -1.50731951e-01 -1.58357918e-01 4.34987992e-01
-2.53192913e-02 4.16935265e-01 -1.64881945e+00 7.80348033e-02
3.42550546e-01 -6.34801993e-03 -9.62243080e-01 9.05136943e-01
9.28435549e-02 6.63906932e-01 4.17630933e-02 -1.87427282e-01
5.30709960e-02 1.19381227e-01 1.77447647e-01 1.09111428e+00
4.44777340e-01 2.66217351e-01 1.85404256e-01 9.99064445e-01
-4.35490459e-01 3.90209615e-01 -6.10678375e-01 -5.83088696e-01
5.23837984e-01 1.19308138e+00 -2.01185435e-01 -4.86043990e-01
-6.70102835e-02 6.89952314e-01 4.08991188e-01 6.05640352e-01
-5.60638487e-01 3.41002196e-01 2.80793071e-01 1.42601699e-01
-3.90314251e-01 1.20284267e-01 -2.92204469e-01 -9.25899804e-01
-1.12691715e-01 -9.25134063e-01 5.06966352e-01 -1.51821637e+00
-1.19265354e+00 6.40521228e-01 1.33307874e-01 -1.15349531e+00
-8.32090855e-01 -3.38898599e-01 -3.83337319e-01 8.60562027e-01
-1.49388754e+00 -9.89852250e-01 -5.89900851e-01 5.42502940e-01
5.46599209e-01 -2.44469911e-01 8.22190106e-01 -8.45602155e-02
-2.16909811e-01 6.63918018e-01 -6.25530556e-02 -1.23285361e-01
1.12005174e+00 -1.01305950e+00 4.02781293e-02 3.76970232e-01
6.78968191e-01 6.84816003e-01 6.97381735e-01 -5.40364861e-01
-1.65754402e+00 -6.63225472e-01 9.49222207e-01 -7.90603876e-01
4.48527455e-01 1.62149742e-01 -7.42178082e-01 5.43798506e-01
1.33578867e-01 -4.12798107e-01 9.16457057e-01 3.33229452e-01
-5.26931643e-01 1.10441269e-02 -6.53216720e-01 4.65459406e-01
7.07601249e-01 -7.11042166e-01 -3.82423460e-01 7.81012118e-01
8.44303310e-01 -2.16926619e-01 -9.46099520e-01 5.35361767e-01
6.14330053e-01 -7.67864585e-01 1.26791692e+00 -8.34135592e-01
6.59348786e-01 4.14541751e-01 -1.93251207e-01 -1.15496540e+00
-5.61778724e-01 -2.51285046e-01 -1.82509214e-01 8.33751798e-01
4.93549645e-01 -1.09691478e-01 9.03138459e-01 1.36816251e+00
3.43694314e-02 -6.23235762e-01 -2.98034340e-01 -5.25383353e-01
-2.46078104e-01 -3.03951025e-01 5.45703650e-01 5.30387044e-01
-3.35907042e-01 6.59916878e-01 -2.41136014e-01 2.49664835e-03
3.81137252e-01 4.90938067e-01 1.00204718e+00 -1.29558372e+00
-4.79702681e-01 -6.30942166e-01 -8.74530897e-02 -1.31398559e+00
1.87817737e-01 -9.45963144e-01 2.07886562e-01 -1.18270755e+00
9.26806271e-01 -6.40789866e-01 -3.60049456e-01 2.15872988e-01
-6.62115037e-01 -3.39750089e-02 1.73027396e-01 8.60170007e-01
-1.09431553e+00 4.51936305e-01 1.11237049e+00 -1.58986777e-01
-3.06041509e-01 9.97277871e-02 -8.63930702e-01 1.74350217e-02
4.96553570e-01 -4.51475441e-01 -6.56424224e-01 -7.47079432e-01
4.86105204e-01 4.21034664e-01 5.12128294e-01 -4.77037907e-01
3.71184468e-01 -2.97179461e-01 2.70777404e-01 -5.41443408e-01
4.39923316e-01 -3.12547982e-01 1.28519595e-01 1.49936721e-01
-1.33680093e+00 4.41802651e-01 -2.71120578e-01 6.83915615e-01
-3.30165565e-01 -2.22004384e-01 2.15915665e-01 3.70120741e-02
-7.50383258e-01 4.39414680e-01 7.55287334e-02 8.68667960e-02
7.55470455e-01 4.02267337e-01 -6.70150876e-01 -7.31991410e-01
-4.21011984e-01 1.78053573e-01 1.85864836e-01 6.43255651e-01
7.85734951e-01 -1.26931643e+00 -8.04416835e-01 -1.19615011e-01
7.30529666e-01 -5.98146081e-01 1.37815133e-01 3.88793349e-01
1.56592578e-01 1.01438308e+00 4.75270450e-01 -8.65565956e-01
-1.52382886e+00 6.40232682e-01 -5.51326387e-02 -7.01100588e-01
-2.28073988e-02 6.64107859e-01 5.63148677e-01 1.14251681e-01
7.59865865e-02 4.30306792e-01 -5.75890169e-02 -2.99288601e-01
6.57247305e-01 -1.78324170e-02 1.87764280e-02 -3.06694865e-01
2.23341584e-01 1.59648079e-02 -4.09995049e-01 -4.29618716e-01
1.18527400e+00 -4.02153246e-02 1.83229782e-02 1.25550747e-01
1.38611901e+00 -4.39608037e-01 -9.39413011e-01 -6.29369557e-01
3.09629709e-01 -7.50659466e-01 3.31022710e-01 -1.27595103e+00
-7.62095451e-01 4.97646362e-01 6.95744634e-01 2.47127339e-01
1.14378595e+00 5.82730711e-01 5.34471750e-01 7.92250574e-01
2.14975134e-01 -1.12787831e+00 8.26413989e-01 2.18836650e-01
1.08367443e+00 -1.57906401e+00 -2.65820205e-01 -2.42505185e-02
-8.97780120e-01 8.01405549e-01 5.85464597e-01 2.25365832e-01
2.83613861e-01 -1.49609730e-01 2.11826026e-01 -5.43208241e-01
-1.49603367e+00 -4.55573499e-02 9.52085078e-01 2.61286795e-01
7.89348185e-01 2.19391301e-01 -1.93550393e-01 3.06231976e-01
-2.23157272e-01 7.64313266e-02 2.63601784e-02 7.75084436e-01
-4.28357214e-01 -1.01343870e+00 -3.16031516e-01 8.49139154e-01
-5.73686123e-01 -3.83331150e-01 -6.73267365e-01 1.44067556e-01
-8.54336739e-01 1.05417502e+00 9.97641832e-02 -6.01969838e-01
2.83707649e-01 3.79973978e-01 2.86100090e-01 -6.10004544e-01
-5.21588206e-01 -4.83615547e-02 -4.48605120e-02 -8.00122142e-01
-1.53881103e-01 -5.01894534e-01 -6.32445514e-01 -1.13084309e-01
-2.88354278e-01 3.81916076e-01 1.00804281e+00 6.03649974e-01
8.29383373e-01 2.22032174e-01 1.12939763e+00 -3.32501590e-01
-1.01403892e+00 -1.12741351e+00 -5.79624735e-02 8.58701289e-01
-1.82791412e-01 -4.03023779e-01 -1.82273552e-01 2.16123372e-01] | [11.579974174499512, 7.690279483795166] |
58d5abd5-245c-4818-a68c-1a28cb758ada | evaluation-of-machine-learning-architectures | 2306.15159 | null | https://arxiv.org/abs/2306.15159v1 | https://arxiv.org/pdf/2306.15159v1.pdf | Evaluation of machine learning architectures on the quantification of epistemic and aleatoric uncertainties in complex dynamical systems | Machine learning methods for the construction of data-driven reduced order model models are used in an increasing variety of engineering domains, especially as a supplement to expensive computational fluid dynamics for design problems. An important check on the reliability of surrogate models is Uncertainty Quantification (UQ), a self assessed estimate of the model error. Accurate UQ allows for cost savings by reducing both the required size of training data sets and the required safety factors, while poor UQ prevents users from confidently relying on model predictions. We examine several machine learning techniques, including both Gaussian processes and a family UQ-augmented neural networks: Ensemble neural networks (ENN), Bayesian neural networks (BNN), Dropout neural networks (D-NN), and Gaussian neural networks (G-NN). We evaluate UQ accuracy (distinct from model accuracy) using two metrics: the distribution of normalized residuals on validation data, and the distribution of estimated uncertainties. We apply these metrics to two model data sets, representative of complex dynamical systems: an ocean engineering problem in which a ship traverses irregular wave episodes, and a dispersive wave turbulence system with extreme events, the Majda-McLaughlin-Tabak model. We present conclusions concerning model architecture and hyperparameter tuning. | ['Themistoklis P. Sapsis', 'Alireza Mojahed', 'Stephen Guth'] | 2023-06-27 | null | null | null | null | ['gaussian-processes'] | ['methodology'] | [ 2.23008431e-02 -6.12304360e-02 5.35958767e-01 -9.59717557e-02
-7.19665289e-01 -4.44615811e-01 6.88249290e-01 2.72192150e-01
-3.94200265e-01 1.16961670e+00 9.84652061e-03 -6.49815321e-01
-8.24811757e-01 -6.92815006e-01 -4.07318115e-01 -9.85688090e-01
-2.92698741e-01 5.53400815e-01 -4.44743186e-02 -7.48618692e-02
3.80367309e-01 6.68158174e-01 -1.34532022e+00 -6.27455413e-01
9.80576754e-01 1.12479568e+00 -4.74402159e-01 5.97933769e-01
2.84429282e-01 7.20981777e-01 -3.60649586e-01 -2.66052037e-02
2.25239396e-02 -2.31917068e-01 -3.06990951e-01 -2.86019117e-01
-2.03877494e-01 -1.02937900e-01 -1.07852519e-02 9.31128681e-01
6.72482193e-01 7.12530434e-01 1.24862444e+00 -1.15409815e+00
-2.24671945e-01 1.44508243e-01 7.16559142e-02 1.30967826e-01
-3.05434346e-01 5.49812794e-01 5.92019081e-01 -9.67598259e-01
1.92243345e-02 1.10068619e+00 9.28956628e-01 2.22136438e-01
-1.40934682e+00 -4.88992095e-01 -3.43434989e-01 -4.51352119e-01
-1.37957013e+00 -5.94329655e-01 4.40174192e-01 -1.19016588e+00
8.66129875e-01 1.42268032e-01 3.85068208e-01 7.75889754e-01
5.79455435e-01 -9.41179320e-02 1.01451874e+00 -2.37972856e-01
8.70548129e-01 1.15060478e-01 3.39882761e-01 2.79234707e-01
6.51186943e-01 9.18184102e-01 -2.77623057e-01 -7.77461827e-01
8.33892226e-01 -3.08362335e-01 -3.40388387e-01 -2.59119570e-01
-5.16025722e-01 9.30878282e-01 -3.67253125e-01 -1.96727142e-01
-4.06133652e-01 1.25242382e-01 2.71513134e-01 2.21981630e-01
9.12942410e-01 8.38228226e-01 -9.05133843e-01 -2.96403855e-01
-7.26672888e-01 5.18541753e-01 1.03241074e+00 5.90463400e-01
5.64590871e-01 7.58199573e-01 1.95372120e-01 8.79110813e-01
6.07137859e-01 8.68718922e-01 2.13601783e-01 -1.11121321e+00
-7.59741142e-02 1.05793923e-01 5.80800474e-01 -5.91358960e-01
-5.14589667e-01 -4.43558931e-01 -9.08621371e-01 6.39561951e-01
3.93778980e-01 -7.44377732e-01 -7.64195263e-01 1.47611141e+00
1.14663661e-01 1.82693467e-01 3.23010892e-01 6.05147541e-01
3.80137026e-01 7.61710703e-01 2.06656158e-02 -1.96447358e-01
8.31515133e-01 -3.40399027e-01 -5.30359924e-01 -5.84830306e-02
6.15485728e-01 -4.89002377e-01 5.86773336e-01 5.41855037e-01
-1.05117786e+00 -3.62528145e-01 -8.43510449e-01 5.78881204e-01
-3.42248559e-01 -3.15140367e-01 2.52181232e-01 6.47360146e-01
-6.57926142e-01 1.32813632e+00 -9.99013603e-01 8.10849071e-02
2.12146472e-02 3.29593569e-01 -2.69849449e-01 3.90032589e-01
-1.27253878e+00 1.12031472e+00 1.64546445e-01 6.02133393e-01
-9.23095942e-01 -1.03090048e+00 -9.90796149e-01 1.30753621e-01
-8.91270675e-03 -4.73391443e-01 1.30724037e+00 -2.84077168e-01
-1.63668454e+00 5.55037782e-02 2.62194037e-01 -4.03372228e-01
4.74618316e-01 -4.40321624e-01 -5.27565598e-01 -3.47530872e-01
-3.85827124e-01 -1.82848975e-01 8.26089263e-01 -1.27925849e+00
-4.63452637e-01 -1.07128657e-01 -6.47111535e-01 -1.91095918e-01
2.63206929e-01 -2.08815828e-01 4.36083943e-01 -5.72476089e-01
1.50698289e-01 -1.03673041e+00 -5.42154074e-01 -1.04818121e-01
-3.36676776e-01 -2.08475944e-02 5.58761179e-01 -7.82663524e-01
1.25373161e+00 -2.03704906e+00 -1.58236697e-02 5.88941157e-01
-1.57901511e-01 1.54738799e-01 1.14232741e-01 6.03774726e-01
-9.55382511e-02 2.23239928e-01 -9.18208778e-01 -2.27228373e-01
5.83488643e-02 4.06419307e-01 -4.40683603e-01 5.63807905e-01
5.13593972e-01 1.79787278e-01 -7.10845470e-01 7.06453249e-02
3.56776595e-01 3.80089581e-01 -3.54110003e-01 2.00565010e-01
-4.52854820e-02 6.21279240e-01 -1.59262240e-01 3.88385683e-01
4.18119162e-01 -6.79193065e-02 -3.80185366e-01 1.74124628e-01
-3.41890723e-01 -1.47230495e-02 -1.42792547e+00 6.44163847e-01
-6.93743765e-01 5.78554451e-01 5.93069233e-02 -6.02735102e-01
1.21467316e+00 3.55854392e-01 2.46467918e-01 7.41351908e-03
2.60040998e-01 3.88820350e-01 8.88809562e-02 -4.76786256e-01
4.53145534e-01 -3.89467895e-01 1.05629168e-01 3.42260659e-01
-4.58494388e-03 -5.69148004e-01 -2.35155430e-02 -3.36168289e-01
8.19378018e-01 1.79968998e-01 1.65344238e-01 -8.45974147e-01
3.05730134e-01 -8.52270275e-02 6.39902413e-01 6.81453288e-01
3.13963816e-02 5.42645276e-01 7.06267834e-01 -3.79788309e-01
-1.08707690e+00 -8.47114384e-01 -6.94260240e-01 5.51348448e-01
-2.06331119e-01 1.50704250e-01 -2.94825971e-01 -1.04984883e-02
4.20201421e-01 1.21673799e+00 -5.67806423e-01 -6.12509966e-01
-3.66806477e-01 -1.11878693e+00 3.49286228e-01 6.48292065e-01
-1.78934023e-01 -8.14794481e-01 -4.83471662e-01 3.50832015e-01
5.27855456e-01 -5.62707365e-01 -2.71045957e-02 5.12892127e-01
-1.05619454e+00 -1.12562287e+00 -4.15973455e-01 -1.62720546e-01
5.67507565e-01 -8.43339503e-01 1.07597959e+00 -2.48728752e-01
1.46103604e-02 2.98979670e-01 1.02707678e-02 -7.52497017e-01
-8.74142826e-01 -4.51246589e-01 5.56715906e-01 -2.52601236e-01
9.08035785e-02 -4.71353650e-01 -3.02971452e-01 3.93050730e-01
-6.99540794e-01 -6.28680468e-01 2.42832899e-01 1.03540325e+00
4.16191131e-01 2.07279131e-01 5.42016745e-01 -7.13958383e-01
8.66405129e-01 -6.28245592e-01 -1.27400565e+00 8.92413780e-03
-1.03374410e+00 3.07209820e-01 5.12205124e-01 -3.31057876e-01
-9.85708654e-01 -2.52957970e-01 -2.46184431e-02 -5.20094156e-01
-1.47452373e-02 7.62743711e-01 2.78580397e-01 -1.54510230e-01
6.44136488e-01 -5.27601950e-02 4.13901508e-01 -6.05085611e-01
-2.30758846e-01 6.46230698e-01 3.27738106e-01 -8.13320458e-01
6.77900434e-01 -1.04462817e-01 4.78090733e-01 -1.16857016e+00
-3.12619537e-01 -1.78686365e-01 -3.47923189e-01 -2.20814869e-01
4.73639786e-01 -7.63961375e-01 -7.15698957e-01 6.88027382e-01
-7.96628594e-01 -5.83439767e-01 -5.17196536e-01 9.32969213e-01
-5.71169674e-01 4.84800106e-03 -6.93029761e-01 -1.39954114e+00
-1.94356725e-01 -1.17243671e+00 6.92237198e-01 3.86298634e-02
-4.35622633e-01 -1.36769092e+00 4.29863125e-01 -4.35512811e-01
6.51034474e-01 7.21823454e-01 1.18082547e+00 -8.58177602e-01
6.68425187e-02 -3.38627189e-01 2.00432435e-01 8.59586120e-01
1.09329641e-01 6.13204718e-01 -9.55870092e-01 -2.23928660e-01
2.44221509e-01 -1.82585880e-01 6.71279848e-01 8.58171523e-01
6.02703571e-01 -2.44054079e-01 3.30671743e-02 3.40465635e-01
1.34140027e+00 4.70461279e-01 2.33870611e-01 1.41610637e-01
2.02944607e-01 8.07060122e-01 1.53111219e-01 6.99472010e-01
-2.12061346e-01 -1.68849472e-02 3.08038414e-01 2.01556206e-01
5.78222454e-01 1.05975993e-01 8.40591043e-02 7.15176344e-01
-2.44673833e-01 -1.41176090e-01 -1.37127233e+00 3.79146755e-01
-1.62586176e+00 -6.18188083e-01 -3.92648011e-01 2.56680131e+00
4.90242988e-01 3.01212728e-01 -2.75595576e-01 2.23112151e-01
5.60107946e-01 -2.74109960e-01 -4.90020037e-01 -5.42130768e-01
9.38662216e-02 1.70289814e-01 6.25697672e-01 8.02670956e-01
-1.20057905e+00 8.64522308e-02 6.85623026e+00 4.35228586e-01
-7.53775001e-01 -2.59108692e-01 7.61314273e-01 1.15095086e-01
-7.04880280e-04 -8.37069526e-02 -7.03666568e-01 5.73277175e-01
1.53006423e+00 -3.36600691e-01 1.23337477e-01 9.13500369e-01
5.38342714e-01 -3.00784409e-01 -1.06736445e+00 5.48333406e-01
-4.62339252e-01 -1.27735054e+00 -2.23384202e-01 7.45990872e-02
8.46305132e-01 -1.33126989e-01 1.03264220e-01 4.71118689e-01
6.24502778e-01 -1.06532907e+00 6.67520523e-01 1.03269720e+00
7.03764498e-01 -8.23849440e-01 1.12613034e+00 1.76769003e-01
-7.05101848e-01 3.16688456e-02 -1.92501798e-01 -2.89501280e-01
3.94121140e-01 8.94796252e-01 -3.12663257e-01 2.65960097e-01
6.19728565e-01 2.44766504e-01 -4.49189730e-02 1.16463768e+00
1.85523838e-01 8.56760383e-01 -6.41509414e-01 -8.91668126e-02
2.10573152e-01 -5.71601272e-01 7.60376275e-01 7.87620723e-01
5.31922460e-01 2.28399158e-01 -2.25850463e-01 7.48878658e-01
4.26257908e-01 -3.40406805e-01 -4.32621181e-01 -7.82382190e-02
5.13990104e-01 8.71900380e-01 -3.63780439e-01 9.07212310e-03
-1.03523590e-01 -7.89869428e-02 -4.32643116e-01 6.49938762e-01
-4.85998631e-01 -4.18160349e-01 1.01641285e+00 1.27457112e-01
-4.53372635e-02 -3.09295356e-01 -2.69846827e-01 -4.70729887e-01
-4.87452775e-01 -5.40403247e-01 8.73222649e-02 -6.26728177e-01
-1.60849106e+00 6.00263059e-01 3.31287742e-01 -1.41933537e+00
-7.45616674e-01 -1.05107439e+00 -5.88369429e-01 1.33472705e+00
-1.23049772e+00 -3.00360680e-01 -1.56013463e-02 -1.73123702e-01
4.30392511e-02 -2.79228270e-01 8.49225819e-01 6.59484863e-02
-7.94560313e-01 -1.21674828e-01 9.01239455e-01 3.20629515e-02
5.15240788e-01 -1.24129784e+00 4.20904040e-01 6.20869637e-01
-6.62537456e-01 5.28731644e-01 1.20942056e+00 -7.16250598e-01
-1.14502561e+00 -1.33441281e+00 2.59427100e-01 -5.98503768e-01
8.37380707e-01 1.87412143e-01 -1.18107486e+00 4.19380754e-01
-3.53680432e-01 1.94109231e-01 7.25317061e-01 4.46681539e-03
2.67093450e-01 1.54105991e-01 -1.16591048e+00 2.83519208e-01
1.70030400e-01 -3.47336680e-01 -5.14318347e-01 1.87578917e-01
4.96172160e-01 -3.21765691e-01 -1.37274349e+00 8.09322178e-01
6.25012159e-01 -7.02823102e-01 4.45045799e-01 -9.38100696e-01
3.46430808e-01 -2.92777479e-01 -1.36304319e-01 -1.63858104e+00
-3.37277204e-01 -7.30440497e-01 -2.73209423e-01 1.05160511e+00
5.89731574e-01 -9.04734731e-01 4.34683561e-01 1.34266460e+00
-2.45651960e-01 -8.30074251e-01 -1.16140258e+00 -1.04974818e+00
5.64867556e-01 -5.36756217e-01 5.13566315e-01 7.06540346e-01
-4.56209093e-01 -1.73739418e-02 -2.54241467e-01 5.52686751e-01
8.35324466e-01 -4.55314785e-01 4.59651977e-01 -1.78500128e+00
-2.25525841e-01 -5.24976909e-01 -2.34686568e-01 -1.72301948e-01
4.87890374e-03 -5.65432720e-02 3.73717844e-01 -1.12737679e+00
-5.79387903e-01 -4.81617182e-01 -2.56889611e-01 3.19551378e-02
-4.18237373e-02 -1.32581621e-01 -2.54983842e-01 1.87493369e-01
3.86609226e-01 8.64461124e-01 6.17449522e-01 1.21365301e-01
-4.98628438e-01 3.72534841e-01 -6.24651872e-02 9.76870120e-01
6.51890218e-01 -5.90251923e-01 -3.02763939e-01 -1.44444336e-03
2.72846282e-01 3.81755590e-01 4.09350365e-01 -1.16904366e+00
1.98613793e-01 -2.87436992e-01 3.80626559e-01 -2.84503341e-01
1.16967343e-01 -7.88580000e-01 6.02532804e-01 5.72404444e-01
-2.71726638e-01 1.62168264e-01 6.40338302e-01 6.69252813e-01
-2.80931026e-01 -4.39722985e-01 9.21559691e-01 1.11616245e-02
-3.32696170e-01 2.86149532e-01 -6.86552405e-01 -1.68101120e-04
8.41169715e-01 -1.03114992e-01 -2.67423123e-01 -3.91777903e-01
-9.40937638e-01 4.23451275e-01 2.55111337e-01 5.97386807e-02
3.92151803e-01 -8.59945953e-01 -7.59113669e-01 3.88682246e-01
-5.36711775e-02 1.21892773e-01 2.44231969e-01 7.69883275e-01
-6.39427483e-01 2.27662519e-01 2.64018476e-01 -4.34085101e-01
-5.44060767e-01 1.41612023e-01 8.08627486e-01 3.81071493e-02
-2.29512259e-01 8.62234235e-01 -3.25920433e-02 -5.47412694e-01
-2.52541136e-02 -4.88681525e-01 3.06437984e-02 5.57249300e-02
3.57294261e-01 1.04589474e+00 2.40708590e-01 -5.29357553e-01
-1.24620296e-01 4.57463622e-01 5.34984112e-01 -1.54369786e-01
1.21393991e+00 9.09732878e-02 -1.86368376e-01 9.25451696e-01
8.92570198e-01 -3.64904821e-01 -1.65045500e+00 -4.53583337e-02
9.71021950e-02 5.96608706e-02 3.91427994e-01 -6.47966564e-01
-4.68447775e-01 7.19986439e-01 5.09229243e-01 4.43624228e-01
7.17476487e-01 -6.13045335e-01 2.17473537e-01 5.19155800e-01
1.62744910e-01 -1.06576169e+00 -4.86435324e-01 7.23666072e-01
9.93599653e-01 -1.12088847e+00 2.85819739e-01 1.10954285e-01
-5.16479075e-01 1.24795938e+00 3.34873796e-01 -1.92794174e-01
1.44912708e+00 5.74067831e-01 6.04725927e-02 5.58120161e-02
-8.06182384e-01 2.78811336e-01 3.43658984e-01 2.25859314e-01
5.20997345e-02 -2.21384853e-01 5.27541749e-02 8.04503322e-01
-8.70343894e-02 -6.57070056e-02 6.11454666e-01 9.69349086e-01
-5.14775455e-01 -4.44922566e-01 -5.60124099e-01 9.63104665e-01
-2.33421743e-01 -1.53150782e-01 1.97239816e-01 9.09074724e-01
-2.84268379e-01 8.61878514e-01 2.92523086e-01 -1.57131523e-01
5.92213333e-01 3.61482233e-01 -1.88737348e-01 -3.09838355e-01
-3.24022859e-01 -7.64196590e-02 2.58889824e-01 -1.80291876e-01
-1.88948095e-01 -7.32494414e-01 -8.82275462e-01 -4.39719677e-01
-6.41333818e-01 6.16925001e-01 6.21048629e-01 9.48417962e-01
1.40083939e-01 5.65173924e-01 5.01639009e-01 -1.08170640e+00
-1.09889579e+00 -1.22490251e+00 -9.86124933e-01 -1.60101607e-01
5.47801495e-01 -9.74331081e-01 -1.15493381e+00 -7.27730766e-02] | [6.5775146484375, 3.277268648147583] |
9896c376-04fe-4232-b704-db5e03b19255 | objective-based-hierarchical-clustering-of | 2012.08466 | null | https://arxiv.org/abs/2012.08466v2 | https://arxiv.org/pdf/2012.08466v2.pdf | Objective-Based Hierarchical Clustering of Deep Embedding Vectors | We initiate a comprehensive experimental study of objective-based hierarchical clustering methods on massive datasets consisting of deep embedding vectors from computer vision and NLP applications. This includes a large variety of image embedding (ImageNet, ImageNetV2, NaBirds), word embedding (Twitter, Wikipedia), and sentence embedding (SST-2) vectors from several popular recent models (e.g. ResNet, ResNext, Inception V3, SBERT). Our study includes datasets with up to $4.5$ million entries with embedding dimensions up to $2048$. In order to address the challenge of scaling up hierarchical clustering to such large datasets we propose a new practical hierarchical clustering algorithm B++&C. It gives a 5%/20% improvement on average for the popular Moseley-Wang (MW) / Cohen-Addad et al. (CKMM) objectives (normalized) compared to a wide range of classic methods and recent heuristics. We also introduce a theoretical algorithm B2SAT&C which achieves a $0.74$-approximation for the CKMM objective in polynomial time. This is the first substantial improvement over the trivial $2/3$-approximation achieved by a random binary tree. Prior to this work, the best poly-time approximation of $\approx 2/3 + 0.0004$ was due to Charikar et al. (SODA'19). | ['Dmitrii Avdiukhin', 'Grigory Yaroslavtsev', 'Stanislav Naumov'] | 2020-12-15 | null | null | null | null | ['2048'] | ['playing-games'] | [-0.32445398 0.17707783 -0.14514005 -0.15077372 -0.95190954 -0.50161207
0.28560966 0.6186595 -0.8887861 0.5546027 0.20934168 -0.11331315
-0.40465292 -0.7677696 -0.7152626 -0.793329 -0.68480027 0.7792713
0.22196846 -0.09742954 0.34052205 0.35896432 -1.2863399 -0.16604407
0.41421303 0.9941652 0.14518906 0.9360582 -0.15952635 0.40590045
-0.36096945 -0.7562877 0.6441969 0.04816559 -1.0603927 0.16165498
0.37540677 -0.1352107 -0.62853575 1.0648483 0.7198217 0.1561158
0.6896717 -1.7246274 -0.72617865 1.0735095 -1.1091355 0.09243865
-0.04554176 0.15170282 1.2857553 -0.6979512 0.8212268 1.2368222
0.7470615 0.38783684 -1.421629 -0.6606773 0.06369001 0.608455
-1.7002326 -0.05804424 0.52188843 -0.23300874 1.2254028 0.4235626
0.59284025 0.47395566 -0.10005542 0.8294418 0.9835517 -0.41588464
0.2235743 0.30366594 0.46233884 1.038488 0.4406773 -0.3883182
-0.2652384 -0.39202273 0.14052886 -0.11753473 0.03868835 -0.3892941
-1.2042556 1.0880821 0.84955424 0.21198171 -0.17859152 0.7051027
0.6057099 0.20420477 0.3743846 0.54223806 -0.59459245 -0.10174356
-1.0629939 0.317882 0.794554 1.1405038 1.0596987 -0.10150912
0.23305498 0.72135293 0.16812383 0.28620523 0.43457454 -1.4138182
0.43812644 0.4630283 -0.1893732 -1.1859615 -0.31025428 -0.0591341
-1.1440247 -0.32987744 -0.0327648 -0.03832654 -0.74566096 1.57117
0.4323829 -0.02207618 -0.18789764 0.40826133 0.54265505 0.92797136
-0.08365526 -0.28478435 1.4481328 -1.1732609 -0.5018993 0.11990607
0.7518178 -0.6270145 0.659703 0.41682863 -1.2851582 -0.31610766
-0.98642313 -0.22666913 -0.8300816 -0.2452318 0.7842432 0.6946545
-1.8246672 0.4339594 -0.78677493 -0.67772454 0.43155026 0.8227559
-0.4480613 -0.33725312 -0.86758167 0.6857708 0.6211725 -0.2862508
-0.66790617 -0.5744885 -0.77033645 0.00901787 0.43135774 -0.67431897
0.58622426 -0.4228675 -1.0247452 0.8230049 -0.06202903 -0.92860174
0.28468344 0.04404789 0.14867625 0.55092806 0.2225916 1.3786076
0.51515234 -1.27162 -0.5462187 -0.4426264 0.07475529 0.13023208
-1.0550004 0.1405069 -0.9082709 -0.4604959 -0.11436814 -1.0380859
-0.84531325 -0.10902341 -0.6032373 -0.40987626 0.6026269 -0.5629971
1.348806 -1.987996 0.48185027 0.34961164 0.7570389 -0.10554209
-0.27754137 0.77931726 0.02850883 0.58313936 -0.41303977 -0.48165336
0.4743257 0.2422392 0.30719343 0.65366596 -0.22792457 0.93367535
-0.6448966 -0.93725795 0.19260152 0.26225972 -0.74995047 -0.08429487
0.15785061 -0.7486314 0.18239473 0.5876172 0.6713337 -0.37825835
0.33494884 -0.2620672 -0.00927462 -0.4551465 -1.2813295 1.5288372
-0.22925135 0.75282836 0.4218189 -1.3406267 0.42839488 0.1182498
1.0622737 -0.14456618 0.27968678 -0.06909275 -0.22017021 -0.2606782
0.78978705 0.18558824 -0.2864496 0.32543862 0.17540786 -0.11782231
0.63975006 0.89566237 1.614264 -0.82026935 0.16249391 -0.5926077
0.12086274 0.34135827 0.4200195 0.75438684 -0.45702308 0.44242907
0.5355648 -0.27362055 -1.2690811 -0.9085873 0.0178294 0.99654406
0.14660797 -0.98935777 -0.8743576 -0.5862329 0.1584313 0.18634406
-0.80933803 0.06443848 -0.483725 -1.1811954 0.5083581 0.5452167
0.615866 -0.70170444 -0.25118035 0.25336254 -0.06332137 -1.0972208
-0.47840452 0.312253 -0.89835274 -1.0190195 -0.7311402 -0.90597534
0.7559117 0.20441853 1.0742244 0.05074609 -0.71211344 0.5117201
-0.40447122 -0.04439743 0.2313568 0.30378842 0.321743 -0.36097863
0.5131903 -0.305864 -0.8115532 -0.04131478 -0.97806287 -0.37448528
0.5435454 0.8985604 0.5137945 0.62261546 0.05290593 -0.77116835
0.559499 -0.6034456 -0.7194202 -0.01240006 -1.0436271 0.03164202
0.59733063 -0.30591565 -0.2090689 0.29988304 -0.21391648 -0.62212056
0.16704164 0.516725 0.1179717 -0.14866531 0.48080444 0.13195376
-0.07244425 -0.2086217 0.77436453 0.68436396 0.29377002 -0.5806097
1.1394742 0.6416036 -0.13252324 -0.99658936 -0.30005231 -0.9919719
-0.49017963 0.10214323 0.9853753 -0.80151296 -0.8899453 0.01485723
-0.9477301 -0.19019458 -0.21075842 0.30368054 -0.5133255 0.75616693
-0.9198118 -0.6259977 -0.5274975 -1.2539455 0.7793538 -0.2356893
0.04844396 -0.77476525 0.06318463 0.48023018 0.27757314 0.28287536
0.98199767 -0.4422671 -0.6311452 -0.09739892 -0.53847724 0.32722202
-0.42498946 0.16700676 -0.30419365 -0.63080174 -0.4700483 -0.35431072
1.1358372 0.57056105 1.3211032 -0.612755 -0.4489813 0.64320946
2.0393732 -0.0148706 0.59335434 0.51300144 0.7707405 0.41929278
0.30536205 0.43101385 0.7219937 0.44551888 0.56245625 0.01740265
0.12144095 0.32690933 0.28568092 1.425095 -0.04860052 -0.17864592
-1.0588387 1.1983366 -1.8298655 -0.81262636 -0.07391722 1.8827496
0.88518804 0.14808159 0.38570103 0.15850937 0.74256337 0.12110442
-0.38801795 -0.6143199 0.1024295 0.44223753 1.1617862 0.52191186
-1.161843 1.0773389 6.197508 1.204077 -0.5115928 0.27886072
0.60695386 -0.3445937 -0.13658409 -0.1376737 -0.8294766 0.44773263
1.2722613 -0.4185314 0.6552167 0.9445547 -0.33946478 0.0947172
-0.8579754 1.2776961 0.21722218 -1.6258016 -0.15542914 0.57603437
0.8818608 0.30598143 0.14250535 0.48000565 0.8531871 -0.8983588
0.2776904 -0.2939003 0.74188924 -1.140756 0.81985146 0.14047079
-1.4674007 -0.26029795 -0.8136399 0.41125086 0.03295166 0.6729562
-0.6904626 0.48236057 1.2757168 0.45232457 -0.6691189 0.74338526
0.5026712 0.554921 -0.6669171 -0.37592912 0.6719861 -0.05311818
0.2059037 1.456465 0.24710101 0.3702662 0.07467142 0.2948992
-0.6100749 0.29895252 -0.61990917 -0.25792867 0.58836913 1.3996211
-0.9933078 -0.6219064 -0.04884857 0.97274286 0.55933857 0.12619033
-0.9761673 -0.8560139 0.5551652 -0.03646696 0.66269034 -0.4678349
-0.18154216 -1.0169789 -0.22932038 -0.630831 0.4814244 -0.35886228
-1.4302826 0.66040146 0.22058405 -0.73305184 0.11062363 -0.7309499
-0.03363341 0.2980084 -1.1274596 -0.78439665 0.08636368 0.6406283
0.56342673 -0.04825818 0.6081073 0.37837142 -0.6039844 0.7255333
0.53158283 0.19593453 0.49619028 -1.7250673 0.26572776 0.66511405
0.09906681 0.59326893 0.7056244 -0.31151104 -1.6940905 -1.138702
0.7964234 -0.4721795 0.81523985 -0.41505203 -0.42504987 0.6208105
0.8104867 -0.10316961 0.7080519 0.01585227 -0.45421335 -0.36924157
-1.322281 0.69731957 1.0497651 -0.2047095 -0.285111 0.5786122
0.937168 0.1306366 -1.2713531 0.26268 0.24261852 -0.7933298
1.3202078 -0.3600844 0.44115603 -0.04325346 -0.44086826 -1.0191197
-0.5935456 -0.5966555 -0.29638273 0.879798 0.33632636 -0.35072294
0.8866645 0.47507647 0.19823255 -0.8952966 -0.99306107 -0.76735175
0.26442873 -0.41754618 0.37818146 1.1934497 0.02983847 0.10149357
-0.20442717 0.06831327 1.1503989 -0.1488521 0.85354424 -0.9721891
-0.25174764 -0.5999113 -0.8106297 -0.8230347 0.05885311 -1.0732337
-0.20100851 -1.8474088 0.54593337 -0.33654764 -0.5032473 0.41629714
0.03252269 0.6310042 0.39870906 0.00982699 -0.99780136 0.46930078
0.5296793 -0.478794 0.09089949 -0.63529384 -0.7609303 0.48529604
0.96113 -0.6274907 -0.18224636 -0.578042 0.21431905 -0.16583955
0.02266006 -0.8474145 0.57630956 0.14507687 0.03328579 -0.73321325
0.5056616 -0.94481647 -0.09777492 0.66264254 -0.33945715 0.76691306
0.13066489 0.8232358 -0.1569726 -0.12180197 0.76114166 -0.39255792
-0.67403865 0.65782887 -0.35968396 0.04125896 1.3410547 -0.15929604
-0.48244384 -0.2799253 -0.7606677 0.6471523 0.35221276 -0.05342792
0.82563174 -1.3112082 -0.7748954 -0.38289514 -0.14354008 -0.10760043
0.18150285 0.9968886 -1.029742 0.39484286 -0.07387137 -0.5481014
-1.4496695 0.96861726 -0.26654103 -0.4185847 -0.46140343 1.1695405
-0.22885866 -0.47009093 0.4094432 0.04303158 0.28592238 0.21130992
0.22970133 0.8404925 -0.1742203 -0.52845854 -0.4847949 0.53210557
-0.34248942 -0.4584724 1.5915709 -0.15982549 -0.6244811 0.06219329
1.874354 -0.17847204 -0.5643485 -0.15058549 0.06720918 -0.2682605
0.07349211 -0.22242425 -1.284407 0.83709127 0.59736043 0.5459781
1.1755619 0.18504824 1.0815184 0.7986123 0.44154352 -1.4846984
0.07704163 0.09946185 0.49631885 -1.1222467 0.29190996 -0.03189568
-0.51058173 0.82928854 0.54231024 -0.2720779 0.98483163 0.28903922
-0.2061517 -0.3995262 -0.9728387 -0.41275537 -0.14704333 0.37056008
-0.02953708 0.1873228 -0.51609945 0.10101906 -0.29972073 -0.47106913
0.42490467 0.8373151 -0.547941 -0.95991355 -0.28044736 0.6655471
-0.4947171 -0.28772724 -0.3958562 1.0010045 0.07921713 0.88952607
-0.05514352 -0.63976157 -0.13043278 -0.17999181 0.26483274 -0.45123172
-0.72141224 0.13433751 -0.20400274 -0.7306963 -0.42031682 -0.5063702
-1.1863359 -0.9149126 -0.3163312 0.3850362 0.9356509 0.33846217
0.3152166 0.19070977 0.680899 -0.8647055 -0.42688727 -0.72955805
-0.7085204 0.21645968 -0.12793748 -0.2637596 -0.68483955 0.08923633] | [7.018632411956787, 5.280104637145996] |
eb7aff25-320b-472c-8aa9-aef9789892d2 | small-object-detection-based-on-modified-fssd | 2108.10503 | null | https://arxiv.org/abs/2108.10503v1 | https://arxiv.org/pdf/2108.10503v1.pdf | Small Object Detection Based on Modified FSSD and Model Compression | Small objects have relatively low resolution, the unobvious visual features which are difficult to be extracted, so the existing object detection methods cannot effectively detect small objects, and the detection speed and stability are poor. Thus, this paper proposes a small object detection algorithm based on FSSD, meanwhile, in order to reduce the computational cost and storage space, pruning is carried out to achieve model compression. Firstly, the semantic information contained in the features of different layers can be used to detect different scale objects, and the feature fusion method is improved to obtain more information beneficial to small objects; secondly, batch normalization layer is introduced to accelerate the training of neural network and make the model sparse; finally, the model is pruned by scaling factor to get the corresponding compressed model. The experimental results show that the average accuracy (mAP) of the algorithm can reach 80.4% on PASCAL VOC and the speed is 59.5 FPS on GTX1080ti. After pruning, the compressed model can reach 79.9% mAP, and 79.5 FPS in detection speed. On MS COCO, the best detection accuracy (APs) is 12.1%, and the overall detection accuracy is 49.8% AP when IoU is 0.5. The algorithm can not only improve the detection accuracy of small objects, but also greatly improves the detection speed, which reaches a balance between speed and accuracy. | ['Qinqin Zhou', 'Xianggong Hong', 'Hao Zhang', 'Qingcai Wang'] | 2021-08-24 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [ 4.60929945e-02 -5.77792048e-01 -6.47191778e-02 -4.82156165e-02
-6.89659491e-02 1.18791901e-01 -5.92450649e-02 5.48739098e-02
-6.66950703e-01 2.97635086e-02 -4.83859926e-01 1.45344064e-02
3.07270825e-01 -9.93475914e-01 -2.58501768e-01 -8.61669660e-01
2.26960093e-01 -1.66967899e-01 1.15719652e+00 1.05698116e-01
1.35987327e-01 5.61832070e-01 -1.73086178e+00 2.42922857e-01
8.74014080e-01 1.49505556e+00 9.22405124e-01 5.35975754e-01
-2.17714980e-01 7.49872744e-01 -8.84791493e-01 1.59144402e-01
2.17786759e-01 -7.73314759e-02 -4.32816073e-02 6.32916614e-02
1.49886921e-01 -8.31684768e-01 -4.91034180e-01 1.39605713e+00
6.13044858e-01 -9.77124050e-02 1.77359760e-01 -1.05838501e+00
-3.07989746e-01 2.79763132e-01 -9.67303038e-01 5.42628706e-01
-2.86665201e-01 1.64021567e-01 3.61205250e-01 -8.33854973e-01
7.59229735e-02 1.41348100e+00 3.24939281e-01 3.34387898e-01
-6.92211270e-01 -1.12475502e+00 -1.18715592e-01 4.82813418e-01
-1.72108126e+00 -2.28685036e-01 1.95122436e-01 -1.46526381e-01
7.01965272e-01 3.28356355e-01 7.35426247e-01 1.63831800e-01
6.15067557e-02 8.92592072e-01 7.81252742e-01 -2.45596349e-01
-2.18935862e-01 1.72896758e-01 6.83351457e-02 6.77693427e-01
7.15304196e-01 -1.45813869e-02 -2.05300003e-01 4.63929951e-01
8.72544289e-01 5.11884093e-01 -1.83540300e-01 2.62913972e-01
-1.09490907e+00 5.66733181e-01 8.59357595e-01 3.56388420e-01
-2.09554911e-01 -1.35768205e-01 5.16694248e-01 4.41792002e-03
-2.81158797e-02 -1.32644013e-01 -1.75102204e-01 -1.78832549e-03
-7.84603000e-01 -2.02033222e-02 3.22527915e-01 9.76536632e-01
7.46894538e-01 1.09324351e-01 -2.23756343e-01 9.48820412e-01
2.29355618e-01 9.63092685e-01 6.32952869e-01 -7.75484800e-01
5.19803286e-01 9.04544592e-01 7.64784366e-02 -1.36054850e+00
-3.75109732e-01 -4.78139371e-01 -8.83732855e-01 6.99797720e-02
1.99762002e-01 -3.90623212e-02 -8.44590247e-01 1.02510238e+00
4.18730974e-01 -6.65559098e-02 -2.09444985e-02 1.12861311e+00
1.05525792e+00 1.05037367e+00 3.49329598e-02 -5.44545531e-01
1.62633467e+00 -9.74955082e-01 -6.78441644e-01 -3.96034598e-01
4.46711987e-01 -1.14193130e+00 9.33489025e-01 3.00251842e-01
-7.58509994e-01 -1.10723031e+00 -1.27952099e+00 1.10981865e-02
-5.15997002e-04 8.20129275e-01 4.13744539e-01 3.22679222e-01
-3.60795408e-01 1.37575850e-01 -8.91355455e-01 -2.32464686e-01
6.45514727e-01 2.50253886e-01 2.54407525e-01 -4.15291458e-01
-9.42208111e-01 5.63307464e-01 1.11923122e+00 1.37971282e-01
-4.83679444e-01 -3.27870071e-01 -4.14314002e-01 2.66832113e-01
7.15986907e-01 -1.50774986e-01 1.15164149e+00 -6.54963255e-01
-9.46152985e-01 4.30222303e-01 -9.19733644e-02 -3.97457629e-01
5.28649628e-01 5.14561832e-02 -6.88598752e-01 3.42244297e-01
2.92211592e-01 7.06591845e-01 6.24146402e-01 -5.68419814e-01
-1.58980966e+00 -2.81975150e-01 -6.51275665e-02 3.26099962e-01
-6.52135491e-01 4.05876160e-01 -8.81993890e-01 -2.77190179e-01
4.13987428e-01 -6.46953523e-01 -7.93794990e-02 1.49516568e-01
-1.40621811e-01 -3.07216942e-01 1.41229963e+00 -5.14135063e-01
1.49218619e+00 -2.52235413e+00 -3.86157840e-01 -3.65605834e-03
2.57604897e-01 7.64678240e-01 1.36431739e-01 -3.22839290e-01
4.89694446e-01 -3.10995549e-01 1.73990250e-01 1.14647098e-01
-4.48183119e-01 1.15597054e-01 -2.95354649e-02 1.68252528e-01
-5.12868613e-02 5.54911375e-01 -5.11277139e-01 -9.27392662e-01
4.42821860e-01 2.88454056e-01 -1.96685240e-01 1.94835305e-01
1.63277134e-01 -1.46619439e-01 -7.01609910e-01 8.06302190e-01
1.16501451e+00 -1.87619433e-01 -3.82194936e-01 -4.55431938e-01
-4.89126801e-01 -1.44874647e-01 -1.66676271e+00 9.95732188e-01
-1.66734263e-01 5.79425335e-01 1.09473065e-01 -6.58089936e-01
1.21144879e+00 -3.14249516e-01 2.41136849e-01 -9.87762868e-01
4.47839290e-01 3.79201859e-01 1.78015545e-01 -5.61154008e-01
4.38176125e-01 2.26842329e-01 2.88223207e-01 -4.77624359e-03
-4.99461025e-01 1.03827432e-01 5.26823938e-01 1.58675730e-01
6.16346121e-01 -3.82967681e-01 1.47852570e-01 -2.15387549e-02
5.74157000e-01 1.70004860e-01 8.01224589e-01 5.06674826e-01
-7.76431635e-02 2.83509701e-01 1.37354091e-01 -3.39872986e-01
-7.71911621e-01 -7.74714172e-01 -4.91155505e-01 8.81866336e-01
8.70863736e-01 -4.36647922e-01 -7.56062388e-01 -2.57978797e-01
3.41816427e-04 1.91736221e-01 -1.73703715e-01 -4.57700223e-01
-5.33999801e-01 -9.57286716e-01 2.80389249e-01 8.22784424e-01
1.22107315e+00 -1.03441584e+00 -1.00195765e+00 3.13071102e-01
-2.51431167e-01 -1.16975999e+00 -2.11596459e-01 -4.28745508e-01
-8.50553155e-01 -1.12057257e+00 -5.06457865e-01 -9.75508630e-01
5.63857794e-01 8.97889376e-01 3.50920618e-01 4.26822692e-01
-5.56349158e-01 -4.38991278e-01 -3.07122827e-01 -8.54191780e-01
-5.01771420e-02 -2.12419942e-01 1.24504171e-01 -9.11861733e-02
6.36710644e-01 8.09636489e-02 -6.84832335e-01 5.32158792e-01
-8.46614897e-01 1.45699382e-01 9.04823780e-01 6.57800317e-01
7.25204170e-01 3.82889569e-01 1.87503502e-01 -2.26268113e-01
1.76615924e-01 1.42715037e-01 -8.97627056e-01 -2.31827553e-02
-6.36509359e-01 -5.81200182e-01 7.75657713e-01 -7.46442795e-01
-1.05979264e+00 -7.85990357e-02 8.76938328e-02 -5.20865858e-01
6.96139038e-02 1.98037967e-01 -1.72406688e-01 -2.58651465e-01
4.98248786e-01 4.22389269e-01 1.70417354e-01 -4.46367234e-01
-1.15549594e-01 1.14365280e+00 5.05291283e-01 1.98407352e-01
5.22740364e-01 2.94086695e-01 -2.22747549e-02 -8.19545627e-01
-6.72434330e-01 -4.92598772e-01 -2.81775326e-01 -1.12641133e-01
6.90855980e-01 -1.17355764e+00 -8.66072655e-01 8.02868068e-01
-9.10168767e-01 1.49858475e-01 1.26024671e-02 8.66795540e-01
1.04975723e-01 4.07115340e-01 -9.82554317e-01 -5.95594585e-01
-5.46071947e-01 -9.68428671e-01 8.42679977e-01 1.00231338e+00
5.59038341e-01 -1.14013590e-01 -9.12672281e-01 1.76506639e-01
4.70641762e-01 -2.01384395e-01 3.57106626e-01 -2.13417128e-01
-8.30228865e-01 -5.18242419e-01 -1.03433013e+00 4.28275079e-01
-3.89593490e-03 3.18103164e-01 -7.12095857e-01 -3.88131768e-01
1.05589055e-01 3.09000760e-02 1.00412083e+00 3.83555532e-01
1.30264831e+00 -1.14001974e-01 -7.57999837e-01 5.70902050e-01
1.46705961e+00 5.55476189e-01 5.46793342e-01 4.95160431e-01
8.25474739e-01 7.87097141e-02 1.40562212e+00 4.82929885e-01
1.33299172e-01 4.84934777e-01 3.95298779e-01 -1.53934583e-01
-3.88235182e-01 -3.54533978e-02 1.94256172e-01 7.60065734e-01
-1.32509768e-01 1.60640925e-01 -6.96807802e-01 1.90119728e-01
-1.68391526e+00 -9.92886901e-01 -4.19227242e-01 2.17610741e+00
5.86062193e-01 5.24328947e-01 1.14219114e-02 1.20520949e-01
8.84270787e-01 6.32615089e-02 -6.54011786e-01 1.92237824e-01
-1.21423066e-01 -2.59371966e-01 6.92792118e-01 -1.40078207e-02
-1.06435645e+00 7.64816940e-01 5.21476030e+00 1.39768302e+00
-1.43014657e+00 -7.44883120e-02 4.57959622e-01 -4.09341455e-01
6.98470473e-01 -2.61137307e-01 -1.49417043e+00 9.07044888e-01
5.47206581e-01 -1.75457805e-01 1.48831353e-01 1.27134669e+00
1.92795694e-01 -3.14089417e-01 -4.13873702e-01 1.31031287e+00
2.65709534e-02 -9.65888023e-01 -1.24107309e-01 -5.47007732e-02
4.24676001e-01 3.78315113e-02 -1.09555557e-01 4.16345477e-01
-2.24962592e-01 -6.30517066e-01 5.97246766e-01 2.76089370e-01
8.95858169e-01 -9.83605027e-01 1.04823351e+00 8.01290751e-01
-1.63076842e+00 -5.48941910e-01 -1.25562537e+00 -1.56656399e-01
2.81270295e-02 6.78733408e-01 -4.94082063e-01 2.31936291e-01
1.15050220e+00 7.30574489e-01 -7.81652868e-01 1.26107299e+00
-8.12752172e-02 3.89094800e-01 -7.45564878e-01 -4.90144700e-01
1.29077360e-01 -2.44214490e-01 3.39800447e-01 1.16619134e+00
4.32879716e-01 3.43028218e-01 6.08386874e-01 4.56182897e-01
2.58970335e-02 3.47497851e-01 6.92909062e-02 3.50503683e-01
7.47368455e-01 1.55933309e+00 -8.08097899e-01 -7.01891482e-01
-4.32616621e-01 7.32078731e-01 1.64221358e-02 1.99825261e-02
-1.06718576e+00 -9.90106046e-01 1.71095327e-01 3.43684759e-03
6.46596313e-01 -9.25968885e-02 2.04131939e-02 -1.07005501e+00
3.17029655e-01 -6.31179035e-01 3.86902243e-01 -9.15756166e-01
-5.92355132e-01 3.10381681e-01 -3.52670044e-01 -1.41753638e+00
3.31344157e-01 -6.59734905e-01 -6.17922843e-01 7.04521835e-01
-1.30510831e+00 -8.86982203e-01 -9.18212116e-01 3.08413625e-01
5.09904683e-01 -7.87902549e-02 2.05385119e-01 5.82808912e-01
-1.04919839e+00 5.32697618e-01 4.19839054e-01 1.78786904e-01
5.50311327e-01 -6.33936822e-01 -8.27760696e-02 1.15042770e+00
-2.24439546e-01 3.24892282e-01 3.11881900e-01 -4.61931914e-01
-1.18402815e+00 -1.28839135e+00 3.88709724e-01 1.75618365e-01
2.17005074e-01 -1.48082316e-01 -1.11163366e+00 1.64202899e-01
-5.03475010e-01 3.52912545e-01 9.19094831e-02 -4.59131688e-01
-9.16112363e-02 -6.12703443e-01 -1.10226560e+00 4.51934338e-01
7.75601447e-01 -8.71358812e-02 -3.02197784e-01 4.10070837e-01
8.45012069e-01 -6.49905920e-01 -7.26921618e-01 5.02412200e-01
5.49269915e-01 -9.29360211e-01 9.87853885e-01 -5.74290119e-02
6.14253469e-02 -8.33682418e-01 -1.02329858e-01 -6.44089937e-01
-5.56834102e-01 1.93368435e-01 -2.50570089e-01 1.22102344e+00
-3.69402803e-02 -7.62901962e-01 5.86020052e-01 2.37796932e-01
-5.49402796e-02 -7.19030380e-01 -6.98511302e-01 -7.87122309e-01
-6.30057633e-01 -1.73724577e-01 6.29446268e-01 5.28227746e-01
-3.99135202e-01 4.36984122e-01 -1.03894070e-01 2.53729403e-01
5.05565047e-01 5.06313145e-01 7.77156830e-01 -1.19388366e+00
-1.03072867e-01 -4.31483746e-01 -5.75318813e-01 -1.55453372e+00
-8.02089751e-01 -3.12131584e-01 -9.52217579e-02 -1.50913393e+00
3.53919655e-01 -5.61655998e-01 -4.08839554e-01 3.82487684e-01
-4.74647015e-01 3.51898372e-01 3.12557161e-01 5.54885805e-01
-7.96716452e-01 3.35237592e-01 1.45526195e+00 -5.92937060e-02
-9.51324329e-02 1.84364408e-01 -4.78407264e-01 8.02705526e-01
8.08672905e-01 -3.62554550e-01 -1.31620377e-01 -2.92903125e-01
-4.52591211e-01 -7.42852986e-02 2.17053726e-01 -1.38917506e+00
4.62856323e-01 -1.68712735e-01 9.94833708e-01 -8.17696095e-01
2.47933269e-01 -7.71665633e-01 -2.41222262e-01 1.24660718e+00
2.96456009e-01 -4.46744144e-01 4.40222859e-01 5.20200074e-01
-2.68664360e-01 -2.93471813e-01 9.26121056e-01 1.59556970e-01
-1.14167213e+00 5.09954929e-01 -1.31621674e-01 -2.14728549e-01
1.31191397e+00 -4.41790521e-01 -3.98344159e-01 2.23604396e-01
-2.81896412e-01 4.44180310e-01 3.30257803e-01 3.99153262e-01
7.27972150e-01 -1.51201975e+00 -7.35613823e-01 4.45307076e-01
5.80218285e-02 3.62368912e-01 5.41261196e-01 9.61670578e-01
-8.74910414e-01 1.93857238e-01 -1.97211459e-01 -8.89570713e-01
-1.73277581e+00 7.06619442e-01 3.07722121e-01 2.87666440e-01
-5.79500258e-01 8.54058146e-01 1.86255962e-01 3.24302912e-01
4.64792587e-02 -3.32446963e-01 -4.34204817e-01 4.46470939e-02
1.22451007e+00 7.45989323e-01 -2.31857359e-01 -5.73036551e-01
-3.53208959e-01 7.58762836e-01 -3.40451568e-01 5.82946479e-01
8.58036339e-01 -1.32934481e-01 -1.13601372e-01 -2.55215373e-02
1.05265653e+00 3.14531513e-02 -1.22703862e+00 -4.34165567e-01
-7.07748294e-01 -9.09083307e-01 2.13022888e-01 -4.40120786e-01
-1.23357987e+00 1.09710526e+00 1.01878655e+00 3.30631942e-01
1.34186542e+00 -8.93978924e-02 1.06499815e+00 3.84532750e-01
2.61599958e-01 -1.15517449e+00 2.47724708e-02 3.63973707e-01
4.00575548e-01 -1.13383400e+00 3.31016421e-01 -6.78652287e-01
-4.80758190e-01 1.25964081e+00 1.12974691e+00 -1.32575393e-01
1.11545525e-01 3.76244575e-01 -3.02858325e-03 9.62127894e-02
-2.80410677e-01 -2.71591485e-01 2.33121738e-01 2.40183428e-01
-1.75050929e-01 1.64319292e-01 -3.46360266e-01 6.68456256e-01
-3.66027690e-02 -6.46264851e-02 3.98399591e-01 8.21331978e-01
-1.29332232e+00 -3.96959782e-01 -4.91281420e-01 8.63411307e-01
-4.40725744e-01 4.15005349e-03 2.67980456e-01 6.79249823e-01
5.11578679e-01 1.15005803e+00 4.33849424e-01 -6.53516114e-01
5.60588598e-01 -7.88998365e-01 5.05972095e-02 -3.87531400e-01
8.57965574e-02 3.68339300e-01 -2.86912203e-01 -5.20036280e-01
-1.29379034e-01 -4.05777514e-01 -1.62741148e+00 -5.63350856e-01
-8.73825967e-01 5.23118004e-02 4.60390329e-01 6.35551095e-01
2.68756390e-01 6.12821519e-01 6.26787364e-01 -4.37743545e-01
-4.11521286e-01 -1.05555618e+00 -6.37870848e-01 1.71859190e-01
-7.55762830e-02 -6.97719514e-01 -2.65519500e-01 -1.15633987e-01] | [8.76840591430664, -0.5999980568885803] |
37b5f24f-45a2-4133-ae75-9c54e3619295 | cgmn-a-contrastive-graph-matching-network-for | 2205.15083 | null | https://arxiv.org/abs/2205.15083v2 | https://arxiv.org/pdf/2205.15083v2.pdf | CGMN: A Contrastive Graph Matching Network for Self-Supervised Graph Similarity Learning | Graph similarity learning refers to calculating the similarity score between two graphs, which is required in many realistic applications, such as visual tracking, graph classification, and collaborative filtering. As most of the existing graph neural networks yield effective graph representations of a single graph, little effort has been made for jointly learning two graph representations and calculating their similarity score. In addition, existing unsupervised graph similarity learning methods are mainly clustering-based, which ignores the valuable information embodied in graph pairs. To this end, we propose a contrastive graph matching network (CGMN) for self-supervised graph similarity learning in order to calculate the similarity between any two input graph objects. Specifically, we generate two augmented views for each graph in a pair respectively. Then, we employ two strategies, namely cross-view interaction and cross-graph interaction, for effective node representation learning. The former is resorted to strengthen the consistency of node representations in two views. The latter is utilized to identify node differences between different graphs. Finally, we transform node representations into graph-level representations via pooling operations for graph similarity computation. We have evaluated CGMN on eight real-world datasets, and the experiment results show that the proposed new approach is superior to the state-of-the-art methods in graph similarity learning downstream tasks. | ['Shirui Pan', 'Wei Lin', 'Fei Jiang', 'Xiang Li', 'Yizhen Zheng', 'Luzhi Wang', 'Di Jin'] | 2022-05-30 | null | null | null | null | ['visual-tracking', 'graph-similarity'] | ['computer-vision', 'graphs'] | [ 1.12609476e-01 3.77530009e-02 -5.87392598e-02 -2.69689053e-01
-1.34174228e-01 -4.39130843e-01 6.10703170e-01 6.62486076e-01
-2.96598282e-02 1.10275045e-01 1.07550733e-01 -7.53948018e-02
-3.46730441e-01 -1.07415962e+00 -2.53348351e-01 -6.33536518e-01
7.88667649e-02 1.95999861e-01 4.75844085e-01 -1.53132647e-01
2.09610164e-01 3.89670372e-01 -1.14858091e+00 -3.16925980e-02
8.85369897e-01 6.97948635e-01 3.12186629e-01 2.75328547e-01
-4.30590391e-01 6.64271653e-01 -2.25946739e-01 -5.27151644e-01
1.61016837e-01 -6.49943352e-01 -5.21323383e-01 2.40216762e-01
4.81330454e-01 2.18700141e-01 -7.62023389e-01 1.49702668e+00
4.70365345e-01 4.30715859e-01 5.44490218e-01 -1.49934733e+00
-9.64455664e-01 5.29792130e-01 -7.32662439e-01 3.76149863e-01
3.97412390e-01 -8.73697773e-02 1.29348195e+00 -4.97630745e-01
6.66889727e-01 1.31278050e+00 5.57309210e-01 2.13864893e-01
-9.97950852e-01 -5.12380421e-01 4.41326648e-01 2.60614157e-01
-1.36255074e+00 -1.02800101e-01 1.41012084e+00 -3.11627418e-01
6.33238673e-01 7.81106055e-02 8.19322467e-01 4.59756523e-01
9.04939473e-02 5.58792472e-01 8.43517005e-01 -1.70550182e-01
1.85636319e-02 -8.33683759e-02 2.15920597e-01 1.00416791e+00
3.28739971e-01 -9.22097340e-02 -2.52858430e-01 -1.48392888e-02
6.88863277e-01 5.26709974e-01 -4.60261315e-01 -8.43832016e-01
-1.13320506e+00 7.78591037e-01 1.11377728e+00 5.83129406e-01
-1.87169567e-01 2.99206991e-02 6.16357625e-01 5.19024909e-01
4.91210222e-01 2.02022463e-01 2.02432781e-01 5.42888880e-01
-4.31298792e-01 -1.23827130e-01 4.69854951e-01 9.51236367e-01
9.92851317e-01 -1.03225298e-01 -7.66721740e-02 7.94065177e-01
5.71217954e-01 1.08333297e-01 6.76764667e-01 -3.02793533e-01
6.11729562e-01 1.27258778e+00 -6.20975792e-01 -1.87888169e+00
-3.18851352e-01 -3.32487553e-01 -1.14156115e+00 -9.93956774e-02
7.51339421e-02 1.75542101e-01 -7.99941897e-01 1.57981312e+00
3.03217471e-01 6.30708277e-01 7.98640251e-02 7.81943798e-01
1.37926829e+00 4.19625163e-01 1.75734945e-02 -1.50246471e-01
1.06531537e+00 -1.01890349e+00 -5.81356883e-01 -2.55310774e-01
7.81839490e-01 -6.49568558e-01 8.53031218e-01 -2.49629885e-01
-8.98521543e-01 -6.93507493e-01 -1.08131158e+00 1.41139656e-01
-5.82982898e-01 -1.89563319e-01 7.89285839e-01 3.68542641e-01
-9.33085382e-01 8.43825221e-01 -6.16003692e-01 -6.80478632e-01
4.42859232e-01 1.62495151e-01 -7.35093832e-01 -2.31907979e-01
-1.05556452e+00 5.29426277e-01 6.55446053e-01 1.48747683e-01
-2.97156632e-01 -2.96676040e-01 -1.23004639e+00 2.85856932e-01
3.70552182e-01 -5.98472834e-01 5.52564323e-01 -9.32932258e-01
-1.01057696e+00 9.89128649e-01 1.68164909e-01 -1.32553011e-01
1.81475446e-01 4.35703963e-01 -6.68717802e-01 1.48717657e-01
1.15583152e-01 2.72594184e-01 6.63384914e-01 -1.14368105e+00
-4.01316196e-01 -6.55802310e-01 5.92194535e-02 5.56569993e-01
-5.12223423e-01 -8.37392062e-02 -7.13014424e-01 -6.97300434e-01
7.31666982e-01 -7.97876596e-01 -1.64893746e-01 4.29024063e-02
-3.24042797e-01 -3.92830104e-01 9.81970310e-01 -5.51705778e-01
1.24398756e+00 -2.13255310e+00 3.04586351e-01 5.39521575e-01
6.64716005e-01 3.34447861e-01 -3.33017617e-01 5.98283768e-01
-3.88012141e-01 1.32864080e-02 -1.84465706e-01 -1.56706527e-01
-2.83942223e-01 3.07412855e-02 2.13033393e-01 4.75328654e-01
-1.33852810e-02 8.85532141e-01 -1.24676728e+00 -6.62921607e-01
2.50592530e-01 2.03143045e-01 -2.13260159e-01 4.89018887e-01
2.09730968e-01 2.75422126e-01 -5.95782638e-01 3.96232247e-01
7.15014815e-01 -4.95910704e-01 5.45801580e-01 -5.11738837e-01
4.33680683e-01 -5.29293828e-02 -1.31655216e+00 1.90472388e+00
-2.54803210e-01 2.60592103e-01 -2.72288352e-01 -1.29631126e+00
1.34378254e+00 2.63652746e-02 4.02263701e-01 -5.12037754e-01
7.80899450e-02 1.88247394e-02 1.21929258e-01 -3.98660481e-01
7.57687613e-02 1.32191822e-01 1.67671353e-01 5.40763438e-01
2.44236484e-01 -1.17478825e-01 3.36173117e-01 6.01211667e-01
1.09182072e+00 -1.58141166e-01 5.31628609e-01 -1.22752182e-01
9.83046949e-01 -5.53762555e-01 4.42108840e-01 2.77974337e-01
-2.35139400e-01 5.54409027e-01 5.47350764e-01 -3.89064103e-01
-5.04899740e-01 -1.18624747e+00 6.58212185e-01 8.03918004e-01
7.09428906e-01 -6.42779827e-01 -4.63420302e-01 -1.06687176e+00
-1.57995857e-02 -2.37585027e-02 -4.24374431e-01 -7.14168370e-01
-5.99486709e-01 -3.45939517e-01 9.32904705e-02 5.63086152e-01
7.32698441e-01 -1.00208294e+00 5.54882549e-02 -4.65870015e-02
7.69801140e-02 -8.24532449e-01 -8.89440358e-01 -2.68873870e-01
-1.04015326e+00 -1.38952684e+00 -5.86807847e-01 -1.33163404e+00
9.87764835e-01 9.58625376e-01 1.03192949e+00 6.74561024e-01
-7.46137723e-02 3.28608304e-01 -2.71213830e-01 2.06121877e-01
-2.53313094e-01 6.89145327e-02 -1.37222305e-01 2.27310866e-01
2.90697485e-01 -8.89044225e-01 -4.69552517e-01 2.17222854e-01
-6.69146240e-01 7.95592666e-02 6.71386778e-01 7.45662928e-01
6.26596987e-01 6.58851266e-02 5.72836637e-01 -1.11472416e+00
9.74271655e-01 -5.19724905e-01 -4.63036716e-01 5.67112744e-01
-7.10010946e-01 1.94840506e-01 9.08601463e-01 -3.08792174e-01
-8.23296666e-01 -8.81344080e-02 2.02392071e-01 -6.57311141e-01
2.96963584e-02 8.28770101e-01 -3.72201025e-01 -2.44208738e-01
3.80380034e-01 3.14868480e-01 1.65730089e-01 -3.18768173e-01
5.19882560e-01 3.47624421e-01 4.88609612e-01 -1.44373372e-01
9.25607741e-01 1.83217153e-01 1.50620744e-01 -5.62582433e-01
-3.89293164e-01 -7.30480671e-01 -6.34289026e-01 -2.44718775e-01
7.35143244e-01 -6.68554842e-01 -6.68939710e-01 3.92582148e-01
-8.92117143e-01 2.45323479e-01 1.02968313e-01 5.29500067e-01
-2.92564988e-01 9.36821282e-01 -3.36851358e-01 -4.01811868e-01
-4.26372647e-01 -8.47472012e-01 8.82700145e-01 5.77489376e-01
2.78533280e-01 -1.33230233e+00 2.05980092e-01 1.72679096e-01
1.28784627e-01 2.65019208e-01 9.89385188e-01 -9.81568217e-01
-4.58992839e-01 -3.58542383e-01 -6.82748675e-01 3.47247869e-02
5.72824240e-01 -1.50176406e-01 -4.40207124e-01 -6.06530368e-01
-4.47239786e-01 -1.26928046e-01 7.56341755e-01 3.98121923e-02
1.02741146e+00 -6.02329075e-02 -6.98049247e-01 5.44145703e-01
1.40473783e+00 1.40005693e-01 4.24252540e-01 -1.36839226e-01
1.24548650e+00 5.59329271e-01 4.26222086e-01 5.05883992e-02
3.05668563e-01 5.49175620e-01 4.03744131e-01 -1.71562240e-01
-1.89685225e-01 -5.49438715e-01 -7.12875724e-02 1.26165843e+00
-1.14016630e-01 -3.18424374e-01 -7.31758893e-01 3.86594206e-01
-2.15157676e+00 -8.58947277e-01 -1.07864529e-01 2.26278949e+00
1.11539029e-01 1.65839955e-01 3.60541455e-02 -1.86405510e-01
1.29121840e+00 5.60254216e-01 -4.76012170e-01 -1.33723477e-02
1.66551188e-01 -3.93743962e-02 1.00267380e-01 1.24205269e-01
-1.00697148e+00 9.37401950e-01 4.81802988e+00 7.05607116e-01
-9.56864536e-01 -7.18664750e-02 3.04970622e-01 5.14622033e-01
-4.82447416e-01 2.01453045e-01 -8.14899919e-04 4.55449253e-01
2.79832333e-01 -4.67688143e-01 4.09401804e-01 8.17371666e-01
-1.65126562e-01 3.98243964e-01 -1.00219941e+00 1.27432871e+00
3.02392602e-01 -1.20916843e+00 3.31943482e-01 -2.56427377e-01
5.40437460e-01 -2.57558435e-01 -3.09102207e-01 2.89959431e-01
2.12366030e-01 -7.65379608e-01 -3.34698223e-02 6.52317047e-01
2.56842583e-01 -8.01469624e-01 8.88170958e-01 1.30287379e-01
-1.95019317e+00 2.15589002e-01 -5.43367147e-01 2.02325210e-01
-1.89912561e-02 3.16424996e-01 -5.79848051e-01 1.14457738e+00
5.15030384e-01 1.17889214e+00 -9.37206626e-01 1.09854007e+00
-2.09333420e-01 9.51986313e-02 -9.43188556e-03 -2.79136263e-02
2.43382201e-01 -6.36144340e-01 4.25748795e-01 6.15057230e-01
1.97733998e-01 -1.18670806e-01 5.07968366e-01 7.97991395e-01
-3.82102400e-01 4.20316815e-01 -1.04534686e+00 -2.65661031e-01
6.15706444e-01 1.54278302e+00 -1.04014134e+00 -2.59117514e-01
-6.15662992e-01 1.18259490e+00 7.52580941e-01 2.35345021e-01
-6.36307359e-01 -7.81714022e-01 3.65913928e-01 -1.68604285e-01
8.83804560e-02 -1.63676754e-01 3.05429071e-01 -1.20863903e+00
9.25074592e-02 -5.87471366e-01 8.10289145e-01 -7.73085117e-01
-1.61757338e+00 5.57002544e-01 -5.34140207e-02 -1.30898428e+00
-6.79034591e-02 -3.51399869e-01 -1.19861388e+00 8.10007632e-01
-1.32869637e+00 -1.41719997e+00 -8.70925784e-01 8.70427608e-01
2.04799175e-01 -3.25126112e-01 4.60324675e-01 3.05114329e-01
-4.71295118e-01 6.43432736e-01 -9.52365473e-02 3.94118369e-01
5.53952873e-01 -1.21897626e+00 6.47186756e-01 8.52150142e-01
4.22262669e-01 6.71754718e-01 1.57717958e-01 -8.31396461e-01
-1.45824981e+00 -1.10092616e+00 6.23734891e-01 6.09719269e-02
7.05081761e-01 -2.05531940e-01 -1.22991717e+00 5.86979568e-01
1.18198864e-01 5.67740381e-01 5.34876525e-01 8.54263827e-03
-4.61295605e-01 -1.96674749e-01 -1.01739645e+00 6.73334956e-01
1.41317940e+00 -7.37161636e-01 -5.46238720e-01 3.48516285e-01
7.88645744e-01 -1.95174262e-01 -9.85805869e-01 3.97773206e-01
4.42620993e-01 -9.57263708e-01 9.71388757e-01 -8.36679280e-01
8.51313993e-02 -5.36496878e-01 8.78866315e-02 -1.55698013e+00
-5.50634384e-01 -3.71858060e-01 1.43865258e-01 1.48226690e+00
-1.38604725e-02 -8.58878553e-01 9.29714382e-01 6.53136820e-02
-1.14363991e-01 -6.08985662e-01 -4.27590638e-01 -6.99395239e-01
-4.42387581e-01 3.01355183e-01 7.18942404e-01 1.37389898e+00
1.10468484e-01 6.39284134e-01 -4.47336398e-02 2.21386537e-01
5.22817552e-01 5.16219616e-01 7.99593866e-01 -1.43383741e+00
-1.93458483e-01 -6.79308414e-01 -1.11032033e+00 -7.08577693e-01
3.67442876e-01 -1.34647036e+00 -2.81958997e-01 -1.77597463e+00
2.77563095e-01 -3.26857358e-01 -3.98181409e-01 2.22994760e-01
-5.49046218e-01 4.98367399e-02 2.58056939e-01 2.10243374e-01
-5.95750391e-01 6.54768884e-01 1.27260196e+00 -3.98193538e-01
-2.36237809e-01 -2.67045870e-02 -6.60527527e-01 6.91896796e-01
7.01915205e-01 -2.46998310e-01 -9.32795405e-01 -2.17552900e-01
-1.98459607e-02 2.50850618e-02 2.26769969e-01 -1.04929888e+00
4.36404854e-01 -3.17777283e-02 6.43489435e-02 -4.11779314e-01
-1.38951138e-01 -8.99202466e-01 2.06650510e-01 4.86200541e-01
-2.15300739e-01 2.91319102e-01 -2.85967946e-01 9.98149097e-01
-4.18603539e-01 -2.32534692e-01 7.30192780e-01 -3.42652917e-01
-9.69599187e-01 6.99730694e-01 2.74321020e-01 2.01032981e-01
9.94833410e-01 -3.52889866e-01 -4.24389392e-01 -3.37129235e-01
-7.67254949e-01 2.76787132e-01 4.55908448e-01 5.56485176e-01
8.34371507e-01 -1.70738804e+00 -4.85554636e-01 1.01978801e-01
4.97705281e-01 -1.20705418e-01 3.29305768e-01 7.09244251e-01
-3.58968645e-01 -8.82996619e-02 -3.34730804e-01 -5.40697217e-01
-1.55475736e+00 8.18173647e-01 2.83692956e-01 -5.00641584e-01
-6.05874300e-01 6.75473928e-01 3.37425351e-01 -6.43277049e-01
1.18805654e-02 -1.43652439e-01 -5.62112987e-01 7.61420233e-03
1.85381949e-01 2.61015594e-01 -5.65793552e-03 -6.82320893e-01
-4.26638246e-01 8.46948206e-01 -2.58491099e-01 5.49190760e-01
9.74381685e-01 -6.72483146e-02 -2.99263895e-01 2.76514322e-01
1.47749078e+00 -2.58818746e-01 -6.60509109e-01 -4.65687811e-01
-3.88182998e-02 -5.86135864e-01 -2.68244091e-02 -9.36301798e-02
-1.51159728e+00 8.82268012e-01 6.20910883e-01 4.67692435e-01
1.09676182e+00 1.27608523e-01 7.93447196e-01 5.78989863e-01
1.29012108e-01 -6.64670944e-01 3.98910522e-01 1.13339990e-01
7.02926040e-01 -1.37020779e+00 1.94167674e-01 -7.20873415e-01
-5.85103452e-01 1.02770627e+00 9.37124074e-01 -3.99835587e-01
6.89202845e-01 -5.06085277e-01 -3.01014900e-01 -6.14039600e-01
-3.20878088e-01 -3.78328443e-01 6.18961334e-01 7.31941521e-01
4.29654390e-01 -6.36968389e-02 -3.58600408e-01 2.43344650e-01
2.17599288e-01 -4.13374990e-01 1.51348546e-01 8.17706704e-01
-3.09769630e-01 -8.77677262e-01 1.20267846e-01 6.05123758e-01
1.55439392e-01 -7.38043338e-02 -7.04697251e-01 7.89157152e-01
-1.69142783e-01 7.86201119e-01 1.46580487e-01 -6.34262443e-01
3.30421418e-01 -3.04698139e-01 4.28867310e-01 -7.47620225e-01
-5.53344250e-01 -2.25998551e-01 -1.32499069e-01 -5.27518451e-01
-5.32192469e-01 -3.18849504e-01 -1.31137562e+00 -2.74355680e-01
-5.94129324e-01 1.56463027e-01 1.68085128e-01 7.81231225e-01
4.04346704e-01 5.07416427e-01 7.98009336e-01 -6.83101118e-01
-1.99849948e-01 -7.77403593e-01 -7.16808677e-01 9.05345559e-01
-1.50608361e-01 -7.47294128e-01 -4.06202167e-01 -2.49063209e-01] | [7.232662200927734, 6.205608367919922] |
cff4c907-0c4a-42fd-9d81-b469fa68bb5e | learning-representations-for-counterfactual | 1605.03661 | null | http://arxiv.org/abs/1605.03661v3 | http://arxiv.org/pdf/1605.03661v3.pdf | Learning Representations for Counterfactual Inference | Observational studies are rising in importance due to the widespread
accumulation of data in fields such as healthcare, education, employment and
ecology. We consider the task of answering counterfactual questions such as,
"Would this patient have lower blood sugar had she received a different
medication?". We propose a new algorithmic framework for counterfactual
inference which brings together ideas from domain adaptation and representation
learning. In addition to a theoretical justification, we perform an empirical
comparison with previous approaches to causal inference from observational
data. Our deep learning algorithm significantly outperforms the previous
state-of-the-art. | ['Fredrik D. Johansson', 'Uri Shalit', 'David Sontag'] | 2016-05-12 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 3.54033321e-01 4.46341187e-01 -8.59940767e-01 -5.13324082e-01
-4.66943473e-01 -9.96249467e-02 9.62459981e-01 5.62747896e-01
-4.13875490e-01 1.54817343e+00 1.10203505e+00 -9.44973648e-01
-5.47258556e-01 -9.16492343e-01 -1.06466150e+00 -4.30324405e-01
-2.82976478e-01 5.11203587e-01 -6.50096536e-01 -1.40333742e-01
3.30665946e-01 1.48462608e-01 -1.32178164e+00 2.00610325e-01
9.33480024e-01 5.28325975e-01 -3.65470767e-01 1.28742754e-01
7.41693005e-02 8.67591202e-01 -2.92586029e-01 -4.67134327e-01
-2.20039815e-01 -6.05446637e-01 -1.04779458e+00 -5.03033042e-01
4.85382915e-01 -5.22587419e-01 -5.87760866e-01 8.66992533e-01
7.43163884e-01 2.96844423e-01 7.15676844e-01 -1.21611524e+00
-1.13944423e+00 9.50166702e-01 -5.40744007e-01 2.77837485e-01
5.30318379e-01 1.60104200e-01 8.71364534e-01 -1.56887054e-01
6.39922559e-01 1.56811678e+00 4.65736449e-01 5.34545422e-01
-1.57155311e+00 -1.07936502e+00 2.62304962e-01 5.03845394e-01
-6.40178382e-01 -3.64278704e-01 7.14591920e-01 -5.13507307e-01
7.68988073e-01 1.32076085e-01 6.40737176e-01 1.89701951e+00
4.98397231e-01 5.36899805e-01 1.41252518e+00 -4.12776232e-01
6.05460584e-01 -5.17582238e-01 -2.50667512e-01 3.17928761e-01
7.95817018e-01 9.02289867e-01 -5.08442223e-01 -4.53109086e-01
7.53303707e-01 7.01033771e-02 -5.90701625e-02 -5.50698042e-01
-1.33719599e+00 1.33349502e+00 4.86125678e-01 -5.38067445e-02
-8.67862999e-01 5.32836616e-01 4.82033372e-01 1.44867733e-01
4.49819177e-01 6.25677824e-01 -7.44947791e-01 3.75573374e-02
-7.43196011e-01 9.72370207e-01 8.29216301e-01 3.17540199e-01
-4.29295450e-02 -1.06314041e-01 -2.11287528e-01 3.52492452e-01
2.68791735e-01 5.12829781e-01 5.46186328e-01 -1.17394471e+00
2.90863752e-01 3.60254705e-01 5.49185455e-01 -5.91318727e-01
-5.95591068e-01 -5.30139767e-02 -9.77119923e-01 4.55650762e-02
5.52169383e-01 -5.93013167e-01 -6.26976073e-01 1.95845509e+00
4.64507222e-01 5.90351403e-01 1.95824072e-01 8.35972309e-01
7.91239560e-01 1.54393807e-01 6.35257959e-01 -7.00328231e-01
1.29079735e+00 7.66377077e-02 -9.99486506e-01 1.57385319e-02
4.17247802e-01 -3.64214778e-01 7.79117286e-01 2.39187092e-01
-8.05541575e-01 -2.03702226e-01 -7.84881592e-01 5.92112169e-02
-4.27964061e-01 -5.84121287e-01 1.29965031e+00 8.21845770e-01
-5.22108793e-01 8.09887469e-01 -4.56738651e-01 -4.73570049e-01
8.73068094e-01 2.83530086e-01 -9.05820578e-02 -3.71409282e-02
-1.67349422e+00 9.39731240e-01 4.50310349e-01 -2.76299894e-01
-8.31305981e-01 -1.32637846e+00 -9.84096587e-01 1.22490779e-01
4.29793596e-01 -1.38038087e+00 1.63092065e+00 -6.79172099e-01
-1.41711092e+00 8.13105404e-01 -1.73006982e-01 -1.02168417e+00
5.58852017e-01 -3.40703875e-01 -6.96125090e-01 -3.36800873e-01
2.37132952e-01 5.00171244e-01 3.77203971e-01 -8.31161439e-01
-6.37143373e-01 -7.15144813e-01 2.27354914e-01 2.03481857e-02
2.42996424e-01 4.57222350e-02 1.00048614e+00 -7.18923569e-01
-5.05842865e-01 -6.51017070e-01 -6.31756485e-01 -3.54101717e-01
-2.65529186e-01 -6.34261966e-01 4.92901742e-01 -2.53900230e-01
1.03851771e+00 -1.53787816e+00 -1.14503168e-01 -2.51499832e-01
1.03033669e-01 -4.52141352e-02 2.04384342e-01 2.38722891e-01
-7.50207365e-01 3.21580410e-01 -2.00571433e-01 4.84944731e-01
7.80181810e-02 7.91137069e-02 -7.53535330e-01 4.68209565e-01
2.04287767e-01 1.03993452e+00 -1.47841525e+00 -5.14462352e-01
4.35598314e-01 -3.86763830e-04 -7.98097789e-01 1.37626916e-01
-5.21786988e-01 5.31020820e-01 -4.32873785e-01 1.84067473e-01
5.87771118e-01 -1.35427281e-01 7.33535826e-01 2.44554967e-01
-2.96991497e-01 1.00079703e+00 -8.83925200e-01 1.61014092e+00
-4.54758346e-01 4.35550541e-01 -5.25498331e-01 -1.48960495e+00
3.75061095e-01 6.86763346e-01 3.64051431e-01 -8.55984986e-01
2.24945486e-01 3.50542762e-03 3.64321977e-01 -8.61042798e-01
1.17452912e-01 -8.42152894e-01 -3.46786946e-01 4.60644782e-01
-1.77126393e-01 -1.37008891e-01 -9.99106020e-02 -8.37818235e-02
8.89968455e-01 4.11773115e-01 1.04687870e+00 -5.24953663e-01
1.90714113e-02 -1.58158578e-02 4.19443697e-01 8.11337411e-01
-2.98883826e-01 6.99484572e-02 6.80745423e-01 -8.51512671e-01
-8.01255047e-01 -1.43161559e+00 -4.15138453e-01 9.06089962e-01
-3.19969237e-01 2.75784224e-01 -4.56391394e-01 -8.47115755e-01
4.35931116e-01 1.39426911e+00 -1.09178734e+00 -2.53416032e-01
-5.92661679e-01 -1.11762965e+00 3.10337543e-01 6.33194566e-01
5.83579957e-01 -1.29748774e+00 -7.72808492e-01 1.32862091e-01
-3.25469613e-01 -4.11452800e-01 -4.31961641e-02 -2.36689169e-02
-1.19052076e+00 -1.20503139e+00 -4.44916815e-01 -1.91296533e-01
-1.16454363e-01 -1.60618126e-01 1.44714952e+00 -5.64246297e-01
-1.52483165e-01 -2.28399988e-02 1.59637481e-01 -1.11129773e+00
-4.12373483e-01 -2.87283093e-01 4.02173430e-01 -3.20736170e-01
8.29524875e-01 -8.84656072e-01 -9.31289911e-01 -4.58421290e-01
-6.54223084e-01 -2.70653218e-01 4.66810197e-01 8.49233866e-01
2.18033656e-01 -2.86947072e-01 1.22235501e+00 -1.28929520e+00
8.69559944e-01 -1.11445773e+00 -5.37040710e-01 8.07991251e-03
-7.44113624e-01 3.73535931e-01 4.81660008e-01 -4.29465413e-01
-1.46555519e+00 -3.55035186e-01 6.07469045e-02 -3.63967828e-02
-5.88841498e-01 6.38808250e-01 -1.32573605e-01 9.35878515e-01
9.65201020e-01 -3.28550845e-01 -1.30260542e-01 -2.34055221e-01
7.09628046e-01 4.68589246e-01 5.49803376e-01 -6.00215554e-01
9.22946036e-02 6.23750865e-01 3.29053909e-01 -5.64819157e-01
-1.13738918e+00 5.29411510e-02 -2.23373353e-01 1.88764423e-01
7.96177030e-01 -8.12465847e-01 -1.46891463e+00 -1.57426089e-01
-1.17912197e+00 -4.34407264e-01 -4.89323556e-01 8.95432591e-01
-9.08913493e-01 -1.07755542e-01 5.19249663e-02 -8.77988100e-01
3.41404974e-02 -5.96373200e-01 8.31803441e-01 2.15168551e-01
-6.87577963e-01 -1.24049807e+00 4.32756603e-01 1.81283250e-01
9.17471051e-02 7.18562841e-01 1.17554843e+00 -4.54353392e-01
-1.42618239e-01 2.82009006e-01 -3.05233419e-01 -3.80909413e-01
1.89878836e-01 -4.78278190e-01 -8.65943670e-01 6.78086132e-02
-1.57301985e-02 -3.71053368e-01 1.02458763e+00 1.20182574e+00
1.52903473e+00 -6.83612585e-01 -6.27228677e-01 1.07484363e-01
1.30792284e+00 2.39614353e-01 6.48198247e-01 1.80339962e-01
3.41494918e-01 7.66982734e-01 5.49919605e-01 6.01320982e-01
5.74371934e-01 3.81543487e-01 4.89121586e-01 1.05130874e-01
5.05875163e-02 -5.81613481e-01 -9.75894630e-02 -4.85126138e-01
-2.80201077e-01 -1.28826588e-01 -8.14030707e-01 8.82325053e-01
-2.05390835e+00 -1.42625999e+00 -2.01839834e-01 2.19133568e+00
9.36938941e-01 -1.32760465e-01 2.38213986e-01 -1.03648461e-01
3.94969314e-01 2.23105237e-01 -7.76490927e-01 -8.68756294e-01
1.33457974e-01 6.87387347e-01 5.92326462e-01 2.88363814e-01
-1.32340622e+00 5.26739120e-01 7.62106991e+00 4.09979463e-01
-7.05387473e-01 2.32387513e-01 8.16131711e-01 -2.70209461e-01
-6.04219913e-01 4.95062172e-02 -2.66276777e-01 5.13506174e-01
1.48201025e+00 -5.13468027e-01 1.73724815e-01 4.89220917e-01
6.96738660e-01 -1.57212406e-01 -1.61324108e+00 6.35504246e-01
-3.59767616e-01 -1.82925713e+00 2.34007195e-01 1.28315508e-01
1.13597369e+00 -5.21057546e-02 1.97657779e-01 2.36978844e-01
1.30389464e+00 -1.45202756e+00 4.42651719e-01 5.37268102e-01
7.61701405e-01 -6.49375677e-01 8.25990975e-01 1.61960647e-01
-2.63054997e-01 -3.10966611e-01 -1.48213923e-01 -8.49864364e-01
4.15377356e-02 7.20936716e-01 -9.03707445e-01 7.22607017e-01
6.88151658e-01 4.62692797e-01 2.69630611e-01 7.78688490e-01
-5.26561677e-01 9.18064773e-01 1.59055039e-01 -1.16174117e-01
4.25022393e-02 7.59170428e-02 1.05342470e-01 1.06801736e+00
1.76830634e-01 3.53517979e-01 -2.43764147e-01 1.12310278e+00
-4.77272928e-01 -9.12733823e-02 -1.28064013e+00 8.05434436e-02
4.44148123e-01 2.94726700e-01 -1.55340973e-02 -5.32002926e-01
-4.93244976e-01 4.10819411e-01 2.94236809e-01 3.40850383e-01
-7.70917952e-01 1.49119884e-01 1.01029325e+00 1.39851004e-01
-3.41254263e-03 3.42147946e-01 -6.55096471e-01 -9.24309373e-01
-4.70087022e-01 -7.76695549e-01 8.91244233e-01 -5.45046389e-01
-1.38883519e+00 -5.53620577e-01 4.97645080e-01 -7.28231788e-01
-6.13751113e-01 -5.64813852e-01 -6.27998531e-01 7.47322142e-01
-1.40235019e+00 -8.19136798e-01 3.77056599e-01 1.22872330e-01
4.36115891e-01 3.29254009e-02 1.15389442e+00 5.06768152e-02
-3.92892122e-01 2.49393985e-01 5.23632653e-02 -9.77525786e-02
6.02698445e-01 -1.37576807e+00 3.62505466e-01 5.20037949e-01
-1.69588149e-01 7.88095474e-01 1.06884503e+00 -8.11492503e-01
-1.04038012e+00 -9.98087287e-01 1.33703506e+00 -6.59036696e-01
7.51946270e-01 6.71189129e-02 -4.15310800e-01 8.47892165e-01
4.08382714e-01 -2.44625434e-01 8.52419853e-01 7.53166497e-01
-4.67393130e-01 4.82931286e-02 -1.23966312e+00 9.42986012e-01
1.22577965e+00 -1.68145806e-01 -1.37785792e+00 4.07556832e-01
9.05000925e-01 -1.80635184e-01 -6.04090571e-01 3.72892141e-01
8.41727376e-01 -7.48536348e-01 1.18445837e+00 -1.58682859e+00
1.21459031e+00 2.35289291e-01 -4.21331972e-02 -1.60004735e+00
-2.56521881e-01 -6.33144081e-01 -2.99420178e-01 6.44352317e-01
2.76064485e-01 -7.90912569e-01 6.85419023e-01 4.84565169e-01
3.41650993e-01 -5.54638565e-01 -1.18658078e+00 -5.43631732e-01
6.79182529e-01 -3.75603825e-01 1.07615387e+00 1.40183151e+00
3.77635330e-01 4.74830151e-01 -3.81394237e-01 8.56919736e-02
7.10079730e-01 6.26401961e-01 5.05048335e-01 -1.60935760e+00
-7.94409867e-03 -5.02823591e-01 2.32808497e-02 -5.44864595e-01
4.17071730e-01 -8.88559401e-01 -3.34278911e-01 -1.57615805e+00
5.15977085e-01 5.70974611e-02 -5.71707964e-01 1.98648989e-01
-3.94495249e-01 -1.70991868e-01 -3.79484832e-01 -5.85356534e-01
-3.05948585e-01 6.58537269e-01 1.05666649e+00 -1.10031456e-01
-9.69610140e-02 -6.36020824e-02 -1.13987172e+00 7.74709344e-01
9.18157697e-01 -6.50801897e-01 -2.48605356e-01 -3.71831954e-02
4.85047698e-01 3.93138409e-01 8.21948946e-01 -2.92482823e-01
-3.10735613e-01 -8.44314992e-01 5.86122274e-01 -1.51094034e-01
-2.44934902e-01 -4.83340293e-01 -5.42931296e-02 1.00563371e+00
-7.50667214e-01 -1.32708132e-01 3.52897972e-01 7.96979189e-01
1.12696022e-01 1.40732661e-01 4.75190699e-01 -2.92657048e-01
-5.56989908e-01 7.41653070e-02 -3.87499392e-01 1.76563680e-01
7.31013060e-01 3.89755547e-01 -5.93026459e-01 -4.12923574e-01
-5.08785546e-01 4.17720944e-01 -1.17176056e-01 7.80276120e-01
4.77451235e-01 -1.31088102e+00 -1.01020002e+00 -2.39851639e-01
8.86038542e-02 -3.46982807e-01 3.20907950e-01 7.25367904e-01
-1.15117572e-01 8.63283813e-01 -3.28665495e-01 1.40263453e-01
-5.99807858e-01 1.12754297e+00 1.54040352e-01 -4.54812914e-01
-4.45410877e-01 4.43459868e-01 5.32766640e-01 -6.11370325e-01
-1.87675983e-01 -5.03789604e-01 -1.41541347e-01 -3.38760056e-02
4.09016609e-01 6.80735528e-01 -3.32900107e-01 1.65999621e-01
-4.17767704e-01 -2.67509878e-01 1.50794491e-01 -8.56897384e-02
1.47755516e+00 1.40582651e-01 4.80818748e-03 5.77229857e-01
8.18543375e-01 -3.18172514e-01 -8.44504595e-01 1.20001532e-01
9.33326855e-02 -4.13647443e-01 2.17793901e-02 -1.09712791e+00
-2.85547286e-01 6.96728408e-01 7.28116155e-01 1.90675944e-01
8.17800403e-01 1.72779307e-01 2.57989109e-01 2.24233031e-01
1.91274658e-01 -9.64884162e-01 -2.99129426e-01 -4.27673198e-02
1.05501866e+00 -1.46846449e+00 3.77602369e-01 1.84407681e-02
-2.15614900e-01 4.67725456e-01 8.86549726e-02 -4.85147417e-01
7.55498588e-01 4.17739823e-02 -2.55282968e-01 -2.42408976e-01
-9.11481678e-01 -8.24540257e-02 7.22613186e-02 7.66201198e-01
1.02737629e+00 7.92808592e-01 -8.33584785e-01 5.10946274e-01
-5.64294159e-01 6.95286095e-01 5.21621823e-01 3.09736282e-01
1.19374685e-01 -9.60340261e-01 -3.05993527e-01 8.31671536e-01
-8.22187304e-01 -1.91330314e-01 -2.36053392e-01 9.95380759e-01
2.21832246e-01 1.05100858e+00 3.20746273e-01 2.47399241e-01
3.76094252e-01 2.75263995e-01 6.95575058e-01 -4.38736945e-01
-1.32499665e-01 -5.07247508e-01 3.02390605e-01 -6.76779747e-01
-8.53069067e-01 -1.16667914e+00 -1.02762306e+00 -5.47100246e-01
3.16987514e-01 -1.05417203e-02 3.92219990e-01 1.11929321e+00
3.02182704e-01 9.00973380e-01 2.08615780e-01 -3.92578065e-01
-6.82543576e-01 -1.09737933e+00 -3.15020502e-01 4.16607648e-01
6.32046938e-01 -9.17009294e-01 -3.20212613e-03 -3.33835511e-03] | [8.070560455322266, 5.377635955810547] |
8e45506e-148f-4c59-a1b4-5a4349330a12 | uvid-net-enhanced-semantic-segmentation-of | 2011.14284 | null | https://arxiv.org/abs/2011.14284v2 | https://arxiv.org/pdf/2011.14284v2.pdf | UVid-Net: Enhanced Semantic Segmentation of UAV Aerial Videos by Embedding Temporal Information | Semantic segmentation of aerial videos has been extensively used for decision making in monitoring environmental changes, urban planning, and disaster management. The reliability of these decision support systems is dependent on the accuracy of the video semantic segmentation algorithms. The existing CNN based video semantic segmentation methods have enhanced the image semantic segmentation methods by incorporating an additional module such as LSTM or optical flow for computing temporal dynamics of the video which is a computational overhead. The proposed research work modifies the CNN architecture by incorporating temporal information to improve the efficiency of video semantic segmentation. In this work, an enhanced encoder-decoder based CNN architecture (UVid-Net) is proposed for UAV video semantic segmentation. The encoder of the proposed architecture embeds temporal information for temporally consistent labelling. The decoder is enhanced by introducing the feature-refiner module, which aids in accurate localization of the class labels. The proposed UVid-Net architecture for UAV video semantic segmentation is quantitatively evaluated on extended ManipalUAVid dataset. The performance metric mIoU of 0.79 has been observed which is significantly greater than the other state-of-the-art algorithms. Further, the proposed work produced promising results even for the pre-trained model of UVid-Net on urban street scene with fine tuning the final layer on UAV aerial videos. | ['Radhika Pai', 'Manohara Pai M M', 'Ujjwal Verma', 'Girisha S'] | 2020-11-29 | null | null | null | null | ['aerial-video-semantic-segmentation'] | ['computer-vision'] | [ 2.61978269e-01 -3.64001021e-02 1.00712314e-01 -2.77724743e-01
-4.76330854e-02 -4.85598654e-01 3.10444355e-01 3.79929543e-02
-6.73221946e-01 6.13278508e-01 -2.31464714e-01 -1.10428900e-01
-1.55561149e-01 -8.11219633e-01 -5.81794739e-01 -6.30376339e-01
-1.35685205e-01 -1.34854969e-02 8.46867740e-01 -1.54700071e-01
2.95609355e-01 3.43332171e-01 -1.59544873e+00 1.94175273e-01
5.70552826e-01 1.13586426e+00 5.81116974e-01 1.00350189e+00
-4.75264601e-02 7.45644569e-01 -3.31377357e-01 2.23613381e-01
5.11930168e-01 -3.26005340e-01 -1.05514646e+00 3.60299885e-01
3.08188617e-01 -5.96860111e-01 -3.49221826e-01 9.90781546e-01
9.81887206e-02 4.89136398e-01 4.85274255e-01 -1.34547925e+00
-2.82306969e-01 3.93338621e-01 -3.87054205e-01 4.75487173e-01
-8.96133780e-02 6.21453077e-02 4.85743493e-01 -3.11131299e-01
6.57340646e-01 1.26000702e+00 6.19787455e-01 1.83408290e-01
-3.94194543e-01 -5.10175467e-01 3.51044357e-01 5.15420914e-01
-1.34263062e+00 -1.47450138e-02 4.78436112e-01 -6.38901353e-01
1.11432433e+00 -1.67607546e-01 8.26238632e-01 2.38559097e-01
3.26462716e-01 5.07814765e-01 7.52991796e-01 -1.88676715e-01
2.30821997e-01 -2.94928432e-01 2.04517663e-01 9.39885020e-01
5.52057140e-02 1.69886783e-01 -7.78535102e-03 5.32200694e-01
7.16907680e-01 2.15154320e-01 -4.85862195e-02 8.70244280e-02
-9.11868691e-01 6.77940428e-01 6.71985865e-01 4.38264966e-01
-3.67366433e-01 5.44777632e-01 6.72349691e-01 4.63209935e-02
3.77167821e-01 3.23210955e-02 -6.17730916e-01 -3.70595790e-02
-1.39106035e+00 1.38888285e-02 3.24819982e-01 9.49901462e-01
7.04187989e-01 5.75137317e-01 8.47406462e-02 4.74134028e-01
4.27985400e-01 3.01792651e-01 4.10772562e-01 -1.24173725e+00
1.91671312e-01 6.97219491e-01 8.22931677e-02 -1.18713129e+00
-5.96333921e-01 -2.70621091e-01 -4.92505044e-01 1.38913482e-01
5.80115169e-02 -3.83305430e-01 -1.38072813e+00 1.48070467e+00
3.81921470e-01 5.77223003e-01 3.33443701e-01 1.02826107e+00
6.88821256e-01 1.03962743e+00 4.12105918e-01 6.26004636e-02
1.07935858e+00 -1.17190421e+00 -6.84881628e-01 -7.52405748e-02
6.13719821e-01 -6.85219109e-01 3.74280423e-01 1.85743064e-01
-7.91399360e-01 -8.87730896e-01 -1.23607790e+00 1.24677993e-01
-6.01636529e-01 3.68261039e-01 3.45203400e-01 3.68950486e-01
-1.21712089e+00 5.98428071e-01 -9.38762307e-01 -8.45547199e-01
4.24061447e-01 6.57595396e-01 -3.33030432e-01 6.05730563e-02
-1.09353065e+00 5.28533936e-01 1.00156307e+00 4.21666890e-01
-1.34987020e+00 -1.94287449e-01 -8.22409213e-01 -1.39069378e-01
3.22877675e-01 -4.53134418e-01 1.09361994e+00 -1.35613275e+00
-1.28478980e+00 4.56936717e-01 1.31849766e-01 -8.58331978e-01
3.11657071e-01 -1.13470830e-01 -2.00311229e-01 6.26884460e-01
2.47568995e-01 1.30097759e+00 6.91678643e-01 -9.88218606e-01
-1.16003704e+00 -2.31779993e-01 2.46775374e-01 4.48011994e-01
-1.50819719e-02 -1.60880640e-01 -3.25464875e-01 -4.75005746e-01
3.14337313e-02 -1.07749891e+00 -2.87621856e-01 -1.60566151e-01
-6.28273040e-02 1.47658318e-01 1.78763151e+00 -8.99320245e-01
1.13200438e+00 -2.00405598e+00 -1.58990487e-01 -1.14658341e-01
-2.83282161e-01 6.91358984e-01 3.51596512e-02 1.96773425e-01
4.73739579e-02 1.03016317e-01 -5.70512235e-01 2.38992497e-02
-4.98402566e-01 3.57113540e-01 8.69968086e-02 5.13526440e-01
2.65086144e-02 6.20111942e-01 -6.40222430e-01 -5.19184470e-01
5.78868806e-01 3.23701203e-01 -5.48410058e-01 2.45085098e-02
-4.08282727e-01 5.36596477e-01 -4.73136306e-01 6.40461862e-01
6.93712234e-01 1.59176543e-01 -1.58042103e-01 -2.81113714e-01
-5.33228636e-01 -4.83926535e-01 -1.16122591e+00 1.79327941e+00
-1.36972770e-01 8.39586318e-01 1.75140902e-01 -1.44709826e+00
7.70464778e-01 4.76533592e-01 7.88272917e-01 -4.56482708e-01
5.87616324e-01 4.59583662e-03 -9.83431339e-02 -8.26399684e-01
8.43620718e-01 1.48612127e-01 1.15013286e-01 -3.21501940e-02
1.65808737e-01 -1.31445795e-01 3.52731198e-01 -2.54544504e-02
7.53503144e-01 5.27082384e-01 -2.03968324e-02 -3.52433980e-01
8.21950912e-01 6.28431559e-01 5.70255280e-01 3.49628031e-01
-5.55555284e-01 3.30408663e-01 5.38086481e-02 -4.60466653e-01
-1.14469218e+00 -4.81278539e-01 1.47135397e-02 6.25699282e-01
5.03270507e-01 -1.16936617e-01 -1.25394166e+00 -5.50671577e-01
-4.76156920e-01 3.37128341e-01 -4.21432078e-01 1.13849178e-01
-4.74193990e-01 -6.09781086e-01 6.99098706e-01 5.64114332e-01
1.41496980e+00 -9.82658327e-01 -1.07118690e+00 3.73344928e-01
-1.97427943e-01 -1.52424777e+00 -4.26012836e-02 3.28041874e-02
-1.11363995e+00 -1.24875236e+00 -3.26966345e-01 -9.65041578e-01
5.11119545e-01 5.02135754e-01 3.22727114e-01 1.12166815e-01
-4.33632493e-01 3.76302153e-01 -5.80622196e-01 -2.38973573e-01
4.36085351e-02 3.28150950e-02 -3.26455832e-01 -1.87125057e-01
1.24435581e-01 -1.79339319e-01 -7.81757057e-01 2.83296496e-01
-1.20291138e+00 1.14823900e-01 1.76693708e-01 6.24193311e-01
5.08889318e-01 5.54190278e-01 3.74577522e-01 -6.49422765e-01
1.47216141e-01 -4.61829305e-01 -7.80580401e-01 -6.81490749e-02
-4.55816448e-01 -1.46740526e-01 5.61909676e-01 8.17712769e-02
-1.05803335e+00 3.12759161e-01 -1.35973021e-01 -3.27385664e-01
-4.27383721e-01 4.57050383e-01 2.22322971e-01 -1.43424392e-01
3.98140587e-02 -1.18696028e-02 -1.94861814e-01 -3.72683331e-02
1.74836405e-02 6.84866548e-01 5.31768858e-01 -1.56227350e-01
3.46751273e-01 5.04453719e-01 2.07247198e-01 -1.10401082e+00
-6.11829519e-01 -6.70851707e-01 -9.45334017e-01 -5.85382223e-01
1.68972015e+00 -1.04715097e+00 -3.98537099e-01 7.72967100e-01
-1.19435263e+00 -4.51236665e-01 1.33294880e-01 5.22833645e-01
-5.38107574e-01 3.54516208e-01 -4.38924462e-01 -6.66014791e-01
-3.62095714e-01 -1.44776034e+00 9.89582717e-01 4.38642770e-01
2.10014373e-01 -1.11622512e+00 -2.53738582e-01 4.83178616e-01
3.76955092e-01 4.81753588e-01 6.73079729e-01 -3.90947789e-01
-8.37203622e-01 -6.51200302e-03 -4.15731847e-01 6.07603729e-01
-5.81534356e-02 2.67079115e-01 -8.16097617e-01 3.17598134e-02
-1.80141732e-01 7.02481270e-02 9.25200164e-01 8.37471962e-01
8.67940545e-01 -2.12260365e-01 -1.55666739e-01 5.87701797e-01
1.83971727e+00 8.95893097e-01 6.20418787e-01 5.08929014e-01
7.63599396e-01 4.91808832e-01 1.00885987e+00 3.39318007e-01
3.78992289e-01 4.57501292e-01 7.81783700e-01 -8.68272260e-02
-1.11693569e-01 4.77853119e-02 4.80822682e-01 6.88381672e-01
3.92169915e-02 -5.43647170e-01 -8.59139562e-01 7.67504215e-01
-2.07914710e+00 -8.91497016e-01 -2.68634647e-01 1.60023856e+00
7.89575577e-02 -1.39475122e-01 -2.13735089e-01 3.50992441e-01
7.89134800e-01 1.54615059e-01 -2.39447877e-01 -5.41216135e-01
1.93178102e-01 2.46163785e-01 9.86344635e-01 7.28340030e-01
-1.54642534e+00 1.56057072e+00 5.44122553e+00 6.59168124e-01
-1.29024529e+00 3.14397335e-01 3.24362218e-01 3.28072608e-01
4.92255181e-01 1.50268286e-01 -7.38257706e-01 3.83679926e-01
9.23039734e-01 3.91464353e-01 8.46755281e-02 7.01919079e-01
7.88330078e-01 -6.94929481e-01 -2.93031573e-01 6.95756257e-01
-2.57930774e-02 -1.17415071e+00 9.89358500e-02 -3.24748635e-01
8.90443385e-01 -2.03791894e-02 -1.46926537e-01 -1.23917475e-01
9.02010947e-02 -7.93232560e-01 7.23636568e-01 5.45524716e-01
4.51040804e-01 -9.54767287e-01 1.11300039e+00 1.13506347e-01
-1.51305711e+00 -3.55569363e-01 -3.95478129e-01 -1.55492067e-01
3.65904033e-01 9.06042848e-03 -9.74971473e-01 5.29406309e-01
7.87570059e-01 1.05382752e+00 -5.54725647e-01 1.01740968e+00
1.09608233e-01 6.98716462e-01 -2.37699911e-01 2.99317837e-01
1.14299607e+00 -3.68670702e-01 5.01021504e-01 1.29820371e+00
5.64041257e-01 3.32905948e-01 4.96028125e-01 3.29028249e-01
2.39609987e-01 -9.69865173e-03 -7.33366251e-01 -2.22103223e-01
7.86534846e-02 1.15311098e+00 -1.26907229e+00 -4.82450426e-01
-1.28941938e-01 9.19061422e-01 -5.16084909e-01 4.02862966e-01
-9.06750321e-01 -4.33376044e-01 4.57528204e-01 1.35073870e-01
5.43385386e-01 -6.44350648e-01 -1.20766433e-02 -5.83734035e-01
-5.84539294e-01 -3.21919739e-01 2.63690233e-01 -1.03552794e+00
-1.58528537e-01 7.85193443e-01 2.46917024e-01 -1.11756587e+00
-8.79431143e-02 -6.50875092e-01 -3.77446294e-01 3.75142276e-01
-1.42310011e+00 -1.33500409e+00 -7.27058530e-01 6.58008695e-01
1.07876682e+00 -2.49169558e-01 3.36372226e-01 3.70849133e-01
-6.81253910e-01 -2.17400923e-01 -1.10004783e-01 1.38113543e-01
3.22035849e-01 -7.58683801e-01 -6.67729527e-02 1.43323755e+00
-2.05115974e-01 -1.44610777e-02 5.97526371e-01 -9.80316758e-01
-7.86237657e-01 -1.62601352e+00 3.71572554e-01 2.30252832e-01
3.58464599e-01 2.70854086e-01 -4.28384542e-01 7.12944567e-01
5.78159213e-01 -1.91349328e-01 2.89543837e-01 -9.92876828e-01
4.60294873e-01 -1.29153222e-01 -1.26136565e+00 1.99680835e-01
7.48596907e-01 -1.85130596e-01 -1.77079558e-01 1.20589189e-01
9.36702132e-01 -2.30231702e-01 -5.52522182e-01 4.72586453e-01
2.47462675e-01 -9.38944221e-01 7.31885910e-01 -2.31143907e-01
3.83810878e-01 -9.23940897e-01 -3.72373164e-01 -8.82185102e-01
-6.84083775e-02 -1.09685682e-01 4.17826921e-01 9.57496881e-01
6.98274150e-02 -1.21058218e-01 7.26626992e-01 3.20578992e-01
-5.48673153e-01 -3.76116633e-01 -5.35826147e-01 -4.61694330e-01
-5.76830089e-01 -5.54944098e-01 2.32419282e-01 6.59248710e-01
-8.02597463e-01 8.80176388e-03 -2.62101233e-01 3.22587848e-01
3.77191246e-01 -3.83902401e-01 3.76037478e-01 -9.84253466e-01
2.08522975e-01 1.87207945e-02 -9.17494833e-01 -7.55321503e-01
1.24281272e-01 -7.36069500e-01 6.01029173e-02 -1.84920847e+00
-1.72534198e-01 -2.29573399e-01 -4.66439500e-02 5.02000093e-01
2.64831781e-01 3.64637434e-01 2.85211802e-01 -1.09856762e-02
-6.85810924e-01 4.37445998e-01 1.22936594e+00 -2.12712735e-01
-1.87130958e-01 -2.29812264e-01 -5.16241454e-02 8.66891503e-01
9.85050321e-01 -4.87300694e-01 -8.57979894e-01 -6.35174632e-01
-2.42886126e-01 7.34277740e-02 4.85451460e-01 -1.53429604e+00
4.26639318e-01 -1.24003038e-01 3.35072905e-01 -9.01323855e-01
1.94901437e-01 -1.11528468e+00 3.56232405e-01 7.73279905e-01
-1.59677565e-01 5.43866634e-01 4.59707797e-01 7.36848116e-01
-6.74191356e-01 -5.38245916e-01 9.81576025e-01 -4.35108066e-01
-1.52943897e+00 3.83486688e-01 -8.27894807e-01 -1.41238436e-01
1.50256896e+00 -6.55494452e-01 -9.98950601e-02 -1.08505160e-01
-6.77795887e-01 2.26748466e-01 3.19999903e-01 2.86649287e-01
7.10798264e-01 -8.28990817e-01 -1.86715156e-01 6.55924827e-02
-3.46431047e-01 1.26956433e-01 4.59577978e-01 8.75410795e-01
-1.34677434e+00 6.90124929e-01 -6.11631751e-01 -6.92954779e-01
-1.32548308e+00 2.23725945e-01 4.24474329e-01 3.78435552e-02
-4.35435355e-01 6.60956383e-01 2.36821212e-02 6.54021837e-03
1.36674762e-01 -5.77107549e-01 -5.33844292e-01 8.49652141e-02
-1.78397484e-02 5.64016938e-01 -1.27713740e-01 -9.06423330e-01
-2.57463723e-01 7.89444208e-01 2.05913246e-01 -2.87723482e-01
1.26644111e+00 -4.94334698e-01 -2.07136273e-01 9.81497094e-02
1.13689649e+00 -7.77335525e-01 -1.47690332e+00 3.09995592e-01
-1.26323774e-02 -1.80182099e-01 5.14758825e-01 -7.06243873e-01
-1.45967460e+00 1.04551125e+00 9.61937308e-01 -2.65206039e-01
1.34453595e+00 -7.56484985e-01 1.07185173e+00 2.96886176e-01
2.68507481e-01 -1.47714257e+00 -3.91527731e-03 8.09374273e-01
3.31087619e-01 -1.15733445e+00 -7.69463778e-02 -1.90670639e-01
-8.08726847e-01 1.30907607e+00 7.97064781e-01 -2.93474495e-01
8.10636044e-01 2.38652557e-01 -6.31419420e-02 -2.25332186e-01
-4.97054994e-01 -3.82271528e-01 7.37527460e-02 6.02280319e-01
1.47368699e-01 -5.97288683e-02 -3.97393316e-01 2.63416529e-01
1.93970770e-01 2.48037845e-01 6.17936552e-01 1.13449359e+00
-8.04658592e-01 -7.08141804e-01 -2.17773020e-01 1.33650124e-01
-6.05290234e-01 2.24825218e-02 -1.58899263e-01 8.51627886e-01
6.19920850e-01 1.11999404e+00 8.08608159e-02 -5.60239673e-01
-3.98978963e-02 -8.21763650e-02 1.98963299e-01 -3.04700464e-01
-6.95998073e-01 5.96112013e-03 1.03509136e-01 -5.97394586e-01
-1.05965674e+00 -6.23226881e-01 -1.60341132e+00 -3.48188542e-03
-1.96405426e-01 2.11411059e-01 9.68967140e-01 1.18324900e+00
1.86603010e-01 7.80931950e-01 2.93469548e-01 -7.27668703e-01
3.72431099e-01 -7.30868459e-01 -3.95031393e-01 -4.22117859e-02
2.53334790e-01 -7.77060628e-01 7.13039488e-02 6.45930946e-01] | [9.05648136138916, -0.9683645367622375] |
3f4c0cb9-a8d5-478b-aea1-4e835048ae92 | a-causal-framework-for-decomposing-spurious | 2306.05071 | null | https://arxiv.org/abs/2306.05071v1 | https://arxiv.org/pdf/2306.05071v1.pdf | A Causal Framework for Decomposing Spurious Variations | One of the fundamental challenges found throughout the data sciences is to explain why things happen in specific ways, or through which mechanisms a certain variable $X$ exerts influences over another variable $Y$. In statistics and machine learning, significant efforts have been put into developing machinery to estimate correlations across variables efficiently. In causal inference, a large body of literature is concerned with the decomposition of causal effects under the rubric of mediation analysis. However, many variations are spurious in nature, including different phenomena throughout the applied sciences. Despite the statistical power to estimate correlations and the identification power to decompose causal effects, there is still little understanding of the properties of spurious associations and how they can be decomposed in terms of the underlying causal mechanisms. In this manuscript, we develop formal tools for decomposing spurious variations in both Markovian and Semi-Markovian models. We prove the first results that allow a non-parametric decomposition of spurious effects and provide sufficient conditions for the identification of such decompositions. The described approach has several applications, ranging from explainable and fair AI to questions in epidemiology and medicine, and we empirically demonstrate its use on a real-world dataset. | ['Elias Bareinboim', 'Drago Plecko'] | 2023-06-08 | null | null | null | null | ['causal-inference', 'epidemiology', 'causal-inference'] | ['knowledge-base', 'medical', 'miscellaneous'] | [ 2.73062795e-01 1.35058433e-01 -8.07602942e-01 -4.26357776e-01
-1.88949093e-01 -3.27095568e-01 4.62550014e-01 8.12405571e-02
1.65842357e-03 1.03255880e+00 2.05009520e-01 -6.66800320e-01
-6.97162330e-01 -8.38074565e-01 -8.15286696e-01 -6.92581415e-01
-3.55845451e-01 3.19976330e-01 -3.08819443e-01 1.97224453e-01
2.13935256e-01 2.07718298e-01 -1.28586590e+00 -2.38601401e-01
1.01369047e+00 3.15456718e-01 -1.56168580e-01 4.16633338e-01
-7.29693323e-02 6.24174416e-01 -3.23530704e-01 -4.14080799e-01
-1.31301925e-01 -8.07597816e-01 -9.27082479e-01 -2.64253438e-01
8.19118768e-02 -6.79889172e-02 -1.34365529e-01 8.62798929e-01
8.88353065e-02 -2.37118557e-01 7.59245217e-01 -1.50911021e+00
-7.95967937e-01 1.00579488e+00 -6.96477473e-01 1.21522531e-01
3.68108481e-01 -9.55118164e-02 1.21053100e+00 -5.09688079e-01
4.93868440e-01 1.60323620e+00 5.05414784e-01 3.29840481e-01
-1.58854210e+00 -8.38795543e-01 4.87092510e-02 3.69819887e-02
-1.06192529e+00 -2.37595662e-01 4.83836919e-01 -6.08510554e-01
4.55206901e-01 5.22574008e-01 4.19229269e-01 1.10579824e+00
3.50722939e-01 3.89349937e-01 1.44868934e+00 -4.67212051e-01
1.28345430e-01 3.06277201e-02 5.56242466e-01 6.49191976e-01
6.03084207e-01 3.93039614e-01 -7.40943074e-01 -4.49438810e-01
1.04990423e+00 -1.55940549e-02 -8.20092857e-02 -2.48277977e-01
-1.23260164e+00 1.10845459e+00 1.65122420e-01 4.09070760e-01
-4.21993881e-01 3.86187017e-01 9.83178709e-03 2.07026139e-01
3.77031207e-01 4.66523886e-01 -5.42004049e-01 -4.33657281e-02
-6.68565810e-01 2.79386371e-01 7.72788346e-01 6.25988364e-01
6.10367596e-01 -1.66869849e-01 1.39611408e-01 3.76418203e-01
9.84362885e-02 7.69381046e-01 2.00472549e-02 -9.51350272e-01
1.32146701e-01 5.56657434e-01 3.43782097e-01 -1.25275111e+00
-4.91936862e-01 -2.17698872e-01 -1.03822005e+00 -3.13304871e-01
6.05790913e-01 -1.75698757e-01 -5.24496794e-01 2.35574412e+00
3.24008912e-01 2.22074673e-01 -3.27095687e-01 8.13872635e-01
3.69895399e-01 4.08186674e-01 4.19110388e-01 -7.11124241e-01
1.15013778e+00 -7.75176138e-02 -9.39229429e-01 -3.26404497e-02
5.69455087e-01 -6.67749047e-01 1.06209731e+00 1.64661527e-01
-1.10770810e+00 -2.69737214e-01 -3.87178361e-01 2.93037295e-02
-5.55569306e-03 -2.60912985e-01 1.19227290e+00 7.48364925e-01
-7.76444137e-01 4.75605845e-01 -7.53165185e-01 -3.35060030e-01
1.07560664e-01 3.99771184e-01 -1.80634975e-01 -6.23792037e-02
-1.39126801e+00 8.27914417e-01 -1.67061746e-01 4.38615563e-04
-7.96809137e-01 -1.03944135e+00 -3.56993973e-01 1.58674285e-01
4.95256543e-01 -1.02170730e+00 8.10901225e-01 -1.06167507e+00
-8.00795197e-01 5.02937734e-01 -5.73811650e-01 -1.81062877e-01
3.68217796e-01 -3.51449810e-02 -2.69744247e-01 -2.60931015e-01
3.50706667e-01 7.70976245e-02 5.49485028e-01 -1.01714277e+00
-4.66613859e-01 -6.42312229e-01 -8.77547637e-03 -2.78728306e-01
-5.17310500e-02 2.14016378e-01 7.73804486e-02 -5.48006892e-01
1.80183947e-01 -8.32216918e-01 -3.86464536e-01 -4.38638955e-01
-5.80940425e-01 -3.51344258e-01 1.86071873e-01 -3.78862411e-01
1.28220344e+00 -2.00578761e+00 3.95300180e-01 3.59818667e-01
3.54847848e-01 -4.67904657e-01 6.55629709e-02 6.18729055e-01
-5.29919207e-01 6.09862804e-01 -3.15066665e-01 9.02844369e-02
1.64966509e-02 2.84229517e-01 -5.13139367e-01 6.44471824e-01
1.04269408e-01 8.51961970e-01 -8.69544148e-01 -2.94659227e-01
4.80274484e-02 2.72600114e-01 -7.17527330e-01 3.38729061e-02
5.58509827e-02 4.39360082e-01 -6.03568494e-01 4.53404009e-01
4.75060076e-01 -3.66675347e-01 5.97358227e-01 3.05857956e-01
-3.36543560e-01 3.73774081e-01 -1.21323776e+00 1.03677130e+00
-1.72084674e-01 3.62630635e-01 1.09447502e-02 -1.45078230e+00
4.84606057e-01 4.05258447e-01 4.99249846e-01 -3.48214626e-01
8.93663466e-02 1.69304937e-01 2.21770346e-01 -5.56273162e-01
3.05278987e-01 -6.75203562e-01 -1.64898977e-01 5.26239038e-01
-3.19986612e-01 3.16166162e-01 8.78303126e-02 1.06426969e-01
1.28134954e+00 -3.47439200e-01 4.67801124e-01 -4.70834404e-01
1.74534976e-01 -7.98440538e-03 5.82965016e-01 8.78970206e-01
7.81973600e-02 2.81105768e-02 8.39748859e-01 -1.83968559e-01
-7.54161954e-01 -1.26966035e+00 -3.63669783e-01 9.38871026e-01
8.76274258e-02 -1.54378757e-01 -5.60794353e-01 -3.53195578e-01
1.89487159e-01 8.16507638e-01 -9.36349630e-01 -2.15582192e-01
-3.50775093e-01 -1.12847674e+00 5.21048665e-01 4.31168705e-01
-8.28981847e-02 -7.86761999e-01 -4.73424256e-01 4.83663045e-02
-4.91398990e-01 -7.20107496e-01 -1.34695768e-01 4.33998033e-02
-1.18748784e+00 -1.38711345e+00 -1.20729461e-01 -8.24070945e-02
7.41861165e-01 3.46968025e-01 1.20637202e+00 7.97824264e-02
-3.39213759e-01 3.73763084e-01 -2.17172299e-02 -3.77910227e-01
-4.43917096e-01 -3.07398707e-01 1.87396914e-01 -1.58708051e-01
5.04630923e-01 -6.56844676e-01 -3.85583669e-01 4.14947510e-01
-8.74890685e-01 -7.12500960e-02 7.70163953e-01 7.10469007e-01
3.40719014e-01 9.08200815e-02 5.34320354e-01 -9.08595502e-01
6.69485092e-01 -8.67155015e-01 -5.37144423e-01 2.10651517e-01
-8.52184951e-01 3.08331698e-01 3.94394457e-01 -3.62210482e-01
-9.02017593e-01 -4.09027398e-01 3.56284112e-01 2.26043202e-02
-2.09910989e-01 9.88630474e-01 -1.94404542e-01 5.60548186e-01
4.29313958e-01 -1.41356751e-01 -7.24334270e-02 -3.65354270e-01
5.36946595e-01 7.73193687e-02 1.39020607e-01 -8.32944572e-01
5.39752722e-01 6.15421951e-01 4.15254384e-01 -7.67703474e-01
-6.81596160e-01 -2.28524685e-01 -4.65649635e-01 -3.82910445e-02
8.63841176e-01 -4.75425005e-01 -9.99105692e-01 2.28633713e-02
-1.02021861e+00 -7.66506940e-02 -1.38075978e-01 7.07437277e-01
-6.36871099e-01 -5.22076972e-02 -3.29438180e-01 -9.54311728e-01
3.53631526e-01 -9.34915185e-01 4.91162747e-01 -1.33521602e-01
-5.52982450e-01 -1.18883359e+00 4.02670681e-01 2.15652809e-01
-1.06843099e-01 1.37808621e-01 1.46002865e+00 -1.76884770e-01
-5.97707868e-01 -3.59636024e-02 -2.84175664e-01 -2.54587650e-01
2.84971714e-01 3.05998564e-01 -5.25712073e-01 1.14087649e-01
-1.02728687e-01 1.75811753e-01 7.45572746e-01 1.10974312e+00
9.31767166e-01 -4.16098565e-01 -5.30358315e-01 2.07623318e-01
1.26675737e+00 2.17261344e-01 5.67397714e-01 -2.05445394e-01
5.08467376e-01 1.03369701e+00 3.51022065e-01 4.22711372e-01
4.42066610e-01 6.12723947e-01 3.07975888e-01 -2.34455869e-01
3.21571261e-01 -3.25235724e-01 2.13292956e-01 5.93288958e-01
-4.90600318e-01 -1.22781917e-02 -8.38797688e-01 6.05026603e-01
-1.89387488e+00 -1.30627632e+00 -8.94734383e-01 2.34535003e+00
1.00372088e+00 -2.65163481e-01 2.91942865e-01 5.78060672e-02
7.22763062e-01 -2.63546050e-01 -2.58816838e-01 -5.05056024e-01
-1.53378099e-01 2.55009532e-01 6.74960017e-01 6.87132239e-01
-6.33459210e-01 6.18323445e-01 7.81088018e+00 3.28199297e-01
-8.63056958e-01 4.98121828e-02 6.59283459e-01 1.05467714e-01
-7.32286215e-01 3.30421656e-01 -5.54506004e-01 2.16389567e-01
1.11467671e+00 -2.98860162e-01 3.57538790e-01 4.62283403e-01
9.35635686e-01 -2.57213533e-01 -1.25929916e+00 4.15964812e-01
-5.29909372e-01 -9.77491975e-01 4.79138866e-02 4.09945011e-01
7.99320281e-01 -6.21329010e-01 1.82055697e-01 -1.09421745e-01
6.85449898e-01 -1.30774057e+00 4.75324541e-01 5.53814173e-01
5.58053017e-01 -7.50286043e-01 3.87390912e-01 4.07424331e-01
-5.61699092e-01 -4.89149019e-02 -4.25176680e-01 -8.27587306e-01
-4.26848941e-02 1.18101287e+00 -6.13960087e-01 5.85913360e-01
4.79012340e-01 5.00196874e-01 1.10531831e-02 4.57804143e-01
-4.66534078e-01 9.99605656e-01 -6.50086254e-02 -5.21735251e-02
-1.68892607e-01 -3.64354819e-01 3.85131598e-01 8.72383595e-01
3.70277911e-01 3.97181302e-01 -4.34123188e-01 1.31269670e+00
8.76468495e-02 7.26533383e-02 -5.61674297e-01 -2.87656695e-01
4.08380270e-01 7.65943050e-01 -5.85568130e-01 -1.86614528e-01
-4.57137197e-01 3.67631435e-01 7.95847327e-02 4.27890688e-01
-1.10033989e+00 3.64684731e-01 9.56251085e-01 2.88202703e-01
-1.10298589e-01 -2.86454380e-01 -7.83447444e-01 -1.14198029e+00
-2.40563303e-01 -8.02447557e-01 4.72881734e-01 -3.86829704e-01
-1.33033836e+00 -2.69450605e-01 3.85993302e-01 -4.75818425e-01
-3.36632103e-01 -3.13437879e-01 -3.53770554e-01 1.10693204e+00
-1.06853235e+00 -7.16163874e-01 1.43532991e-01 6.37466431e-01
8.63113403e-02 5.56056142e-01 8.82066429e-01 1.33729875e-01
-7.70811796e-01 1.52980402e-01 6.53568655e-02 -1.72263637e-01
6.38203740e-01 -1.25002205e+00 -1.08822897e-01 9.12046194e-01
4.05189255e-03 1.16317666e+00 1.11731303e+00 -7.76507914e-01
-1.48000622e+00 -6.10316694e-01 1.28769135e+00 -5.01639545e-01
9.67942894e-01 -1.59510180e-01 -6.83898389e-01 8.45766902e-01
-3.73829044e-02 -4.55741107e-01 8.33772063e-01 9.85684991e-01
-5.27638137e-01 1.57286637e-02 -7.00724065e-01 6.94194376e-01
1.11994076e+00 -1.80012688e-01 -5.55770636e-01 -5.62714525e-02
4.45072085e-01 2.79062361e-01 -7.45464563e-01 4.05986875e-01
7.92770267e-01 -1.14115322e+00 9.57233965e-01 -1.08705044e+00
9.92183506e-01 -2.94265300e-02 -3.90992798e-02 -1.25221562e+00
-3.41603994e-01 -3.67992043e-01 2.41022393e-01 1.04736614e+00
3.93174887e-01 -6.61890924e-01 4.57109660e-01 9.90746379e-01
2.56843358e-01 -4.19000119e-01 -9.17718530e-01 -6.65454090e-01
3.44237089e-01 -7.31892288e-01 4.88213509e-01 1.37287569e+00
1.14034474e-01 3.79865617e-01 -5.69642901e-01 3.01828116e-01
8.71677458e-01 3.67296964e-01 8.80159616e-01 -1.36944687e+00
-3.49571586e-01 -5.42022347e-01 -1.18296124e-01 -7.21129954e-01
1.74255490e-01 -5.37121773e-01 -1.64318547e-01 -1.23837841e+00
6.56633675e-01 -6.01416826e-01 -2.16771930e-01 2.51613975e-01
-4.68230307e-01 -2.22972438e-01 -1.13252275e-01 2.37080783e-01
3.84740680e-02 2.24484980e-01 1.11637712e+00 1.34906173e-01
-2.51534343e-01 5.72769418e-02 -9.68575835e-01 7.85311520e-01
6.09484613e-01 -7.18253434e-01 -4.84560728e-01 -2.21691042e-01
6.86026394e-01 5.00607133e-01 7.96682179e-01 -9.50964391e-02
-2.13112295e-01 -8.47181082e-01 9.83449817e-03 -3.35560083e-01
-2.42119297e-01 -7.45386779e-01 6.31668329e-01 6.19228721e-01
-5.22570431e-01 4.67129648e-02 -9.52240229e-02 4.72534508e-01
-4.57160696e-02 -2.29506530e-02 4.28075254e-01 5.48150018e-02
-2.50509202e-01 1.31769493e-01 -3.88164222e-01 1.16793342e-01
9.05183792e-01 2.03201860e-01 -3.80653381e-01 -4.51848358e-01
-6.39435589e-01 -5.90574183e-02 1.77884519e-01 2.86754698e-01
3.77504766e-01 -1.10325491e+00 -6.95255339e-01 -1.19721055e-01
-3.10041934e-01 -6.60478234e-01 3.91000211e-01 1.40058100e+00
9.71535891e-02 8.55023921e-01 6.68165088e-02 -3.72820228e-01
-1.20086825e+00 7.73946285e-01 9.52240825e-02 -2.63012707e-01
-1.00933954e-01 4.98501509e-01 8.92197311e-01 -2.98610143e-02
-1.61587387e-01 -4.77629155e-01 2.47939572e-01 -3.96099351e-02
4.53952163e-01 6.46442831e-01 -5.64048290e-01 -2.61451513e-01
-4.66011703e-01 3.95357311e-01 4.66563344e-01 -1.30768493e-01
1.32320821e+00 -3.42290550e-01 -6.41682029e-01 7.96801507e-01
8.49638641e-01 2.18526945e-01 -6.89971924e-01 9.63732153e-02
8.16640779e-02 -3.78661215e-01 -2.49719873e-01 -7.66641617e-01
-7.35482812e-01 7.95257568e-01 -1.30877702e-03 7.30910480e-01
1.10345411e+00 2.34707505e-01 -7.12712258e-02 -1.15704365e-01
3.11961591e-01 -5.62491238e-01 -2.77045816e-01 1.76085263e-01
7.28438139e-01 -9.11379278e-01 -3.56267393e-02 -7.41203845e-01
-1.67327985e-01 7.02888250e-01 2.25754566e-02 -1.24618746e-01
7.53947258e-01 2.17474967e-01 -2.04665646e-01 -3.60512108e-01
-9.00604248e-01 -6.86219782e-02 1.61410093e-01 4.19510216e-01
8.52699876e-01 5.36952496e-01 -1.06319797e+00 8.21436763e-01
-2.88602114e-01 2.81508211e-02 5.58630764e-01 2.69959986e-01
-1.49712369e-01 -1.14794898e+00 -6.23032987e-01 3.88023108e-01
-6.75572276e-01 -2.89183497e-01 -5.60105562e-01 1.15332878e+00
1.27697468e-01 1.23988485e+00 5.24547212e-02 -9.26264524e-02
9.13469717e-02 8.53320062e-02 4.89985317e-01 -3.11967015e-01
-1.88086703e-01 1.42028809e-01 4.40223068e-02 -6.76555097e-01
-9.70697761e-01 -9.94238019e-01 -1.10251737e+00 -9.15572107e-01
-3.07717949e-01 3.50823343e-01 4.62818503e-01 1.22034132e+00
9.62389112e-02 6.05515242e-01 6.05232894e-01 -1.83670828e-03
-4.08273667e-01 -7.94829071e-01 -8.01073670e-01 3.30757797e-01
1.89668551e-01 -9.55673337e-01 -5.70009112e-01 1.67077839e-01] | [7.869596004486084, 5.3755903244018555] |
5763216e-f16d-4c34-88e7-31bddb6e8c35 | helpfulness-and-fairness-of-task-oriented | 2205.12554 | null | https://arxiv.org/abs/2205.12554v3 | https://arxiv.org/pdf/2205.12554v3.pdf | Helpfulness and Fairness of Task-Oriented Dialogue Systems | Goal-oriented dialogue systems aim to help users achieve certain goals. Therefore, how humans perceive their helpfulness is important. However, neither the human-perceived helpfulness of goal-oriented dialogue systems nor its fairness implication has been well studied. In this paper, we study computational measurements of helpfulness. We first formally define a dialogue response as helpful if it is relevant & coherent, useful, and informative to a query. Then, we collect human annotations for the helpfulness of dialogue responses based on our definition and build a classifier to automatically determine the helpfulness of a response. We further propose to use the helpfulness level of a dialogue system towards different user queries to measure the fairness of a dialogue system. Experiments with state-of-the-art dialogue systems under three information-seeking scenarios reveal that existing systems tend to be more helpful for questions regarding concepts from highly-developed countries than less-developed countries, uncovering potential fairness concerns underlying the current goal-oriented dialogue systems. | ['Nanyun Peng', 'Jiin Kim', 'Yu Hou', 'Jiao Sun'] | 2022-05-25 | null | null | null | null | ['goal-oriented-dialogue-systems'] | ['natural-language-processing'] | [-3.78190160e-01 8.61502647e-01 -1.22502185e-02 -6.97287798e-01
-4.35127825e-01 -6.82767451e-01 9.38912809e-01 4.96257216e-01
-5.84513664e-01 7.01057732e-01 6.90965056e-01 -2.75332570e-01
1.16545483e-02 -5.20049751e-01 6.72889888e-01 -3.31559330e-01
2.53791511e-01 4.40235108e-01 9.47532058e-03 -9.58083153e-01
8.30035448e-01 3.39214094e-02 -6.86625302e-01 1.39448866e-01
1.22682822e+00 5.77901125e-01 -3.00341696e-01 7.72989750e-01
-3.16229343e-01 1.21811843e+00 -9.61080253e-01 -8.41667414e-01
1.53084695e-02 -7.91665912e-01 -1.61204362e+00 -6.11024760e-02
1.40270904e-01 -6.30035996e-01 1.36134729e-01 1.06966245e+00
5.52687109e-01 3.76497060e-01 7.49114692e-01 -1.36377287e+00
-6.08633757e-01 6.26591742e-01 -2.51425114e-02 1.65664300e-01
6.75542533e-01 1.46638840e-01 1.36494803e+00 -3.65549505e-01
4.70070809e-01 1.60183930e+00 3.24058235e-02 1.07230306e+00
-9.64608073e-01 -2.87351936e-01 -7.62480125e-02 -2.47286260e-01
-5.27847826e-01 -7.94762731e-01 6.51840270e-01 -5.97628534e-01
5.16252935e-01 6.14516258e-01 3.79996240e-01 5.61633527e-01
-3.78478020e-02 4.78299230e-01 1.09388530e+00 -3.75413150e-01
5.23604512e-01 5.53054333e-01 7.23548055e-01 5.39094746e-01
-1.74948603e-01 -4.82181400e-01 -4.66685206e-01 -7.05789864e-01
5.69840968e-02 -6.86966419e-01 -3.31848145e-01 1.89262539e-01
-5.46636701e-01 1.33854783e+00 2.52263546e-01 5.08119702e-01
-5.01824915e-01 -5.12633443e-01 6.47243619e-01 7.23556340e-01
7.87979126e-01 1.10447037e+00 -2.16632962e-01 -4.75927442e-01
-3.46493185e-01 4.81934130e-01 1.52972901e+00 4.97708052e-01
5.83796740e-01 -2.88708419e-01 -6.06783092e-01 1.09290779e+00
2.50297934e-01 -1.68755651e-02 4.89531845e-01 -1.45858705e+00
1.47067308e-01 1.02285576e+00 4.57242548e-01 -1.08243573e+00
-5.06756544e-01 2.79845536e-01 -6.72348797e-01 -5.64106703e-02
6.71639025e-01 -5.76319277e-01 3.13938200e-01 1.76572347e+00
3.27795535e-01 -1.15018904e+00 2.74762809e-01 1.11730766e+00
1.00758529e+00 2.41175428e-01 2.67395884e-01 -4.59072202e-01
1.27832663e+00 -7.71050513e-01 -7.65624225e-01 -2.18148995e-02
8.73928010e-01 -7.23515689e-01 1.40098035e+00 7.02920333e-02
-1.08902407e+00 -1.12088472e-01 -4.51350123e-01 -1.97417602e-01
-7.30265528e-02 -1.94579676e-01 3.85888070e-01 1.04089737e+00
-9.83284533e-01 4.29330111e-01 1.42318726e-01 -5.71842253e-01
-5.97559512e-02 -1.50190562e-01 -1.75621361e-01 4.20414060e-01
-1.40380371e+00 1.23616827e+00 -1.97993331e-02 -4.88730460e-01
-2.85208225e-01 -4.08748657e-01 -6.24046862e-01 1.64586037e-01
3.40362310e-01 -4.82703626e-01 1.67792118e+00 -7.60119915e-01
-1.62616432e+00 1.09689081e+00 -4.47417349e-02 -3.71987939e-01
5.70570350e-01 -1.49896920e-01 -9.68724042e-02 2.70749837e-01
3.46999824e-01 5.12999773e-01 4.47244883e-01 -9.03784215e-01
-6.44897938e-01 -4.31093514e-01 6.23170137e-01 6.60872400e-01
-4.99498963e-01 3.69207323e-01 1.18972100e-01 -1.47208139e-01
-2.82929599e-01 -5.79550385e-01 -2.98591197e-01 -2.36681536e-01
-5.11722207e-01 -9.67072129e-01 4.81524676e-01 -6.21376276e-01
1.36206067e+00 -1.79222333e+00 -2.85079032e-01 -1.10889629e-01
6.50192022e-01 2.60223210e-01 -1.80795677e-02 6.79779232e-01
3.64129961e-01 5.42147696e-01 1.01035938e-01 3.04709729e-02
3.13237965e-01 -9.40794498e-02 -2.40841713e-02 2.03818947e-01
2.00125948e-02 5.81443012e-01 -1.15313876e+00 -5.52814603e-01
6.30300418e-02 -2.07618758e-01 -5.48710883e-01 8.01740170e-01
-2.00691178e-01 3.56021225e-01 -6.67292893e-01 2.52328902e-01
3.57762277e-01 6.75639361e-02 1.35047957e-01 6.93675950e-02
-1.26300663e-01 7.49287486e-01 -3.69389385e-01 9.29584622e-01
-3.83271813e-01 5.43475807e-01 3.11958671e-01 -7.52026200e-01
1.14558887e+00 2.46825323e-01 1.48103833e-01 -9.59842086e-01
2.22755775e-01 -1.65898574e-03 4.23499167e-01 -6.59355700e-01
1.06742477e+00 -2.25001752e-01 -1.87379897e-01 1.04253936e+00
-3.48190069e-01 -5.32624483e-01 2.90920675e-01 8.35606515e-01
7.53344476e-01 -6.17196441e-01 7.08224893e-01 -6.60595536e-01
8.44645321e-01 -1.17507994e-01 2.45662823e-01 7.52743363e-01
-8.53382766e-01 6.50112927e-02 1.21665919e+00 -5.62385693e-02
-6.63426101e-01 -2.14743495e-01 2.89761662e-01 1.67214334e+00
1.69619337e-01 -5.63154101e-01 -1.08952761e+00 -8.95429373e-01
-2.59653300e-01 1.27409983e+00 -3.88511270e-01 -2.64000833e-01
-7.10092112e-02 -3.27570945e-01 7.47162640e-01 -3.58260423e-01
8.14131141e-01 -8.74395072e-01 -9.02714133e-01 5.88751845e-02
-7.30047643e-01 -7.68606305e-01 -5.17780602e-01 -4.29595321e-01
-4.43238735e-01 -1.39657891e+00 -6.88035429e-01 -2.28135586e-01
3.47944051e-01 3.32403094e-01 1.41078711e+00 4.90206063e-01
2.57580847e-01 5.32613814e-01 -5.92348874e-01 -2.83693314e-01
-9.63923752e-01 -8.96302834e-02 2.02241279e-02 -3.13752145e-01
5.79281390e-01 8.02694112e-02 -6.37733221e-01 5.89276850e-01
-4.98888880e-01 -2.31170058e-01 -2.72126406e-01 7.92193472e-01
-6.74194157e-01 -2.18812808e-01 8.96684885e-01 -1.04149854e+00
2.06374049e+00 -6.09140158e-01 -1.08319530e-02 3.77679259e-01
-8.95842016e-01 -4.98746671e-02 7.42541969e-01 1.52724132e-01
-1.47419262e+00 -8.44668329e-01 -3.13094586e-01 7.32135296e-01
-2.36519828e-01 3.34005207e-01 -8.01738054e-02 1.17607620e-02
1.06185806e+00 -3.04564267e-01 1.65454581e-01 -1.53752938e-01
5.99769771e-01 1.15001702e+00 -7.31168268e-03 -9.27947402e-01
2.46254891e-01 -3.07294391e-02 -5.69490433e-01 -1.22990096e+00
-6.91173077e-01 -8.12120378e-01 -1.13709942e-01 -5.87686777e-01
5.10813594e-01 -2.93458223e-01 -1.31072676e+00 2.65597343e-01
-1.25998783e+00 -2.44487584e-01 4.24140394e-02 -5.54802008e-02
-4.90168065e-01 8.40841413e-01 -7.07844257e-01 -1.46050727e+00
-9.25034344e-01 -8.82051885e-01 4.75111216e-01 4.70557332e-01
-1.06806636e+00 -1.07444572e+00 1.67582870e-01 6.43809438e-01
6.42474949e-01 -2.17170328e-01 1.06155705e+00 -1.14344084e+00
4.71779704e-01 -8.64775851e-03 -2.92879283e-01 1.94073394e-01
3.69624346e-01 -1.55088142e-01 -7.91718006e-01 7.41850724e-03
5.83129227e-01 -8.41106713e-01 2.67668873e-01 1.23267375e-01
4.48718220e-01 -1.05863965e+00 3.25069696e-01 -4.72124696e-01
7.29719400e-01 2.13183556e-02 4.72225159e-01 1.74525604e-01
-3.50624807e-02 1.53451324e+00 1.12101150e+00 7.78488636e-01
6.05341196e-01 6.03460312e-01 1.81873277e-01 2.64436871e-01
5.23986757e-01 -7.94187337e-02 2.18628854e-01 3.23246539e-01
1.16603523e-01 -2.10891262e-01 -8.43837857e-01 3.64163190e-01
-1.72464085e+00 -8.65572393e-01 -2.94080764e-01 2.05748129e+00
1.17110717e+00 -1.73880368e-01 4.99912232e-01 -9.13868025e-02
6.01903856e-01 2.61214644e-01 -3.57107460e-01 -9.45189834e-01
1.79592848e-01 -3.70685220e-01 -3.01577467e-02 8.78014207e-01
-6.74777925e-01 1.11244893e+00 6.07936811e+00 3.40384692e-01
-6.54536605e-01 -2.00729147e-02 1.18680596e+00 3.43298316e-01
-4.31784004e-01 -1.02402292e-01 -2.50219792e-01 2.55603611e-01
7.78665185e-01 -9.27923799e-01 1.23989739e-01 9.15857971e-01
6.26631856e-01 -4.90501165e-01 -1.09005535e+00 6.02481365e-01
-6.45262897e-02 -7.82748163e-01 -1.75971642e-01 -1.19458824e-01
1.93250909e-01 -6.13716185e-01 -1.85358420e-01 3.71168286e-01
7.15030372e-01 -8.55192423e-01 1.85953781e-01 2.95304030e-01
2.24751517e-01 -5.07100403e-01 6.85016692e-01 8.58767271e-01
-3.57546270e-01 2.99058110e-01 -5.08548617e-01 -5.24427116e-01
1.33103624e-01 3.77764314e-01 -1.13079906e+00 1.70586422e-01
3.23891610e-01 2.29002789e-01 -5.78068793e-01 4.03495044e-01
-2.17536777e-01 4.52467114e-01 6.48756400e-02 -5.75624883e-01
2.60315955e-01 -4.04749930e-01 5.16011655e-01 1.03148723e+00
-2.48235419e-01 6.02440476e-01 9.41854641e-02 9.20163989e-01
-1.45346880e-01 5.94157875e-01 -7.28946567e-01 -2.05535695e-01
5.31410694e-01 1.50962663e+00 -3.56317192e-01 -3.60135794e-01
-7.86743537e-02 8.19359958e-01 3.92442852e-01 5.53615466e-02
-1.58916697e-01 -1.43820390e-01 5.52250922e-01 -2.07938075e-01
-9.98325169e-01 2.16054082e-01 -5.03263056e-01 -9.11002755e-01
-2.11660862e-01 -1.19148135e+00 5.07958472e-01 -4.90314454e-01
-1.41263986e+00 4.72637445e-01 -3.59290659e-01 -6.75931513e-01
-5.87659955e-01 -2.45660692e-01 -8.40480983e-01 1.21681821e+00
-1.10165930e+00 -4.38043624e-01 -3.91501069e-01 4.01806831e-01
6.26070559e-01 -2.66900212e-01 1.05059648e+00 -3.13089281e-01
-3.04412127e-01 4.56414729e-01 -5.10292411e-01 2.60354489e-01
1.12109613e+00 -1.45618510e+00 3.26459557e-01 5.65620184e-01
-5.23201764e-01 9.37334478e-01 9.34327543e-01 -5.27718544e-01
-7.96912789e-01 -3.91684145e-01 1.24552250e+00 -3.44983637e-01
5.51522672e-01 3.35463405e-01 -9.44905579e-01 -2.33432911e-02
7.56223440e-01 -9.35198843e-01 1.03079557e+00 4.11253422e-01
-3.42214227e-01 5.16050875e-01 -1.49558616e+00 7.53483713e-01
5.94666362e-01 -6.77040815e-01 -9.38138664e-01 4.22452360e-01
4.39142555e-01 -3.15044150e-02 -6.69080317e-01 -2.42086291e-01
2.84385949e-01 -1.49513292e+00 6.19693160e-01 -9.16698575e-01
6.96583450e-01 4.46020037e-01 -2.60176100e-02 -1.56533217e+00
-1.86775669e-01 -7.42879868e-01 3.63525480e-01 1.31918001e+00
1.37928292e-01 -6.37254834e-01 6.49846375e-01 1.80149901e+00
2.47462824e-01 -2.58225918e-01 -5.67276239e-01 -2.03163922e-01
5.24457991e-01 1.36236474e-01 3.02671015e-01 1.31172037e+00
1.10541415e+00 9.20622647e-01 -4.59317982e-01 -4.16603625e-01
4.55602020e-01 -1.63832083e-01 1.02277958e+00 -1.31490576e+00
1.58680797e-01 -8.98361444e-01 2.09483460e-01 -9.67765570e-01
4.22839314e-01 -3.33137602e-01 9.24356654e-02 -1.45516491e+00
2.35157803e-01 -3.88346881e-01 5.17053187e-01 2.16921255e-01
-6.98074162e-01 -4.11924183e-01 2.72982925e-01 3.04852366e-01
-7.02122986e-01 5.67177355e-01 9.73987639e-01 -1.40919924e-01
-4.02886212e-01 1.99651927e-01 -1.39229810e+00 6.46095097e-01
1.09501266e+00 -4.47742231e-02 -4.31393415e-01 2.04598352e-01
1.04218021e-01 3.81777227e-01 3.54352444e-02 -2.54560173e-01
1.79777429e-01 -6.27468646e-01 -4.94880676e-01 -2.68021300e-02
-9.66530219e-02 -3.85976225e-01 -7.03772128e-01 5.11393368e-01
-1.07773089e+00 -1.43230110e-01 -4.68725353e-01 1.61468059e-01
-2.96263039e-01 -7.23332465e-01 8.49639654e-01 -3.23462278e-01
-4.91433084e-01 -8.97514448e-02 -7.60647774e-01 5.80908477e-01
7.19418228e-01 2.44100943e-01 -5.99581838e-01 -1.19890487e+00
-1.17144808e-01 6.29443288e-01 5.35984635e-01 4.71828431e-01
5.12755334e-01 -7.86656082e-01 -9.54092324e-01 -4.81544018e-01
3.63582879e-01 -6.88903332e-01 6.92593157e-02 5.07292628e-01
-3.16543996e-01 5.80717623e-01 -3.28899354e-01 -6.85147345e-02
-1.52477777e+00 6.35168031e-02 5.16318917e-01 -2.71146744e-01
-9.92469117e-02 7.46766746e-01 2.48782739e-01 -6.21089041e-01
3.00242454e-01 4.16593105e-01 -7.98677385e-01 4.68378961e-01
7.14838266e-01 7.16249704e-01 -1.78499043e-01 -7.23728776e-01
-9.57927629e-02 -2.28458390e-01 -1.00137770e-01 -4.96798962e-01
6.10812187e-01 -4.66012061e-01 -3.63290697e-01 2.72803575e-01
9.31491435e-01 -3.86933312e-02 -4.81119126e-01 -2.30139524e-01
3.09082836e-01 -6.32239461e-01 -2.98339278e-01 -8.75951231e-01
-5.37219167e-01 8.25798273e-01 -5.66571280e-02 1.09970915e+00
7.13425159e-01 -2.79158503e-01 3.54310840e-01 6.73930705e-01
2.85246789e-01 -1.55054569e+00 2.42034063e-01 6.66021168e-01
1.04956317e+00 -1.57071805e+00 -1.98216870e-01 -3.18682134e-01
-1.25691605e+00 1.16663909e+00 9.27661002e-01 5.42442262e-01
3.64095658e-01 -5.52719831e-01 5.23534596e-01 -3.49565715e-01
-8.04868758e-01 -1.40413061e-01 1.50665134e-01 4.32297111e-01
1.16971552e+00 2.24751845e-01 -1.12644970e+00 5.68233967e-01
-5.28809726e-01 -2.86100745e-01 8.95341814e-01 5.74684918e-01
-6.75065637e-01 -8.41960311e-01 -1.01059355e-01 4.06733960e-01
-4.03365701e-01 -1.43992841e-01 -1.42797792e+00 1.71380982e-01
-8.59200835e-01 1.93506420e+00 -3.49733382e-01 -3.95316780e-01
3.73971194e-01 1.49193585e-01 -1.98664814e-01 -7.58059442e-01
-1.21553516e+00 -3.97103816e-01 8.75894010e-01 -4.48017538e-01
-5.45465350e-01 -2.33237147e-01 -1.02789962e+00 -8.14251304e-01
-3.75913769e-01 8.18122387e-01 1.92860171e-01 8.57949674e-01
1.83407754e-01 -2.34143794e-01 8.90168071e-01 -1.94680709e-02
-9.52832162e-01 -1.28612506e+00 -4.69921291e-01 8.76704991e-01
8.65970030e-02 -1.58713982e-01 -4.75042760e-01 -5.02951086e-01] | [12.770588874816895, 8.126909255981445] |
ca164e46-7f5e-4d32-9cef-dd9f2050dc8c | hybrid-transformers-for-music-source | 2211.08553 | null | https://arxiv.org/abs/2211.08553v1 | https://arxiv.org/pdf/2211.08553v1.pdf | Hybrid Transformers for Music Source Separation | A natural question arising in Music Source Separation (MSS) is whether long range contextual information is useful, or whether local acoustic features are sufficient. In other fields, attention based Transformers have shown their ability to integrate information over long sequences. In this work, we introduce Hybrid Transformer Demucs (HT Demucs), an hybrid temporal/spectral bi-U-Net based on Hybrid Demucs, where the innermost layers are replaced by a cross-domain Transformer Encoder, using self-attention within one domain, and cross-attention across domains. While it performs poorly when trained only on MUSDB, we show that it outperforms Hybrid Demucs (trained on the same data) by 0.45 dB of SDR when using 800 extra training songs. Using sparse attention kernels to extend its receptive field, and per source fine-tuning, we achieve state-of-the-art results on MUSDB with extra training data, with 9.20 dB of SDR. | ['Alexandre Défossez', 'Francisco Massa', 'Simon Rouard'] | 2022-11-15 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 2.79858232e-01 -2.12566108e-01 9.77528095e-03 5.08403108e-02
-1.54536283e+00 -7.23140001e-01 2.48134226e-01 -9.44958031e-02
-3.00071150e-01 5.47759116e-01 6.53699398e-01 -2.12614685e-02
-1.07632391e-01 -3.69417012e-01 -8.05197001e-01 -5.30127287e-01
-2.58355960e-02 1.85299829e-01 3.03571582e-01 -3.29035044e-01
-2.37108245e-01 -7.12973773e-02 -1.48926365e+00 4.60105747e-01
6.96600676e-01 1.43320680e+00 3.68037939e-01 9.25015271e-01
2.39109993e-01 9.54500258e-01 -5.70849478e-01 -2.71609336e-01
1.86901703e-01 -7.18909740e-01 -8.59997153e-01 -2.96723753e-01
7.11674929e-01 -2.10898280e-01 -2.78688133e-01 8.62076402e-01
9.88264322e-01 2.35932067e-01 4.79562372e-01 -8.70231688e-01
-8.11259747e-01 1.23301840e+00 -3.65663350e-01 3.58019829e-01
2.75945365e-01 -2.33950064e-01 1.63264620e+00 -8.82852912e-01
1.02793790e-01 1.17218065e+00 1.01410878e+00 3.11999232e-01
-1.54114139e+00 -8.86170745e-01 -1.23837434e-01 3.43275517e-01
-1.14703131e+00 -8.16638470e-01 7.77498186e-01 -3.09466094e-01
1.12336957e+00 2.36612782e-01 2.58915454e-01 1.27191508e+00
-2.03718275e-01 9.57277000e-01 9.20972288e-01 -5.19416332e-01
1.33800626e-01 -1.28152862e-01 1.97534889e-01 7.37168863e-02
-4.76895690e-01 1.02451190e-01 -9.67798948e-01 -1.10149972e-01
6.39061153e-01 -5.63782752e-01 -4.21289474e-01 1.42547250e-01
-1.22108185e+00 6.90881670e-01 5.11160493e-01 6.06640935e-01
-2.35259831e-01 2.75416821e-01 5.20837843e-01 6.25181973e-01
5.57897568e-01 6.60381377e-01 -6.42306507e-01 -5.34283280e-01
-1.18126440e+00 1.89235918e-02 4.34674233e-01 8.06514561e-01
4.71255988e-01 5.29063880e-01 -3.89052153e-01 1.48763227e+00
-1.87763944e-01 5.82561135e-01 8.69978547e-01 -1.04607117e+00
2.45978892e-01 -1.40313089e-01 -2.82135487e-01 -4.93733227e-01
-8.07073340e-02 -9.71706986e-01 -7.99646974e-01 -1.04601644e-01
1.73171818e-01 -2.83796012e-01 -8.56340051e-01 2.07862377e+00
-3.05041164e-01 4.74344850e-01 8.98178518e-02 8.73445570e-01
7.54918158e-01 8.29668760e-01 -3.34132373e-01 -7.08985999e-02
1.11492443e+00 -1.16984499e+00 -6.96949601e-01 -3.72934043e-01
9.38443933e-03 -9.08729553e-01 1.12186587e+00 6.81974053e-01
-1.29025507e+00 -1.03475773e+00 -1.15526736e+00 -2.35696793e-01
-1.19233429e-01 2.99400270e-01 1.22835435e-01 4.62038338e-01
-1.26937711e+00 8.29954207e-01 -4.61314619e-01 -1.74980149e-01
1.47444129e-01 3.86138529e-01 -8.79278183e-02 2.60269731e-01
-1.51594663e+00 3.32497507e-01 1.56019673e-01 -3.15276325e-01
-1.12160075e+00 -9.04078364e-01 -7.31944144e-01 3.73490870e-01
1.36035144e-01 -4.00274307e-01 1.42109251e+00 -1.09219754e+00
-1.74138534e+00 4.95259464e-01 -1.60864834e-02 -8.97407532e-01
-9.27491172e-04 -6.22992635e-01 -6.26169682e-01 6.53996691e-02
2.42621347e-01 6.81066155e-01 1.01257861e+00 -8.54610980e-01
-6.47749007e-01 -1.89107597e-01 -8.72391686e-02 2.04257637e-01
-4.06041056e-01 1.26401797e-01 -3.69803041e-01 -1.15140915e+00
-7.68928230e-02 -9.39200997e-01 1.32990792e-01 -6.10402167e-01
-2.59356171e-01 -3.69253121e-02 7.37245619e-01 -9.70696330e-01
1.32946742e+00 -2.64346862e+00 4.45221871e-01 -1.72391430e-01
-1.81013450e-01 2.66044706e-01 -5.11356175e-01 2.17796028e-01
-1.65555730e-01 -2.89084297e-02 -3.84439290e-01 -3.92442316e-01
1.41219765e-01 -1.33827806e-01 -5.86055815e-01 3.91752236e-02
1.21452853e-01 7.19209850e-01 -7.69645095e-01 -1.03283040e-01
-1.91673730e-02 6.43592656e-01 -7.37976372e-01 1.64917111e-01
-1.69045404e-02 3.08018804e-01 1.90412655e-01 3.85397792e-01
4.38455582e-01 -7.83515070e-03 9.99221578e-02 -2.02134699e-01
1.14113055e-01 6.96160436e-01 -1.04485083e+00 2.14556122e+00
-7.71221936e-01 1.03614485e+00 3.03483874e-01 -8.30109775e-01
9.13800120e-01 5.82647085e-01 5.33492684e-01 -7.73506939e-01
3.76909114e-02 5.55215955e-01 1.17524818e-01 7.15460330e-02
4.26485240e-01 -3.87063473e-01 -1.41631380e-01 3.52722198e-01
8.12810659e-01 -3.32809240e-02 1.91491190e-02 8.67894590e-02
1.25695479e+00 6.79698288e-02 -4.83163260e-02 -2.48018324e-01
2.59274930e-01 -4.72204447e-01 4.10766602e-01 6.69077277e-01
-1.11354347e-02 9.36289549e-01 4.04774636e-01 2.16600418e-01
-9.40728843e-01 -1.27825701e+00 -3.02069098e-01 1.53242493e+00
-1.78926468e-01 -6.16947234e-01 -7.14403212e-01 -3.10032129e-01
-2.29503848e-02 6.59634471e-01 -5.41502357e-01 -1.65323749e-01
-2.93234915e-01 -4.62394983e-01 8.59099030e-01 6.82006836e-01
3.91759187e-01 -1.04926777e+00 -3.84606481e-01 6.06962681e-01
-5.87021470e-01 -1.05286503e+00 -6.84248984e-01 8.12993288e-01
-5.92043996e-01 -5.44914603e-01 -1.16699541e+00 -6.84267879e-01
-3.45694244e-01 1.82887837e-01 1.29642320e+00 -8.26373875e-01
1.67540595e-01 1.62365526e-01 -6.47510827e-01 -2.18943328e-01
-1.89439029e-01 4.04292524e-01 1.72112420e-01 1.17765181e-01
7.25387111e-02 -1.15575182e+00 -1.98454022e-01 2.76786834e-01
-6.33361995e-01 -2.88706779e-01 6.01435304e-01 1.11328971e+00
5.04846036e-01 -4.89493236e-02 1.03120792e+00 -4.98210937e-01
6.86080635e-01 -3.54167789e-01 -1.92602083e-01 -1.32474482e-01
-2.73971856e-01 1.11415103e-01 5.68335176e-01 -5.06753743e-01
-8.11250925e-01 -1.95555538e-01 -4.59481210e-01 -8.35813463e-01
-5.66133410e-02 3.20166469e-01 -5.24348915e-02 2.19945952e-01
9.10008252e-01 2.85371542e-01 -3.00386786e-01 -9.68213201e-01
2.64694750e-01 1.10923779e+00 7.43237019e-01 -5.27404010e-01
3.29466701e-01 -6.52344823e-02 -6.52334213e-01 -9.16200876e-01
-1.06816030e+00 -7.35593617e-01 -4.58033532e-01 1.29290491e-01
9.62362647e-01 -1.11940014e+00 -4.20814812e-01 3.49719763e-01
-8.54484200e-01 -6.48945153e-01 -4.80252773e-01 5.62269509e-01
-7.52639949e-01 -5.95815666e-02 -8.80185246e-01 -7.81766176e-01
-4.19975042e-01 -9.33305919e-01 1.25584066e+00 -2.87266433e-01
-3.35537642e-01 -5.18295288e-01 4.06525701e-01 3.69730949e-01
6.05776191e-01 -1.76473677e-01 5.32236338e-01 -7.15085983e-01
-1.73426479e-01 2.35676467e-01 -5.80118075e-02 8.93804789e-01
2.71566004e-01 -6.24281824e-01 -1.66282964e+00 -1.99767068e-01
-1.42824888e-01 -5.78818440e-01 1.24854600e+00 6.19974792e-01
1.11424375e+00 -2.60240287e-01 1.07524469e-01 7.11785197e-01
1.16036189e+00 3.24853182e-01 6.17655814e-01 2.00744290e-02
4.91280705e-01 1.40486673e-01 2.49903291e-01 3.21516454e-01
-4.23146859e-02 8.56680930e-01 2.03157842e-01 -1.09605253e-01
-5.16011894e-01 -2.55845368e-01 8.22432876e-01 1.13277125e+00
-5.25485314e-02 -1.82689413e-01 -5.26314378e-01 7.85388231e-01
-1.62962437e+00 -1.02598906e+00 2.64450729e-01 2.21428561e+00
1.14160109e+00 1.24350049e-01 3.78563523e-01 5.44444561e-01
4.83266711e-01 3.36044908e-01 -5.22132158e-01 -3.85338247e-01
-4.69485879e-01 5.95833361e-01 2.49519020e-01 5.22416711e-01
-1.23094416e+00 7.76223540e-01 6.30046892e+00 1.32411075e+00
-1.20264983e+00 3.72712255e-01 4.52438742e-01 -3.83008868e-01
-1.90445095e-01 -3.80799860e-01 -4.37745124e-01 4.20695215e-01
1.29200351e+00 9.08669680e-02 7.87770033e-01 5.97163916e-01
-2.76581049e-01 1.02712288e-01 -1.07957685e+00 1.02158833e+00
1.96672156e-01 -1.01325881e+00 -3.37095529e-01 -1.22280948e-01
7.13288724e-01 4.09899056e-01 2.95192540e-01 5.78550935e-01
4.77682561e-01 -9.22100663e-01 9.34888899e-01 5.98806106e-02
1.07605970e+00 -7.71220326e-01 7.02391148e-01 1.12947322e-01
-1.41548431e+00 -1.05199128e-01 -3.19156110e-01 1.04406148e-01
-4.11310382e-02 4.99445111e-01 -7.49935925e-01 6.84558690e-01
1.14836669e+00 9.92925882e-01 -4.00153309e-01 8.26403797e-01
3.33187804e-02 1.09447801e+00 -4.51355219e-01 3.55600268e-01
2.03473032e-01 2.07300752e-01 7.49976814e-01 1.47769988e+00
5.15691459e-01 -2.85510093e-01 -1.82168350e-01 5.98309755e-01
-1.34632230e-01 1.06691971e-01 -3.72046739e-01 -2.54895110e-02
2.22315431e-01 9.32871222e-01 -7.58449361e-02 -3.20946455e-01
-2.21479073e-01 1.18558812e+00 2.33695701e-01 4.43850905e-01
-8.05557609e-01 -5.40520072e-01 8.58581066e-01 -4.70639998e-03
6.65074170e-01 4.39083725e-02 1.95250809e-02 -1.10108471e+00
-2.54208922e-01 -1.11932421e+00 4.34473962e-01 -1.07407570e+00
-1.33968627e+00 9.84826207e-01 -4.71987605e-01 -1.40204382e+00
-3.14009309e-01 -4.39480275e-01 -3.09295386e-01 9.35035646e-01
-1.31119931e+00 -1.12846208e+00 1.87789723e-01 7.40970910e-01
7.08281100e-01 -2.51744866e-01 9.82154906e-01 6.79981828e-01
-1.70060545e-01 8.14736009e-01 4.16655511e-01 3.27883840e-01
1.09865427e+00 -1.48865199e+00 2.75974751e-01 6.45134866e-01
8.06625903e-01 2.63797373e-01 4.86665845e-01 -2.07616389e-01
-9.34054434e-01 -9.96942461e-01 7.34509587e-01 -1.22634672e-01
7.74615884e-01 -5.55766761e-01 -8.29830348e-01 7.39403605e-01
6.99640512e-01 -7.84249902e-02 8.79934490e-01 6.79655254e-01
-7.29519367e-01 -3.87850493e-01 -6.27659678e-01 2.97324777e-01
1.05017567e+00 -9.57916021e-01 -6.66961908e-01 -1.34314820e-01
1.12652469e+00 -2.83258259e-01 -9.26533937e-01 4.08080518e-01
6.13237798e-01 -1.16661847e+00 1.11668694e+00 -3.30015838e-01
4.31295574e-01 -1.66479304e-01 -7.56215155e-01 -1.64776576e+00
-6.35571063e-01 -7.34088838e-01 5.42934313e-02 1.48512030e+00
6.59174442e-01 -3.18095446e-01 2.39435628e-01 -4.90078211e-01
-5.38619757e-01 -4.40103233e-01 -1.02244115e+00 -1.05946505e+00
1.84607551e-01 -8.54952037e-01 4.23751920e-01 9.09489155e-01
2.67961144e-01 7.86698937e-01 -7.39345610e-01 -5.88133335e-02
4.38054949e-01 2.10285187e-01 3.96796882e-01 -1.12274170e+00
-8.58174741e-01 -5.32473564e-01 -2.50990421e-01 -1.14851725e+00
4.09415066e-02 -9.01526630e-01 1.36530623e-01 -1.32909334e+00
-5.45230284e-02 -2.95641154e-01 -8.72215509e-01 5.89711368e-01
5.78420274e-02 5.50709128e-01 3.95769954e-01 -2.07213648e-02
-6.36720598e-01 7.73985922e-01 8.19702625e-01 -4.00545657e-01
-2.82096356e-01 3.98720726e-02 -7.82765508e-01 5.07637739e-01
7.59648263e-01 -2.28930771e-01 -3.55287850e-01 -4.73844528e-01
-1.64768454e-02 2.25463167e-01 1.92099527e-01 -1.66603911e+00
1.00583799e-01 4.60773319e-01 1.75590321e-01 -5.96791089e-01
8.20072949e-01 -6.20972812e-01 1.80072673e-02 2.83561666e-02
-6.51591659e-01 -4.64839101e-01 5.25674760e-01 5.23876190e-01
-8.12060833e-01 5.18060029e-02 8.14848423e-01 1.38718992e-01
-4.49731529e-01 4.06642333e-02 -3.15780371e-01 3.99684310e-01
4.01081055e-01 2.10190818e-01 1.16797745e-01 -6.51239395e-01
-1.07120025e+00 -1.79342657e-01 -1.07289795e-02 4.68010545e-01
2.65625924e-01 -1.70916021e+00 -8.70764792e-01 2.56472737e-01
6.49343133e-02 -3.39487731e-01 5.43518364e-01 6.05641186e-01
1.70452878e-01 4.93898779e-01 -1.47979081e-01 -6.63326085e-01
-1.19518757e+00 5.35689116e-01 3.59370649e-01 -2.46151924e-01
-3.07449251e-01 1.27597713e+00 3.96292955e-01 -3.03806156e-01
4.14506465e-01 -4.62173164e-01 -1.04407288e-01 4.35326219e-01
4.40184742e-01 3.11439097e-01 2.27930725e-01 -5.97558796e-01
-3.63262564e-01 6.93078279e-01 3.68674934e-01 -6.17684722e-01
1.29475224e+00 -1.58131525e-01 2.04440072e-01 9.01437700e-01
1.24133992e+00 5.82914412e-01 -1.20612490e+00 -6.96551204e-01
-3.15434366e-01 -1.89012825e-01 3.29949528e-01 -1.04929507e+00
-1.06769574e+00 1.24043918e+00 5.58509171e-01 4.84778136e-01
1.64587367e+00 1.54363573e-01 9.13647830e-01 5.42601012e-02
1.25945851e-01 -9.36746955e-01 2.19710276e-01 7.74767697e-01
1.24415648e+00 -1.00697112e+00 -5.07509589e-01 9.43668559e-02
-8.15691054e-01 8.04356813e-01 2.85272300e-01 -1.70839518e-01
8.59276533e-01 5.40069342e-01 5.46882451e-02 4.20496374e-01
-7.81056464e-01 -5.64347267e-01 5.45018256e-01 5.18291652e-01
5.95034480e-01 1.45503104e-01 4.36886400e-01 1.04816401e+00
-5.90235353e-01 -2.01833397e-01 7.00837672e-02 3.32327425e-01
-5.16838253e-01 -1.07242727e+00 -3.85271996e-01 3.69635880e-01
-9.04067516e-01 -6.34516001e-01 -4.25538421e-01 2.15088889e-01
1.42279938e-01 1.13536215e+00 1.10131666e-01 -7.40990758e-01
2.34718278e-01 1.69876978e-01 3.19456339e-01 -5.48615217e-01
-8.60782564e-01 9.17204082e-01 2.46626526e-01 -3.56478035e-01
-3.94132376e-01 -6.63551569e-01 -8.22934747e-01 2.78956629e-02
-1.97759509e-01 3.41726601e-01 1.39870688e-01 6.20754838e-01
4.62878793e-01 1.10902786e+00 5.88146806e-01 -8.47628355e-01
-3.95525396e-01 -1.33964014e+00 -8.34980607e-01 6.30141050e-02
8.46790135e-01 -4.43030536e-01 -3.55821073e-01 2.28432238e-01] | [15.403169631958008, 5.467497825622559] |
497511f8-4d30-472d-9ede-b2fd3b8ac5cc | fmri-from-eeg-is-only-deep-learning-away-the | 2211.02024 | null | https://arxiv.org/abs/2211.02024v2 | https://arxiv.org/pdf/2211.02024v2.pdf | fMRI from EEG is only Deep Learning away: the use of interpretable DL to unravel EEG-fMRI relationships | The access to activity of subcortical structures offers unique opportunity for building intention dependent brain-computer interfaces, renders abundant options for exploring a broad range of cognitive phenomena in the realm of affective neuroscience including complex decision making processes and the eternal free-will dilemma and facilitates diagnostics of a range of neurological deceases. So far this was possible only using bulky, expensive and immobile fMRI equipment. Here we present an interpretable domain grounded solution to recover the activity of several subcortical regions from the multichannel EEG data and demonstrate up to 60% correlation between the actual subcortical blood oxygenation level dependent sBOLD signal and its EEG-derived twin. Then, using the novel and theoretically justified weight interpretation methodology we recover individual spatial and time-frequency patterns of scalp EEG predictive of the hemodynamic signal in the subcortical nuclei. The described results not only pave the road towards wearable subcortical activity scanners but also showcase an automatic knowledge discovery process facilitated by deep learning technology in combination with an interpretable domain constrained architecture and the appropriate downstream task. | ['Alexei Ossadtchi', 'Ilia Mikheev', 'Alexander Kovalev'] | 2022-10-23 | null | null | null | null | ['eeg-decoding', 'eeg-decoding'] | ['medical', 'time-series'] | [ 1.78363442e-01 1.69185847e-01 2.69690126e-01 -4.98209506e-01
-1.56903431e-01 -3.95706683e-01 2.88835645e-01 5.45773059e-02
-6.83417916e-01 1.13613284e+00 3.32306117e-01 1.18706021e-02
-5.84835649e-01 -3.30861002e-01 -1.14487953e-01 -8.14286470e-01
-5.68477571e-01 5.93988895e-01 -1.86892375e-01 -1.15085587e-01
2.52115816e-01 5.27273297e-01 -1.23690450e+00 2.42253512e-01
8.44867051e-01 1.28797746e+00 1.72461569e-01 1.33434847e-01
2.97138870e-01 1.96316928e-01 -2.55698353e-01 -2.68545955e-01
-1.14282072e-02 -5.79943478e-01 -5.69217026e-01 -2.78273284e-01
-3.07040811e-01 -1.64253622e-01 5.64009659e-02 7.10358500e-01
6.64306223e-01 2.90390670e-01 7.22205877e-01 -8.01492274e-01
-2.09088311e-01 5.17681479e-01 -4.91756916e-01 8.38306487e-01
1.75248787e-01 4.77243006e-01 5.19046664e-01 -9.32379246e-01
5.82384825e-01 5.60492396e-01 4.36611444e-01 3.04627568e-01
-1.45297647e+00 -7.87583292e-01 -5.78776300e-02 5.56602776e-01
-1.32380450e+00 -6.21004581e-01 6.42016053e-01 -5.51317394e-01
9.91460681e-01 3.50328833e-02 1.15789509e+00 1.48531365e+00
4.52449232e-01 -1.35011271e-01 1.66801584e+00 -3.04590389e-02
4.26897079e-01 2.74468869e-01 1.62778944e-01 5.06783545e-01
2.66740382e-01 5.25180660e-02 -8.21647286e-01 -2.67019961e-02
4.85351831e-01 -3.41811538e-01 -4.31328595e-01 -1.26922965e-01
-1.46792126e+00 6.46489263e-01 3.67609531e-01 7.59711564e-01
-8.70082676e-01 -1.74433038e-01 4.78256822e-01 5.98181225e-02
3.27761710e-01 6.60446525e-01 -5.02991855e-01 4.56809625e-02
-1.08024430e+00 -2.44952902e-01 6.27770305e-01 4.49080803e-02
6.15598857e-01 3.22428077e-01 -9.75524187e-02 5.17455816e-01
1.83963314e-01 3.94314319e-01 6.14014566e-01 -9.15035784e-01
2.17218712e-01 2.38801792e-01 4.64555398e-02 -8.51437330e-01
-1.13298786e+00 -5.01055062e-01 -8.58697712e-01 3.35542113e-01
4.33108240e-01 -5.23594439e-01 -2.07138807e-01 1.56016052e+00
1.45362332e-01 -1.32281855e-01 -5.74328840e-01 1.29659414e+00
2.60735214e-01 -5.00907712e-02 3.27390552e-01 -4.55101937e-01
1.46943796e+00 5.53194210e-02 -6.37294114e-01 -6.03060246e-01
1.08108893e-01 2.01356381e-01 9.49427187e-01 4.87462461e-01
-9.97635722e-01 -3.56329471e-01 -1.15387428e+00 1.08471118e-01
-2.37506256e-01 -2.43256360e-01 9.38955128e-01 7.39846706e-01
-1.11737669e+00 5.68868160e-01 -8.91861975e-01 -2.41178647e-01
8.11912477e-01 7.18866348e-01 -6.67934537e-01 2.98037469e-01
-1.23813903e+00 1.45318842e+00 3.86121243e-01 4.21104491e-01
-6.10371768e-01 -8.52413118e-01 -3.04602146e-01 2.94107646e-01
5.33216074e-02 -6.91836238e-01 5.25607347e-01 -9.39510524e-01
-1.58920729e+00 1.16342664e+00 -4.62095961e-02 -3.32286119e-01
2.48114675e-01 3.76093596e-01 -3.59642982e-01 3.14614743e-01
-1.46631137e-01 4.83333707e-01 7.55147934e-01 -6.31886899e-01
2.72723019e-01 -8.83277893e-01 -6.60649598e-01 1.95869416e-01
-2.87311643e-01 -2.10357770e-01 5.85218966e-01 -2.30319217e-01
7.94963315e-02 -4.87971693e-01 8.50942507e-02 6.86792955e-02
-1.35570895e-02 1.79181084e-01 -2.42428742e-02 -8.81454051e-01
6.27454758e-01 -2.04667258e+00 3.16362500e-01 6.74387097e-01
5.68018138e-01 -1.68144673e-01 -6.82575703e-02 3.31891403e-02
-5.28669238e-01 -1.79286703e-01 -2.93264538e-01 3.75619382e-01
3.38364989e-01 -4.12173152e-01 -3.59783508e-02 1.06525266e+00
9.24473181e-02 1.08002555e+00 -7.50596821e-01 -1.29200891e-01
2.54282027e-01 4.64330286e-01 -4.94804591e-01 1.23290367e-01
1.83977276e-01 8.21479857e-01 -1.85439140e-01 2.91458309e-01
4.53367651e-01 1.97828427e-01 6.14565849e-01 -3.94321620e-01
-2.31250450e-01 2.30078533e-01 -6.78691745e-01 2.06226969e+00
-1.78941131e-01 9.16739881e-01 3.37901950e-01 -1.31217408e+00
1.16100538e+00 4.31197792e-01 5.56802034e-01 -1.25891721e+00
6.47784948e-01 2.48660818e-01 8.09588850e-01 -5.47704577e-01
-3.44033480e-01 -8.77999008e-01 1.70830369e-01 5.53786993e-01
4.64674830e-01 -1.94536224e-01 -2.83484638e-01 -3.75680864e-01
1.10456610e+00 1.61376655e-01 3.82628947e-01 -8.23237300e-01
3.43406618e-01 -5.59728503e-01 4.54662368e-02 3.23117942e-01
-4.37964201e-01 1.46346837e-01 7.20413566e-01 -1.86266944e-01
-5.59849143e-01 -1.16686440e+00 -3.41097206e-01 7.94285297e-01
-1.61878496e-01 3.27937424e-01 -6.73232019e-01 1.33045226e-01
-2.60940731e-01 6.96980774e-01 -7.82396317e-01 -6.81717932e-01
-1.39160365e-01 -9.71216857e-01 1.34789243e-01 3.71412843e-01
4.35808837e-01 -1.08758616e+00 -9.63133037e-01 3.06682020e-01
3.30095887e-02 -1.10396123e+00 3.87450129e-01 5.90562165e-01
-9.66241479e-01 -9.21340823e-01 -3.37840974e-01 -3.75484169e-01
1.50644407e-01 -4.40484673e-01 9.89067614e-01 -2.20867530e-01
-8.20064962e-01 1.43288895e-01 1.35381699e-01 -4.11333948e-01
1.81525216e-01 1.02435298e-01 2.42926508e-01 1.79798484e-01
5.53759634e-01 -1.29191017e+00 -1.08085585e+00 5.15678804e-03
-3.71977806e-01 -8.27094316e-02 5.79430163e-01 5.25010049e-01
3.82594287e-01 -4.21633154e-01 1.05145943e+00 -2.94876426e-01
9.18560743e-01 -8.85339975e-01 -3.62348348e-01 -7.93477222e-02
-6.31718099e-01 1.09628491e-01 5.10134935e-01 -2.73148745e-01
-1.22356224e+00 -2.39736170e-01 -3.90887298e-02 9.17655453e-02
-3.77870381e-01 2.71938711e-01 -1.64519966e-01 1.26734609e-02
1.01063454e+00 6.13984242e-02 -7.42993206e-02 -4.87299934e-02
5.65992743e-02 4.61828142e-01 5.94635844e-01 -5.02269506e-01
1.52685881e-01 4.86622125e-01 2.34235987e-01 -7.37073004e-01
-3.22622716e-01 2.96815392e-03 -8.88182282e-01 -3.86520714e-01
1.08662927e+00 -6.95283711e-01 -1.19067311e+00 -8.90690982e-02
-9.14056420e-01 -4.21286464e-01 -2.52697021e-01 7.53482640e-01
-5.42148709e-01 4.64679487e-03 -2.79047370e-01 -9.09049928e-01
-5.63087225e-01 -8.10735881e-01 5.57956755e-01 -9.56149027e-02
-8.87200236e-01 -8.25765073e-01 7.30916634e-02 3.64632934e-01
5.35440981e-01 4.67352569e-01 1.16970062e+00 -5.94525814e-01
-5.26124313e-02 -1.54537469e-01 -5.98902553e-02 -6.97896704e-02
-2.08036393e-01 -7.44238973e-01 -1.18931937e+00 2.40854248e-01
4.87053514e-01 -3.55342925e-01 2.93084770e-01 4.56521481e-01
1.19668388e+00 2.38656849e-01 3.41944546e-02 4.83719230e-01
1.25167835e+00 1.69202864e-01 8.14228714e-01 -6.41056150e-02
2.14974448e-01 9.25214171e-01 -2.19327286e-01 5.60239255e-01
9.61657688e-02 3.94181967e-01 2.80336231e-01 8.21140632e-02
8.77713487e-02 2.94580460e-01 1.55353174e-01 3.81471217e-01
-5.57533622e-01 3.08137715e-01 -8.93963277e-01 3.33775133e-01
-1.25708425e+00 -1.14697599e+00 -3.59334469e-01 2.29143500e+00
7.31050611e-01 -1.78439543e-02 2.39472035e-02 4.45961952e-02
3.68780643e-01 -2.37903148e-01 -9.12431300e-01 -6.15459323e-01
5.79905137e-02 8.97587478e-01 1.18371204e-01 2.69284666e-01
-3.71245593e-01 4.31515902e-01 6.68403339e+00 2.06910267e-01
-1.02267480e+00 3.98396254e-01 6.19482100e-01 -6.11234486e-01
-1.72241926e-01 -5.24540603e-01 1.34426817e-01 5.08476615e-01
1.65393126e+00 -2.69832432e-01 1.09355438e+00 2.25418419e-01
6.48357928e-01 -6.14569008e-01 -1.29922545e+00 1.20993149e+00
-1.79444179e-01 -1.13251793e+00 -7.66686201e-01 9.16636549e-03
1.73967421e-01 3.04081235e-02 -5.16825402e-03 -1.07478917e-01
-6.90173984e-01 -1.17385077e+00 5.92213988e-01 9.90663528e-01
1.17119133e+00 -4.33337986e-01 3.92620832e-01 3.13657820e-01
-5.05163968e-01 -2.06772178e-01 3.27617526e-02 -4.29925293e-01
-1.62907690e-01 5.80255687e-01 -4.42606181e-01 -4.79293130e-02
5.08395493e-01 4.63068634e-01 -1.49409607e-01 8.24191034e-01
-1.43395841e-01 4.89957064e-01 -2.43404210e-01 -4.33461443e-02
-1.57029942e-01 -3.86701643e-01 2.82115370e-01 1.03818429e+00
4.52846050e-01 4.18934733e-01 -7.25865424e-01 1.39182818e+00
9.20320116e-03 -3.69803384e-02 -4.48625177e-01 -1.93628401e-01
1.87536761e-01 1.54705989e+00 -9.76712883e-01 1.33577436e-01
-1.03474811e-01 8.98804307e-01 2.73871481e-01 4.82653677e-01
-6.97208464e-01 -2.62793869e-01 5.28187931e-01 2.88496882e-01
-1.26658902e-01 -3.67140323e-01 -7.85065413e-01 -1.25645685e+00
-1.15616240e-01 -6.38896465e-01 -4.17098254e-02 -1.03917074e+00
-9.48622763e-01 4.86647516e-01 5.40383905e-02 -4.26514626e-01
-1.90212026e-01 -7.82658577e-01 -6.27858281e-01 1.36342096e+00
-1.39695573e+00 -1.46274030e-01 -4.12692368e-01 8.14662218e-01
6.54023811e-02 1.56811014e-01 1.26288819e+00 3.80116433e-01
-4.56246167e-01 -5.33663742e-02 -1.52500868e-01 -3.27329546e-01
5.35395563e-01 -8.57677400e-01 -3.83921623e-01 5.19331694e-01
-1.54312149e-01 3.82625550e-01 5.57460308e-01 -2.14937568e-01
-1.35714209e+00 -2.69849777e-01 4.12865549e-01 -2.97319204e-01
8.72268796e-01 -8.77206445e-01 -7.82238126e-01 2.32716396e-01
3.86611938e-01 -1.18867837e-01 9.60098863e-01 1.68325439e-01
1.13086924e-02 -4.74073768e-01 -1.30272865e+00 2.95949191e-01
1.06997681e+00 -8.12731087e-01 -8.52559507e-01 4.30537403e-01
-2.50184268e-01 1.14728816e-01 -8.99726808e-01 -1.31727502e-01
6.63958848e-01 -1.28337300e+00 7.91208684e-01 -5.34704804e-01
2.03348100e-01 3.00400555e-01 1.75576866e-01 -1.52613485e+00
-4.20603603e-01 -4.86295223e-01 -2.67750546e-02 7.47056663e-01
3.41045201e-01 -7.31210053e-01 6.33991182e-01 1.21725333e+00
-6.98647648e-02 -4.84261215e-01 -1.24365914e+00 -2.65136778e-01
2.83910371e-02 -4.79814619e-01 2.27821589e-01 7.68728137e-01
6.55841589e-01 3.47256988e-01 7.13082775e-02 -2.95129240e-01
5.24098694e-01 -1.05882533e-01 -3.61301929e-01 -1.43790543e+00
-2.19765633e-01 -4.69633222e-01 -4.98426974e-01 1.44461989e-01
4.31045681e-01 -1.18827093e+00 -1.22390375e-01 -1.31802452e+00
1.88792109e-01 2.88335606e-02 -4.55340296e-01 3.73478413e-01
2.64053077e-01 2.64083028e-01 -2.77402192e-01 -3.07843834e-01
-8.56568143e-02 4.64662492e-01 7.49558687e-01 6.81709722e-02
-1.44068569e-01 -5.57159722e-01 -1.00654674e+00 5.02467394e-01
9.07455564e-01 -3.38429779e-01 -6.51294827e-01 -1.69519648e-01
3.95990312e-01 2.32693911e-01 5.26960433e-01 -1.09267688e+00
-1.40431583e-01 1.54663831e-01 1.07764733e+00 4.89361972e-01
4.39602852e-01 -1.06815660e+00 3.10995162e-01 5.31330585e-01
-2.71808892e-01 -4.16016281e-01 2.46251374e-01 3.48493248e-01
3.26321810e-01 8.00881982e-02 9.38919485e-01 -3.48775297e-01
-5.40517151e-01 1.73672616e-01 -7.26362884e-01 1.43195942e-01
1.16164064e+00 -4.51456547e-01 -2.78764606e-01 -4.43185270e-02
-1.14605999e+00 -1.65829912e-01 -2.82556027e-01 5.78336157e-02
5.22475183e-01 -9.39032257e-01 -6.14002645e-01 4.81519014e-01
-4.88265872e-01 -7.35906065e-01 5.46419382e-01 1.58643782e+00
-1.09544680e-01 4.04154301e-01 -9.46502566e-01 -1.66276008e-01
-6.26917005e-01 2.82621771e-01 8.12819123e-01 1.80004656e-01
-7.32066929e-01 7.16037810e-01 8.66843387e-03 3.37746710e-01
-1.43969625e-01 -8.13881606e-02 -1.82632416e-01 7.44002461e-01
5.65132022e-01 4.26200926e-01 3.74761999e-01 -2.32399240e-01
-6.35513902e-01 -9.43827182e-02 4.85650271e-01 -3.30332786e-01
1.83142102e+00 -1.68238595e-01 -3.88882697e-01 6.14848375e-01
9.77333724e-01 -5.19366801e-01 -1.25908911e+00 4.23438966e-01
-5.74274687e-03 -2.50603676e-01 3.97308618e-01 -1.26700926e+00
-1.06963825e+00 1.29518449e+00 9.38133180e-01 -2.69668698e-02
1.29175699e+00 -1.13426805e-01 2.32878774e-01 4.12943512e-01
5.28890610e-01 -1.37273240e+00 -2.10327819e-01 -9.93539765e-02
1.06797242e+00 -9.06559944e-01 2.52813011e-01 2.33497381e-01
-6.56877935e-01 1.20282900e+00 5.26832163e-01 -2.70269483e-01
6.64797306e-01 3.71450692e-01 -3.76799792e-01 -5.25174439e-01
-5.81238806e-01 -2.42340681e-03 2.95015544e-01 9.31070030e-01
5.32775044e-01 1.22399814e-03 -6.30344033e-01 1.09276056e+00
-3.26925255e-02 5.71866855e-02 2.11825609e-01 3.04296523e-01
-4.39468265e-01 -6.73475266e-01 -2.89647490e-01 9.37242687e-01
-2.64684469e-01 -7.14515969e-02 -2.31729954e-01 7.25787401e-01
2.44395316e-01 6.60334170e-01 1.12620845e-01 -1.44578740e-02
3.31702590e-01 8.59069407e-01 8.90052736e-01 -5.03289402e-01
-7.91673541e-01 -1.09617643e-01 8.89935717e-02 -7.59314477e-01
-4.10891473e-01 -8.71643960e-01 -1.44360733e+00 -6.04859777e-02
2.25291431e-01 -1.63775027e-01 8.20617378e-01 1.21844196e+00
5.00394523e-01 5.08754790e-01 2.14491427e-01 -1.21041095e+00
-8.53577405e-02 -9.72556233e-01 -1.05329144e+00 1.43990323e-01
2.22612962e-01 -8.09728324e-01 -1.65140092e-01 -1.37621120e-01] | [12.973811149597168, 3.415456771850586] |
63e8cbde-0207-461a-a807-35aa28ca29cb | conditional-image-generation-with-score-based | 2111.13606 | null | https://arxiv.org/abs/2111.13606v1 | https://arxiv.org/pdf/2111.13606v1.pdf | Conditional Image Generation with Score-Based Diffusion Models | Score-based diffusion models have emerged as one of the most promising frameworks for deep generative modelling. In this work we conduct a systematic comparison and theoretical analysis of different approaches to learning conditional probability distributions with score-based diffusion models. In particular, we prove results which provide a theoretical justification for one of the most successful estimators of the conditional score. Moreover, we introduce a multi-speed diffusion framework, which leads to a new estimator for the conditional score, performing on par with previous state-of-the-art approaches. Our theoretical and experimental findings are accompanied by an open source library MSDiff which allows for application and further research of multi-speed diffusion models. | ['Christian Etmann', 'Carola-Bibiane Schönlieb', 'Jan Stanczuk', 'Georgios Batzolis'] | 2021-11-26 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [-1.70088604e-01 -7.20018446e-02 -1.23988636e-01 -3.06751132e-01
-9.02996361e-01 -3.34648460e-01 1.00798273e+00 2.32580174e-02
-2.93720812e-01 5.84217429e-01 7.16652796e-02 -2.92161852e-01
-5.16617239e-01 -1.10302758e+00 -3.12607437e-01 -8.91340077e-01
-1.44674733e-01 7.76347816e-01 4.54647750e-01 -9.68141388e-03
1.89183787e-01 3.54642451e-01 -1.34537125e+00 -6.99040815e-02
8.04940581e-01 8.17852259e-01 1.42220736e-01 9.13877249e-01
-7.59245157e-02 8.69925439e-01 -6.77853465e-01 -1.02905178e+00
-1.35028750e-01 -6.00451767e-01 -6.44890547e-01 -2.85853297e-01
8.53242353e-02 -3.58800799e-01 -2.35106483e-01 1.30910277e+00
3.18135321e-01 -1.52495936e-01 1.10079777e+00 -1.21373987e+00
-9.45217490e-01 7.48915613e-01 -3.97450984e-01 5.24198651e-01
1.92358106e-01 -2.75526196e-01 1.07497203e+00 -6.98127389e-01
7.76598215e-01 1.18155134e+00 5.42046487e-01 4.74113643e-01
-1.12760627e+00 -4.07732725e-01 1.19452618e-01 3.12040985e-01
-1.28828776e+00 6.20359369e-03 8.55381608e-01 -4.94969368e-01
7.74209201e-01 8.78814608e-02 1.04803801e+00 1.33547175e+00
4.17302549e-01 1.16991854e+00 1.35859895e+00 -4.34080482e-01
4.12646621e-01 1.38751894e-01 2.00641945e-01 8.17403972e-01
3.06123585e-01 2.63866484e-01 -4.33909118e-01 -3.31453651e-01
1.08138192e+00 6.62783608e-02 2.35546634e-01 -6.36473000e-01
-9.20767486e-01 1.26883483e+00 1.83673710e-01 5.58959067e-01
-4.81570065e-01 4.62203860e-01 -8.52016266e-03 2.31198266e-01
7.96671867e-01 -5.00177324e-01 -5.99886328e-02 -4.15138811e-01
-1.38925517e+00 4.55403775e-01 9.56617415e-01 7.66197026e-01
5.47345638e-01 1.46993771e-01 -1.68098956e-01 5.44978440e-01
8.53663206e-01 7.12240219e-01 1.81406528e-01 -8.05720508e-01
5.45552187e-02 2.39303932e-01 -1.19883411e-01 -7.51462638e-01
-1.25118732e-01 -7.96469569e-01 -8.51560831e-01 1.89101458e-01
4.59223539e-01 -2.92839646e-01 -5.35549223e-01 1.67793143e+00
8.48853216e-02 3.33392859e-01 7.48908371e-02 4.63858515e-01
4.10436988e-01 5.69028616e-01 -1.52218090e-02 -1.70104817e-01
8.68120611e-01 -8.27896774e-01 -6.16777360e-01 3.26623470e-01
4.88779724e-01 -6.95194066e-01 6.74237072e-01 5.16305387e-01
-1.16968286e+00 -2.71890968e-01 -6.80846393e-01 3.93468142e-01
-1.45529881e-01 -1.51884794e-01 1.14307535e+00 1.17254186e+00
-1.48119068e+00 6.29354894e-01 -1.14689708e+00 -4.18105572e-01
4.91276652e-01 8.19641799e-02 2.06854925e-01 -8.64521787e-02
-9.59616482e-01 8.86459589e-01 -5.32089919e-02 -3.18929940e-01
-1.23519766e+00 -4.71732765e-01 -5.13958037e-01 -5.60460426e-03
-1.94329545e-02 -8.03562999e-01 1.52240014e+00 -7.80470371e-01
-1.58400893e+00 7.39969730e-01 -1.56208977e-01 -7.38453269e-01
7.66246438e-01 -1.36685476e-01 -3.76012832e-01 7.81899393e-02
-6.82471378e-04 3.87563258e-01 7.25370824e-01 -1.06411660e+00
-4.20809746e-01 -3.09333771e-01 1.09037578e-01 -1.20352231e-01
-4.71472889e-01 2.32591137e-01 -3.18802059e-01 -7.07407534e-01
-3.25418532e-01 -7.31624305e-01 -2.19759613e-01 -2.16314718e-01
-3.38260710e-01 -4.91689086e-01 4.04961050e-01 -3.20418417e-01
1.52292800e+00 -1.84281027e+00 5.92443764e-01 2.45452791e-01
3.97351325e-01 -6.25301003e-02 1.09113947e-01 7.74426043e-01
4.01682943e-01 -1.09727094e-02 -3.67798150e-01 -6.10625029e-01
2.64673829e-01 1.74243137e-01 -5.33695929e-02 4.91602987e-01
7.65609220e-02 9.62739944e-01 -9.30092514e-01 -3.72184426e-01
1.49501950e-01 8.67490172e-01 -4.44854736e-01 -1.11268200e-01
-3.05665992e-02 2.88818806e-01 -5.10513246e-01 4.78103310e-01
5.86329103e-01 -3.77314597e-01 1.63622200e-01 4.33596492e-01
-9.55223069e-02 1.65124983e-01 -1.24845481e+00 1.60745966e+00
-4.04686838e-01 6.57586813e-01 -1.44561633e-01 -8.66731942e-01
1.01792431e+00 1.09194443e-01 5.58171988e-01 -3.58835787e-01
2.34395191e-01 8.25970396e-02 -3.99222858e-02 3.99013050e-02
3.78711015e-01 -5.41319549e-01 -4.42661941e-02 8.46105158e-01
4.09281254e-01 -9.60482359e-02 3.78065765e-01 5.66428840e-01
1.07754135e+00 -5.49287349e-02 2.07460970e-01 -5.44123411e-01
4.84927118e-01 -3.91161859e-01 -2.84342654e-02 9.96627450e-01
-2.41856083e-01 2.21903279e-01 6.62925780e-01 -3.65157664e-01
-9.59010363e-01 -1.50572002e+00 -2.60413259e-01 9.45038319e-01
-5.73381186e-02 -5.62650084e-01 -1.10321939e+00 -6.68969929e-01
-7.99528658e-02 6.49570525e-01 -9.13665116e-01 1.18865751e-01
-2.34888077e-01 -1.15318072e+00 5.46523929e-01 7.51473188e-01
2.86481380e-01 -8.73546183e-01 -2.09019795e-01 2.11740881e-01
1.94128349e-01 -5.47925830e-01 7.67933875e-02 8.94719064e-02
-1.11503971e+00 -1.01059806e+00 -1.16395450e+00 -3.72428715e-01
2.09758446e-01 3.05399690e-02 1.13118029e+00 1.46185858e-02
8.77847970e-02 5.88811278e-01 -1.80926606e-01 -4.98688996e-01
-7.72331953e-01 3.42218816e-01 -1.79151177e-01 -3.20568606e-02
5.35257161e-01 -4.51683789e-01 -6.52630270e-01 1.54447109e-01
-1.02028906e+00 -2.27566019e-01 6.31086409e-01 3.97629499e-01
4.80340242e-01 5.08456677e-02 6.34564221e-01 -9.38950241e-01
1.10737431e+00 -6.44597769e-01 -7.39752471e-01 -8.07991996e-03
-8.66106391e-01 1.95154205e-01 2.10479304e-01 -4.33521986e-01
-1.22052062e+00 -3.99521381e-01 -5.80348909e-01 -6.31853193e-03
-9.42104831e-02 5.12696862e-01 2.63410240e-01 7.15702921e-02
3.33560944e-01 3.02216530e-01 -3.68598402e-01 -4.68471080e-01
4.95955497e-01 3.09186131e-01 2.52655800e-02 -4.28772032e-01
6.29173160e-01 7.22784877e-01 9.34950635e-03 -8.02538633e-01
-8.31638277e-01 -2.56239265e-01 -8.37747812e-01 -4.34304655e-01
8.36938798e-01 -7.73902237e-01 -2.96537071e-01 8.78793836e-01
-1.03916550e+00 -4.65563446e-01 -3.86956632e-01 3.77102256e-01
-7.67305791e-01 2.71224171e-01 -1.00264215e+00 -1.14352155e+00
-8.38105381e-02 -1.05811369e+00 1.07252598e+00 3.47478718e-01
-5.71773313e-02 -1.77422380e+00 6.77000046e-01 1.07847683e-01
8.02795947e-01 -3.85802463e-02 6.71077073e-01 -5.58156967e-01
-6.32152498e-01 -2.35807627e-01 -8.25195834e-02 3.73180062e-01
-4.00307894e-01 2.39956617e-01 -7.70735323e-01 -1.85441434e-01
6.62344173e-02 9.14461538e-02 1.08450317e+00 7.59176254e-01
5.08284569e-01 1.66769817e-01 -5.42995691e-01 4.33724582e-01
1.46414626e+00 -3.59239206e-02 6.19113863e-01 3.05483639e-01
3.91626000e-01 2.50306845e-01 1.98779538e-01 5.57817042e-01
5.45478165e-01 5.88522851e-01 2.59564906e-01 1.46216512e-01
-1.86257377e-01 -2.27145359e-01 3.04192305e-01 1.25836360e+00
-3.12536329e-01 -3.66775006e-01 -8.03414881e-01 5.42196572e-01
-1.76169729e+00 -1.18754852e+00 -6.33879483e-01 2.09991217e+00
5.40610611e-01 2.63665676e-01 5.44165850e-01 1.69603348e-01
5.25366366e-01 2.98995703e-01 -2.14229375e-01 -2.59907693e-01
-9.14580375e-02 5.27833343e-01 2.84391254e-01 6.74348533e-01
-9.95469272e-01 8.16983521e-01 8.27368927e+00 9.45434034e-01
-7.50753701e-01 7.12694645e-01 2.65759051e-01 4.89812624e-03
-6.15351737e-01 -9.10713244e-03 -9.12212610e-01 4.99574929e-01
1.19124293e+00 -3.44888270e-01 1.35841906e-01 8.52452159e-01
-2.11409181e-01 -1.77061781e-01 -7.52932966e-01 6.92327917e-01
2.76064295e-02 -1.26387680e+00 -1.04062296e-01 4.45134014e-01
1.12017655e+00 2.63608456e-01 2.49443084e-01 1.03568234e-01
8.52758169e-01 -6.27810657e-01 7.71663368e-01 9.10507560e-01
3.74813646e-01 -8.85327578e-01 8.61904204e-01 4.61382747e-01
-9.64070320e-01 1.18022978e-01 -5.64113140e-01 -1.86862692e-01
5.31150699e-01 1.02129376e+00 -2.94002533e-01 4.29148406e-01
4.80135232e-01 9.68764782e-01 -4.51554030e-01 1.08524990e+00
-5.50527930e-01 8.98563325e-01 -1.71279311e-01 -4.93405819e-01
1.57112062e-01 -9.40786004e-02 4.78560716e-01 1.47470748e+00
5.83906054e-01 -3.44521582e-01 -2.37730905e-01 1.11483228e+00
7.17473216e-03 2.57858075e-02 -6.24350905e-01 -4.66435067e-02
2.24846587e-01 1.00190365e+00 -9.16550100e-01 -5.96934736e-01
-4.81298417e-01 8.95807505e-01 5.04380763e-01 1.42408758e-01
-8.50852430e-01 9.18955505e-02 6.13814831e-01 6.92242309e-02
5.29824555e-01 -6.72026277e-01 -1.58684894e-01 -1.40085936e+00
-3.98949862e-01 -2.19951466e-01 2.85784245e-01 -3.64677787e-01
-1.49881983e+00 5.90414345e-01 3.73439044e-01 -6.94522977e-01
-4.41223085e-01 -5.51073074e-01 -6.56116486e-01 8.12503994e-01
-1.59483325e+00 -8.86243522e-01 3.46067594e-03 5.49605310e-01
5.13865769e-01 -1.96582556e-01 9.09751058e-01 2.66277373e-01
-2.80898958e-01 1.93207741e-01 4.50313300e-01 -1.34455174e-01
2.45768934e-01 -1.50488341e+00 7.08632529e-01 7.53668964e-01
6.18291259e-01 5.47083020e-01 5.88821530e-01 -7.06911385e-01
-1.21359634e+00 -5.33508301e-01 8.54448438e-01 -7.60623872e-01
1.00707674e+00 -2.74246544e-01 -7.73684502e-01 4.97483373e-01
2.78983414e-01 -4.57506537e-01 8.85667384e-01 2.73736626e-01
-9.72626507e-02 8.53511319e-02 -8.50605905e-01 2.64781654e-01
9.51735556e-01 -4.40864980e-01 -3.77212286e-01 1.82357043e-01
1.93210378e-01 5.60617261e-02 -9.41351354e-01 -5.04404074e-03
6.24903440e-01 -1.53266382e+00 9.29003000e-01 -2.34994218e-01
3.53428811e-01 1.45295456e-01 -2.17508823e-01 -1.30355930e+00
-4.69403207e-01 -3.88259500e-01 -5.17110527e-01 1.16005945e+00
3.41293961e-01 -4.60645974e-01 8.08763504e-01 1.01014465e-01
2.66688854e-01 -8.52328002e-01 -9.17265236e-01 -8.09943378e-01
5.77652991e-01 -1.01094508e+00 3.72122794e-01 4.61553544e-01
-3.04034859e-01 1.91790789e-01 -3.63396764e-01 -2.47734129e-01
9.33519483e-01 -7.84557015e-02 5.28981507e-01 -1.70469940e+00
-4.04813409e-01 -8.97988856e-01 -4.89357740e-01 -1.24432766e+00
1.12897813e-01 -8.83663654e-01 -4.36373234e-01 -1.94633305e+00
4.84170139e-01 -2.52823502e-01 -5.24883568e-01 -2.18306527e-01
-2.76905783e-02 4.34768081e-01 1.33892000e-01 2.99025357e-01
-8.16286743e-01 5.48703074e-01 1.17001116e+00 1.22743338e-01
1.65085077e-01 4.81734842e-01 -6.55029297e-01 7.55595505e-01
9.11630511e-01 -8.09765697e-01 -4.97890413e-01 -4.29490447e-01
6.13173962e-01 1.12790940e-02 4.81039941e-01 -1.04722619e+00
2.43168563e-01 2.61008833e-03 1.93880588e-01 -6.14053667e-01
2.99980313e-01 -4.38324362e-01 2.33361428e-03 6.65417135e-01
-2.61578947e-01 2.11566538e-01 -1.24906309e-01 8.73819828e-01
-1.68478236e-01 -5.88464081e-01 7.15053797e-01 -1.38850778e-01
-2.86548406e-01 2.19023600e-01 -6.72151685e-01 -6.15321621e-02
1.06570613e+00 5.33877686e-02 -2.70709485e-01 -6.44329071e-01
-7.69989073e-01 -4.42907751e-01 3.93038005e-01 3.21119636e-01
5.89715660e-01 -1.22858930e+00 -7.92675018e-01 2.24189281e-01
-2.35595673e-01 -6.01412416e-01 3.04358006e-01 8.93234849e-01
-4.35157776e-01 4.27793592e-01 -1.18331492e-01 -5.59707046e-01
-9.49192405e-01 4.70562309e-01 9.88179073e-02 -6.55360699e-01
-4.53868747e-01 8.55409384e-01 8.78992584e-03 -1.36312112e-01
9.47567448e-03 -2.13965312e-01 -6.54314756e-02 -2.80942377e-02
5.93540370e-01 6.23689115e-01 -9.20508355e-02 -5.89050949e-01
-3.08264613e-01 4.43441480e-01 3.74045894e-02 -5.92526019e-01
1.39441586e+00 -1.26034841e-01 -6.52405918e-02 6.62862659e-01
8.42959285e-01 3.34827928e-03 -1.28051448e+00 -2.03384861e-01
-3.37321497e-02 -3.24228138e-01 2.38272458e-01 -8.52586746e-01
-1.06205225e+00 1.17911291e+00 5.68283141e-01 8.71046007e-01
8.51887405e-01 3.84432673e-01 5.61105728e-01 -3.49041224e-02
5.87173998e-01 -8.03132415e-01 1.12044066e-01 4.73663509e-01
4.60162371e-01 -9.22800839e-01 -1.75077036e-01 -1.74363002e-01
-5.00067651e-01 9.66520369e-01 2.37509813e-02 -4.10383433e-01
1.07179213e+00 5.92821956e-01 -1.24652520e-01 -3.68036270e-01
-8.39921057e-01 -3.45153034e-01 2.81302005e-01 1.01001322e+00
6.72677159e-01 3.74714404e-01 -4.10484254e-01 5.07979453e-01
-2.56959140e-01 -3.55785638e-02 3.75589103e-01 7.13719130e-01
-5.82982659e-01 -1.45345795e+00 -1.45653561e-01 3.15761983e-01
-5.48997343e-01 -1.78776845e-01 -3.34080428e-01 6.97541177e-01
-2.37224609e-01 8.94651711e-01 -1.61193058e-01 -1.59696251e-01
-2.18460318e-02 2.33106837e-01 8.30873191e-01 -3.67085040e-01
-3.90440822e-01 1.81229666e-01 -3.84182155e-01 -1.62868068e-01
-7.16140866e-01 -1.06665421e+00 -8.06061447e-01 -4.91859198e-01
-4.73869681e-01 1.08554242e-02 9.41037834e-01 1.02865136e+00
9.76961479e-02 4.58099604e-01 1.29800975e-01 -6.85659826e-01
-4.42491859e-01 -1.19297707e+00 -7.68334985e-01 9.80719104e-02
1.73230600e-02 -7.63021231e-01 -1.32749930e-01 -1.64070949e-01] | [11.25130558013916, -0.10414239764213562] |
b4cb1e46-4cc8-4290-a469-7c47309e14a5 | conceptron-a-probabilistic-deep-one-class | null | null | https://openreview.net/forum?id=q58E59ZPLp | https://openreview.net/pdf?id=q58E59ZPLp | Conceptron: a probabilistic deep one-class classification method | One-class learning through deep architectures is a particularly challenging task; in this scenario the crasis of kernel methods and deep networks can represent a viable strategy to empower already effective methods. In this contribution we present Conceptron, a probabilistic and deep one-class classification method. The proposed algorithm is a hybridization of the Nystrom version of the Import Vector Domain Description (IVDD) to deep learning layers rendering the approach highly scalable (via batch stochastic gradient optimization) and automatically learning the underlying feature space. Further we modify the cost function to allow to get a Laplace distribution of the samples probabilities. Experiments on MNIST, CIFAR-10 and other benchmark datasets show that Conceptron (and/or variations) performs comparably or better with competing state-of-the-art methods with the additional capability of providing probabilities (through a logistic model) and avoiding any degeneracy in the training process. | ['Sergio Decherchi', 'Andrea Cavalli', 'Erika Gardini'] | 2021-09-29 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [-5.49085811e-03 -3.13409418e-02 -2.56886929e-01 -6.35454118e-01
-6.77214265e-01 -3.75498712e-01 1.09338367e+00 9.62124318e-02
-7.87196696e-01 1.11920238e+00 -2.85069913e-01 -3.48186076e-01
-5.30378699e-01 -6.88058257e-01 -6.51888192e-01 -1.08142221e+00
-9.03599113e-02 1.01992583e+00 2.16797531e-01 1.01982392e-01
3.08119923e-01 8.62908423e-01 -1.49654603e+00 1.82308212e-01
5.19405782e-01 1.39756906e+00 -7.83968791e-02 6.59767747e-01
-2.80623347e-01 7.00220346e-01 -4.29197699e-01 -3.81286800e-01
6.82837889e-02 1.68184981e-01 -7.25726128e-01 -4.98038352e-01
3.42128158e-01 -1.87647551e-01 -9.59791094e-02 8.03978622e-01
5.89129448e-01 2.86661655e-01 1.38640392e+00 -9.86422420e-01
-3.66230816e-01 4.74156022e-01 -2.69049525e-01 1.28428221e-01
-1.57313235e-02 -2.58654237e-01 8.95728052e-01 -1.23813379e+00
2.21893370e-01 1.23174453e+00 9.37541187e-01 3.46343905e-01
-1.53052652e+00 -4.22393113e-01 -2.10624740e-01 3.54894370e-01
-1.52468479e+00 -6.12156540e-02 7.05257833e-01 -5.92435539e-01
1.09657502e+00 -3.89821008e-02 2.30324030e-01 1.51306462e+00
2.67438591e-01 8.27239573e-01 1.14843357e+00 -5.39980590e-01
6.54477417e-01 6.50757372e-01 2.69181430e-01 5.01966715e-01
4.40381020e-01 2.29067892e-01 -4.53235537e-01 -7.38350689e-01
4.93647158e-01 1.02266163e-01 1.48594603e-01 -8.43458414e-01
-7.73410976e-01 1.31172574e+00 2.19118416e-01 1.38862163e-01
-5.37891448e-01 3.15902889e-01 4.30817038e-01 2.03118715e-02
5.46095908e-01 2.16240183e-01 -8.35570276e-01 -1.16135299e-01
-1.04912639e+00 6.13905609e-01 1.10987866e+00 3.70710880e-01
6.43845916e-01 2.00082943e-01 -1.55546755e-01 5.54959118e-01
3.94539952e-01 3.48421007e-01 3.87798637e-01 -6.98134601e-01
8.95727873e-02 2.47920826e-01 2.82593399e-01 -4.77356464e-01
-5.55060387e-01 -7.61959493e-01 -8.46671462e-01 6.95235550e-01
6.82639122e-01 -8.28093439e-02 -9.33583081e-01 1.26541388e+00
3.48125249e-01 8.33653379e-03 1.07362017e-01 7.11724877e-01
5.19607604e-01 7.22709596e-01 2.88796782e-01 1.05483010e-01
1.17450368e+00 -4.78133947e-01 -3.54379833e-01 3.55052799e-02
3.69728863e-01 -7.39337146e-01 8.67908478e-01 8.93246293e-01
-7.18905985e-01 -6.30679309e-01 -1.06720054e+00 1.35112002e-01
-7.72171974e-01 3.86925429e-01 9.58880782e-01 9.74981368e-01
-8.30158770e-01 1.13365698e+00 -8.11583638e-01 5.83785921e-02
7.30872452e-01 6.74033761e-01 -4.09917504e-01 4.32118848e-02
-1.15708995e+00 9.85158086e-01 6.55497730e-01 -2.02924181e-02
-8.17107499e-01 -7.78685153e-01 -7.52088428e-01 3.80518258e-01
5.50056510e-02 -4.54218030e-01 1.24260557e+00 -8.43009949e-01
-1.88934124e+00 6.05066836e-01 1.02345534e-01 -8.90688598e-01
7.91810036e-01 -5.76445460e-01 1.54186502e-01 1.40533715e-01
-3.38618398e-01 6.48014486e-01 1.09116530e+00 -8.71247113e-01
-5.22748053e-01 -1.29695699e-01 -3.74825060e-01 -1.28702909e-01
-3.57381552e-01 -7.78415203e-02 3.67962718e-01 -5.46477139e-01
-2.53883094e-01 -6.13733113e-01 -2.25879028e-01 9.91855189e-02
-3.58706653e-01 -6.62467062e-01 6.41410828e-01 -4.25558478e-01
7.81418443e-01 -1.96356809e+00 1.79532245e-01 3.40184778e-01
2.05986664e-01 4.94705975e-01 1.64179713e-01 5.01944482e-01
-1.51187509e-01 -1.19429946e-01 -4.04132426e-01 -4.33908492e-01
3.70656997e-01 1.18147776e-01 -3.77013475e-01 8.44336092e-01
4.07706410e-01 5.08083403e-01 -6.61283135e-01 -2.11620688e-01
4.11371648e-01 1.06169355e+00 -3.27630013e-01 1.23835437e-01
-2.39330634e-01 -6.96578668e-03 -2.04090223e-01 3.46335232e-01
9.14050281e-01 -1.75437331e-01 -2.91887186e-02 -9.14001837e-03
2.21019424e-02 2.50160486e-01 -1.37345874e+00 1.31011879e+00
-3.73779982e-01 5.98641038e-01 -2.72812963e-01 -1.41374552e+00
1.07683694e+00 2.92951316e-01 2.88681388e-01 -5.68970889e-02
2.06077620e-01 5.01073718e-01 -9.01705027e-02 -4.60764440e-03
1.95808664e-01 -4.04528171e-01 9.34647024e-02 1.12404495e-01
5.84603906e-01 3.00814211e-01 -1.35537302e-02 -7.92784691e-02
6.52515948e-01 4.00325507e-01 4.34161454e-01 -5.23574054e-01
6.91576719e-01 -1.67638436e-01 1.37626126e-01 1.03656340e+00
-1.05383527e-02 4.16983545e-01 6.79358006e-01 -6.88283563e-01
-1.05464530e+00 -1.49263406e+00 -5.67719996e-01 9.97166514e-01
-5.21482646e-01 1.07750008e-02 -6.27040505e-01 -5.48161566e-01
2.16679513e-01 9.34025407e-01 -7.46263981e-01 -2.56346285e-01
-2.38622159e-01 -1.19754469e+00 5.22427619e-01 6.47275865e-01
3.77616793e-01 -8.99901927e-01 -5.92652202e-01 4.28167433e-01
5.14140427e-01 -8.06301177e-01 4.42133158e-01 8.32381368e-01
-7.66377926e-01 -7.76063204e-01 -1.02058184e+00 -4.76791471e-01
1.19844573e-02 -4.62491155e-01 8.71124446e-01 -6.73407614e-01
-5.60667217e-01 -6.20018095e-02 -1.53625101e-01 -5.06589293e-01
-3.09959948e-01 3.37548077e-01 1.96052566e-01 1.44614875e-01
6.77925348e-01 -6.46254420e-01 -5.57832181e-01 -2.32643083e-01
-6.40199125e-01 -3.52307618e-01 6.27632022e-01 9.72310364e-01
2.74626911e-01 2.08015993e-01 5.86049974e-01 -9.16739404e-01
5.28891027e-01 -5.39352000e-01 -9.24073279e-01 -1.50221139e-01
-8.15672398e-01 3.28073502e-01 8.44684243e-01 -6.26818776e-01
-1.10017395e+00 3.26207459e-01 -3.22534859e-01 -3.94236147e-01
-3.56834918e-01 1.87121123e-01 -1.26138655e-03 -2.10527480e-01
8.99261475e-01 3.12592536e-01 -8.21690038e-02 -9.49811578e-01
5.16110003e-01 6.35953367e-01 3.30968261e-01 -5.92079818e-01
6.89072013e-01 6.07156873e-01 2.46475860e-01 -6.78274870e-01
-7.67750859e-01 -5.61527073e-01 -8.11504602e-01 1.77739933e-01
7.38280237e-01 -7.42489457e-01 -8.40551674e-01 5.61014593e-01
-1.24774206e+00 -2.81765908e-01 -3.05981278e-01 7.38504291e-01
-6.12433851e-01 1.09212115e-01 -7.02859402e-01 -1.04693854e+00
-2.72913605e-01 -9.91963208e-01 9.31511283e-01 4.39645857e-01
7.41486475e-02 -1.19822884e+00 8.24353173e-02 -1.67638913e-01
6.51052713e-01 1.54992089e-01 1.00865400e+00 -1.28136981e+00
-1.35145590e-01 -4.89362746e-01 -3.79488677e-01 7.08932042e-01
-3.57455283e-01 -5.05650826e-02 -1.44447398e+00 -1.82566985e-01
-1.53308913e-01 -4.60024953e-01 1.19239819e+00 5.70669532e-01
1.28204799e+00 -2.02250376e-01 -2.66555816e-01 6.14475191e-01
1.65104783e+00 -1.27822757e-01 4.75368530e-01 4.55612361e-01
4.17065084e-01 5.32033682e-01 2.46543467e-01 7.35415459e-01
1.32119618e-02 5.12268603e-01 3.16645294e-01 -1.21008851e-01
1.67241633e-01 5.31943850e-02 1.98759541e-01 6.60398137e-03
2.24314593e-02 -6.45710230e-02 -9.14975464e-01 2.25821406e-01
-1.80429769e+00 -9.33583736e-01 -2.47696966e-01 2.29205251e+00
8.03419352e-01 4.65024740e-01 2.25561842e-01 3.52807522e-01
3.61668050e-01 -8.35375711e-02 -5.05175292e-01 -6.09323859e-01
-1.24011740e-01 8.15977752e-01 6.26354873e-01 5.18728912e-01
-1.41670215e+00 8.00457954e-01 6.22855854e+00 1.21167481e+00
-1.11193919e+00 1.27338275e-01 6.70165777e-01 1.37516379e-01
2.50542194e-01 -1.02067977e-01 -1.21142232e+00 2.77592212e-01
1.31658173e+00 1.77189693e-01 2.13158205e-01 1.26460385e+00
-2.21930757e-01 -2.10755214e-01 -1.04264629e+00 1.07916653e+00
-7.24070668e-02 -1.57408690e+00 7.67363533e-02 7.97949731e-02
3.23633850e-01 2.10337579e-01 2.34572321e-01 6.88804507e-01
2.80019671e-01 -1.27422929e+00 6.98578537e-01 6.62445962e-01
6.16771519e-01 -9.43496644e-01 1.02392018e+00 3.37090433e-01
-7.29583800e-01 -1.53652534e-01 -7.19141304e-01 -8.61822516e-02
-1.20723039e-01 8.93426359e-01 -9.62675452e-01 4.63784039e-01
7.06796646e-01 4.13184285e-01 -5.28639972e-01 1.15168405e+00
-7.45867239e-03 8.13082814e-01 -5.34652889e-01 -3.07829827e-01
5.03635585e-01 -3.96697335e-02 3.99281770e-01 1.45216978e+00
7.78393168e-03 -4.93467212e-01 7.31009245e-02 1.06727850e+00
1.49610788e-01 -7.77634755e-02 -3.86315167e-01 -1.53196026e-02
1.61302432e-01 1.35094941e+00 -6.88434422e-01 -4.41994697e-01
-1.85806170e-01 1.01546216e+00 4.54882562e-01 4.07050997e-01
-6.73222125e-01 -7.68538415e-01 6.63175762e-01 -8.30445737e-02
7.02337384e-01 -1.31909296e-01 -2.48364761e-01 -8.29221427e-01
-7.93080777e-02 -4.78624463e-01 2.80887723e-01 -4.04499471e-01
-1.52015114e+00 6.40943646e-01 2.12651744e-01 -7.10139334e-01
-3.90550137e-01 -1.35206175e+00 -5.14072299e-01 1.02390337e+00
-1.84259808e+00 -9.21419084e-01 2.33385593e-01 3.62359405e-01
2.32126683e-01 -4.07430828e-01 1.01956999e+00 1.73900068e-01
-2.34089807e-01 5.16771555e-01 7.72015035e-01 -9.78668407e-02
4.99681860e-01 -1.56023955e+00 1.79177269e-01 2.15159953e-01
1.09316707e-01 3.88205260e-01 9.13933337e-01 -2.11511552e-01
-9.58026648e-01 -8.32668066e-01 6.98888481e-01 -4.24270809e-01
7.47352779e-01 -6.97351992e-01 -9.50660765e-01 1.43054903e-01
-1.07808135e-01 7.32869357e-02 8.24276626e-01 3.65446471e-02
-5.20783126e-01 -9.95788798e-02 -1.23399663e+00 8.84746686e-02
3.64922494e-01 -4.29288030e-01 -6.58655524e-01 4.48745340e-01
4.09526438e-01 -9.80590582e-02 -6.52504861e-01 4.09574538e-01
6.18807375e-01 -1.13590193e+00 1.22615349e+00 -7.19979703e-01
1.01284653e-01 5.93917146e-02 -2.74391592e-01 -1.01790810e+00
-2.18233407e-01 -6.07303202e-01 -5.11165857e-01 1.09900796e+00
2.75364816e-01 -8.51381600e-01 9.33432698e-01 2.05690533e-01
2.78048366e-01 -1.09277010e+00 -1.15127802e+00 -8.49228561e-01
4.54734623e-01 -4.30702209e-01 1.72947749e-01 4.49323058e-01
-4.64238226e-01 1.98751599e-01 -1.49849534e-01 3.46437544e-02
9.64981735e-01 -2.35124394e-01 4.38645780e-01 -1.81150496e+00
-4.73803192e-01 -6.28276765e-01 -6.45352900e-01 -4.97335225e-01
4.12035108e-01 -7.73665309e-01 -2.41199702e-01 -1.34004974e+00
1.15954846e-01 -4.32953477e-01 -4.96107787e-01 3.08802545e-01
1.13290794e-01 7.77901486e-02 -1.86918259e-01 -5.72195649e-03
-3.01186442e-01 7.07662940e-01 3.76303613e-01 1.15891412e-01
1.41739817e-02 2.99777865e-01 -3.37561578e-01 8.30958068e-01
9.46753263e-01 -7.36124933e-01 -2.28257194e-01 1.99207306e-01
2.38055035e-01 -3.89755934e-01 7.50124454e-01 -1.16765916e+00
3.91333923e-02 2.08624437e-01 9.33343172e-01 -5.72523832e-01
7.16327488e-01 -5.25806963e-01 -3.61486167e-01 6.12507939e-01
-3.80016446e-01 -3.57841492e-01 2.22109288e-01 7.69590735e-01
-3.30802496e-03 -4.50253546e-01 1.12608743e+00 2.19465606e-02
-5.08608282e-01 1.07369967e-01 -6.42596662e-01 1.59249716e-02
8.46891940e-01 7.96454586e-03 -7.58962054e-03 -4.47772555e-02
-8.40769291e-01 -1.88957557e-01 5.04760519e-02 2.10662913e-02
5.42431295e-01 -1.06455135e+00 -6.34509444e-01 1.39451444e-01
-1.50898620e-01 -7.26267621e-02 -6.38107359e-02 5.65595865e-01
-7.03517497e-01 6.69645071e-01 -1.28469884e-01 -5.99070668e-01
-7.21551359e-01 3.67279440e-01 5.91253817e-01 -4.75631177e-01
-5.34443319e-01 1.15944159e+00 8.24900344e-03 -6.48120046e-01
5.96201003e-01 -8.15779790e-02 -3.00663531e-01 1.25017576e-02
6.11950457e-01 3.16936374e-01 1.58200949e-01 -2.20843002e-01
-5.25272906e-01 1.89292774e-01 -3.59510243e-01 -2.07491174e-01
1.56755984e+00 4.51720923e-01 1.51962548e-01 5.24096727e-01
1.28569520e+00 -4.95939732e-01 -1.55459070e+00 -2.17356086e-01
2.57669151e-01 -2.88937334e-02 2.91906148e-01 -8.71020913e-01
-4.24685836e-01 1.34048140e+00 1.02926886e+00 -5.86343035e-02
5.17578185e-01 -2.52260454e-02 4.30047870e-01 6.62096322e-01
5.37629463e-02 -1.02668715e+00 -1.51320593e-02 4.39512342e-01
4.18302149e-01 -1.25688493e+00 1.33612499e-01 -3.90779451e-02
-4.20223832e-01 1.58276379e+00 2.70295024e-01 -5.39189637e-01
1.03605819e+00 1.07804969e-01 -1.92294478e-01 -7.93847814e-02
-6.28375351e-01 -1.17608368e-01 4.90287960e-01 6.78308845e-01
4.47033763e-01 9.47406143e-02 -2.32794151e-01 7.67060339e-01
8.77329409e-02 -7.43691549e-02 2.59676874e-01 6.34766221e-01
-5.33667743e-01 -1.19927478e+00 -3.58814865e-01 4.85539496e-01
-7.03114927e-01 -2.29331285e-01 -1.17576405e-01 8.83505344e-01
-7.20626339e-02 4.72622335e-01 -2.17771694e-01 3.81989330e-02
-8.90925601e-02 3.80101472e-01 4.17143196e-01 -4.37135071e-01
-5.37868917e-01 -1.24591596e-01 -1.78647414e-02 -4.24335986e-01
-8.95589069e-02 -5.41778147e-01 -1.15416312e+00 -1.44886926e-01
-3.40788364e-01 2.03498244e-01 1.13472939e+00 1.09517395e+00
1.53985128e-01 2.47713551e-01 3.79638165e-01 -1.36732316e+00
-1.30968273e+00 -8.92241240e-01 -7.46291757e-01 -3.79129797e-02
3.18860888e-01 -1.12133694e+00 -5.77197790e-01 -2.51784146e-01] | [7.361118316650391, 3.7742092609405518] |
d65be4fb-e49d-405e-ac91-824cb9615e9d | multi-label-network-classification-via | 1902.09294 | null | http://arxiv.org/abs/1902.09294v1 | http://arxiv.org/pdf/1902.09294v1.pdf | Multi-Label Network Classification via Weighted Personalized Factorizations | Multi-label network classification is a well-known task that is being used in
a wide variety of web-based and non-web-based domains. It can be formalized as
a multi-relational learning task for predicting nodes labels based on their
relations within the network. In sparse networks, this prediction task can be
very challenging when only implicit feedback information is available such as
in predicting user interests in social networks. Current approaches rely on
learning per-node latent representations by utilizing the network structure,
however, implicit feedback relations are naturally sparse and contain only
positive observed feedbacks which mean that these approaches will treat all
observed relations as equally important. This is not necessarily the case in
real-world scenarios as implicit relations might have semantic weights which
reflect the strength of those relations. If those weights can be approximated,
the models can be trained to differentiate between strong and weak relations.
In this paper, we propose a weighted personalized two-stage multi-relational
matrix factorization model with Bayesian personalized ranking loss for network
classification that utilizes basic transitive node similarity function for
weighting implicit feedback relations. Experiments show that the proposed model
significantly outperforms the state-of-art models on three different real-world
web-based datasets and a biology-based dataset. | ['Lars Schmidt-Thieme', 'Josif Grabocka', 'Ahmed Rashed'] | 2019-02-25 | null | null | null | null | ['implicit-relations'] | ['natural-language-processing'] | [ 2.90623993e-01 4.69144315e-01 -8.48968208e-01 -6.38308287e-01
2.89097447e-02 -3.47180992e-01 4.71816778e-01 6.30840898e-01
-2.57110000e-01 7.20268786e-01 4.20734584e-01 -1.27486721e-01
-8.20670784e-01 -1.14399219e+00 -3.01835775e-01 -5.11666000e-01
-1.86799332e-01 8.71889114e-01 3.95774722e-01 -2.95668304e-01
1.81653962e-01 2.92520106e-01 -1.35284662e+00 5.03125250e-01
6.94966137e-01 7.22792327e-01 -2.56307662e-01 4.15873706e-01
-2.75227129e-01 1.05769408e+00 -2.32271314e-01 -7.03668177e-01
5.04330657e-02 -1.29746124e-01 -1.11611557e+00 -3.73571754e-01
-4.89827693e-02 1.20241418e-01 -3.70504856e-01 9.11584795e-01
2.13187382e-01 2.00511530e-01 1.03786862e+00 -1.48767531e+00
-5.06831765e-01 8.22574496e-01 -7.29075313e-01 1.97202101e-01
3.33770812e-01 -7.51182854e-01 1.78013444e+00 -5.22795022e-01
5.91970026e-01 1.48423100e+00 5.25842667e-01 2.10227340e-01
-1.42413414e+00 -7.46899009e-01 4.17459220e-01 4.41858590e-01
-1.22863245e+00 5.92011139e-02 8.62562239e-01 -5.31803012e-01
6.97152972e-01 3.60906303e-01 3.44663978e-01 6.79723740e-01
-4.34412658e-02 4.96020734e-01 8.94525468e-01 -3.23756248e-01
-1.43361881e-01 3.19455892e-01 4.64154094e-01 5.48000276e-01
8.14712346e-02 -3.80828500e-01 -6.75397336e-01 -5.34638286e-01
3.82759035e-01 4.03196931e-01 -5.34320101e-02 -5.63633621e-01
-9.08261776e-01 1.07790363e+00 9.48807001e-01 4.90110189e-01
-2.94710785e-01 -8.06800276e-02 5.24163723e-01 4.99133825e-01
9.41781878e-01 1.56522885e-01 -5.13199568e-01 3.69255930e-01
-5.05950749e-01 1.21222539e-02 7.63984382e-01 6.13543332e-01
1.02221441e+00 -6.92345381e-01 -6.57215714e-02 1.29937792e+00
7.66771436e-01 -3.16402167e-01 4.11270320e-01 -8.44178379e-01
2.68977612e-01 1.16711831e+00 -2.18007252e-01 -1.47928846e+00
-4.80511367e-01 -4.84666020e-01 -1.05385113e+00 -2.40453497e-01
2.70323157e-01 1.61136672e-01 -4.17547584e-01 1.70857906e+00
5.76400399e-01 4.45588529e-01 -2.12657541e-01 6.94542646e-01
9.93762255e-01 5.55474639e-01 1.94305807e-01 -3.76751781e-01
1.14053011e+00 -9.01689351e-01 -7.30679333e-01 4.17178757e-02
8.58689368e-01 -5.45347393e-01 6.78439558e-01 2.30230957e-01
-3.84889811e-01 -1.56658337e-01 -7.59759724e-01 7.65919015e-02
-5.19230366e-01 -3.36462885e-01 1.03886855e+00 4.11928326e-01
-9.05335963e-01 7.11738646e-01 -2.39690274e-01 -7.04733968e-01
3.60017896e-01 4.23444897e-01 -4.66995567e-01 -2.20212013e-01
-1.58153915e+00 8.43077064e-01 2.61498660e-01 6.09516129e-02
-2.93192357e-01 -4.58696902e-01 -6.30930603e-01 1.13246687e-01
5.46573460e-01 -4.74096417e-01 6.84269488e-01 -8.33441198e-01
-1.10247457e+00 7.19641626e-01 -2.77686298e-01 -1.63120255e-01
1.31567001e-01 -9.13729072e-02 -4.65001762e-01 6.24536015e-02
1.53366834e-01 3.61884892e-01 5.54903984e-01 -1.16302967e+00
-4.31692153e-01 -3.83998960e-01 5.55437803e-01 4.49475288e-01
-7.68322825e-01 -9.20454040e-03 -2.00561747e-01 -4.72329736e-01
4.27301973e-01 -9.91112053e-01 -3.72557402e-01 2.30136737e-01
-3.46607178e-01 -7.40928352e-01 8.80692124e-01 -2.63160877e-02
1.14102352e+00 -1.61812198e+00 3.18127126e-01 6.10887706e-01
4.46832269e-01 1.61545202e-01 -2.72409528e-01 4.99216676e-01
-2.77823478e-01 2.55364776e-01 1.87987089e-01 -1.27264671e-02
-3.70144218e-01 3.68676573e-01 -1.08972237e-01 3.15184981e-01
-2.38754880e-03 5.74548364e-01 -1.20892382e+00 -6.32668197e-01
-8.89470503e-02 5.08072317e-01 -4.72028583e-01 1.81246996e-01
-5.25050126e-02 2.76689470e-01 -5.52750230e-01 3.39534730e-01
2.44736567e-01 -6.19210899e-01 7.95809269e-01 -4.15851176e-01
4.90309596e-01 2.75329769e-01 -1.03816247e+00 1.33511031e+00
-4.41311896e-01 3.55612010e-01 -8.03389251e-02 -1.49422371e+00
9.23951685e-01 4.13743913e-01 8.86536658e-01 -1.70122504e-01
-1.85618345e-02 -5.29435091e-02 2.24729091e-01 -2.60448754e-01
2.10232675e-01 -1.30767509e-01 3.07781428e-01 7.01009870e-01
1.81053311e-01 2.23490551e-01 3.35921377e-01 7.30730355e-01
1.23589277e+00 -1.51667729e-01 4.00470316e-01 -1.21253908e-01
7.34882653e-01 -4.76195127e-01 8.45009029e-01 4.20654804e-01
1.04302160e-01 2.05965355e-01 8.92326415e-01 -3.05816472e-01
-4.26767796e-01 -7.99994886e-01 -5.34120714e-03 1.51214826e+00
-5.12153991e-02 -8.80877554e-01 1.82721820e-02 -1.09107959e+00
4.54585105e-02 2.65927501e-02 -8.31931949e-01 -2.79364258e-01
-8.93287510e-02 -9.22382653e-01 1.34761542e-01 9.81896073e-02
-1.18776143e-01 -8.69475901e-01 3.14091563e-01 1.79671526e-01
-2.66135901e-01 -9.34113920e-01 -1.33496821e-01 3.49023283e-01
-9.58750427e-01 -1.41972458e+00 -5.17545879e-01 -5.66673458e-01
8.69978011e-01 4.34561253e-01 1.19020081e+00 4.07800466e-01
-6.55976608e-02 3.39767694e-01 -7.01348007e-01 5.78836277e-02
-9.24757868e-02 3.05638880e-01 1.57006621e-01 4.09148306e-01
3.76347095e-01 -9.05475736e-01 -5.82194984e-01 6.14092946e-01
-8.66208196e-01 -9.38319340e-02 4.53275204e-01 9.52504694e-01
4.02763367e-01 2.50806123e-01 1.04859090e+00 -1.73421502e+00
7.49578476e-01 -9.98280585e-01 -2.69380156e-02 5.11118829e-01
-8.41431141e-01 1.12501346e-01 3.61392200e-01 -5.22079468e-01
-8.80947948e-01 -6.13657869e-02 2.46000197e-03 -6.40307665e-02
2.96922952e-01 9.24700320e-01 5.27887940e-02 -5.34146614e-02
6.74638033e-01 -3.32857460e-01 -2.93779999e-01 -3.18394572e-01
5.32579005e-01 8.46206665e-01 -3.92701238e-01 -6.21864021e-01
6.64706230e-01 2.78248906e-01 3.80915552e-01 -5.94827294e-01
-1.36514342e+00 -8.64704072e-01 -7.41718948e-01 -4.45724994e-01
4.41585094e-01 -7.65757620e-01 -8.74911427e-01 -1.45572694e-02
-9.15312707e-01 9.67281163e-02 -3.73037495e-02 3.38665873e-01
-1.35342658e-01 4.37326372e-01 -6.65505469e-01 -7.71097600e-01
-2.84353703e-01 -9.02296841e-01 5.19783795e-01 -2.15140469e-02
-4.25943077e-01 -1.22990668e+00 2.57132202e-01 5.18608689e-01
4.29218769e-01 -4.22717705e-02 1.23807490e+00 -7.43770480e-01
-3.68112624e-01 -4.42317352e-02 -4.89003807e-01 1.36352420e-01
5.17022312e-01 -5.78106828e-02 -7.98776150e-01 -4.33819033e-02
-6.29122257e-01 -7.10509300e-01 1.06262994e+00 1.08769126e-01
9.58558440e-01 -3.26223612e-01 -5.06731808e-01 -1.19700488e-02
1.11635506e+00 -4.40130711e-01 2.91208714e-01 -1.07974425e-01
8.56256366e-01 1.22134626e+00 7.80516028e-01 4.02279735e-01
7.05799282e-01 6.86951280e-01 4.95867968e-01 1.10824518e-01
2.54317343e-01 -3.21493775e-01 1.34707466e-01 9.68223155e-01
-5.76144829e-02 -2.49552831e-01 -6.51793361e-01 3.09713840e-01
-2.25965047e+00 -9.01566684e-01 -4.66739565e-01 1.98052752e+00
1.02981651e+00 2.13840619e-01 1.95619598e-01 4.10151809e-01
9.28184927e-01 2.90537447e-01 -5.99101186e-01 -2.86223263e-01
5.60111366e-02 -6.03026226e-02 7.59289861e-02 5.72144568e-01
-9.24747169e-01 7.52932310e-01 5.57341719e+00 7.62175620e-01
-7.85726130e-01 2.66429186e-01 7.90163279e-01 1.81526899e-01
-5.49071670e-01 1.12449192e-01 -5.54859221e-01 1.05761901e-01
9.22891498e-01 -9.68163982e-02 1.03055120e-01 5.80224335e-01
-2.35805493e-02 -9.54936817e-02 -9.89051998e-01 7.90474296e-01
-1.55454874e-01 -1.03681207e+00 1.33029580e-01 -1.79236697e-03
6.14441812e-01 8.12259763e-02 -2.12793678e-01 2.39771709e-01
6.33878767e-01 -8.64262104e-01 -3.17333639e-02 5.78912973e-01
5.64572692e-01 -6.56153023e-01 7.01536119e-01 4.71088827e-01
-1.38546431e+00 -3.97061966e-02 -4.72777426e-01 -1.96676582e-01
5.91698196e-03 1.15997887e+00 -8.05921495e-01 6.63571239e-01
6.62979603e-01 1.32425475e+00 -4.25482631e-01 9.13921952e-01
-3.53006750e-01 5.56145072e-01 -3.81241024e-01 -1.15494065e-01
-1.23917259e-01 -2.82268107e-01 1.36701822e-01 8.83140981e-01
4.52480540e-02 1.76911026e-01 2.19981432e-01 3.22033226e-01
-3.99194628e-01 4.61603492e-01 -6.23367548e-01 -9.69794914e-02
2.29943350e-01 1.81162739e+00 -9.21474099e-01 2.07869336e-02
-5.48305809e-01 7.04841197e-01 8.27148914e-01 1.93184778e-01
-3.72672975e-01 -4.96346429e-02 4.95876551e-01 2.45075285e-01
-1.94704816e-01 1.22768231e-01 1.54329166e-01 -1.15015471e+00
-4.64377165e-01 -2.97106534e-01 8.90701056e-01 -4.84527349e-01
-1.87874949e+00 5.38736522e-01 -7.96292201e-02 -1.12275207e+00
-1.19480290e-01 -4.37945127e-01 -3.02767545e-01 5.03759503e-01
-1.67523634e+00 -1.26168180e+00 -1.10153876e-01 5.87038755e-01
6.97267875e-02 -5.32562584e-02 1.04002941e+00 5.62006295e-01
-3.89678061e-01 4.99708921e-01 -7.88105279e-02 1.38284177e-01
9.41064298e-01 -1.14889920e+00 -2.80586600e-01 2.00909838e-01
4.27297026e-01 7.37870216e-01 4.78563756e-01 -6.62777126e-01
-8.27172935e-01 -1.09304309e+00 9.58149970e-01 -1.01768278e-01
7.99745440e-01 -2.86054581e-01 -9.31158423e-01 6.44905806e-01
-7.17747733e-02 5.82891345e-01 1.36625719e+00 9.73523438e-01
-7.26475894e-01 -3.21946442e-01 -1.13400936e+00 3.03583950e-01
1.27176964e+00 -5.71935713e-01 -1.56147867e-01 7.41126060e-01
5.51152229e-01 2.98724115e-01 -1.21177566e+00 5.28393269e-01
5.84507823e-01 -6.78896904e-01 9.65279043e-01 -7.53333390e-01
4.04978991e-01 -2.39314005e-01 -3.24018449e-02 -1.41121626e+00
-5.31900883e-01 -1.37989551e-01 -3.92847449e-01 1.48442006e+00
6.14435613e-01 -6.83391750e-01 1.01150286e+00 3.45149159e-01
6.40163183e-01 -1.06925845e+00 -5.77753484e-01 -2.15251267e-01
-3.76712263e-01 -1.66875988e-01 2.00322375e-01 1.31475556e+00
3.47289354e-01 9.72357512e-01 -4.80946064e-01 1.07850451e-02
5.27542591e-01 1.12852052e-01 3.39293271e-01 -2.00110602e+00
-2.26464376e-01 -1.73160583e-01 -6.01953328e-01 -8.75011683e-01
5.67088604e-01 -1.20995605e+00 -3.44922304e-01 -1.61728585e+00
4.16215271e-01 -1.13871467e+00 -7.31716216e-01 5.32564223e-01
-2.23833069e-01 3.87426049e-01 -1.80937514e-01 2.62367934e-01
-9.51816976e-01 5.32614291e-01 9.47014093e-01 -3.13619107e-01
7.76852891e-02 3.66618186e-01 -7.80351281e-01 6.90704882e-01
5.93566597e-01 -7.80732691e-01 -9.77212667e-01 4.65227850e-02
1.06333888e+00 1.63716242e-01 -1.74409583e-01 -3.95121187e-01
3.90258551e-01 -1.83777556e-01 7.00122267e-02 -4.30369526e-01
5.07822514e-01 -1.06994331e+00 1.46123722e-01 2.64197439e-01
-8.70472074e-01 -3.90010238e-01 -6.50090694e-01 1.08665276e+00
-1.94975108e-01 -3.10794979e-01 5.06591499e-01 -4.58693244e-02
-3.34834546e-01 5.08934438e-01 -1.31800741e-01 -9.15578529e-02
8.99243534e-01 1.56636938e-01 -2.34363914e-01 -6.59094095e-01
-8.90886784e-01 3.50976676e-01 -8.75435323e-02 6.33283496e-01
6.95476592e-01 -1.43002379e+00 -6.82624280e-01 -3.18148702e-01
3.72528285e-01 -1.31042466e-01 1.53281912e-01 7.36625254e-01
1.70614958e-01 1.60258532e-01 7.39697693e-03 -4.17306215e-01
-1.54699910e+00 3.15440476e-01 -1.31344292e-02 -8.13530147e-01
-3.27617936e-02 9.19200599e-01 1.00081414e-01 -8.67777646e-01
1.03201613e-01 2.34343842e-01 -9.79230821e-01 5.29836535e-01
3.78806621e-01 2.49141976e-01 1.93194114e-02 -8.39215517e-01
-3.58811349e-01 4.52907741e-01 -3.97130698e-01 2.65595675e-01
1.60102034e+00 -7.02832863e-02 -7.08025277e-01 7.34930754e-01
1.17257702e+00 -1.90273225e-01 -5.29472172e-01 -9.20302749e-01
2.80857384e-01 -3.51652712e-01 -8.29735696e-02 -4.94672835e-01
-1.29796731e+00 7.47995019e-01 2.28534520e-01 4.56803411e-01
8.47360075e-01 1.80368781e-01 2.43576452e-01 5.11204243e-01
4.31015700e-01 -8.20944071e-01 3.15112174e-01 5.05830646e-01
5.98705769e-01 -1.26844203e+00 3.97808552e-01 -9.64964092e-01
-4.57635015e-01 9.49581265e-01 7.26441920e-01 -7.76933432e-02
1.21586096e+00 -1.73998967e-01 -2.86340922e-01 -2.26738781e-01
-9.80347753e-01 -1.58138618e-01 6.05912864e-01 3.87130231e-01
9.54454243e-01 2.98106801e-02 -5.58701813e-01 2.63091177e-01
2.73926407e-01 -2.64229804e-01 2.20571280e-01 6.28317714e-01
-3.64587516e-01 -1.68302250e+00 1.02673713e-02 1.14015210e+00
-4.62430388e-01 -1.79443397e-02 -4.83846068e-01 3.90602984e-02
-2.75326353e-02 1.09496033e+00 -3.29516172e-01 -5.52231729e-01
-1.17804304e-01 -7.99228922e-02 2.33328611e-01 -9.89363015e-01
-5.81330597e-01 -3.01630378e-01 2.31490746e-01 -4.12911981e-01
-9.32066381e-01 -3.67566109e-01 -1.19895804e+00 -2.60810673e-01
-7.72368193e-01 4.11958367e-01 3.27662319e-01 9.79452133e-01
2.92105108e-01 5.93541980e-01 6.00619197e-01 -4.87940341e-01
-4.95474637e-02 -1.07971287e+00 -6.67053163e-01 5.10348260e-01
-2.30588522e-02 -1.01860070e+00 -2.13048920e-01 -3.82037848e-01] | [7.5199713706970215, 6.488358974456787] |
6d9c523a-0af0-4ab9-8b1c-19ed8b7f0337 | ddr-id-dual-deep-reconstruction-networks | 2007.09431 | null | https://arxiv.org/abs/2007.09431v1 | https://arxiv.org/pdf/2007.09431v1.pdf | DDR-ID: Dual Deep Reconstruction Networks Based Image Decomposition for Anomaly Detection | One pivot challenge for image anomaly (AD) detection is to learn discriminative information only from normal class training images. Most image reconstruction based AD methods rely on the discriminative capability of reconstruction error. This is heuristic as image reconstruction is unsupervised without incorporating normal-class-specific information. In this paper, we propose an AD method called dual deep reconstruction networks based image decomposition (DDR-ID). The networks are trained by jointly optimizing for three losses: the one-class loss, the latent space constrain loss and the reconstruction loss. After training, DDR-ID can decompose an unseen image into its normal class and the residual components, respectively. Two anomaly scores are calculated to quantify the anomalous degree of the image in either normal class latent space or reconstruction image space. Thereby, anomaly detection can be performed via thresholding the anomaly score. The experiments demonstrate that DDR-ID outperforms multiple related benchmarking methods in image anomaly detection using MNIST, CIFAR-10 and Endosome datasets and adversarial attack detection using GTSRB dataset. | ['Yiqun Li', 'Shudong Xie', 'Tin Lay Nwe', 'Sheng Dong', 'Dongyun Lin'] | 2020-07-18 | null | null | null | null | ['adversarial-attack-detection', 'adversarial-attack-detection'] | ['computer-vision', 'knowledge-base'] | [ 3.89156938e-01 -1.45552069e-01 1.66106477e-01 -2.37895653e-01
-6.38167381e-01 -2.48664021e-01 3.79103452e-01 2.46572196e-01
-4.15139884e-01 1.56619936e-01 -6.20579049e-02 -2.72772282e-01
1.01155221e-01 -6.50643170e-01 -9.27026510e-01 -1.04589593e+00
-1.06427610e-01 1.93197161e-01 -1.52134985e-01 -1.60438493e-02
3.56815964e-01 6.64160967e-01 -1.31369615e+00 5.19224823e-01
6.46354854e-01 1.45387232e+00 -2.58322448e-01 6.66029274e-01
5.71864955e-02 9.38477576e-01 -5.03636777e-01 -2.00385258e-01
7.12112963e-01 -3.74848366e-01 -6.08402193e-01 2.15298012e-01
5.25323868e-01 -6.25383198e-01 -6.11855030e-01 1.48953843e+00
5.27629077e-01 8.07006583e-02 7.49027789e-01 -1.52965224e+00
-6.54702365e-01 4.78745066e-02 -6.28733933e-01 7.03648984e-01
2.51228034e-01 4.06677961e-01 8.12569439e-01 -8.18339050e-01
6.06067002e-01 1.16171932e+00 5.98946869e-01 5.92152119e-01
-1.40509641e+00 -3.54380518e-01 2.51318574e-01 3.43028247e-01
-1.04162836e+00 -1.34205222e-01 9.08467829e-01 -6.83780789e-01
8.34303141e-01 2.84261644e-01 2.96762735e-01 1.32449031e+00
4.20607358e-01 8.92581165e-01 8.91669035e-01 -1.26099259e-01
1.45777583e-01 -2.04010516e-01 -4.16044481e-02 7.32539952e-01
2.06822842e-01 1.46432295e-01 -4.31291938e-01 -2.54333615e-01
5.86772740e-01 4.80626792e-01 -3.50285500e-01 -3.91794950e-01
-1.04921842e+00 7.88696826e-01 3.12697113e-01 -8.99290144e-02
-5.63708425e-01 1.19903274e-01 8.44047606e-01 7.51379550e-01
4.97311741e-01 2.00477958e-01 -3.16936672e-01 2.20255747e-01
-6.42676711e-01 1.65915623e-01 2.36906067e-01 3.20422441e-01
4.86669272e-01 4.32973981e-01 -2.06472129e-01 6.69633687e-01
5.98374784e-01 2.73125529e-01 9.75215912e-01 -7.76361942e-01
2.34608755e-01 6.87039256e-01 -2.09679678e-01 -1.35184634e+00
-5.25238030e-02 -4.42009896e-01 -1.07787585e+00 6.38509095e-01
3.97480845e-01 2.77318448e-01 -1.21653950e+00 1.70363462e+00
2.80340821e-01 3.32835436e-01 3.13588858e-01 9.00578320e-01
7.43853331e-01 5.61612010e-01 -7.09481537e-02 7.03341365e-02
1.05806088e+00 -7.35018432e-01 -6.30703330e-01 -3.24252658e-02
6.08046055e-01 -7.98878133e-01 1.06824589e+00 7.09010005e-01
-9.03724253e-01 -2.83627719e-01 -1.18521500e+00 4.39757034e-02
-2.28435948e-01 1.45409657e-02 3.60949129e-01 3.02403837e-01
-6.71001792e-01 6.74134910e-01 -9.68861163e-01 5.91232777e-02
6.17660463e-01 1.93682566e-01 -5.81749201e-01 -1.42192692e-01
-8.42280626e-01 4.64900523e-01 2.66508013e-01 -1.19062059e-01
-1.55302846e+00 -6.85249150e-01 -8.22570205e-01 -1.68393224e-01
-3.12539539e-03 -4.49892402e-01 5.65864265e-01 -1.14055407e+00
-1.23471045e+00 1.30960155e+00 3.57485026e-01 -9.63889360e-01
7.10543096e-01 -1.42581195e-01 -4.33727562e-01 3.25574458e-01
7.99491853e-02 3.32133085e-01 1.27412295e+00 -1.22843528e+00
-3.07710558e-01 -6.02224827e-01 -1.45721301e-01 4.08243872e-02
-6.37873113e-02 -2.24079803e-01 3.75946634e-03 -1.15607858e+00
6.38961852e-01 -7.62736917e-01 -3.16492975e-01 1.65566310e-01
-6.71602845e-01 5.02726398e-02 1.16857398e+00 -9.74833071e-01
8.12622190e-01 -2.56279254e+00 8.08939859e-02 3.23570698e-01
3.62144262e-01 6.22314736e-02 -3.56925935e-01 -1.46111369e-01
-5.63485622e-01 -1.57138154e-01 -3.47395450e-01 -3.32165986e-01
-1.24430262e-01 2.85901070e-01 -7.30766416e-01 9.50320601e-01
1.30245194e-01 5.20441532e-01 -6.20050371e-01 -1.35980114e-01
1.34000525e-01 4.02826041e-01 -7.54931867e-01 3.23746383e-01
-2.34386832e-01 7.57617295e-01 -2.05004171e-01 8.56757104e-01
8.69478941e-01 -4.15995605e-02 -3.35571587e-01 -3.70591044e-01
2.82770425e-01 -8.12003016e-02 -1.00711501e+00 1.72062480e+00
-7.43876398e-02 6.04953527e-01 -4.16897312e-02 -1.52757800e+00
7.88479149e-01 1.41378433e-01 7.04717100e-01 -7.98340142e-01
3.79264401e-03 2.19248176e-01 8.67996588e-02 -5.71660757e-01
8.80920738e-02 1.98767453e-01 2.61377633e-01 4.92203861e-01
2.65298456e-01 6.45227730e-01 -8.92305151e-02 4.08888817e-01
1.28596783e+00 -2.74061449e-02 -6.03797324e-02 -1.99092224e-01
6.16220117e-01 -2.98990995e-01 9.33699489e-01 7.11250246e-01
-4.56531644e-01 6.52919352e-01 7.52060592e-01 -9.39438820e-01
-1.27490234e+00 -1.39900196e+00 -2.10768893e-01 8.25069368e-01
1.77640602e-01 -8.69798586e-02 -7.00895488e-01 -1.02663422e+00
-1.90991821e-04 4.63046253e-01 -6.38670683e-01 -5.29060125e-01
-5.97316146e-01 -1.01699293e+00 6.47360623e-01 1.74025729e-01
6.00042105e-01 -9.85381484e-01 -3.28398645e-01 -9.55393389e-02
-3.13610673e-01 -1.01870286e+00 -4.30759192e-01 1.23055525e-01
-9.05896246e-01 -1.08736992e+00 -3.49862337e-01 -5.09593606e-01
1.10626101e+00 -1.97553262e-02 9.53729212e-01 2.82013535e-01
-8.41265798e-01 5.60968101e-01 -2.23299563e-01 -1.28264189e-01
-5.96357703e-01 -7.02252269e-01 2.20607787e-01 4.26093608e-01
3.10307771e-01 -7.41523206e-01 -1.00962222e+00 2.95184046e-01
-1.18102849e+00 -3.07026505e-01 6.15486383e-01 1.06430483e+00
1.15526950e+00 1.24072298e-01 2.91946262e-01 -6.72566950e-01
2.47191727e-01 -6.58218384e-01 -6.43121600e-01 -2.88665611e-02
-7.13014066e-01 2.05886260e-01 6.05051577e-01 -4.60971951e-01
-7.36727655e-01 1.74042225e-01 -3.44194412e-01 -9.24345732e-01
-3.30268979e-01 2.32146889e-01 -2.86265969e-01 -1.53661430e-01
7.79147625e-01 7.39737093e-01 1.24154322e-01 -4.14035052e-01
-3.53884548e-02 1.32557780e-01 7.94042706e-01 -4.28197980e-01
7.62600958e-01 8.06977630e-01 2.44923234e-01 -6.28941119e-01
-8.60660672e-01 -5.98058105e-01 -2.66272217e-01 -7.97519386e-02
9.40416098e-01 -8.78748834e-01 -5.18995583e-01 7.67107785e-01
-8.70387912e-01 -1.04805753e-01 -5.32419205e-01 3.76419753e-01
-5.24845302e-01 9.01483536e-01 -6.51935220e-01 -4.72112089e-01
-4.89442199e-01 -1.45729399e+00 9.21221554e-01 -2.93166399e-01
3.54830116e-01 -8.17731440e-01 1.96666140e-02 2.99468726e-01
2.05849573e-01 7.34350026e-01 1.01691973e+00 -9.58230495e-01
-6.34777427e-01 -3.81993622e-01 -4.27882038e-02 9.61368263e-01
-1.35702327e-01 -2.37828732e-01 -8.44785392e-01 -6.35369182e-01
4.49882686e-01 -3.29658478e-01 1.19200742e+00 4.04489011e-01
1.70684934e+00 -7.56541133e-01 6.95371106e-02 1.14473045e+00
1.49826157e+00 1.07356943e-02 9.27325249e-01 5.53483248e-01
8.24430048e-01 1.76570237e-01 3.90629590e-01 4.53937829e-01
-1.17189810e-01 5.40482461e-01 9.67612684e-01 3.70791219e-02
-2.14416217e-02 3.74511741e-02 7.37548828e-01 4.19288903e-01
1.75875783e-01 -1.78147957e-01 -7.19082773e-01 3.40560257e-01
-1.79668748e+00 -1.06807733e+00 2.34902035e-02 2.25618315e+00
4.10091788e-01 1.16657004e-01 2.73376852e-02 3.29596937e-01
7.10985124e-01 5.21504804e-02 -9.08175945e-01 -2.76411772e-01
-4.22984630e-01 -1.93589896e-01 4.03533995e-01 1.27529711e-01
-1.42022121e+00 4.22357410e-01 5.49119329e+00 8.53000343e-01
-1.21015561e+00 5.51560111e-02 1.01767302e+00 5.48246279e-02
-1.66972324e-01 -1.72015890e-01 -4.10518438e-01 6.63892806e-01
6.15071654e-01 3.13819766e-01 3.35630059e-01 9.44814801e-01
-1.30922422e-01 1.81477487e-01 -9.99736011e-01 1.19290173e+00
2.05704153e-01 -1.18867743e+00 5.07206619e-01 1.72881171e-01
5.41807890e-01 2.49362320e-01 6.43351257e-01 8.31787363e-02
3.03168129e-02 -1.05118895e+00 4.16174978e-01 5.02907693e-01
5.79505742e-01 -8.06169331e-01 6.44132137e-01 1.14480637e-01
-7.30330706e-01 -1.67834356e-01 -5.19890189e-01 4.47782367e-01
-2.49928430e-01 6.44282043e-01 -3.99535775e-01 3.18835825e-01
9.97591317e-01 7.76625037e-01 -5.57357192e-01 8.02661717e-01
-4.65004072e-02 5.23974836e-01 -2.58068025e-01 9.55378056e-01
2.21868128e-01 -4.25797701e-01 1.30986583e+00 9.22614276e-01
2.09447578e-01 -2.28236407e-01 2.66932160e-01 8.62990081e-01
-2.13856712e-01 -1.17894590e-01 -6.81609690e-01 7.71094114e-02
-7.39900097e-02 1.26596224e+00 -7.43580222e-01 -1.63679570e-02
-2.12137580e-01 1.48758435e+00 -2.38168589e-03 3.26415837e-01
-7.34376431e-01 -1.59409001e-01 8.25747132e-01 -3.57610546e-02
2.11530268e-01 1.92563713e-01 -2.13152338e-02 -1.47065306e+00
1.63512543e-01 -1.01132905e+00 7.45088577e-01 -3.30104113e-01
-1.56234670e+00 6.79212511e-01 -4.31474835e-01 -1.45411634e+00
-1.33556500e-01 -8.12894821e-01 -8.00553024e-01 4.34049189e-01
-1.40149009e+00 -1.02741599e+00 -3.24164271e-01 8.16717863e-01
4.92662996e-01 -7.90385246e-01 6.74287200e-01 5.27744055e-01
-8.10537517e-01 9.26149130e-01 2.31599793e-01 5.23227394e-01
5.39818645e-01 -1.48811483e+00 2.55168349e-01 1.28090370e+00
4.93118614e-02 3.50004494e-01 8.37100506e-01 -4.98112738e-01
-1.03931201e+00 -1.33615386e+00 -2.39193793e-02 -3.34275603e-01
4.28438216e-01 -1.42802730e-01 -1.09952748e+00 7.10028708e-01
-1.15583062e-01 7.81004488e-01 6.81587577e-01 -5.61498046e-01
-8.10754955e-01 -2.98004337e-02 -1.54109633e+00 4.01533306e-01
6.22041345e-01 -4.85735953e-01 -2.61198133e-01 5.75872481e-01
7.56704211e-01 -3.92880738e-01 -8.69411051e-01 5.44892728e-01
1.98683813e-01 -1.01814938e+00 1.37345469e+00 -8.55911136e-01
5.11211574e-01 -4.70006108e-01 -4.84008908e-01 -1.04666197e+00
4.81563667e-03 -4.80903000e-01 -4.64829087e-01 8.22791576e-01
6.39024228e-02 -6.07547224e-01 8.28954935e-01 -3.03954389e-02
-3.48587006e-01 -9.12890792e-01 -1.05201983e+00 -7.77364194e-01
1.46236755e-02 -3.39688838e-01 4.28614676e-01 1.03201091e+00
-6.78545952e-01 -2.66260514e-03 -3.77529442e-01 5.00507414e-01
1.02865303e+00 -7.70198852e-02 5.75584948e-01 -1.05345678e+00
-3.52526516e-01 -2.58692890e-01 -1.14742517e+00 -6.28746629e-01
8.88971016e-02 -1.06265664e+00 -1.59069046e-01 -8.99776459e-01
1.68124750e-01 -2.60659844e-01 -7.48465061e-01 4.31983083e-01
8.19779187e-02 4.73817855e-01 -1.28179997e-01 5.07218301e-01
-7.58768737e-01 6.01684630e-01 9.82783735e-01 -2.97326893e-01
1.56539217e-01 -6.99111521e-02 -2.19648793e-01 1.00424635e+00
8.96968722e-01 -6.03991210e-01 -2.30775356e-01 -1.70248613e-01
3.12273741e-01 -2.96388954e-01 7.79400706e-01 -1.10133708e+00
6.29414693e-02 9.13548246e-02 6.14892542e-01 -6.78654134e-01
5.66110127e-02 -7.59620905e-01 -8.66220817e-02 6.51316524e-01
-3.87883097e-01 2.29455560e-01 5.94192110e-02 1.01514924e+00
-3.35040957e-01 -1.09238252e-01 1.26845336e+00 -1.29429623e-01
-6.60089433e-01 6.26622856e-01 -3.82783085e-01 2.98915748e-02
1.02265298e+00 -2.54548401e-01 -2.09925771e-01 -7.49328062e-02
-9.39482212e-01 -1.25520587e-01 3.03879291e-01 2.18853086e-01
1.15003169e+00 -1.48359644e+00 -7.75945365e-01 5.61161458e-01
3.42872828e-01 1.23267151e-01 4.88695651e-01 1.01329517e+00
-8.06148529e-01 -6.06597178e-02 -5.59311807e-01 -9.39445674e-01
-1.21503878e+00 7.43250132e-01 5.71610868e-01 -4.74365175e-01
-1.07681859e+00 9.12681162e-01 4.97178197e-01 -4.54078048e-01
4.07905310e-01 2.03644801e-02 -9.29434896e-02 -3.21731925e-01
5.01971841e-01 3.11583877e-01 8.56699720e-02 -7.72782326e-01
-2.21288294e-01 1.82924256e-01 -4.65087175e-01 2.43705750e-01
1.31737101e+00 -1.46554917e-01 -4.42306459e-01 3.13717965e-03
1.43146908e+00 -3.22060585e-01 -1.33924639e+00 -2.57546335e-01
-5.73130436e-02 -5.41181326e-01 2.78058261e-01 -6.34235144e-01
-1.46119416e+00 7.76494026e-01 1.32397664e+00 -7.66042098e-02
1.45157015e+00 -2.76481897e-01 1.05069530e+00 4.12945360e-01
-3.55030149e-01 -1.03431690e+00 8.27619791e-01 3.35106224e-01
1.00612533e+00 -1.46137261e+00 -9.23524052e-03 8.27151760e-02
-4.56531495e-01 9.37684536e-01 7.77252078e-01 -5.52795053e-01
6.81449533e-01 -6.79313019e-02 1.84472978e-01 -4.39480424e-01
-4.73424435e-01 2.14743257e-01 4.01302129e-01 5.11706173e-01
-3.34668234e-02 -2.79700726e-01 2.25104734e-01 4.13610876e-01
1.62391514e-01 -6.62720740e-01 4.54262108e-01 6.43229306e-01
-1.75377160e-01 -8.60418558e-01 -3.97855401e-01 6.77541554e-01
-1.17371798e+00 -2.72597838e-02 -8.20018202e-02 3.73875737e-01
1.43738762e-01 3.55038375e-01 1.61180392e-01 -3.55212271e-01
1.13826590e-02 -1.22139953e-01 2.77841419e-01 -2.20310479e-01
-2.52926320e-01 -4.81792577e-02 -6.01975501e-01 -9.54666138e-01
-9.62413996e-02 -6.98715448e-01 -1.13546240e+00 -2.20696896e-01
-1.78491250e-01 -4.62098457e-02 7.24465668e-01 7.65106857e-01
3.80978018e-01 5.57858765e-01 7.90592611e-01 -4.89558607e-01
-6.27012968e-01 -5.67492247e-01 -5.17391026e-01 9.61554229e-01
8.12376380e-01 -5.16823888e-01 -8.45960855e-01 8.43571723e-02] | [7.697807312011719, 2.0797438621520996] |
5fa704b0-6fd0-483f-b38d-4aa104892a86 | avt-audio-video-transformer-for-multimodal | null | null | https://openreview.net/pdf?id=yFuHxmSwGus | https://openreview.net/pdf?id=yFuHxmSwGus | AVT: Audio-Video Transformer for Multimodal Action Recognition | Action recognition is an essential field for video understanding. To learn from heterogeneous data sources effectively, in this work, we propose a novel multimodal action recognition approach termed Audio-Video Transformer (AVT). AVT uses a combination of video and audio signals to improve action recognition accuracy, leveraging the effective spatio-temporal representation by the video Transformer. For multimodal fusion, simply concatenating multimodal tokens in a cross-modal Transformer requires large computational and memory resources, instead we reduce the cross-modality complexity through an audio-video bottleneck Transformer. To improve the learning efficiency of multimodal Transformer, we integrate self-supervised objectives, i.e., audio-video contrastive learning, audio-video matching, and masked audio and video learning, into AVT training, which maps diverse audio and video representations into a common multimodal representation space. We further propose a masked audio segment loss to learn semantic audio activities in AVT. Extensive experiments and ablation studies on three public datasets and two in-house datasets consistently demonstrate the effectiveness of the proposed AVT. Specifically, AVT outperforms its previous state-of-the-art counterparts on Kinetics-Sounds and Epic-Kitchens-100 datasets by 8% and 1%, respectively, without external training data. AVT also surpasses one of the previous state-of-the-art video Transformers by 10% on the VGGSound dataset by leveraging the audio signal. Compared to one of the previous state-of-the-art multimodal Transformers, AVT is 1.3x more efficient in terms of FLOPs and improves the accuracy by 4.2% on Epic-Kitchens-100. Visualization results further demonstrate that the audio provides complementary and discriminative features, and our AVT can effectively understand the action from a combination of audio and video. | ['Mohamed Omar', 'Linda Liu', 'Xiang Hao', 'Xiaohang Sun', 'Kevin Hsu', 'Jingru Yi', 'Wentao Zhu'] | 2022-09-22 | null | null | null | submitted-to-iclr-2022-9 | ['audio-classification', 'video-understanding', 'multi-modal-classification'] | ['audio', 'computer-vision', 'miscellaneous'] | [ 5.44478476e-01 -4.19963509e-01 -1.95054621e-01 -5.73687516e-02
-1.38051140e+00 -4.89642471e-01 3.80621433e-01 -9.58342031e-02
-4.18544829e-01 3.74731034e-01 3.45224857e-01 1.20016702e-01
-1.46641746e-01 -4.17649865e-01 -9.51428413e-01 -6.75048172e-01
9.60826427e-02 2.91544665e-02 3.90622020e-01 5.71292713e-02
-2.78674588e-02 -1.29257128e-01 -1.91494274e+00 9.27361965e-01
6.51117086e-01 1.40951550e+00 3.41304615e-02 7.31706619e-01
-1.66192636e-01 1.15245605e+00 -4.13550586e-01 -2.00405106e-01
8.42508152e-02 -3.81735355e-01 -6.51432633e-01 6.60333037e-02
6.27007902e-01 -3.49746287e-01 -5.54967701e-01 5.76271236e-01
6.78401351e-01 9.85175967e-02 3.52853298e-01 -1.59765542e+00
-3.60052317e-01 5.51671386e-01 -6.54615998e-01 3.18815522e-02
6.91992640e-01 2.61582226e-01 9.65035319e-01 -8.84981334e-01
3.44555169e-01 1.39394104e+00 7.88404167e-01 3.82454067e-01
-1.19244158e+00 -8.93461227e-01 2.99589008e-01 5.40978730e-01
-1.35408521e+00 -7.28504062e-01 6.62128627e-01 -4.56931680e-01
9.38170552e-01 3.69881243e-01 7.39761114e-01 1.41785657e+00
-2.63408482e-01 1.28555429e+00 8.01560163e-01 -1.62098229e-01
3.28599066e-02 -3.22690070e-01 -2.04561129e-01 7.89771855e-01
-4.13996667e-01 -1.09659377e-02 -1.19677567e+00 -1.25100881e-01
5.91169119e-01 1.77217662e-01 -2.79289424e-01 -1.87387213e-01
-1.37636781e+00 5.07523894e-01 1.63555723e-02 1.17051072e-01
-2.03521892e-01 5.45370162e-01 7.56366730e-01 3.12407434e-01
3.23735148e-01 1.45921838e-02 -3.39085668e-01 -8.03631186e-01
-8.51325035e-01 -4.11870554e-02 3.72837275e-01 6.16114438e-01
5.31927466e-01 2.49530151e-01 -3.35683525e-01 8.57989311e-01
3.24485958e-01 6.90555334e-01 5.38371384e-01 -1.13656652e+00
7.06301689e-01 7.04824090e-01 -3.10156763e-01 -7.42530525e-01
-1.46765813e-01 -6.56809583e-02 -5.47394872e-01 -1.83854863e-01
3.58730406e-01 9.47705656e-02 -9.64768231e-01 1.70215714e+00
2.56508768e-01 7.56456017e-01 1.36893257e-01 8.44236195e-01
9.44848120e-01 6.93855166e-01 8.38581994e-02 -5.32290377e-02
1.25877070e+00 -1.04655254e+00 -6.01812899e-01 -4.11099456e-02
5.79030514e-01 -6.30009592e-01 1.18798459e+00 5.00101566e-01
-9.03762162e-01 -6.85529113e-01 -9.42541182e-01 3.96822654e-02
-1.07558183e-01 2.97421873e-01 6.17061317e-01 4.02379543e-01
-6.62251651e-01 3.43681127e-01 -1.13370490e+00 -1.98251531e-01
5.96986890e-01 3.06272656e-01 -6.28571451e-01 -1.59311011e-01
-1.02491736e+00 2.24580824e-01 2.91667044e-01 -1.23874657e-01
-1.42046690e+00 -1.01242316e+00 -9.98293757e-01 7.03261718e-02
7.10925221e-01 -5.72355270e-01 1.21726739e+00 -1.15573919e+00
-1.66004884e+00 4.09954339e-01 -2.59158105e-01 -3.09143543e-01
4.28395241e-01 -4.90495056e-01 -6.01132393e-01 7.37705290e-01
-2.89385510e-03 7.85162628e-01 1.15049887e+00 -9.46000934e-01
-7.74209142e-01 -2.57336438e-01 4.79513928e-02 4.00484204e-02
-7.68890202e-01 1.71390234e-03 -8.98339927e-01 -7.92603374e-01
-9.13010463e-02 -9.19357657e-01 3.30591768e-01 1.52360916e-01
-1.70395762e-01 7.12799327e-03 1.17237401e+00 -6.36352658e-01
1.31597185e+00 -2.37501097e+00 2.63338715e-01 -1.53846011e-01
2.83198684e-01 2.09881678e-01 -4.24376309e-01 3.75371367e-01
-1.42820507e-01 -1.22679219e-01 -1.95132867e-01 -4.66297388e-01
8.07203427e-02 1.51633814e-01 -1.77215040e-01 3.81337553e-01
2.51300305e-01 8.76349449e-01 -7.70896792e-01 -5.91256201e-01
3.17456126e-01 7.71343827e-01 -6.53229833e-01 8.13237578e-02
-1.27741292e-01 4.81297314e-01 -3.54477793e-01 9.95496929e-01
2.58106589e-01 -2.46203691e-01 1.50255844e-01 -4.87929344e-01
9.80419219e-02 1.34092852e-01 -1.06385815e+00 2.11352491e+00
-4.78204101e-01 6.17945492e-01 -4.67305854e-02 -1.15665400e+00
5.05278528e-01 5.69605052e-01 8.69470656e-01 -8.69853556e-01
-1.06763385e-01 1.72431260e-01 -4.12938446e-01 -6.87407792e-01
1.79045886e-01 1.59407705e-01 -1.46589309e-01 4.05292839e-01
4.20014262e-01 5.08452058e-01 2.22212344e-01 3.28309178e-01
1.37091100e+00 4.42524910e-01 -2.05377579e-01 3.54216814e-01
5.64939737e-01 -3.38659942e-01 6.33089900e-01 4.58700687e-01
-5.61833009e-02 5.58373272e-01 4.94960666e-01 -1.52668908e-01
-5.33225298e-01 -1.19779801e+00 2.30733946e-01 1.49392009e+00
1.31752178e-01 -9.43051338e-01 -6.19098008e-01 -7.97998548e-01
8.66134539e-02 1.19420506e-01 -5.56434512e-01 -2.68014789e-01
-5.93183815e-01 -3.79476130e-01 8.61665249e-01 8.88501167e-01
7.60448217e-01 -7.49063969e-01 -5.00631154e-01 1.20364591e-01
-6.01007760e-01 -1.40812087e+00 -6.68696761e-01 -1.72394723e-01
-6.95371926e-01 -1.12457287e+00 -6.25119388e-01 -3.82558167e-01
1.80779278e-01 4.72431839e-01 7.89272428e-01 -1.70186654e-01
-3.64045769e-01 9.09846544e-01 -6.20308161e-01 1.18571240e-03
-1.61007971e-01 -3.64123955e-02 5.18498458e-02 5.81552863e-01
9.31120813e-02 -8.55039477e-01 -5.92005610e-01 5.08164465e-01
-9.95200038e-01 1.40722215e-01 6.77594960e-01 8.61881971e-01
6.52663112e-01 -5.00174053e-02 4.50276375e-01 -3.38545203e-01
-7.36958757e-02 -4.61214751e-01 -3.36305290e-01 2.77786911e-01
-1.50087878e-01 7.76767433e-02 4.36196327e-01 -8.13641667e-01
-1.00076306e+00 3.15147340e-01 6.64758310e-02 -1.08687556e+00
-6.10808609e-03 5.66345513e-01 -3.93500924e-01 -1.11166202e-01
3.45905781e-01 3.69615436e-01 5.65161705e-02 -6.29333556e-01
2.31652126e-01 6.06666446e-01 8.13004076e-01 -6.05564833e-01
5.75182557e-01 5.88322401e-01 -1.32545456e-01 -7.74442494e-01
-5.19676387e-01 -5.28942645e-01 -3.40200603e-01 -5.00600159e-01
9.01764512e-01 -1.23272943e+00 -1.06976748e+00 7.28459358e-01
-8.04651797e-01 -2.25611567e-01 -1.10008240e-01 6.83377564e-01
-6.13575995e-01 4.46065813e-01 -4.69234943e-01 -7.85878897e-01
-1.87413484e-01 -1.23315501e+00 1.51150334e+00 -6.11504763e-02
-1.30843237e-01 -6.78940117e-01 -1.10538594e-01 8.86968255e-01
1.91239238e-01 2.72411764e-01 5.43940008e-01 -4.23905492e-01
-7.06077576e-01 -1.57035187e-01 -1.95322990e-01 4.08387899e-01
4.16760892e-02 1.37990830e-03 -1.18043745e+00 -3.02973270e-01
-4.35496122e-01 -5.88287830e-01 1.29664266e+00 1.30129382e-01
1.30582452e+00 -1.34324431e-01 -3.53626817e-01 7.11841762e-01
1.11514020e+00 2.25287572e-01 6.13967657e-01 1.73115492e-01
9.82169688e-01 2.54777431e-01 6.48621023e-01 6.76792681e-01
4.37011480e-01 9.84389186e-01 5.85867107e-01 -2.03740597e-03
-1.79945722e-01 -4.89391506e-01 9.98102546e-01 7.77244687e-01
-1.05015017e-01 -1.12694561e-01 -7.60609150e-01 5.15979767e-01
-2.17372632e+00 -1.20158005e+00 1.89831018e-01 2.07440829e+00
6.66919470e-01 -6.00603335e-02 4.83351380e-01 4.31894451e-01
4.82062906e-01 2.72931159e-01 -5.57828665e-01 3.99844438e-01
-1.06707402e-01 1.95388705e-01 1.97771773e-01 1.63489655e-01
-1.38439775e+00 7.60594547e-01 5.53872061e+00 1.21504581e+00
-1.26292074e+00 2.51174867e-01 1.43405333e-01 -6.07989073e-01
-1.47792757e-01 -1.22737564e-01 -5.70542216e-01 6.40074074e-01
1.05781937e+00 1.09594353e-01 4.78736103e-01 5.93112826e-01
2.26064444e-01 -2.27296539e-02 -1.22048783e+00 1.40475810e+00
2.81640947e-01 -1.31851208e+00 8.41481462e-02 -1.11813009e-01
2.54071832e-01 -8.06666315e-02 1.11943260e-01 4.99677777e-01
-6.29502460e-02 -9.90626812e-01 9.59842741e-01 5.60158670e-01
8.98010671e-01 -6.12319767e-01 5.13636529e-01 -6.93950802e-02
-1.78609431e+00 -3.82075906e-01 2.55759507e-01 1.86204806e-01
3.15706611e-01 2.54991889e-01 -2.24035993e-01 8.19890916e-01
1.06307030e+00 1.20123100e+00 -6.34427190e-01 8.54463935e-01
-2.79368423e-02 8.11864555e-01 -3.29242080e-01 4.73103493e-01
2.18800038e-01 1.90623015e-01 5.67540407e-01 1.14344203e+00
4.15406853e-01 -4.21538725e-02 4.29427058e-01 4.12682742e-01
-1.70989349e-01 -8.60781297e-02 -2.82503158e-01 -4.58212078e-01
4.45987225e-01 1.02642572e+00 -2.42994696e-01 -4.00446475e-01
-5.69028854e-01 9.64398026e-01 -1.30875446e-02 2.46813715e-01
-1.27409530e+00 -2.96379298e-01 9.06228423e-01 -8.10442269e-02
5.95088601e-01 8.54129493e-02 2.88801312e-01 -1.28931713e+00
3.64874780e-01 -1.20682454e+00 5.92429817e-01 -7.82726467e-01
-9.67385590e-01 4.88773048e-01 1.24977276e-01 -1.57548928e+00
-1.98309988e-01 -4.56966966e-01 -2.27728024e-01 2.46376142e-01
-1.23684072e+00 -1.45210600e+00 -5.48771977e-01 1.02931213e+00
5.96376419e-01 -3.40455443e-01 7.78445125e-01 7.39368618e-01
-6.95776880e-01 7.72936821e-01 -1.09571956e-01 1.10742815e-01
8.50513220e-01 -8.37571383e-01 6.60883933e-02 6.66022003e-01
3.85326028e-01 1.69561699e-01 2.17634231e-01 -3.36175412e-01
-1.85169399e+00 -1.11061060e+00 3.08383554e-01 -3.98686498e-01
7.86354780e-01 -4.27684128e-01 -8.97498071e-01 5.20971417e-01
5.97821586e-02 1.75950438e-01 8.47283185e-01 -8.56506750e-02
-8.39617014e-01 -4.63712424e-01 -6.92871213e-01 4.10119951e-01
1.25497961e+00 -9.58755255e-01 -3.81940186e-01 1.33613750e-01
8.97102892e-01 -3.02561909e-01 -1.03011119e+00 5.11650503e-01
9.85782266e-01 -8.65979433e-01 1.09682679e+00 -5.54126084e-01
4.38645154e-01 -4.67884928e-01 -5.37771940e-01 -1.03761530e+00
3.81033570e-02 -8.68856370e-01 -3.59216571e-01 1.36806607e+00
1.89444304e-01 -4.18203354e-01 6.17808461e-01 6.05412312e-02
-2.68992931e-01 -6.71153963e-01 -1.16339087e+00 -8.41470361e-01
-4.89094108e-01 -9.08649266e-01 4.66185510e-01 8.43646407e-01
8.59506652e-02 3.10588211e-01 -6.84951723e-01 3.60090807e-02
4.71317738e-01 4.21011001e-02 8.72203767e-01 -9.05767262e-01
-6.63948894e-01 -2.85298437e-01 -6.70258284e-01 -1.15746295e+00
2.80241281e-01 -7.67953515e-01 -2.39264548e-01 -1.19456542e+00
2.75236160e-01 6.68521523e-02 -4.68408465e-01 1.01901531e+00
-1.93986017e-02 4.81476635e-01 4.60201114e-01 8.93196836e-02
-9.41635489e-01 8.67074788e-01 9.36249614e-01 -5.70660591e-01
-2.11871207e-01 -3.50827783e-01 -5.13093412e-01 6.79234803e-01
3.41980815e-01 -3.79074544e-01 -5.88290215e-01 -5.01672268e-01
-6.87470958e-02 2.84551084e-01 6.41738176e-01 -1.17200363e+00
2.07223698e-01 1.04836755e-01 2.47334272e-01 -5.75389624e-01
8.99557710e-01 -7.79687941e-01 3.67675036e-01 2.53960341e-01
-3.05669010e-01 -8.62851515e-02 5.43510914e-01 7.89869726e-01
-5.24972320e-01 4.21019554e-01 4.79282171e-01 2.08272234e-01
-8.10370028e-01 2.72926867e-01 -4.08882827e-01 2.62175258e-02
1.01964462e+00 -3.41907233e-01 -5.51808059e-01 -4.79364514e-01
-7.36890793e-01 2.56508112e-01 1.26564294e-01 5.70516706e-01
8.31003249e-01 -1.57284153e+00 -5.69145918e-01 2.79964447e-01
3.73035491e-01 -5.18405616e-01 6.79373384e-01 1.41742241e+00
-1.55775458e-01 2.57992417e-01 -1.77767858e-01 -9.85045493e-01
-1.75823545e+00 3.99334729e-01 1.99871331e-01 -1.13017857e-01
-6.17756367e-01 6.64930761e-01 2.76910126e-01 -1.88777760e-01
4.44731832e-01 -2.04361215e-01 4.78479266e-02 2.36989513e-01
7.40450561e-01 5.85149825e-01 5.60032064e-03 -6.85803413e-01
-5.45091212e-01 7.45221853e-01 1.01377502e-01 -2.59122938e-01
1.29223871e+00 3.28892730e-02 2.00841486e-01 5.04453480e-01
1.41350162e+00 -1.37534127e-01 -1.40859973e+00 -2.57312566e-01
-3.19578677e-01 -6.21120095e-01 7.09574372e-02 -7.43557096e-01
-1.37703300e+00 9.24557388e-01 6.85153306e-01 -2.06832260e-01
1.60518944e+00 -2.13562530e-02 9.99091506e-01 3.00489277e-01
2.34885350e-01 -9.20974433e-01 5.94851911e-01 2.49613434e-01
7.89201319e-01 -1.08663642e+00 -1.59305438e-01 -3.98103148e-01
-8.37239325e-01 9.56402600e-01 5.79082727e-01 3.59922588e-01
4.08543766e-01 2.94837594e-01 -2.29453146e-02 -4.35923524e-02
-1.06996334e+00 -1.73335209e-01 4.40369874e-01 4.86878902e-01
1.72592416e-01 -2.98457712e-01 3.15176755e-01 7.58444250e-01
2.36085191e-01 2.24210516e-01 5.69261014e-02 1.05307317e+00
-1.34693280e-01 -9.94479001e-01 -3.68815124e-01 1.57540113e-01
-4.07205671e-01 -5.87819964e-02 -3.75094116e-01 7.32782066e-01
4.95666228e-02 1.08214140e+00 5.12205735e-02 -9.15356874e-01
3.63503397e-01 1.26626939e-01 5.03543615e-01 -9.63791534e-02
-6.01653993e-01 4.26926374e-01 1.42419845e-01 -1.17280507e+00
-8.04231584e-01 -7.08562970e-01 -1.19585812e+00 -2.64913440e-01
-1.02130607e-01 -3.03683132e-02 3.08676511e-01 9.39195693e-01
7.19138145e-01 7.73934543e-01 4.71417487e-01 -9.91613090e-01
-2.67603070e-01 -7.53689170e-01 -3.53401333e-01 3.46591532e-01
3.70081455e-01 -1.01729453e+00 -4.25354093e-01 2.99099177e-01] | [9.559859275817871, 1.073599934577942] |
c7dec5f7-883e-4bca-8e11-9996f6c4e148 | hope-net-a-graph-based-model-for-hand-object | 2004.0006 | null | https://arxiv.org/abs/2004.00060v1 | https://arxiv.org/pdf/2004.00060v1.pdf | HOPE-Net: A Graph-based Model for Hand-Object Pose Estimation | Hand-object pose estimation (HOPE) aims to jointly detect the poses of both a hand and of a held object. In this paper, we propose a lightweight model called HOPE-Net which jointly estimates hand and object pose in 2D and 3D in real-time. Our network uses a cascade of two adaptive graph convolutional neural networks, one to estimate 2D coordinates of the hand joints and object corners, followed by another to convert 2D coordinates to 3D. Our experiments show that through end-to-end training of the full network, we achieve better accuracy for both the 2D and 3D coordinate estimation problems. The proposed 2D to 3D graph convolution-based model could be applied to other 3D landmark detection problems, where it is possible to first predict the 2D keypoints and then transform them to 3D. | ['Bardia Doosti', 'Majid Mirbagheri', 'David Crandall', 'Shujon Naha'] | 2020-03-31 | hope-net-a-graph-based-model-for-hand-object-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Doosti_HOPE-Net_A_Graph-Based_Model_for_Hand-Object_Pose_Estimation_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Doosti_HOPE-Net_A_Graph-Based_Model_for_Hand-Object_Pose_Estimation_CVPR_2020_paper.pdf | cvpr-2020-6 | ['hand-object-pose'] | ['computer-vision'] | [-5.03641903e-01 5.19811995e-02 -7.70441517e-02 1.57204773e-02
-5.52921951e-01 -6.71822190e-01 4.75125313e-01 -1.63925320e-01
-3.52422357e-01 -5.45642264e-02 3.51248160e-02 -5.24230264e-02
1.00772612e-01 -3.31716210e-01 -6.29410982e-01 -2.46561170e-01
-3.17860991e-01 1.03866768e+00 4.06312555e-01 1.68307319e-01
1.93997145e-01 1.32424819e+00 -1.01699972e+00 -2.52575666e-01
-1.41115161e-02 1.03319681e+00 3.29990759e-02 1.00646043e+00
2.98030347e-01 4.90861058e-01 -3.07210684e-01 -5.67329489e-02
5.69205046e-01 1.40513703e-01 -5.99530399e-01 2.33705148e-01
8.33189130e-01 -9.60256636e-01 -6.25544488e-01 8.11830938e-01
7.51167238e-01 6.36588186e-02 5.98386288e-01 -1.45135391e+00
-3.12509447e-01 2.78322212e-02 -1.03044105e+00 -3.47565144e-01
5.63061416e-01 1.92004740e-01 8.20445478e-01 -1.10178232e+00
7.98915029e-01 1.58304417e+00 8.24877799e-01 3.11584085e-01
-8.90517235e-01 -4.60004210e-01 1.95873767e-01 -8.50221664e-02
-1.59316969e+00 -3.53003386e-03 1.13071501e+00 -4.94609505e-01
1.02903903e+00 -2.25262910e-01 8.58028471e-01 9.30576324e-01
1.35443330e-01 9.42426264e-01 5.37296474e-01 -4.68165487e-01
-1.95140064e-01 -3.97481024e-01 -9.63165462e-02 1.20037878e+00
1.35809854e-01 -7.09412992e-02 -2.10490361e-01 -2.17211321e-01
1.46133423e+00 4.47463900e-01 1.52022252e-02 -7.92225540e-01
-1.28387392e+00 5.43278456e-01 8.66499543e-01 2.61223949e-02
-7.30672002e-01 6.52735054e-01 -6.30938336e-02 -2.21521527e-01
3.83833230e-01 1.12785451e-01 -4.98920739e-01 -3.27962749e-02
-6.80773556e-01 5.54141104e-01 7.68629014e-01 1.19843781e+00
4.58575517e-01 -3.72058570e-01 -1.78283855e-01 3.39519113e-01
7.77085781e-01 5.76069772e-01 -1.09019630e-01 -9.72100317e-01
5.40640235e-01 6.61382377e-01 3.57557833e-01 -1.04960132e+00
-8.31735730e-01 -1.81611985e-01 -4.38313663e-01 4.76123840e-01
5.83360910e-01 -2.54766583e-01 -1.24544406e+00 1.42198873e+00
6.49673760e-01 -6.44493243e-03 -6.05337620e-01 1.31208849e+00
5.95506251e-01 2.01898500e-01 -2.06539303e-01 5.34626007e-01
1.39166176e+00 -1.09961081e+00 -3.47924173e-01 -4.36717480e-01
2.87194550e-01 -8.37161422e-01 7.23019183e-01 1.80360507e-02
-1.24447787e+00 -5.85185111e-01 -6.26514137e-01 -3.55595589e-01
-2.34902233e-01 5.22477865e-01 5.96830428e-01 7.58385882e-02
-8.78674448e-01 4.22950000e-01 -1.23609948e+00 -4.27947700e-01
5.54619193e-01 6.57975495e-01 -4.54406738e-01 1.77842706e-01
-5.11905491e-01 9.43184733e-01 1.59385324e-01 5.91653228e-01
-7.66764879e-01 -2.56409049e-01 -8.67874563e-01 -5.30837220e-04
4.94515836e-01 -8.48655820e-01 1.33884418e+00 -3.54868807e-02
-1.43582141e+00 8.35463405e-01 -1.66457742e-01 6.40372708e-02
8.83827448e-01 -6.09210074e-01 4.68160003e-01 1.58015877e-01
-3.84271927e-02 7.10815072e-01 1.03003168e+00 -1.10516536e+00
-4.17735487e-01 -9.95820403e-01 1.19517781e-01 1.83038116e-01
1.46656498e-01 -1.32661462e-01 -1.03640151e+00 -5.45135975e-01
6.02547288e-01 -1.16518795e+00 -1.20610557e-01 7.75411665e-01
-8.36320102e-01 -5.80279291e-01 9.84828472e-01 -8.16655636e-01
3.52558285e-01 -1.84447134e+00 3.44224989e-01 4.77757603e-01
5.17131031e-01 1.62743747e-01 -2.52090305e-01 2.02313557e-01
1.35106578e-01 -3.13081622e-01 4.33364421e-01 -6.89138412e-01
5.80394305e-02 -1.92773536e-01 5.83642088e-02 7.18924701e-01
3.86682779e-01 1.27444816e+00 -7.97365487e-01 -4.18076664e-01
3.71949434e-01 9.76206481e-01 -3.58456939e-01 4.00488913e-01
-2.96080232e-01 3.05184573e-01 -6.38139129e-01 6.91670775e-01
7.10572183e-01 -3.32359314e-01 1.22543961e-01 -4.38750595e-01
9.87041444e-02 1.91603974e-01 -1.40314519e+00 1.78782606e+00
-3.36559027e-01 6.34639561e-01 1.16904452e-01 -2.79592335e-01
7.72948325e-01 3.19035947e-01 4.55689728e-01 -8.01369324e-02
4.68143046e-01 -4.71246056e-02 -2.69384980e-01 -2.52579033e-01
1.94350839e-01 2.11036831e-01 2.53708720e-01 5.23261428e-01
7.30792955e-02 -3.43916118e-01 -2.77605027e-01 -1.32931918e-01
6.90470695e-01 3.61653030e-01 1.58315599e-01 3.08689237e-01
1.72904164e-01 -1.67551324e-01 -1.57795459e-01 5.44140577e-01
-9.22023579e-02 6.13865733e-01 5.16071498e-01 -5.52208722e-01
-1.12807345e+00 -1.07594562e+00 5.22008777e-01 7.08385170e-01
3.35519947e-02 -1.57576740e-01 -5.53569615e-01 -8.95214319e-01
5.54649055e-01 1.43797591e-01 -5.31938493e-01 1.60530448e-01
-8.69046748e-01 2.91496575e-01 1.88041598e-01 1.08415282e+00
4.02267009e-01 -6.92547441e-01 -7.72793591e-01 3.81630496e-03
1.03768341e-01 -1.17983854e+00 -1.05069935e+00 4.03430350e-02
-8.38862181e-01 -1.33124650e+00 -1.01146448e+00 -9.96740699e-01
7.92122841e-01 3.64631951e-01 4.42983896e-01 9.50681865e-02
-3.48301828e-01 5.62770844e-01 -1.06133763e-02 -5.25280952e-01
1.36116579e-01 1.72618717e-01 1.62769929e-01 -1.53690815e-01
1.84710562e-01 -5.20208001e-01 -7.25225449e-01 3.28189731e-01
-1.79484233e-01 -1.74155682e-01 5.20124674e-01 3.01415771e-01
5.52370906e-01 -2.69174367e-01 -2.13731200e-01 -2.48929113e-01
4.44336146e-01 4.10678297e-01 -9.16768372e-01 3.34810525e-01
-1.79949507e-01 1.43259898e-01 1.63434193e-01 -7.43153811e-01
-5.33725023e-01 8.32536638e-01 -1.26073822e-01 -9.99590695e-01
-8.02573264e-02 2.02537671e-01 7.35572651e-02 -3.72056484e-01
3.31829280e-01 -3.50320227e-02 2.15304330e-01 -7.64922023e-01
5.65941811e-01 5.75189710e-01 5.07009983e-01 -3.54784459e-01
1.04074860e+00 4.78486925e-01 2.97887355e-01 -5.18290758e-01
-7.37802088e-01 -4.30850953e-01 -1.40549147e+00 -2.69862115e-01
9.22670245e-01 -9.14959192e-01 -1.30404472e+00 6.47087336e-01
-1.65083337e+00 -4.00360107e-01 -5.20459451e-02 7.04101503e-01
-6.77685678e-01 2.67348200e-01 -5.23497939e-01 -8.97966146e-01
-5.96571982e-01 -1.22274745e+00 1.84162915e+00 -5.35869636e-02
-3.08354974e-01 -8.14255476e-01 -1.28621936e-01 -2.15339392e-01
-3.53998505e-02 2.72711754e-01 6.67933881e-01 -4.43450481e-01
-6.36874199e-01 -9.62882936e-01 -3.97967339e-01 -1.07670367e-01
8.84001106e-02 -6.42413124e-02 -6.78943634e-01 -5.28872848e-01
-3.42735171e-01 -1.86280161e-01 4.06530678e-01 5.22497356e-01
8.36528599e-01 -2.26237312e-01 -5.17237186e-01 5.18107891e-01
9.39644516e-01 -6.48899376e-02 1.71886861e-01 7.10588768e-02
1.12784314e+00 3.68557543e-01 3.67091417e-01 4.40608770e-01
4.43062574e-01 9.87576067e-01 6.71765864e-01 -3.53115723e-02
-4.64385986e-01 -5.52780092e-01 -1.04726516e-01 2.09077045e-01
-4.69413280e-01 8.53600129e-02 -9.55878198e-01 3.68906528e-01
-1.69901490e+00 -4.08136249e-01 -2.96253022e-02 2.10775566e+00
3.77422571e-01 1.08876593e-01 5.53310215e-01 -2.40590632e-01
7.48051882e-01 1.02151804e-01 -7.56756186e-01 3.77162635e-01
6.40862644e-01 7.05649406e-02 4.62029696e-01 4.69676465e-01
-1.00760937e+00 1.12756598e+00 6.21548891e+00 1.19265743e-01
-1.11495614e+00 -9.29924846e-02 -1.83654383e-01 -8.75310972e-02
4.25843000e-01 -7.11689740e-02 -8.87411714e-01 -1.44951016e-01
7.39330985e-03 4.09916043e-01 4.23859715e-01 1.04610491e+00
3.10117397e-02 2.76522338e-01 -1.47888410e+00 1.10037816e+00
2.87710857e-02 -1.01439691e+00 -2.11307686e-02 1.57617643e-01
3.12397659e-01 9.61220264e-02 -1.73110321e-01 -2.00895458e-01
2.05256686e-01 -6.38163686e-01 9.92458522e-01 4.21704918e-01
7.55645692e-01 -5.55050254e-01 3.99835825e-01 4.83522654e-01
-1.39737117e+00 1.48127303e-01 -1.26147509e-01 -1.23572834e-01
1.38642251e-01 2.27614958e-02 -1.34012556e+00 7.37293512e-02
6.03148818e-01 5.12920022e-01 -4.92587119e-01 1.08651662e+00
-5.85462689e-01 -1.60382584e-01 -6.02454305e-01 -2.51135498e-01
1.13384269e-01 3.37819576e-01 6.53737485e-01 7.78674841e-01
2.68719018e-01 -2.19696611e-02 3.83310944e-01 9.17136550e-01
-6.36830032e-02 -4.15086210e-01 -4.53831106e-01 -4.91743125e-02
4.09209758e-01 1.18686461e+00 -9.41170335e-01 -9.80538949e-02
1.06778508e-02 1.29043853e+00 4.36217636e-01 4.16477859e-01
-6.53863251e-01 -6.36847675e-01 6.04111969e-01 1.60070762e-01
6.19091272e-01 -8.96838367e-01 -1.61968358e-02 -1.05221128e+00
2.98058540e-01 -2.79424489e-01 1.58851847e-01 -1.14518261e+00
-1.18186140e+00 4.41932887e-01 -1.50258958e-01 -9.49488759e-01
-3.36531907e-01 -9.89794731e-01 -4.45123851e-01 1.00187719e+00
-1.16122627e+00 -1.59981942e+00 -5.63320279e-01 7.88746893e-01
2.92541474e-01 1.03164203e-01 6.29070342e-01 -1.30792603e-01
-1.67791665e-01 5.75317383e-01 -3.37033719e-01 8.95598829e-01
5.65484941e-01 -1.11431062e+00 9.59222972e-01 4.07820672e-01
3.95697266e-01 6.78027391e-01 3.72170061e-01 -7.42776334e-01
-1.87536323e+00 -1.00253868e+00 7.75974035e-01 -8.01867902e-01
4.21734333e-01 -5.96326053e-01 -3.59495819e-01 1.23266089e+00
-4.13420260e-01 2.71410704e-01 -2.79170007e-01 2.72005528e-01
-6.28928721e-01 5.40001616e-02 -1.13081872e+00 6.44123018e-01
1.17112255e+00 -6.66663170e-01 -4.17741150e-01 7.97038853e-01
5.26378453e-01 -1.02960813e+00 -7.26007283e-01 -1.17017694e-01
1.04291272e+00 -4.15774286e-01 1.22771466e+00 -8.30126584e-01
9.64306593e-02 -1.92288116e-01 8.90069827e-02 -1.10207021e+00
-4.66673017e-01 -5.19403398e-01 -5.15126467e-01 5.91409624e-01
8.80600233e-03 -4.62185413e-01 1.02322423e+00 5.51808655e-01
3.11232954e-01 -5.69843292e-01 -1.01690841e+00 -7.40693271e-01
-2.54365444e-01 -4.70005155e-01 5.07641852e-01 3.08986545e-01
-1.90514728e-01 4.64777082e-01 -1.30464181e-01 7.02650845e-01
8.13002944e-01 4.33385402e-01 1.11560571e+00 -1.32982981e+00
-6.62789419e-02 -5.09053469e-01 -7.63427317e-01 -1.51700556e+00
1.16271280e-01 -7.69087136e-01 1.61666587e-01 -1.68203092e+00
1.31429911e-01 -1.14555404e-01 1.47222355e-01 7.23722339e-01
4.14951853e-02 2.47358710e-01 5.08098543e-01 2.42377192e-01
-2.66915023e-01 1.98170885e-01 1.56736743e+00 -1.68683663e-01
-4.36245024e-01 2.83620417e-01 -1.37984589e-01 8.01730156e-01
3.71815562e-01 -4.83115315e-01 7.17348680e-02 -8.28119457e-01
-2.61762112e-01 1.63741842e-01 9.39212978e-01 -7.26289093e-01
4.01517183e-01 1.04220286e-02 8.37029278e-01 -1.09476435e+00
6.73098445e-01 -1.18903124e+00 -1.33247882e-01 5.87922037e-01
-2.69529790e-01 1.50756121e-01 1.11416541e-02 4.51518625e-01
3.99958283e-01 7.20611587e-02 6.08793139e-01 -2.72211611e-01
-3.64490926e-01 7.73970842e-01 1.04845166e-01 -4.90628421e-01
1.12833905e+00 -2.35432133e-01 1.42197639e-01 -6.43630207e-01
-9.25689638e-01 2.71986008e-01 2.88464695e-01 5.62570691e-01
9.11972106e-01 -1.50500607e+00 -6.34039104e-01 2.95163721e-01
-1.32095173e-01 4.22888815e-01 -8.90168771e-02 9.33354616e-01
-6.45331204e-01 5.55290997e-01 3.28139365e-02 -9.62576687e-01
-1.50787437e+00 5.33271790e-01 5.02868891e-01 -3.32946889e-02
-8.46036434e-01 1.06554806e+00 -1.37466714e-01 -6.48806810e-01
8.21103573e-01 -4.75784481e-01 2.90234327e-01 -1.74253285e-01
5.12040675e-01 3.11531067e-01 2.14554695e-03 -6.59514606e-01
-6.27162576e-01 1.09550250e+00 -1.22865953e-01 4.37428392e-02
1.56851459e+00 2.25664154e-01 -1.33840770e-01 3.01847439e-02
1.46060216e+00 -2.20248505e-01 -1.63005292e+00 -3.21326762e-01
-1.38393894e-01 -6.78631186e-01 7.93987736e-02 -6.28195941e-01
-1.35351920e+00 1.11817169e+00 7.19426811e-01 -2.82552570e-01
6.51592016e-01 3.37765485e-01 6.92468405e-01 7.12254941e-01
3.86661202e-01 -7.48856127e-01 3.62099558e-01 5.81016362e-01
1.20294142e+00 -1.01017892e+00 1.88558951e-01 -3.70477945e-01
-1.83973730e-01 1.58994436e+00 5.47561526e-01 -3.61274898e-01
7.78757215e-01 1.33218825e-01 2.54427958e-02 -4.77990299e-01
-1.21358246e-01 -3.01543027e-01 7.15339243e-01 6.60390377e-01
1.52444020e-01 -6.06484488e-02 3.03559303e-01 6.57893792e-02
1.96985584e-02 1.42830878e-01 -1.20028824e-01 1.14408219e+00
-1.73369557e-01 -8.68291676e-01 -6.83875322e-01 1.33173034e-01
3.45285609e-02 2.81764477e-01 -7.51333654e-01 1.08875561e+00
-8.47784728e-02 4.54388678e-01 2.26234719e-01 -4.66087729e-01
8.66917491e-01 9.87098739e-02 1.05884993e+00 -5.24881721e-01
-1.94961593e-01 4.80329812e-01 -3.04960757e-01 -7.10841000e-01
-1.86223343e-01 -4.91335869e-01 -1.36998200e+00 -2.57244140e-01
-3.94471288e-01 -4.76360798e-01 9.52681899e-01 9.53271508e-01
4.84794229e-01 3.11517000e-01 3.77861321e-01 -1.68230307e+00
-8.19225013e-01 -1.03617084e+00 -6.39434218e-01 3.83798480e-02
6.26249373e-01 -9.52972651e-01 -1.19001074e-02 -2.39881009e-01] | [6.587416172027588, -0.8649260401725769] |
8f6e8417-4b5d-46fd-be6d-ace06cecc3ac | real-time-visual-tracking-using-spatial-aware | 1908.00692 | null | https://arxiv.org/abs/1908.00692v1 | https://arxiv.org/pdf/1908.00692v1.pdf | Real Time Visual Tracking using Spatial-Aware Temporal Aggregation Network | More powerful feature representations derived from deep neural networks benefit visual tracking algorithms widely. However, the lack of exploitation on temporal information prevents tracking algorithms from adapting to appearances changing or resisting to drift. This paper proposes a correlation filter based tracking method which aggregates historical features in a spatial-aligned and scale-aware paradigm. The features of historical frames are sampled and aggregated to search frame according to a pixel-level alignment module based on deformable convolutions. In addition, we also use a feature pyramid structure to handle motion estimation at different scales, and address the different demands on feature granularity between tracking losses and deformation offset learning. By this design, the tracker, named as Spatial-Aware Temporal Aggregation network (SATA), is able to assemble appearances and motion contexts of various scales in a time period, resulting in better performance compared to a single static image. Our tracker achieves leading performance in OTB2013, OTB2015, VOT2015, VOT2016 and LaSOT, and operates at a real-time speed of 26 FPS, which indicates our method is effective and practical. Our code will be made publicly available at \href{https://github.com/ecart18/SATA}{https://github.com/ecart18/SATA}. | ['Xian-Ming Liu', 'Lichao Huang', 'Tao Hu', 'Han Shen'] | 2019-08-02 | null | null | null | null | ['real-time-visual-tracking'] | ['computer-vision'] | [-4.08854961e-01 -7.25830078e-01 -8.01479369e-02 -8.68754834e-02
-2.52508551e-01 -5.17642200e-01 4.86439407e-01 -3.72745782e-01
-4.81868833e-01 6.02379382e-01 -1.02450773e-01 9.25609916e-02
1.81212667e-02 -5.60236037e-01 -6.19479060e-01 -6.01061881e-01
-2.46429503e-01 -1.35953590e-01 6.61976159e-01 -4.37767506e-02
-1.57031104e-01 8.00264001e-01 -1.70969987e+00 8.55762288e-02
7.40771353e-01 1.32918167e+00 3.71808827e-01 7.47146130e-01
1.63019627e-01 3.57584536e-01 -4.43036556e-01 -3.32598954e-01
6.33743644e-01 -1.01494938e-01 -2.49209136e-01 -1.43878818e-01
9.57335293e-01 -3.70925099e-01 -6.79806769e-01 1.10557973e+00
5.95300794e-01 9.05564055e-02 2.34670147e-01 -1.26373374e+00
-5.35329282e-01 -2.86145173e-02 -5.61038673e-01 4.64940459e-01
4.39460091e-02 5.87002873e-01 4.54079658e-01 -8.88016105e-01
7.57391155e-01 1.21940088e+00 9.90373194e-01 7.46694863e-01
-9.26018536e-01 -8.99709404e-01 3.63328576e-01 3.03727418e-01
-1.21464252e+00 -4.11802977e-01 6.26764894e-01 -5.63309252e-01
5.87970853e-01 2.37904578e-01 1.00330758e+00 1.16727519e+00
4.23875272e-01 8.00514519e-01 8.68233681e-01 7.92658404e-02
-5.44881336e-02 -2.96529591e-01 -1.93559006e-01 5.77338815e-01
4.24823135e-01 5.66035688e-01 -7.37980068e-01 1.75308421e-01
1.17308128e+00 5.71809672e-02 -3.79266709e-01 -4.13197249e-01
-1.47079003e+00 4.65881884e-01 6.06539488e-01 3.42045575e-01
-4.03331041e-01 4.78933156e-01 3.53766024e-01 3.12611431e-01
4.37194049e-01 -8.82104263e-02 -5.80381811e-01 -2.12084204e-01
-1.07059622e+00 2.60556221e-01 2.25074485e-01 1.04873109e+00
5.93612254e-01 4.61246550e-01 -3.88518482e-01 4.71586734e-01
2.73922890e-01 8.79407883e-01 4.94399011e-01 -1.11806321e+00
5.79313189e-02 2.63802558e-01 4.41261649e-01 -1.01343322e+00
-5.62507212e-01 -3.93860877e-01 -8.61683965e-01 6.05897665e-01
5.49436450e-01 -1.67167574e-01 -8.27181101e-01 1.87646711e+00
6.73929930e-01 4.79956269e-01 -2.96714962e-01 1.09678864e+00
7.86702514e-01 3.77413750e-01 3.26410085e-02 -2.00063780e-01
1.45099378e+00 -8.73437405e-01 -7.42056608e-01 5.60289510e-02
1.96898907e-01 -1.04767370e+00 7.18820274e-01 2.50296652e-01
-1.03877068e+00 -1.15952528e+00 -9.74247336e-01 7.29302391e-02
-2.65684366e-01 3.85531455e-01 6.73569083e-01 4.89338070e-01
-1.25534415e+00 8.23479772e-01 -1.22597182e+00 -3.81084144e-01
4.34510916e-01 4.04533446e-01 -1.01512291e-01 3.46703261e-01
-1.14836061e+00 7.10859060e-01 2.66001999e-01 2.53284305e-01
-6.50893331e-01 -7.14023709e-01 -6.38891697e-01 -3.32485676e-01
1.42901316e-01 -8.92526805e-01 1.19865119e+00 -9.13416326e-01
-1.51644540e+00 4.10815477e-01 -2.56381452e-01 -5.85119605e-01
8.77420962e-01 -5.24806261e-01 -5.78807592e-01 -1.52515145e-02
-1.01808041e-01 8.47714722e-01 1.08478796e+00 -7.98292160e-01
-8.52787614e-01 -1.55041248e-01 -1.62549645e-01 -1.79945379e-02
-4.00820702e-01 4.65671606e-02 -6.97420537e-01 -8.73998582e-01
-1.13586381e-01 -9.90411401e-01 -7.76921399e-04 5.68340600e-01
1.60891131e-01 -1.71394691e-01 1.26659262e+00 -4.78101164e-01
1.23024905e+00 -2.11480665e+00 -2.69964896e-02 -3.47081274e-01
1.84431210e-01 6.05650246e-01 -1.22636132e-01 1.04998760e-01
1.83070898e-01 -3.38031203e-01 2.14227408e-01 -3.78100634e-01
-7.36132339e-02 -2.73207370e-02 -1.92786515e-01 6.15543723e-01
1.26544520e-01 1.07573044e+00 -9.66193855e-01 -5.41541278e-01
6.61731005e-01 7.17062056e-01 -3.05194944e-01 -1.22768573e-01
-1.63058817e-01 6.80407703e-01 -5.22451997e-01 9.38195467e-01
9.09095526e-01 -2.06450939e-01 -1.42019331e-01 -4.92062777e-01
-5.97072482e-01 -1.51856706e-01 -1.31000721e+00 1.72709954e+00
-2.41528779e-01 8.69767725e-01 6.82204366e-02 -4.18588042e-01
9.00660872e-01 2.62306392e-01 8.66955936e-01 -7.96818018e-01
2.10207239e-01 1.62151739e-01 -1.13555484e-01 -6.33791238e-02
6.18911624e-01 4.52621281e-01 1.68413758e-01 5.29927351e-02
8.04216415e-02 2.68576771e-01 3.06983203e-01 -1.06016368e-01
9.77989137e-01 6.79138184e-01 1.42191395e-01 -1.80858523e-01
5.18107772e-01 -8.31460655e-02 9.15452421e-01 7.32136250e-01
-6.30645990e-01 4.89847153e-01 -2.65624046e-01 -9.47429240e-01
-1.02508521e+00 -1.18699992e+00 -3.19998443e-01 7.88507700e-01
4.52959001e-01 -3.69852394e-01 -3.31851095e-01 -4.15753812e-01
2.29741439e-01 -4.81931539e-03 -5.53164303e-01 -8.45160484e-02
-8.15515101e-01 -3.57346624e-01 3.22980434e-01 6.17779553e-01
7.61504531e-01 -1.16066229e+00 -1.08892226e+00 2.79762059e-01
-4.77985516e-02 -1.17791641e+00 -8.31916869e-01 -1.90523729e-01
-9.73233581e-01 -9.38013673e-01 -9.38098133e-01 -4.99931604e-01
3.82718444e-01 3.39562595e-01 7.91201115e-01 3.63010168e-02
-4.22934443e-01 3.89646322e-01 -2.99145013e-01 -3.49846840e-01
1.49744144e-02 -1.76701546e-01 4.27469134e-01 1.02894850e-01
1.86822623e-01 -5.62140048e-01 -1.09408343e+00 4.69716042e-01
-5.81166685e-01 1.76214829e-01 4.33743536e-01 6.92318022e-01
5.63581705e-01 -4.12783206e-01 1.89651921e-01 -2.48932056e-02
1.60720006e-01 2.54274113e-03 -1.21508896e+00 1.34198397e-01
-4.13904220e-01 -2.21753791e-01 5.73012590e-01 -9.06868041e-01
-8.30270588e-01 3.55211169e-01 7.95194507e-02 -9.43708122e-01
1.00287020e-01 -1.98579803e-01 4.42303009e-02 -4.18426961e-01
5.36061645e-01 3.12126935e-01 3.23896110e-02 -3.77883732e-01
3.59719902e-01 3.00250024e-01 7.45042682e-01 -4.13076788e-01
1.08547187e+00 7.63240814e-01 -9.32256058e-02 -5.69085240e-01
-5.12113094e-01 -3.85424554e-01 -8.10416341e-01 -7.36302137e-01
7.95130849e-01 -9.60514784e-01 -9.19191360e-01 8.61947298e-01
-1.09127760e+00 -4.64121372e-01 -3.59877974e-01 7.57058680e-01
-4.97059017e-01 3.90449941e-01 -6.05141938e-01 -6.22335374e-01
-5.48726618e-01 -9.75819707e-01 9.41630960e-01 7.17978358e-01
3.48840915e-02 -8.10551405e-01 8.38774890e-02 -1.98084801e-01
6.88247204e-01 4.77642268e-01 -1.79740682e-01 -1.08253919e-02
-9.29322124e-01 -1.08002752e-01 -1.98336974e-01 2.17690080e-01
2.85947889e-01 2.68343091e-01 -8.57920527e-01 -7.02367246e-01
-2.49616832e-01 1.85465217e-02 8.47979963e-01 7.68210292e-01
8.88431907e-01 -1.67813420e-01 -4.15404260e-01 1.00532031e+00
1.26432395e+00 2.28726178e-01 4.92736489e-01 5.27481616e-01
6.68006361e-01 8.78873765e-02 8.32080901e-01 4.80599642e-01
1.73622504e-01 1.13463449e+00 4.23865616e-01 -5.40446155e-02
-7.25020885e-01 5.64726181e-02 5.61457098e-01 6.90505505e-01
-3.06404710e-01 5.26961349e-02 -6.05420113e-01 6.16756618e-01
-1.92266512e+00 -1.21597624e+00 -1.80306986e-01 2.36971331e+00
7.32721329e-01 7.52619281e-02 2.27175623e-01 -4.85513985e-01
8.16294909e-01 2.65121132e-01 -7.28326738e-01 6.27579773e-03
-1.78796038e-01 1.06038891e-01 7.77858019e-01 2.98656493e-01
-1.35504830e+00 1.03516948e+00 4.95157433e+00 8.68350029e-01
-1.51133192e+00 2.36311823e-01 8.15175399e-02 -2.99242526e-01
2.79871851e-01 -7.02145770e-02 -9.55833137e-01 8.46628129e-01
7.11216569e-01 -2.59262830e-01 1.96138769e-01 7.35014200e-01
3.57305288e-01 1.02103435e-01 -6.54931724e-01 1.01345825e+00
-3.68390352e-01 -1.44102192e+00 -2.71671206e-01 8.73859972e-02
7.02823818e-01 3.11496943e-01 2.52527356e-01 1.75124183e-01
2.28080332e-01 -5.22345603e-01 9.91925955e-01 7.14763761e-01
8.25303674e-01 -4.17436719e-01 5.00940621e-01 -2.09274273e-02
-1.89107394e+00 -2.40566079e-02 -4.62712348e-01 1.68936625e-01
2.54037142e-01 4.32309121e-01 -3.47219855e-01 8.39914024e-01
1.16714835e+00 1.01741767e+00 -6.65524840e-01 1.37683535e+00
4.46285345e-02 2.66646266e-01 -5.02463758e-01 1.35704959e-02
6.80222735e-02 5.22949807e-02 7.11040556e-01 1.15369344e+00
5.29548049e-01 -1.64251432e-01 1.62182450e-01 5.27706861e-01
2.18678474e-01 -2.14194715e-01 -3.73165131e-01 2.40333617e-01
5.83605528e-01 1.39655232e+00 -7.52154291e-01 -2.66990095e-01
-5.04029572e-01 8.31934571e-01 1.37780800e-01 3.34809154e-01
-1.29010594e+00 -1.87067732e-01 8.66417229e-01 7.29765743e-02
8.29284728e-01 -4.02840406e-01 1.19431689e-01 -1.26040602e+00
2.19849691e-01 -5.48981965e-01 2.29899645e-01 -6.80383325e-01
-1.08715236e+00 8.12798381e-01 -1.27637431e-01 -1.96374416e+00
-9.23858723e-04 -5.44924915e-01 -5.08712173e-01 6.90880418e-01
-1.28935838e+00 -1.26120090e+00 -5.62618792e-01 7.59995759e-01
5.46625435e-01 -1.32714540e-01 3.26865792e-01 5.91164470e-01
-3.04072112e-01 6.68231070e-01 1.98752910e-01 1.59789950e-01
8.89811099e-01 -9.25101936e-01 6.64152384e-01 1.03402233e+00
1.05396204e-01 4.56785470e-01 6.54490471e-01 -7.28596210e-01
-1.39771914e+00 -1.31450307e+00 5.19582927e-01 -4.20804173e-01
6.21944666e-01 -2.03720674e-01 -9.35905099e-01 4.33855593e-01
9.99210477e-02 5.87373376e-01 3.34232002e-02 -3.12829167e-01
-1.07081465e-01 -4.14078772e-01 -8.53701115e-01 5.18754303e-01
1.28200924e+00 -1.86154053e-01 -2.30493113e-01 1.00867756e-01
6.19015515e-01 -7.87314653e-01 -1.01452744e+00 5.78236103e-01
9.01690364e-01 -9.87671793e-01 1.13076258e+00 -1.12185031e-01
-1.99105576e-01 -8.68862748e-01 3.96219790e-02 -8.26965988e-01
-5.35350978e-01 -8.44974220e-01 -5.62016964e-01 1.01195240e+00
-1.28006518e-01 -7.83578515e-01 6.43311024e-01 1.35138229e-01
-7.68820941e-02 -7.08168685e-01 -1.29079509e+00 -1.26969373e+00
-1.98861599e-01 -2.42432877e-01 5.54379582e-01 7.60370553e-01
-7.84956515e-01 -3.66776466e-01 -6.25108182e-01 2.30374634e-01
8.81175756e-01 3.49798918e-01 8.91420603e-01 -1.12013912e+00
-1.32450327e-01 -6.06619716e-01 -6.49774611e-01 -1.22744930e+00
-2.81915367e-01 -5.56617916e-01 -1.67413667e-01 -1.26451528e+00
-1.14700340e-01 -3.73188823e-01 -4.34025228e-01 4.31162238e-01
-1.00555427e-01 5.21516263e-01 6.08260572e-01 4.98371094e-01
-7.13886142e-01 6.12003624e-01 1.43538249e+00 -7.11597428e-02
-1.57998115e-01 1.73745245e-01 -8.64590108e-02 6.31303906e-01
8.90874684e-01 -4.26302373e-01 -5.80831617e-02 -4.29373026e-01
-3.34334314e-01 -4.70030345e-02 6.80409610e-01 -1.45229423e+00
5.71067631e-01 -6.60215020e-02 8.56726289e-01 -9.83483315e-01
4.41935122e-01 -7.14651823e-01 6.25311375e-01 8.98666382e-01
2.18333736e-01 2.95804232e-01 5.82661927e-01 4.70354676e-01
-1.08036898e-01 2.01747835e-01 7.81677067e-01 1.26501620e-01
-1.06182754e+00 7.48322904e-01 -1.85777806e-02 -1.29346326e-01
1.04078460e+00 -6.29009306e-01 -2.89992273e-01 -1.18165031e-01
-5.73293865e-01 3.14412326e-01 7.37145483e-01 7.09243417e-01
4.78186429e-01 -1.68538558e+00 -5.75602531e-01 2.86417399e-02
-7.29631931e-02 -2.18207151e-01 5.92643857e-01 1.14259899e+00
-4.22909051e-01 4.16816890e-01 -5.20609915e-01 -9.37919796e-01
-1.29561400e+00 5.46070993e-01 4.00215864e-01 -1.14842318e-01
-9.08865154e-01 6.78647757e-01 1.74317092e-01 1.02232412e-01
3.20579469e-01 -2.38446727e-01 -5.51994182e-02 -5.01540862e-02
7.00136364e-01 2.05575570e-01 -1.88587949e-01 -8.30486059e-01
-5.11430502e-01 9.50134397e-01 4.82396558e-02 2.01937273e-01
1.13292754e+00 -2.07966387e-01 3.55698705e-01 2.05631942e-01
9.34150040e-01 -4.44934368e-02 -2.05119801e+00 -2.39365220e-01
-7.63585940e-02 -8.20097089e-01 -1.00691244e-01 -5.51441848e-01
-1.37857211e+00 5.13979852e-01 1.12236154e+00 3.10265161e-02
1.22055387e+00 -1.74353167e-01 9.33598518e-01 -5.76136112e-02
4.70931947e-01 -9.74440753e-01 9.59791541e-02 5.19195437e-01
8.21693599e-01 -1.15109730e+00 2.29369387e-01 4.60261889e-02
-3.28676969e-01 1.26979721e+00 8.40687811e-01 -1.99824020e-01
3.98074955e-01 3.84434968e-01 1.25069171e-01 6.60647228e-02
-6.64878547e-01 -4.02769178e-01 5.70458412e-01 7.09165454e-01
1.69414803e-01 -9.05903280e-02 -3.40205848e-01 6.80179447e-02
-4.53977548e-02 1.41871259e-01 5.29271811e-02 9.27999794e-01
-2.68902987e-01 -1.12454033e+00 -3.70045990e-01 2.68127173e-02
-2.18420774e-01 1.21249661e-01 2.83806715e-02 1.07154346e+00
4.24615353e-01 4.11549151e-01 6.39158040e-02 -3.70504528e-01
3.00218761e-01 -3.55538875e-01 5.36694169e-01 7.90533349e-02
-5.45477211e-01 2.80318081e-01 -3.13861549e-01 -9.64231133e-01
-7.53240705e-01 -9.15289581e-01 -1.04574001e+00 -4.55797821e-01
-2.71529883e-01 -2.42063671e-01 5.16824365e-01 4.15070802e-01
5.27012289e-01 6.17143452e-01 4.47845340e-01 -1.15670311e+00
-3.36369038e-01 -7.87993908e-01 -1.91718966e-01 2.04435125e-01
4.81682837e-01 -9.88867879e-01 -1.12974957e-01 2.01462567e-01] | [6.317809581756592, -2.1318702697753906] |
16c53109-16ec-444c-ba53-abc3d07618b6 | evaluating-xgboost-for-balanced-and | 2303.15218 | null | https://arxiv.org/abs/2303.15218v1 | https://arxiv.org/pdf/2303.15218v1.pdf | Evaluating XGBoost for Balanced and Imbalanced Data: Application to Fraud Detection | This paper evaluates XGboost's performance given different dataset sizes and class distributions, from perfectly balanced to highly imbalanced. XGBoost has been selected for evaluation, as it stands out in several benchmarks due to its detection performance and speed. After introducing the problem of fraud detection, the paper reviews evaluation metrics for detection systems or binary classifiers, and illustrates with examples how different metrics work for balanced and imbalanced datasets. Then, it examines the principles of XGBoost. It proposes a pipeline for data preparation and compares a Vanilla XGBoost against a random search-tuned XGBoost. Random search fine-tuning provides consistent improvement for large datasets of 100 thousand samples, not so for medium and small datasets of 10 and 1 thousand samples, respectively. Besides, as expected, XGBoost recognition performance improves as more data is available, and deteriorates detection performance as the datasets become more imbalanced. Tests on distributions with 50, 45, 25, and 5 percent positive samples show that the largest drop in detection performance occurs for the distribution with only 5 percent positive samples. Sampling to balance the training set does not provide consistent improvement. Therefore, future work will include a systematic study of different techniques to deal with data imbalance and evaluating other approaches, including graphs, autoencoders, and generative adversarial methods, to deal with the lack of labels. | ['Vaibhav Joshi', 'Khushboo Sharma', 'Anuj Deshmunkh', 'Sanjay Deshmane', 'Anindya Sudhir', 'Gissel Velarde'] | 2023-03-27 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [-3.27332407e-01 -5.25142774e-02 -5.95673621e-01 -3.91228437e-01
-4.75306690e-01 -2.87537307e-01 4.69963729e-01 6.11346543e-01
-3.50270301e-01 8.79276633e-01 4.54633944e-02 -4.53669965e-01
-3.67311329e-01 -1.25677657e+00 -3.54905307e-01 -5.73521316e-01
-1.37038991e-01 8.36703658e-01 -2.54531465e-02 -3.22820365e-01
3.76123339e-01 3.48348528e-01 -1.45347238e+00 3.96213859e-01
8.12034607e-01 9.15478170e-01 -8.39921832e-01 6.93037927e-01
-1.49761856e-01 7.84493387e-01 -1.46954167e+00 -7.87876189e-01
4.77468818e-01 -5.02122343e-01 -7.29372799e-01 -1.73131153e-01
4.62570608e-01 -2.39338771e-01 1.65531054e-01 7.67680645e-01
7.83157766e-01 -1.55579582e-01 5.76259375e-01 -1.72980762e+00
-5.56942225e-01 6.83037639e-01 -7.35608697e-01 3.62112522e-01
4.14384902e-01 2.24431112e-01 8.33756864e-01 -1.90530643e-01
4.44293737e-01 1.10796046e+00 1.15484822e+00 4.54770744e-01
-1.20248473e+00 -7.42669761e-01 -1.00335479e-01 5.27800061e-02
-8.79030466e-01 1.38436005e-01 6.34337366e-01 -3.73578072e-01
1.05496311e+00 5.80078304e-01 1.07481515e+00 1.01126361e+00
2.85067260e-01 7.58270502e-01 1.19531703e+00 -7.94770658e-01
4.16498274e-01 2.94614553e-01 7.00054944e-01 1.75750226e-01
9.71605778e-01 3.83621663e-01 -5.20650089e-01 -5.78505099e-01
1.15497209e-01 1.44377515e-01 5.82276136e-02 -2.61353731e-01
-6.75271928e-01 1.11761630e+00 4.32858974e-01 2.41127208e-01
-3.96875948e-01 1.21209949e-01 7.90793955e-01 5.80833733e-01
8.10869813e-01 7.48979688e-01 -3.92040551e-01 -7.63779953e-02
-1.12003648e+00 6.35099590e-01 8.93879175e-01 8.39080751e-01
4.11300391e-01 2.02218458e-01 -2.90420890e-01 8.56031418e-01
2.06608921e-02 2.13688359e-01 9.07509327e-01 -4.07197356e-01
4.86512363e-01 9.38556910e-01 9.23851728e-02 -1.07701743e+00
-7.24091470e-01 -8.28543127e-01 -6.49833322e-01 4.17405218e-01
6.13517106e-01 -1.52594179e-01 -1.09669924e+00 1.10104132e+00
2.12326363e-01 -4.28094387e-01 -4.61465046e-02 7.95854032e-01
7.63020337e-01 2.65785217e-01 2.23787919e-01 -1.16992436e-01
1.25630426e+00 -7.20912099e-01 -6.54751003e-01 -3.72532398e-01
6.07752323e-01 -6.36922300e-01 9.90834355e-01 5.87078393e-01
-9.31752026e-01 -4.95943516e-01 -1.28754580e+00 4.71458018e-01
-6.80845916e-01 -2.24825740e-01 9.54038501e-01 1.39812446e+00
-6.52713835e-01 7.18158185e-01 -7.33906984e-01 -3.58249992e-01
7.66027927e-01 3.82635891e-01 1.19970553e-02 -8.12177509e-02
-1.22723496e+00 9.22736347e-01 4.66648489e-01 -4.78052884e-01
-1.79928169e-01 -7.25769222e-01 -5.08382440e-01 3.50238681e-02
-2.46141821e-01 -8.96501124e-01 9.22519207e-01 -9.69291925e-01
-9.46305692e-01 8.65672827e-01 4.36136425e-01 -8.55567336e-01
9.76433814e-01 -2.65132010e-01 -6.41609967e-01 -3.53474259e-01
8.01635087e-02 1.54194340e-01 5.51411510e-01 -1.10601676e+00
-7.41055310e-01 -6.45205915e-01 -7.19094798e-02 -6.78838268e-02
-3.31546843e-01 1.28552401e-02 5.77377155e-02 -7.60339141e-01
3.34979780e-02 -5.91251135e-01 -5.50871305e-02 -7.66525209e-01
-6.08956456e-01 5.37243299e-02 4.17572826e-01 -4.82915908e-01
1.31415725e+00 -1.84689319e+00 -7.41787314e-01 4.78787869e-01
2.88879909e-02 2.68376797e-01 2.68423170e-01 5.94708800e-01
-3.66432399e-01 3.49742442e-01 1.46019310e-01 2.21615745e-04
-2.96781845e-02 -4.35741022e-02 -1.90096512e-01 4.15071756e-01
1.15315698e-01 7.98048973e-01 -9.47346091e-01 -1.14432447e-01
1.53657958e-01 7.32465759e-02 -5.46317399e-01 -1.65429279e-01
1.79047391e-01 -1.41366482e-01 -1.64999411e-01 1.08105421e+00
8.20018947e-01 -1.90123275e-01 5.81905805e-02 7.27050826e-02
4.66667086e-01 1.11514866e-01 -1.25017405e+00 7.65064836e-01
-5.72708026e-02 5.95833957e-01 -6.36975408e-01 -1.12467873e+00
1.34196007e+00 -2.84523107e-02 5.50702393e-01 -1.04947352e+00
9.20787007e-02 3.38836551e-01 -5.95762096e-02 -2.06666306e-01
6.02626681e-01 -4.28130090e-01 -2.33443990e-01 4.11868691e-01
-6.70800954e-02 -1.38009548e-01 4.38358456e-01 -4.68412265e-02
1.28462040e+00 -2.72682995e-01 5.17798066e-01 2.58284193e-02
-4.83391620e-02 5.39478660e-01 4.48352158e-01 1.20100272e+00
-3.33394974e-01 5.94443321e-01 6.16524816e-01 -1.05058515e+00
-8.74691546e-01 -8.06068838e-01 -3.22266191e-01 1.07396519e+00
1.34401172e-01 -3.65448982e-01 -7.83449352e-01 -8.53520393e-01
7.07217574e-01 8.89162540e-01 -7.53169000e-01 -5.27672112e-01
-2.93848991e-01 -1.91291428e+00 6.45008564e-01 4.68136132e-01
4.69351619e-01 -1.17085266e+00 -6.55604303e-01 2.76696503e-01
1.08790740e-01 -3.10736686e-01 3.30179393e-01 3.82260799e-01
-1.14699423e+00 -1.44495559e+00 -4.01882678e-01 -1.90372601e-01
5.32160878e-01 -1.69492647e-01 1.60968196e+00 3.33309948e-01
-4.68664199e-01 2.97110844e-02 -6.11793816e-01 -1.08415341e+00
-4.73716199e-01 7.74495676e-03 -1.77386194e-01 -4.26730216e-01
1.05274975e+00 -1.47713616e-01 -6.62796259e-01 4.20583397e-01
-7.22155690e-01 -4.85080987e-01 2.01615244e-01 1.28013289e+00
3.20278972e-01 2.32686028e-01 8.50601256e-01 -1.24140036e+00
1.00601959e+00 -5.89674532e-01 -3.40851605e-01 9.85460803e-02
-1.23510349e+00 -3.28825414e-01 4.16950732e-01 -2.40783975e-01
-5.77954888e-01 -6.01942420e-01 2.09731050e-02 3.58380675e-02
-1.30285889e-01 3.03968936e-01 2.05576718e-01 1.43723130e-01
1.33958995e+00 -3.46508533e-01 -1.05958216e-01 -3.44497234e-01
1.20388322e-01 8.25876653e-01 2.48765096e-01 -2.45115787e-01
2.44156972e-01 5.80964327e-01 -3.88816595e-01 -7.93959647e-02
-4.45220947e-01 -5.69366395e-01 -1.61511093e-01 -7.27617592e-02
1.91348150e-01 -6.41263664e-01 -5.26027918e-01 7.21651196e-01
-6.37195826e-01 -2.74204791e-01 -7.59887695e-01 4.73579943e-01
-4.96170551e-01 -1.97875336e-01 -6.55192614e-01 -6.86182559e-01
-7.28802860e-01 -7.25380123e-01 6.57887757e-01 9.66534987e-02
-6.32269323e-01 -9.01105285e-01 2.39497542e-01 3.83281559e-01
4.10534322e-01 6.80530250e-01 8.06565166e-01 -1.13869798e+00
8.20030496e-02 -1.02690232e+00 6.43992171e-05 4.05780673e-01
2.31565222e-01 1.40215114e-01 -8.59202266e-01 -5.49202859e-01
-1.10086113e-01 -2.38536060e-01 8.78930867e-01 4.39560741e-01
9.32578385e-01 -3.70531619e-01 -5.63040376e-01 4.89686400e-01
1.59038377e+00 4.82310712e-01 8.03740442e-01 1.11919570e+00
2.77020246e-01 5.29279172e-01 8.93471301e-01 4.88728285e-01
1.18809320e-01 5.11477530e-01 4.97459769e-01 -3.68594706e-01
-3.58548984e-02 -5.91247827e-02 -8.84026363e-02 2.01209232e-01
-1.32492036e-01 -4.90769714e-01 -1.10293353e+00 3.51409346e-01
-1.52879906e+00 -1.05297518e+00 -3.07601720e-01 2.32882929e+00
7.08026171e-01 4.82943296e-01 5.95070660e-01 7.52729714e-01
7.59906471e-01 -2.37807080e-01 -5.65300107e-01 -7.20492363e-01
-1.15484722e-01 2.92036504e-01 7.69059122e-01 1.42665684e-01
-1.09362721e+00 4.43647623e-01 7.57618809e+00 4.97637957e-01
-1.21391273e+00 3.27220485e-02 1.05590844e+00 -4.32765812e-01
6.09618463e-02 -2.24172905e-01 -7.50725567e-01 1.03705084e+00
8.72587204e-01 -7.03739375e-02 -1.55600443e-01 1.07177567e+00
-6.88030869e-02 -2.25219950e-01 -8.51839662e-01 9.40082371e-01
7.03936741e-02 -1.22770643e+00 -1.45343930e-01 5.03029749e-02
9.08970416e-01 1.09944977e-01 -2.49814376e-01 4.39446986e-01
7.43301690e-01 -8.44351113e-01 5.94374239e-01 4.05706614e-01
2.20442772e-01 -1.01302242e+00 1.33724713e+00 1.18166856e-01
-1.69016644e-01 -2.81059951e-01 -4.22422290e-01 -2.86909968e-01
-2.66048878e-01 1.20355284e+00 -8.70940208e-01 5.47237396e-01
1.11333525e+00 3.01255286e-01 -8.76995623e-01 1.53028202e+00
-1.30848764e-02 8.53360951e-01 -3.63213062e-01 -2.46208802e-01
-2.53294438e-01 1.88802272e-01 2.98869193e-01 1.32309663e+00
1.35654002e-01 -4.45051074e-01 5.10201529e-02 2.93173283e-01
8.00426528e-02 2.63999909e-01 -3.75998914e-01 4.59172010e-01
5.73159695e-01 8.38030696e-01 -8.84621084e-01 -6.20116234e-01
-1.71864107e-01 4.43971843e-01 1.25965208e-01 2.48876885e-01
-5.96265912e-01 -5.38266242e-01 4.29103881e-01 3.59356880e-01
-6.08090274e-02 5.96115232e-01 -8.93202722e-01 -7.25894153e-01
5.06840609e-02 -1.17642665e+00 9.15170074e-01 -5.33998549e-01
-1.52547538e+00 4.18069959e-01 -3.13883461e-02 -1.34478986e+00
-4.82285172e-01 -5.04354298e-01 -6.63730025e-01 7.28447556e-01
-1.02852356e+00 -8.08203101e-01 -7.19956756e-01 8.06889161e-02
-1.35893160e-02 -2.83640832e-01 8.33209693e-01 4.37931389e-01
-4.55119759e-01 8.84678185e-01 3.73404652e-01 1.65179253e-01
8.99528325e-01 -1.43494475e+00 6.50317788e-01 3.51827711e-01
8.86685308e-03 4.58053350e-01 8.73693645e-01 -7.31323481e-01
-7.48128176e-01 -9.99807477e-01 5.93067348e-01 -6.13441348e-01
1.29533425e-01 -1.45190597e-01 -8.84596407e-01 4.47469801e-01
-4.51496057e-02 1.01942785e-01 8.47830832e-01 4.35146064e-01
-2.29988798e-01 -4.65624452e-01 -1.64295936e+00 9.15621445e-02
7.71006405e-01 4.24056910e-02 -4.96271938e-01 5.37612438e-01
1.60756353e-02 -5.86364269e-01 -9.31712389e-01 5.67237020e-01
6.26899898e-01 -1.44597101e+00 9.82627690e-01 -8.64136934e-01
3.49355638e-01 -8.49563349e-03 2.73789644e-01 -1.61540353e+00
-3.13841015e-01 -1.83479771e-01 9.82302241e-03 1.25941646e+00
3.04109246e-01 -9.75458503e-01 1.17101908e+00 3.55928749e-01
2.48545304e-01 -9.38531756e-01 -5.45818150e-01 -9.97118354e-01
9.03993696e-02 -5.12514114e-01 1.14514720e+00 1.21401060e+00
1.94971606e-01 -8.27782601e-02 -2.42086336e-01 -4.27852035e-01
3.54549736e-01 3.53944838e-01 1.08728707e+00 -1.12867904e+00
-4.62148219e-01 -6.16926670e-01 -1.05072486e+00 -1.84443429e-01
-4.73452568e-01 -8.17783237e-01 -4.26445603e-01 -1.47577596e+00
8.23886320e-02 -7.17203140e-01 -3.57715100e-01 6.40518248e-01
-4.82670218e-01 7.18095839e-01 4.75246385e-02 2.82340944e-01
-1.53220743e-01 -1.10780038e-01 7.27054894e-01 -4.36155885e-01
-4.19932842e-01 2.73487985e-01 -9.87228692e-01 4.85127598e-01
9.92290020e-01 -6.61156178e-01 -9.77995172e-02 -5.81882708e-02
2.98926204e-01 -3.50635320e-01 2.54898846e-01 -1.17163944e+00
-7.36879110e-02 5.86795509e-02 1.01694751e+00 -5.77577949e-01
-1.62521645e-01 -4.17340428e-01 2.21366942e-01 6.91117704e-01
-2.06302956e-01 2.54520386e-01 3.69389564e-01 2.58943945e-01
-3.37908208e-01 -3.56309503e-01 7.07352340e-01 -1.02193743e-01
-3.58582526e-01 -1.55092776e-01 5.31196482e-02 1.82382718e-01
1.16682827e+00 -6.07551455e-01 -6.14265323e-01 -2.68214375e-01
-6.88952506e-01 2.44005650e-01 5.26868820e-01 4.47020888e-01
1.39286622e-01 -1.31980300e+00 -7.72145033e-01 3.85464966e-01
3.73768717e-01 -2.14778081e-01 -1.34215295e-01 4.80255812e-01
-7.77738571e-01 1.10995844e-01 -5.04060090e-01 -5.11632085e-01
-1.19601130e+00 4.84465837e-01 4.59234148e-01 -5.41393995e-01
-2.41328925e-01 7.09446132e-01 -4.00682211e-01 -4.24487561e-01
1.51392475e-01 -2.17449978e-01 -1.10804908e-01 4.14194465e-01
4.69047457e-01 9.50966179e-01 8.26200724e-01 -3.26564647e-02
-3.77491027e-01 2.10847244e-01 -2.75817096e-01 3.55118871e-01
1.24710512e+00 4.05674994e-01 2.34342497e-02 2.68553704e-01
8.46721590e-01 -4.50207517e-02 -6.64762139e-01 3.61908168e-01
1.97578073e-01 -6.42767608e-01 -2.01702908e-01 -1.17489707e+00
-1.10677373e+00 4.03421640e-01 8.27563465e-01 8.33233058e-01
1.11368656e+00 -2.28452057e-01 2.36419320e-01 -9.92241502e-02
4.62752193e-01 -1.19201207e+00 -3.23256589e-02 9.55004841e-02
7.93800533e-01 -1.18706644e+00 5.75146616e-01 -6.19848147e-02
-3.32180470e-01 9.56878304e-01 6.17610455e-01 -3.99238050e-01
5.08720517e-01 1.23769544e-01 2.72778243e-01 -4.08257782e-01
-4.41695958e-01 2.50295997e-01 4.38232645e-02 7.46629775e-01
3.45190525e-01 3.27759176e-01 -6.35522723e-01 4.85992998e-01
-7.53872037e-01 4.92916964e-02 3.08519959e-01 8.76369834e-01
-1.74350757e-02 -1.22546875e+00 -7.93160260e-01 1.21501505e+00
-7.36231327e-01 1.30278200e-01 -6.87777877e-01 1.26588976e+00
3.53277415e-01 9.84189093e-01 4.55475509e-01 -3.62206191e-01
6.60671115e-01 1.67387396e-01 1.87827170e-01 -3.19309860e-01
-1.06823552e+00 -3.39369148e-01 2.15774596e-01 -5.17410338e-01
-2.93725193e-01 -6.49753809e-01 -9.82166886e-01 -5.75453103e-01
-7.34722793e-01 2.16845810e-01 8.42619956e-01 4.93379712e-01
2.50961721e-01 6.36245131e-01 4.69009787e-01 -4.09454107e-01
-7.24845946e-01 -1.13210332e+00 -6.72985017e-01 8.64360571e-01
1.27672479e-01 -5.74706674e-01 -4.69472259e-01 -2.94257641e-01] | [8.756759643554688, 4.27420711517334] |
ecfd0ecf-b724-41b2-aada-5ddb4a8052a6 | ab3dmot-a-baseline-for-3d-multi-object | 2008.08063 | null | https://arxiv.org/abs/2008.08063v1 | https://arxiv.org/pdf/2008.08063v1.pdf | AB3DMOT: A Baseline for 3D Multi-Object Tracking and New Evaluation Metrics | 3D multi-object tracking (MOT) is essential to applications such as autonomous driving. Recent work focuses on developing accurate systems giving less attention to computational cost and system complexity. In contrast, this work proposes a simple real-time 3D MOT system with strong performance. Our system first obtains 3D detections from a LiDAR point cloud. Then, a straightforward combination of a 3D Kalman filter and the Hungarian algorithm is used for state estimation and data association. Additionally, 3D MOT datasets such as KITTI evaluate MOT methods in 2D space and standardized 3D MOT evaluation tools are missing for a fair comparison of 3D MOT methods. We propose a new 3D MOT evaluation tool along with three new metrics to comprehensively evaluate 3D MOT methods. We show that, our proposed method achieves strong 3D MOT performance on KITTI and runs at a rate of $207.4$ FPS on the KITTI dataset, achieving the fastest speed among modern 3D MOT systems. Our code is publicly available at http://www.xinshuoweng.com/projects/AB3DMOT. | ['Jianren Wang', 'David Held', 'Kris Kitani', 'Xinshuo Weng'] | 2020-08-18 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-3.15632433e-01 -5.47567368e-01 -6.52964711e-02 -2.36848533e-01
-4.34665263e-01 -3.66836607e-01 5.45068204e-01 -2.06630290e-01
-6.06089652e-01 5.92905998e-01 -8.19343388e-01 -5.98725736e-01
4.51563904e-03 -8.47409785e-01 -6.16119266e-01 -5.59115112e-01
-5.21437563e-02 9.98746157e-01 9.61024165e-01 -3.12178761e-01
1.55585706e-01 6.83441162e-01 -1.93577933e+00 -6.31719887e-01
6.28604174e-01 9.79610205e-01 3.84844303e-01 1.05091274e+00
-6.44553602e-02 1.91799328e-01 -5.61329663e-01 3.40449154e-01
5.69811046e-01 4.13919576e-02 -8.31590891e-02 -3.38377267e-01
7.43022978e-01 -4.86656398e-01 -3.94991666e-01 7.75371790e-01
7.24390268e-01 -8.87895003e-02 3.61317813e-01 -1.70721900e+00
1.17330626e-01 4.07543182e-02 -4.80150968e-01 3.83783072e-01
2.97675431e-01 4.02303457e-01 5.54590881e-01 -7.71076620e-01
3.91598642e-01 1.23786283e+00 7.12099731e-01 2.84898371e-01
-1.05842018e+00 -1.15792835e+00 -2.87965328e-01 1.83579460e-01
-1.53205490e+00 -5.60849905e-01 3.92345816e-01 -3.36555868e-01
1.09902072e+00 2.78058261e-01 8.17220867e-01 6.28757536e-01
7.50176191e-01 5.71477175e-01 1.22632813e+00 8.91935080e-02
1.62927270e-01 -1.27947824e-02 3.58254403e-01 8.21174502e-01
7.22019851e-01 5.14023781e-01 -6.92600489e-01 6.26777261e-02
5.23363292e-01 8.82866234e-02 5.41716099e-01 -6.00690007e-01
-1.23289955e+00 4.30633336e-01 2.17906490e-01 -5.74764125e-02
-8.92132744e-02 5.81453741e-01 1.06319383e-01 4.87247258e-01
2.53381103e-01 -1.31384209e-01 -9.83767584e-02 -4.20251191e-01
-6.76544011e-01 4.36204553e-01 5.99919140e-01 1.58304977e+00
1.12515080e+00 -3.49285640e-02 2.46562839e-01 2.04239622e-01
5.81821203e-01 1.53730631e+00 -6.72167614e-02 -1.08848989e+00
1.88169122e-01 6.36917174e-01 2.70142466e-01 -6.94183826e-01
-5.62979460e-01 -9.17152241e-02 -4.36972618e-01 8.30741286e-01
2.03542858e-01 3.67323644e-02 -6.26374483e-01 1.11716592e+00
7.93425024e-01 2.77271003e-01 1.53991720e-02 7.32747316e-01
1.03873992e+00 4.22574043e-01 -5.10925829e-01 3.27719301e-02
1.22224069e+00 -5.29487669e-01 -6.98613405e-01 -4.65236783e-01
6.81862593e-01 -7.37234414e-01 6.42861843e-01 4.79188450e-02
-9.55457330e-01 -6.55382633e-01 -1.24201834e+00 -1.93958610e-01
-3.20218235e-01 -8.29744861e-02 5.24697661e-01 6.32905006e-01
-1.15808272e+00 9.79726911e-02 -1.49062097e+00 -7.86927640e-01
2.51555622e-01 7.45555162e-01 -6.90732673e-02 -9.45917442e-02
-6.89559579e-01 1.35592723e+00 2.90924549e-01 -1.93724945e-01
-9.17764664e-01 -2.63963968e-01 -7.08262503e-01 -5.45380294e-01
4.08632308e-01 -7.61569977e-01 1.33920753e+00 5.28075635e-01
-1.49112654e+00 8.87147427e-01 -4.37517703e-01 -6.34293735e-01
3.48398924e-01 -4.81079519e-01 -3.93811971e-01 -2.99742430e-01
3.62098604e-01 5.90182841e-01 5.88020205e-01 -9.20427203e-01
-1.14058828e+00 -4.14889455e-01 -1.23822160e-01 1.26525342e-01
1.88143864e-01 -2.03817844e-01 -5.15090883e-01 4.44320828e-01
2.87024498e-01 -1.31040382e+00 -1.41088665e-01 2.26255149e-01
-1.18921585e-01 -4.20460194e-01 1.37085199e+00 3.75812590e-01
8.52758527e-01 -2.05175281e+00 -3.97822142e-01 -1.66041985e-01
4.22586411e-01 2.39063248e-01 2.75668532e-01 3.07059199e-01
1.02291119e+00 -4.83085960e-01 2.93195434e-02 -5.34718215e-01
2.68618226e-01 4.23638165e-01 1.65939461e-02 7.33211994e-01
3.58165316e-02 8.04786265e-01 -1.02478242e+00 -6.95632935e-01
8.04662108e-01 4.30661708e-01 -3.56417239e-01 -1.25448808e-01
6.03171885e-02 2.68714756e-01 -6.26835942e-01 9.82403696e-01
9.82235610e-01 -2.22519159e-01 -1.28270045e-01 1.45494372e-01
-8.02451193e-01 6.03891611e-01 -1.20037270e+00 1.63914931e+00
-2.11095527e-01 8.46463919e-01 8.18121880e-02 -5.01567185e-01
1.23581111e+00 -1.97077498e-01 6.53800845e-01 -6.07956290e-01
3.61593843e-01 5.22546470e-01 -1.41074255e-01 -1.07861429e-01
1.02569282e+00 1.56255156e-01 -4.05562758e-01 3.34243178e-01
-2.10730642e-01 -7.56548762e-01 1.83603182e-01 1.62812293e-01
1.56226945e+00 2.82915860e-01 2.87193894e-01 -4.79614407e-01
2.47129545e-01 6.11413479e-01 6.72742903e-01 8.82596970e-01
-7.07802951e-01 -1.65419757e-01 -3.37909281e-01 -2.73132443e-01
-7.26892769e-01 -1.28579378e+00 -3.98074210e-01 5.99059463e-01
1.16425157e+00 -5.82551122e-01 1.14417285e-01 -3.93280089e-01
6.90743029e-01 3.86721551e-01 -1.79938793e-01 9.27886367e-02
-7.17478275e-01 -4.59735185e-01 5.83663642e-01 2.73421943e-01
6.93586171e-01 -4.34043348e-01 -1.28529465e+00 3.01918477e-01
1.56943977e-01 -1.23686767e+00 -1.61937967e-01 3.72528613e-01
-1.02765584e+00 -9.66776788e-01 5.20112924e-02 -5.49533904e-01
3.13579857e-01 9.97868061e-01 9.85267460e-01 1.40286282e-01
-1.60558864e-01 2.22525835e-01 -1.76074386e-01 -7.63022602e-01
-1.36046931e-01 -1.45932650e-02 7.46245801e-01 -7.62261689e-01
9.60943222e-01 -4.63282019e-01 -3.32903504e-01 8.64575446e-01
-1.26906365e-01 7.79352263e-02 6.33502901e-01 3.54285032e-01
8.68635654e-01 -7.99589604e-02 -1.03507951e-01 -2.94540554e-01
-1.05144612e-01 -4.03928161e-01 -1.31152976e+00 -5.55185914e-01
-7.10533381e-01 -6.75323606e-02 3.27020586e-02 -2.52389729e-01
-4.87586170e-01 4.43620920e-01 -9.43406820e-02 -5.10760725e-01
-1.79953039e-01 -1.14209458e-01 2.35005811e-01 -4.13405836e-01
5.67747295e-01 1.51566684e-01 1.40107006e-01 -4.78739232e-01
1.10519372e-01 8.36942136e-01 5.49362838e-01 -2.74786711e-01
1.30589414e+00 9.38591659e-01 5.51229835e-01 -1.07578683e+00
-4.30820853e-01 -6.95284963e-01 -7.16464996e-01 -3.80363554e-01
6.06964409e-01 -1.08019423e+00 -1.09868324e+00 3.54151130e-01
-7.95029819e-01 -4.58899379e-01 -9.82736275e-02 7.81954527e-01
-6.18807912e-01 -4.48429435e-02 -1.18770212e-01 -9.18924928e-01
-1.73340246e-01 -1.19406009e+00 1.38497961e+00 2.63810635e-01
2.51126941e-02 -4.98166472e-01 4.19071108e-01 2.11893246e-01
5.47808588e-01 1.60902038e-01 -1.25452638e-01 -1.26175955e-01
-1.20734787e+00 -2.20688760e-01 -2.14207903e-01 -4.19418037e-01
1.90250590e-01 5.11295302e-03 -8.63936067e-01 -5.69994450e-01
-2.53236860e-01 5.28649762e-02 6.21739507e-01 4.13718462e-01
3.73003334e-01 5.11144519e-01 -9.34879363e-01 6.21314228e-01
1.29771304e+00 3.61303061e-01 1.69183642e-01 4.80398357e-01
4.75391030e-01 -2.23598573e-02 1.22810435e+00 2.24368706e-01
8.44239831e-01 9.78917181e-01 6.88203096e-01 3.14079583e-01
-3.28493118e-01 -1.44032136e-01 5.37901938e-01 9.53282952e-01
-6.82670102e-02 -2.36217886e-01 -1.07245231e+00 5.55607438e-01
-1.83739448e+00 -7.29926825e-01 -6.32458210e-01 2.19547439e+00
1.23462163e-01 6.23013914e-01 1.77923247e-01 -5.70166914e-04
5.34032702e-01 1.36809945e-02 -7.87624538e-01 1.09390177e-01
1.35923550e-01 3.84640656e-02 1.07171154e+00 6.20110393e-01
-1.07663476e+00 1.13445783e+00 5.96607304e+00 5.49007058e-01
-9.18836951e-01 1.31668746e-01 -6.57131374e-01 -4.81827080e-01
3.20060492e-01 2.90214121e-01 -1.77465785e+00 3.81658643e-01
1.22529519e+00 -3.85159165e-01 6.81894794e-02 8.52631569e-01
2.15843067e-01 -3.11520875e-01 -8.71444345e-01 1.05476475e+00
-3.28893334e-01 -1.31616294e+00 -5.49801588e-01 6.65058732e-01
3.22956681e-01 8.00945640e-01 -2.09812418e-01 4.68239963e-01
6.40028119e-01 -4.26225126e-01 7.61712730e-01 2.63865530e-01
8.86647046e-01 -5.14531910e-01 6.60390735e-01 5.88648081e-01
-1.74704587e+00 2.74786115e-01 -5.98097444e-01 -4.99436915e-01
3.73482496e-01 5.36178589e-01 -1.01778710e+00 5.47165573e-01
1.14054096e+00 9.52220857e-01 -4.10294771e-01 1.10820246e+00
6.99231997e-02 2.87250578e-01 -9.78079438e-01 -5.68253934e-01
9.23141092e-02 -7.25390166e-02 1.06028688e+00 8.34156632e-01
6.53008640e-01 1.03390418e-01 3.36126834e-01 6.06298625e-01
1.60247058e-01 -5.02532780e-01 -1.20937479e+00 2.00199649e-01
8.83851588e-01 1.20380557e+00 -5.54520547e-01 -4.30820435e-01
-3.85533571e-01 2.95172989e-01 -1.59639582e-01 -4.26070988e-01
-9.16486681e-01 -4.85216111e-01 1.20740378e+00 1.61851436e-01
4.30069923e-01 -1.00474477e+00 -3.00320953e-01 -8.81245673e-01
-2.36799464e-01 -1.94158718e-01 -1.71567723e-02 -6.58059418e-01
-7.36093044e-01 3.18428814e-01 1.43635184e-01 -1.76813602e+00
-1.11621559e-01 -5.55858493e-01 -2.57056445e-01 4.98845339e-01
-1.82365692e+00 -7.93475091e-01 -7.72823811e-01 4.67561781e-01
3.66920620e-01 6.06840998e-02 3.65626425e-01 6.32633924e-01
-4.43414241e-01 2.69095451e-01 1.58390447e-01 -3.98186326e-01
6.94654524e-01 -1.16777587e+00 9.30278718e-01 8.78441691e-01
-7.53676668e-02 3.76903266e-01 8.44240189e-01 -9.85186636e-01
-2.28319144e+00 -1.02483082e+00 8.86800945e-01 -1.03894436e+00
6.57582521e-01 -5.60912251e-01 -4.74804878e-01 7.98036754e-01
-4.93692188e-03 1.13376871e-01 1.47399157e-01 -2.62640953e-01
3.85100871e-01 -3.66712242e-01 -1.13462460e+00 2.87315279e-01
1.50861216e+00 -4.40558270e-02 -3.68194342e-01 2.44205222e-01
6.58656836e-01 -9.19819355e-01 -8.36927891e-01 4.67683077e-01
7.47897148e-01 -8.64716470e-01 8.69813859e-01 2.39219502e-01
-6.09413266e-01 -1.10600817e+00 -4.18036044e-01 -7.21944988e-01
-2.29981378e-01 -7.25095510e-01 -4.32687759e-01 7.19934106e-01
1.17836865e-02 -8.92418742e-01 8.82547021e-01 -8.87135640e-02
-5.68298817e-01 -1.04524605e-01 -1.29186988e+00 -1.33886611e+00
-4.36269969e-01 -5.98120689e-01 7.58258760e-01 3.00187796e-01
-3.43688548e-01 4.10862535e-01 -2.91497260e-01 6.35508478e-01
1.20245981e+00 4.18308318e-01 1.73204911e+00 -1.59191179e+00
2.46495411e-01 -2.84506306e-02 -8.40218484e-01 -1.62759221e+00
-2.06598446e-01 -8.14897597e-01 3.72609973e-01 -1.39679039e+00
-1.92830518e-01 -8.84685576e-01 -1.82508036e-01 3.48898411e-01
3.25458050e-01 5.23252487e-01 8.60646218e-02 4.26283836e-01
-9.31636810e-01 5.43368340e-01 1.06944394e+00 -1.52176201e-01
-2.59966552e-01 2.17299432e-01 -2.13171467e-01 4.92773563e-01
8.20160568e-01 -7.61746287e-01 -2.48434581e-02 -5.92626214e-01
-1.33203745e-01 1.73609350e-02 4.75563079e-01 -1.62129152e+00
6.50203168e-01 -2.37533554e-01 -9.64827463e-03 -1.62154448e+00
9.07911658e-01 -9.63886559e-01 3.76293570e-01 8.68837297e-01
6.89100504e-01 2.57753700e-01 5.13047814e-01 3.90705228e-01
2.70644009e-01 3.30104113e-01 7.54737675e-01 1.28324807e-01
-1.12186348e+00 4.50458914e-01 -4.46673900e-01 -2.41119981e-01
1.23985612e+00 -5.50498903e-01 -6.89606369e-01 6.77501857e-02
9.74838901e-03 7.54627943e-01 8.80103886e-01 3.39444369e-01
5.91842234e-01 -1.42613959e+00 -5.64242542e-01 3.24445695e-01
2.78385490e-01 9.65900943e-02 -1.90560177e-01 1.11167097e+00
-6.42908394e-01 7.02443123e-01 -3.47463936e-01 -1.42736113e+00
-1.60476673e+00 1.02022216e-01 1.30616128e-01 2.43409932e-01
-9.07243133e-01 5.11622965e-01 -3.15775365e-01 -5.34296632e-01
6.54229522e-02 -6.03663087e-01 3.38874161e-01 -4.43235815e-01
3.96981955e-01 5.72319746e-01 5.73014840e-02 -7.42244720e-01
-9.05716538e-01 8.38059902e-01 2.18130961e-01 7.74808526e-02
1.12401056e+00 -4.59831804e-01 3.73800159e-01 6.07069075e-01
5.71308553e-01 -1.01368122e-01 -1.31472468e+00 -2.24019498e-01
-4.95761782e-02 -7.37507105e-01 2.85504431e-01 -1.83146954e-01
-3.52950215e-01 7.75974751e-01 1.10646307e+00 1.91497296e-01
6.72834218e-01 2.17873812e-01 9.89205182e-01 7.69259155e-01
1.10208285e+00 -7.46137023e-01 -3.39428604e-01 8.43577087e-01
1.80893820e-02 -1.52642691e+00 2.34636992e-01 -4.67938632e-02
2.19953451e-02 5.74279428e-01 7.36454129e-01 -2.09381789e-01
4.71972853e-01 6.65536284e-01 2.55059361e-01 -4.45412725e-01
-8.14768076e-01 -7.47642159e-01 -9.71589014e-02 6.41889632e-01
-1.30367115e-01 8.82136226e-02 -9.82000753e-02 -3.42962384e-01
-3.95266801e-01 5.42331487e-02 4.58561540e-01 1.51046228e+00
-9.85590756e-01 -1.01723182e+00 -4.72818404e-01 3.79478157e-01
1.48764789e-01 2.94263482e-01 -1.06268570e-01 1.16190732e+00
4.08143774e-02 1.10222304e+00 2.89332747e-01 -8.23800862e-01
5.48504114e-01 -2.48245806e-01 5.03204048e-01 -5.92922926e-01
-1.63643822e-01 -7.11280704e-02 2.50684861e-02 -7.69206882e-01
-4.76185441e-01 -9.10884857e-01 -1.38881159e+00 -8.17238510e-01
-6.01384521e-01 3.28907706e-02 1.11287212e+00 7.35388100e-01
5.77213287e-01 3.05076301e-01 4.31707889e-01 -1.13878441e+00
-9.62307751e-02 -7.03038335e-01 -5.40462077e-01 -4.09159601e-01
6.19381189e-01 -1.26047993e+00 -3.45486581e-01 -4.96531427e-01] | [6.6661763191223145, -2.2181155681610107] |
9df03d62-d71b-46e5-91fb-d4d555e611a9 | quantum-machine-learning-for-image | 2304.09224 | null | https://arxiv.org/abs/2304.09224v1 | https://arxiv.org/pdf/2304.09224v1.pdf | Quantum machine learning for image classification | Image recognition and classification are fundamental tasks with diverse practical applications across various industries, making them critical in the modern world. Recently, machine learning models, particularly neural networks, have emerged as powerful tools for solving these problems. However, the utilization of quantum effects through hybrid quantum-classical approaches can further enhance the capabilities of traditional classical models. Here, we propose two hybrid quantum-classical models: a neural network with parallel quantum layers and a neural network with a quanvolutional layer, which address image classification problems. One of our hybrid quantum approaches demonstrates remarkable accuracy of more than 99% on the MNIST dataset. Notably, in the proposed quantum circuits all variational parameters are trainable, and we divide the quantum part into multiple parallel variational quantum circuits for efficient neural network learning. In summary, our study contributes to the ongoing research on improving image recognition and classification using quantum machine learning techniques. Our results provide promising evidence for the potential of hybrid quantum-classical models to further advance these tasks in various fields, including healthcare, security, and marketing. | ['Alexey Melnikov', 'Asel Sagingalieva', 'Alexander Sedykh', 'Arsenii Senokosov'] | 2023-04-18 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 5.70105135e-01 -1.46198899e-01 -3.92587185e-01 -2.61734277e-01
-6.89197421e-01 -2.10768953e-01 5.27885079e-01 -2.27208771e-02
-5.67645490e-01 5.93957841e-01 -5.33221185e-01 -3.27061743e-01
-9.88752916e-02 -1.00801671e+00 -5.84475160e-01 -1.42060709e+00
3.79957348e-01 2.47887284e-01 9.58226547e-02 -4.87293810e-01
6.55681849e-01 2.22789824e-01 -1.54595411e+00 3.54877293e-01
1.15095472e+00 1.20638502e+00 -2.00458035e-01 6.49504423e-01
1.67795926e-01 5.16235113e-01 -5.15611358e-02 -6.85540736e-01
1.27108201e-01 -7.24043429e-01 -4.91425693e-01 -4.99384314e-01
1.13338754e-01 -1.13173388e-01 -8.73630345e-01 1.62402070e+00
4.74776417e-01 1.30826518e-01 6.88571930e-01 -6.96112812e-01
-1.24617493e+00 6.70250714e-01 -2.63765037e-01 1.40242323e-01
-2.77820647e-01 4.31367278e-01 1.30700159e+00 -2.25443095e-01
3.01538140e-01 8.57219338e-01 2.57968426e-01 9.16878283e-01
-1.48746121e+00 -9.14643764e-01 -8.63328755e-01 9.17882919e-01
-1.38800812e+00 -2.87463129e-01 7.10174263e-01 -2.05326900e-01
1.14965713e+00 -2.02176511e-01 4.63278532e-01 9.11901891e-01
5.61070561e-01 3.58606845e-01 1.26218569e+00 -7.25539923e-01
2.61462331e-01 2.70925045e-01 3.85717154e-01 8.42261493e-01
9.54477787e-02 7.85662353e-01 -8.33758771e-01 -9.40945968e-02
5.52631438e-01 3.12232655e-02 -6.75474852e-02 -1.69937033e-02
-1.09130120e+00 1.13036990e+00 9.04344916e-01 3.26040447e-01
-3.57872933e-01 4.98750538e-01 2.64561564e-01 7.33526573e-02
7.72240385e-02 6.58968270e-01 -6.46266267e-02 -2.53348388e-02
-5.46382129e-01 -4.11994867e-02 3.68766308e-01 3.83357495e-01
9.98676360e-01 -1.30879074e-01 8.59188661e-03 4.52540278e-01
1.49191737e-01 8.43696117e-01 2.51791954e-01 -1.17021096e+00
-1.37350708e-01 3.46112341e-01 -2.15056866e-01 -5.28586328e-01
-3.17310512e-01 -4.05515105e-01 -1.24342453e+00 -5.69101907e-02
1.45504802e-01 1.44244790e-01 -5.28589189e-01 1.76334667e+00
1.18285976e-01 2.28258088e-01 4.44314897e-01 6.98487222e-01
7.31990159e-01 7.89754450e-01 6.57289103e-03 -1.30415574e-01
1.37257183e+00 -7.67535627e-01 -7.53634393e-01 3.21462005e-01
9.31460381e-01 -5.49609363e-01 5.32430470e-01 5.09725273e-01
-7.75605142e-01 -5.63239515e-01 -1.39302623e+00 -1.45618290e-01
-4.00760412e-01 -1.13949262e-01 1.16710126e+00 1.05987358e+00
-7.36428022e-01 1.27092314e+00 -9.34164822e-01 -1.12549067e-01
4.94060308e-01 8.14433873e-01 -1.89721018e-01 -3.32854316e-02
-1.53218937e+00 8.53316247e-01 6.33929610e-01 1.72618330e-01
-2.98007101e-01 -6.93203583e-02 -4.28696036e-01 1.98282301e-02
-2.90034469e-02 -9.29298639e-01 1.13721466e+00 -4.97885138e-01
-2.00705385e+00 6.93818212e-01 -1.19985484e-01 -5.35304010e-01
-1.94337323e-01 3.03738594e-01 -3.13152850e-01 3.86834681e-01
-3.16068321e-01 6.48486316e-01 6.30925000e-01 -4.60491717e-01
-3.04147571e-01 -5.99262893e-01 7.31210187e-02 -2.02893928e-01
-3.77159119e-01 -3.19351584e-01 1.24012031e-01 1.75037608e-01
1.44279703e-01 -1.27345383e+00 -2.42247969e-01 -3.25363666e-01
-3.14257979e-01 -5.38611114e-01 2.30655849e-01 1.13474905e-01
8.95006418e-01 -2.14970040e+00 4.23882961e-01 1.19260222e-01
1.42254099e-01 3.30832422e-01 -7.74808824e-02 3.63382369e-01
1.02235526e-02 1.39235705e-01 -1.55732885e-01 1.64550081e-01
9.13901702e-02 2.80145645e-01 -1.37361944e-01 5.19895434e-01
2.39810839e-01 1.37768233e+00 -6.20439768e-01 -2.03691378e-01
2.04280183e-01 6.26919091e-01 -9.40260112e-01 -2.49091119e-01
-4.21168804e-02 7.41490841e-01 -5.95583856e-01 2.98966020e-01
7.33730793e-01 -7.36284494e-01 1.75584823e-01 -3.90781969e-01
-9.08623859e-02 2.21222743e-01 -6.70762718e-01 1.67794001e+00
-7.08107427e-02 4.70523536e-01 -2.21826807e-01 -1.30697525e+00
3.21105510e-01 8.04540291e-02 3.20253134e-01 -1.33793783e+00
3.98562878e-01 5.12510419e-01 5.43554664e-01 -6.67665601e-01
4.77787942e-01 -7.38523483e-01 2.97105182e-02 3.52776378e-01
3.87982577e-01 -2.88972586e-01 1.44231126e-01 3.35573666e-02
6.77124441e-01 -1.21134236e-01 -8.30294937e-03 -6.95960000e-02
6.48374140e-01 -1.26186952e-01 2.19813585e-01 1.03035569e+00
-6.71179891e-01 2.82298476e-01 4.73606944e-01 -2.48065218e-01
-1.10382843e+00 -8.38808954e-01 -6.81695104e-01 6.70168519e-01
5.28341174e-01 -3.44139129e-01 -8.28556836e-01 -8.30019414e-02
-3.62086684e-01 4.50099230e-01 -5.04159570e-01 -8.72368455e-01
-4.36030179e-01 -1.65211904e+00 7.89481282e-01 7.94328675e-02
7.57807612e-01 -7.88465679e-01 -2.32930928e-01 2.95135733e-02
-3.84153351e-02 -1.33707702e+00 7.23995939e-02 4.26605850e-01
-9.00632560e-01 -1.09252393e+00 -3.79198760e-01 -4.99542683e-01
2.67671973e-01 6.70484677e-02 4.77806687e-01 1.16246054e-03
-4.99997079e-01 6.20921552e-02 -2.97241509e-01 -4.11449447e-02
-7.83067763e-01 2.64935285e-01 2.67779499e-01 2.08201222e-02
5.88639975e-01 -3.86730909e-01 -8.17608535e-01 -2.11577788e-01
-9.33463216e-01 1.44028634e-01 8.77657294e-01 1.15089726e+00
5.10136724e-01 1.24894813e-01 4.13107723e-01 -7.69567907e-01
2.33713001e-01 -7.77505934e-02 -6.21501148e-01 3.02749097e-01
-6.79738104e-01 5.95628083e-01 7.17802525e-01 -1.26813576e-01
-9.20897126e-01 -2.49522269e-01 -5.06323576e-01 1.92502946e-01
2.04130430e-02 5.25812447e-01 4.56908464e-01 -6.79281592e-01
8.97691488e-01 3.42611820e-01 1.43525079e-01 1.03106230e-01
4.47528303e-01 8.21118355e-01 2.60801166e-01 -3.87422919e-01
6.91623211e-01 4.25882280e-01 6.77103817e-01 -1.18286562e+00
-9.17470992e-01 -2.29469046e-01 -6.13732755e-01 1.38460919e-01
1.41278398e+00 -6.67391956e-01 -1.29202878e+00 6.37793124e-01
-1.18637300e+00 3.52936909e-02 5.03905714e-02 9.66793954e-01
-4.21516061e-01 4.35454458e-01 -1.07932246e+00 -1.06333888e+00
-3.22128594e-01 -1.54914618e+00 9.03134704e-01 6.89040601e-01
5.80421984e-01 -8.20115566e-01 -4.28650863e-02 7.15093315e-01
3.68793547e-01 -1.13959320e-01 1.35261416e+00 -2.96217918e-01
-9.45692241e-01 -2.06743047e-01 -4.06700462e-01 5.62093496e-01
-3.63417566e-01 -8.51757452e-02 -1.26300907e+00 -7.37227127e-02
-2.63082869e-02 -7.50075281e-01 1.30567193e+00 2.42540717e-01
1.11077285e+00 3.52427274e-01 -1.64203405e-01 5.67446887e-01
1.42798448e+00 4.08007681e-01 7.15710521e-01 -2.03381583e-01
5.85888147e-01 2.75175184e-01 -2.01228172e-01 5.59365638e-02
2.20160872e-01 6.13605618e-01 4.15594071e-01 3.05801868e-01
2.01154977e-01 2.49806419e-01 3.79531264e-01 1.34202611e+00
-4.33421582e-01 2.58155942e-01 -6.28991663e-01 -3.35552603e-01
-1.72128820e+00 -1.25266135e+00 -4.50906962e-01 2.27489448e+00
5.81493020e-01 4.06976715e-02 -4.73464966e-01 -2.97538489e-02
5.34875631e-01 -4.19383384e-02 -7.46663630e-01 -4.20282632e-01
-3.11709255e-01 7.54050493e-01 3.52452338e-01 2.13179216e-01
-1.07119823e+00 8.97719383e-01 6.17284679e+00 9.02679145e-01
-1.37251210e+00 3.13694417e-01 4.27125037e-01 2.15238869e-01
-6.11214079e-02 1.02206126e-01 -6.88560367e-01 -8.37181322e-03
1.40321255e+00 1.92972392e-01 8.00556839e-01 3.66736352e-01
-1.68335870e-01 -2.05856651e-01 -1.10005820e+00 1.35402322e+00
-7.11578280e-02 -1.71517992e+00 9.19118822e-02 3.93782854e-01
8.76102984e-01 5.25746644e-01 5.04922986e-01 4.40986812e-01
-4.90992725e-01 -1.02590752e+00 2.78294861e-01 4.46907699e-01
8.09426188e-01 -8.35376263e-01 8.85307550e-01 3.92039329e-01
-5.96863627e-01 -1.61476240e-01 -5.07802069e-01 -2.28066772e-01
-1.29079208e-01 2.82453984e-01 1.50322750e-01 7.58224368e-01
5.04470468e-01 6.84331775e-01 -4.49199468e-01 6.01644218e-01
4.91959527e-02 4.44591016e-01 -1.31287232e-01 -4.42043364e-01
3.46214473e-01 -8.01534712e-01 1.48818091e-01 5.16934872e-01
1.56131908e-01 3.34009826e-01 -2.58132458e-01 1.29434133e+00
-2.72576898e-01 2.23525334e-02 -4.80013639e-01 -7.81502008e-01
-1.04404055e-01 1.14242041e+00 -6.47980094e-01 -3.38196129e-01
-3.75119805e-01 9.97170329e-01 2.04300463e-01 1.84327990e-01
-9.32495415e-01 -6.67774379e-01 3.30405653e-01 -5.84766924e-01
4.04344052e-02 -2.92689443e-01 -3.39948446e-01 -1.77493727e+00
-4.14075613e-01 -5.93757868e-01 -1.17276259e-01 -2.28737861e-01
-1.13043916e+00 3.33354443e-01 -4.25454110e-01 -8.85322213e-01
3.42267542e-03 -1.19660795e+00 -3.09497237e-01 7.34894812e-01
-1.59911323e+00 -8.98269594e-01 1.53184339e-01 2.22601205e-01
-2.59027898e-01 -4.21441495e-02 1.23050761e+00 3.94317925e-01
-1.04739857e+00 4.99932259e-01 8.65178883e-01 6.69204146e-02
4.08006728e-01 -1.02457952e+00 -8.05019774e-03 5.81059039e-01
5.39513111e-01 8.87232900e-01 4.54555750e-01 -6.05816394e-02
-1.84896255e+00 -5.35541058e-01 5.20913959e-01 -3.82596970e-01
7.06158459e-01 -3.94221038e-01 -8.22083056e-01 3.62444557e-02
3.47214133e-01 -9.52080637e-02 9.92275834e-01 1.96123004e-01
-4.96404380e-01 -1.17138125e-01 -8.30542982e-01 3.14664632e-01
6.25560760e-01 -1.18699586e+00 -2.35805094e-01 6.18512630e-01
5.02974927e-01 -9.89507064e-02 -8.40885699e-01 3.62586319e-01
8.56273115e-01 -1.35575902e+00 6.72739565e-01 -4.20396239e-01
5.91502428e-01 -3.67756076e-02 -2.39919409e-01 -1.10600209e+00
-1.40585363e-01 -6.44279063e-01 1.53578054e-02 3.21677446e-01
4.69986767e-01 -1.00876868e+00 7.98467755e-01 4.69293475e-01
-7.20660612e-02 -6.94847286e-01 -1.23949027e+00 -5.57342768e-01
5.62534750e-01 -6.38836801e-01 4.11136933e-02 5.95704675e-01
1.53337121e-01 6.45598173e-01 -1.35805830e-01 -1.33972419e-02
8.76502931e-01 1.37746766e-01 2.13455543e-01 -1.03808355e+00
-5.13473153e-01 -8.37141752e-01 -5.86194515e-01 -8.96887481e-01
2.41175652e-01 -1.33386481e+00 -2.20719010e-01 -8.84911001e-01
8.20935726e-01 -2.13515505e-01 -6.91998303e-01 -6.67774081e-02
3.05702668e-02 6.31938577e-01 1.37194637e-02 3.69369298e-01
-5.95919788e-01 7.93491364e-01 1.38747597e+00 -3.56865883e-01
1.28633991e-01 -2.25616321e-01 -4.26665395e-01 3.92877162e-01
5.70427954e-01 -5.02731681e-01 1.40921120e-02 -3.10190946e-01
6.24992251e-01 -9.95389000e-02 4.47871596e-01 -1.07696998e+00
4.48897123e-01 1.22441158e-01 1.65825918e-01 -3.50775234e-02
4.56358463e-01 -1.90843359e-01 -3.43952775e-01 1.02691758e+00
-2.36161605e-01 -4.44834232e-01 -2.37173960e-01 6.70367062e-01
-2.82354414e-01 -2.40636557e-01 1.29702151e+00 -1.12856723e-01
-4.70813394e-01 3.73607725e-01 -1.85702711e-01 -4.01887238e-01
1.00820029e+00 1.42591417e-01 -5.99561632e-01 -2.39485707e-02
-6.13975883e-01 -1.10818826e-01 2.24769115e-01 -3.21467333e-02
3.90118599e-01 -1.10607994e+00 -2.24431157e-01 3.05357158e-01
3.11376095e-01 -4.56571668e-01 6.90366983e-01 1.24866700e+00
-6.63159966e-01 9.36474085e-01 -1.72441259e-01 -9.11733627e-01
-5.65176010e-01 5.01428366e-01 6.65710747e-01 -6.15999699e-02
-5.81868552e-02 7.26058245e-01 1.76908955e-01 -5.07796705e-01
-1.99814022e-01 -1.20350488e-01 3.08871781e-03 -2.24944040e-01
4.45750177e-01 3.50550771e-01 3.23123597e-02 -8.76404583e-01
-1.74353391e-01 8.04063499e-01 -6.20517135e-02 -1.54433310e-01
1.11515820e+00 2.22652018e-01 -4.79938269e-01 4.88369882e-01
1.38761985e+00 -6.78503752e-01 -6.08762980e-01 -2.75963843e-01
-9.50534791e-02 4.07802641e-01 2.73428947e-01 -3.88219148e-01
-6.03646517e-01 1.52733612e+00 8.84223580e-01 3.57026458e-01
9.13436532e-01 1.37678459e-01 1.01646888e+00 1.09078276e+00
6.81533873e-01 -1.04105842e+00 9.51229632e-02 5.84857225e-01
1.19777262e-01 -1.65920961e+00 -2.56576538e-01 -3.20341773e-02
-1.07988968e-01 1.40762591e+00 2.51286864e-01 -2.91390151e-01
6.64545536e-01 -5.23924351e-01 -1.88388735e-01 -2.43684620e-01
-6.63344085e-01 -4.01508838e-01 3.43559533e-01 6.99391365e-02
5.04118204e-01 3.04178387e-01 -2.57839561e-01 3.31666142e-01
-6.43945411e-02 -1.81081861e-01 5.37519097e-01 7.12407470e-01
-4.06306297e-01 -1.52295065e+00 -2.28252411e-02 4.31026906e-01
-5.67937434e-01 -2.20101446e-01 -1.80424824e-01 4.53859985e-01
2.88000852e-01 8.84865522e-01 -1.63514882e-01 -7.47859597e-01
-7.93236941e-02 2.82775104e-01 1.17738509e+00 -5.32057941e-01
-4.15823400e-01 -7.30320290e-02 -5.53109527e-01 -3.78634214e-01
-7.70181596e-01 -4.28090155e-01 -1.44248939e+00 -4.37125236e-01
-9.94734049e-01 2.45505407e-01 9.82240677e-01 1.31130552e+00
4.45271701e-01 4.31244850e-01 5.24330378e-01 -8.01328957e-01
-1.08122313e+00 -7.37598300e-01 -5.36092043e-01 2.19999835e-01
3.57619524e-01 -5.61954200e-01 -1.71264067e-01 -3.24769318e-01] | [5.556402683258057, 4.974667072296143] |
44ab959b-2be9-4305-ac85-c2a8e1095501 | self-supervised-geometric-correspondence-for | 2210.07199 | null | https://arxiv.org/abs/2210.07199v3 | https://arxiv.org/pdf/2210.07199v3.pdf | Self-Supervised Geometric Correspondence for Category-Level 6D Object Pose Estimation in the Wild | While 6D object pose estimation has wide applications across computer vision and robotics, it remains far from being solved due to the lack of annotations. The problem becomes even more challenging when moving to category-level 6D pose, which requires generalization to unseen instances. Current approaches are restricted by leveraging annotations from simulation or collected from humans. In this paper, we overcome this barrier by introducing a self-supervised learning approach trained directly on large-scale real-world object videos for category-level 6D pose estimation in the wild. Our framework reconstructs the canonical 3D shape of an object category and learns dense correspondences between input images and the canonical shape via surface embedding. For training, we propose novel geometrical cycle-consistency losses which construct cycles across 2D-3D spaces, across different instances and different time steps. The learned correspondence can be applied for 6D pose estimation and other downstream tasks such as keypoint transfer. Surprisingly, our method, without any human annotations or simulators, can achieve on-par or even better performance than previous supervised or semi-supervised methods on in-the-wild images. Our project page is: https://kywind.github.io/self-pose . | ['Xiaolong Wang', 'Fatih Porikli', 'Hong Cai', 'Shubhankar Borse', 'Yang Fu', 'Kaifeng Zhang'] | 2022-10-13 | null | null | null | null | ['6d-pose-estimation-1', '6d-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-1.74758881e-02 1.07654274e-01 -1.16575234e-01 -2.79674232e-01
-9.62957382e-01 -8.43679368e-01 5.92358291e-01 -6.39477149e-02
-3.93403709e-01 2.32419088e-01 -7.35333264e-02 8.47026780e-02
-1.17181055e-02 -2.84081697e-01 -1.18017209e+00 -4.54412669e-01
-1.09528877e-01 9.18888509e-01 3.83929640e-01 1.79828219e-02
2.43387952e-01 6.63209856e-01 -1.54498732e+00 -2.23329484e-01
4.54317361e-01 8.72453451e-01 3.28188300e-01 6.05597019e-01
1.88118592e-01 1.46476492e-01 -1.32961214e-01 -3.69619913e-02
6.57254338e-01 -2.13210732e-01 -6.30387485e-01 3.27722698e-01
8.28891397e-01 -4.61104304e-01 -3.14509094e-01 9.42620397e-01
4.42178607e-01 -9.88149270e-03 6.99470997e-01 -1.37258232e+00
-1.54462337e-01 -1.60557717e-01 -4.88745958e-01 -2.50346959e-01
5.71570933e-01 2.37122014e-01 7.01180041e-01 -1.03556049e+00
9.97741461e-01 1.20616984e+00 8.27636480e-01 6.88572228e-01
-1.30234253e+00 -5.47998369e-01 1.86841667e-01 4.97542694e-02
-1.27173734e+00 -2.62375534e-01 8.35049987e-01 -7.10005224e-01
8.72400224e-01 -1.86925754e-01 8.02539468e-01 1.11511421e+00
1.25406608e-01 7.56135166e-01 9.21947300e-01 -3.31127435e-01
1.63023621e-01 -6.33445531e-02 -3.96609664e-01 7.18897343e-01
1.94633201e-01 1.06872916e-01 -5.36210775e-01 9.49469805e-02
1.19359875e+00 1.56129524e-01 -2.77057499e-01 -1.33769929e+00
-1.76653981e+00 5.53541064e-01 5.38156569e-01 -7.00264275e-02
-9.03128311e-02 3.60451490e-01 1.69707268e-01 2.20993161e-01
5.24312913e-01 5.26488721e-01 -7.40663111e-01 -1.26298770e-01
-4.46508139e-01 4.42812115e-01 7.09947288e-01 1.21161580e+00
8.38446498e-01 -3.06532145e-01 3.54268581e-01 3.56636465e-01
3.80697966e-01 6.63601518e-01 1.37514666e-01 -1.31859565e+00
2.76325434e-01 5.25835335e-01 4.07136738e-01 -8.87967288e-01
-3.43170613e-01 -3.33238542e-01 -4.25926805e-01 5.23303211e-01
6.32806599e-01 2.32533246e-01 -1.04508066e+00 1.74614561e+00
7.85708129e-01 4.50601459e-01 -3.50821018e-01 1.29381752e+00
5.10156214e-01 4.09305543e-01 -2.70870328e-01 1.32220209e-01
1.12173092e+00 -9.44755793e-01 -2.24843070e-01 -3.62138957e-01
5.35666704e-01 -7.56064534e-01 1.00017595e+00 3.67884547e-01
-9.06166136e-01 -3.93664479e-01 -8.89432371e-01 -2.34735280e-01
-2.49787077e-01 -4.75684442e-02 5.24751127e-01 1.11989632e-01
-8.61948550e-01 5.13967812e-01 -1.05383873e+00 -5.81078529e-01
4.06862706e-01 4.54801768e-01 -7.67230451e-01 -1.33775502e-01
-6.53171062e-01 9.95458186e-01 7.79584721e-02 -5.81641160e-02
-1.04569149e+00 -1.03843379e+00 -1.17906868e+00 -5.31534612e-01
6.78148329e-01 -9.21326578e-01 1.38141990e+00 -5.66155791e-01
-1.51193118e+00 1.28267002e+00 1.50565982e-01 -2.03611076e-01
8.80941391e-01 -5.55424333e-01 3.75088751e-01 1.60044625e-01
1.50357500e-01 8.26734066e-01 1.06007648e+00 -1.52860820e+00
-1.90006346e-01 -7.52110004e-01 2.04177976e-01 3.80239666e-01
3.37648652e-02 -4.42306280e-01 -6.77518785e-01 -5.38446546e-01
4.85651404e-01 -1.23272991e+00 -3.18531930e-01 7.79883683e-01
-1.55262440e-01 -1.95702046e-01 9.01273012e-01 -4.10660684e-01
2.65478641e-01 -2.04771590e+00 6.09415293e-01 -1.06460959e-01
1.08001783e-01 8.38476866e-02 -1.88303038e-01 2.60369629e-01
1.04797184e-01 -3.65421712e-01 -2.58344948e-01 -5.08438647e-01
6.48779571e-02 1.36925802e-01 -1.57028511e-01 9.23535824e-01
2.82654434e-01 1.04528582e+00 -1.14093339e+00 -3.78949851e-01
5.40302873e-01 6.49863601e-01 -7.13345408e-01 3.88181239e-01
-3.44155610e-01 8.66016924e-01 -5.04437745e-01 5.59553742e-01
6.54525757e-01 -3.50714147e-01 -1.05608955e-01 -4.07517642e-01
1.27347425e-01 1.06080510e-01 -1.17840254e+00 2.49824238e+00
-6.88629568e-01 4.30351228e-01 2.30203345e-01 -1.19048226e+00
6.62129819e-01 5.61197251e-02 6.38375401e-01 -2.77919441e-01
1.87863052e-01 3.13493907e-01 -2.80432343e-01 -4.44290906e-01
1.37971148e-01 -2.48737782e-02 -2.95677245e-01 3.23753506e-01
3.22468877e-01 -8.69052172e-01 -3.76114517e-01 4.22700010e-02
9.43981886e-01 7.61851013e-01 2.01511726e-01 -1.06333233e-01
2.35942990e-01 1.68195769e-01 3.34244281e-01 2.70334542e-01
-1.19242787e-01 8.23876619e-01 1.62824556e-01 -2.76798218e-01
-1.33654094e+00 -1.35600197e+00 -2.16737747e-01 6.00852787e-01
5.75375676e-01 -2.30928585e-01 -6.53263748e-01 -7.64531970e-01
4.57662374e-01 2.02187940e-01 -5.32959163e-01 -1.26959965e-01
-6.39947355e-01 -6.79619936e-03 6.95792660e-02 5.23399532e-01
2.10937947e-01 -7.17262983e-01 -8.58163476e-01 1.64939970e-01
9.84122455e-02 -1.44706535e+00 -5.38812041e-01 5.19073457e-02
-9.85023975e-01 -1.05164814e+00 -8.74957383e-01 -8.14457119e-01
8.29931974e-01 3.20386559e-01 9.46744025e-01 -2.93253899e-01
-5.65565765e-01 7.91264653e-01 -2.51962453e-01 -3.08341235e-01
-1.87704265e-01 -5.30454405e-02 3.37190539e-01 -1.62204430e-01
4.94164340e-02 -8.15817237e-01 -8.16985846e-01 5.39466560e-01
-5.29092312e-01 2.47954875e-02 5.44010699e-01 7.05807328e-01
8.55359197e-01 -4.75565404e-01 2.97313958e-01 -5.60159445e-01
-2.09080249e-01 -1.71355456e-01 -7.35276759e-01 -9.52088162e-02
-2.90449798e-01 5.02682105e-02 3.25132489e-01 -7.27930844e-01
-6.51764452e-01 5.58986902e-01 1.03218496e-01 -9.00329471e-01
-2.96738446e-01 -4.14938107e-02 -2.37819895e-01 -2.69497633e-01
4.80648190e-01 -6.15761848e-03 3.13658178e-01 -5.88696063e-01
3.60235780e-01 3.16725314e-01 5.45072734e-01 -5.66180587e-01
1.22297680e+00 7.19406009e-01 1.20993719e-01 -6.90120399e-01
-9.65718091e-01 -6.79992557e-01 -9.61939275e-01 -3.37336063e-01
9.43024158e-01 -1.12500393e+00 -7.15659738e-01 5.30472696e-01
-1.10469520e+00 -6.29795849e-01 -3.17125022e-01 7.31872320e-01
-1.06512988e+00 2.76609957e-01 -3.09209317e-01 -4.91101772e-01
-6.85291588e-02 -1.04429114e+00 1.66056263e+00 -2.01068908e-01
-1.14581712e-01 -7.83052266e-01 1.94467362e-02 3.88559103e-01
7.54817214e-04 4.86592919e-01 4.33793545e-01 -1.69918865e-01
-9.38786685e-01 -3.80195737e-01 -3.02657038e-02 2.49049589e-01
3.71435890e-03 -3.83110911e-01 -7.55479991e-01 -4.40591544e-01
-9.53615680e-02 -6.45047605e-01 5.78898132e-01 2.55553305e-01
1.18629515e+00 -5.96309751e-02 -3.58079791e-01 7.77335823e-01
1.36798501e+00 -3.87318283e-01 2.37591237e-01 2.69238472e-01
8.97330225e-01 7.39016950e-01 8.53622913e-01 1.08649820e-01
4.04508412e-01 1.03840554e+00 7.52628863e-01 1.81413114e-01
-2.52257109e-01 -5.33962727e-01 3.10817480e-01 7.71944284e-01
-1.37458548e-01 7.64175951e-02 -1.08915663e+00 6.07140064e-01
-1.69155514e+00 -5.23316801e-01 6.20932914e-02 2.35423255e+00
5.99232018e-01 2.27136880e-01 3.54001336e-02 -2.21684240e-02
4.93659407e-01 -1.04570992e-01 -9.37107623e-01 2.68411398e-01
3.00647765e-01 -9.78224631e-03 4.84742045e-01 5.37743211e-01
-9.70734656e-01 9.24080610e-01 5.11629152e+00 5.30019104e-01
-1.23331583e+00 2.55250067e-01 1.78537846e-01 -1.59339570e-02
-1.30041137e-01 7.65643567e-02 -6.27982795e-01 4.65480946e-02
3.49011749e-01 4.73855734e-02 2.81800747e-01 9.38986838e-01
-7.02363253e-02 -5.80617562e-02 -1.53714120e+00 1.25072849e+00
3.25207233e-01 -1.15161872e+00 -9.41473544e-02 1.39151260e-01
7.80150652e-01 1.62023246e-01 -1.55309914e-02 7.18937814e-02
3.80909890e-02 -8.07567060e-01 8.86725068e-01 3.98560256e-01
9.55710828e-01 -3.66053820e-01 3.71565163e-01 6.35752738e-01
-1.11854291e+00 2.16894686e-01 -1.67070866e-01 -3.08906138e-02
2.75377959e-01 4.14303780e-01 -8.80264044e-01 4.97587860e-01
8.32592905e-01 1.07351637e+00 -4.78119910e-01 1.02062440e+00
-1.83900252e-01 2.29058251e-01 -5.43858588e-01 1.05206899e-01
-6.10265695e-03 -2.41431855e-02 8.06075931e-01 8.41668487e-01
3.80893230e-01 -1.22864135e-02 2.45785236e-01 6.99699700e-01
6.80154655e-03 -2.04891697e-01 -8.30962539e-01 2.00227126e-01
2.49691114e-01 1.16540253e+00 -7.65095413e-01 -6.31639734e-02
-1.30242884e-01 1.14339197e+00 2.62158215e-01 2.38833651e-01
-6.94761097e-01 -1.41231447e-01 7.02888310e-01 4.92962182e-01
3.97976637e-01 -6.34037316e-01 -7.46668279e-02 -1.42011440e+00
3.58763009e-01 -4.51409280e-01 -3.81803624e-02 -9.22903657e-01
-1.25396371e+00 2.71891147e-01 3.87555957e-01 -1.67184079e+00
-3.30749154e-01 -8.87012124e-01 -2.58531511e-01 4.92124945e-01
-1.28144228e+00 -1.21649170e+00 -5.49618006e-01 5.06497562e-01
6.31709993e-01 1.31485000e-01 6.29198551e-01 9.74518731e-02
1.28549963e-01 4.80354458e-01 -2.00913146e-01 3.13019268e-02
9.09422815e-01 -1.19189382e+00 4.21401113e-01 4.27445024e-01
2.02513188e-01 3.21073383e-01 7.45605946e-01 -2.40753397e-01
-1.84839118e+00 -1.05343258e+00 5.41139543e-01 -1.07594621e+00
6.12093449e-01 -9.45640981e-01 -7.93878198e-01 7.04833984e-01
-3.67245257e-01 5.38473189e-01 -1.59294885e-02 -2.23694816e-01
-5.43476939e-01 -5.89985922e-02 -1.13714898e+00 4.19886649e-01
1.66673040e+00 -5.91191530e-01 -5.90006769e-01 4.53078419e-01
9.28127050e-01 -8.11947167e-01 -9.45757926e-01 5.47782481e-01
6.18382633e-01 -5.47761858e-01 1.30151165e+00 -4.45276201e-01
4.20586139e-01 -5.38028419e-01 -2.09863022e-01 -1.26585114e+00
6.02356978e-02 -6.42028749e-01 -1.97259933e-01 7.52226651e-01
9.57819670e-02 -5.93522787e-01 8.98503065e-01 1.03851996e-01
-2.14333057e-01 -8.17943811e-01 -1.09223652e+00 -1.01980925e+00
1.70921028e-01 -5.42403579e-01 3.18238050e-01 7.66458511e-01
-3.75531673e-01 3.86624098e-01 -1.02904931e-01 3.39983404e-01
1.10883462e+00 3.65453690e-01 1.29765284e+00 -1.18929648e+00
-2.47824728e-01 -1.90644637e-01 -9.81839538e-01 -1.45569766e+00
3.64548236e-01 -9.45101857e-01 3.06175172e-01 -1.40116501e+00
2.39497945e-02 -5.01148224e-01 2.32339635e-01 3.02373260e-01
2.34373838e-01 5.44577003e-01 2.01811686e-01 2.31405422e-01
-6.48604870e-01 6.79309309e-01 1.59486747e+00 8.92239735e-02
-5.39548520e-04 -1.16333827e-01 -1.28126591e-01 8.60739946e-01
6.21023178e-01 -3.66795957e-01 -3.16469282e-01 -6.20873690e-01
1.59074798e-01 5.38789220e-02 1.02646601e+00 -9.73077774e-01
1.22150727e-01 -1.03592061e-01 2.56826609e-01 -6.46748483e-01
5.59290469e-01 -9.83923793e-01 2.39013508e-01 3.77817482e-01
-1.68127999e-01 -1.35471717e-01 1.56295761e-01 9.07843709e-01
8.16682130e-02 4.85548228e-02 6.52981281e-01 -1.50220916e-01
-7.12580621e-01 6.82646453e-01 2.59492308e-01 2.48662964e-01
1.19485795e+00 -4.46961045e-01 2.11035386e-01 -2.95978189e-01
-7.96238124e-01 1.97247967e-01 9.69498396e-01 6.60677016e-01
6.56737924e-01 -1.31937242e+00 -6.40881658e-01 2.44889632e-01
5.38780332e-01 8.14932883e-01 1.42615154e-01 8.86754334e-01
-5.61457336e-01 1.25216588e-01 -9.22219902e-02 -1.31247580e+00
-9.96011853e-01 4.66626287e-01 3.91681254e-01 1.18322827e-01
-9.85374928e-01 7.47514546e-01 4.51402009e-01 -1.00210571e+00
3.76818448e-01 -3.21350038e-01 3.84690583e-01 -3.21869254e-01
2.36149266e-01 1.00706331e-01 8.20765644e-02 -8.02088678e-01
-4.90244508e-01 1.27098250e+00 1.44954517e-01 -1.15325369e-01
1.59078681e+00 -8.73430744e-02 1.77045971e-01 6.73770726e-01
1.56758690e+00 -1.94298312e-01 -1.88171434e+00 -1.79684952e-01
-2.83734441e-01 -7.43727148e-01 -1.14336289e-01 -4.29576278e-01
-8.44166338e-01 9.39600229e-01 6.55433655e-01 -4.23468500e-01
7.46328890e-01 5.88803530e-01 6.79066956e-01 4.76402640e-01
9.05900240e-01 -8.23014975e-01 5.23403525e-01 4.90465671e-01
1.13363814e+00 -1.48439407e+00 -4.95142117e-02 -5.73107958e-01
-2.69982874e-01 8.96679580e-01 6.26758516e-01 -6.92802608e-01
8.73440802e-01 1.40346244e-01 4.78196517e-02 -1.29826963e-01
-5.83521187e-01 2.68950276e-02 4.43042696e-01 8.91796350e-01
1.07970005e-02 -7.47411996e-02 1.27976179e-01 9.70947444e-02
-1.19531088e-01 -6.38917983e-02 3.60207349e-01 9.82202530e-01
-1.55924380e-01 -9.59473252e-01 -2.87575811e-01 2.50498980e-01
-7.09216222e-02 5.04916728e-01 -1.88052043e-01 8.59997332e-01
9.46619436e-02 2.96201319e-01 1.40786856e-01 -3.27335268e-01
5.50042391e-01 -1.76928177e-01 1.12532949e+00 -9.29321170e-01
-2.81237550e-02 5.19144572e-02 -2.44325280e-01 -8.67311656e-01
-7.51688480e-01 -9.24612641e-01 -1.24045897e+00 1.71825010e-02
-2.87041336e-01 -1.86638504e-01 7.98195183e-01 6.67694449e-01
4.17129159e-01 9.90110561e-02 5.05093813e-01 -1.66027343e+00
-6.95475876e-01 -7.30156124e-01 -3.10566515e-01 6.42825603e-01
5.64956427e-01 -1.16041970e+00 -6.22278810e-01 2.12670445e-01] | [7.74999475479126, -2.627852201461792] |
36401bc5-a926-48b6-8ffa-a3c2db7628fb | resource-allocation-of-federated-learning-for | 2211.08705 | null | https://arxiv.org/abs/2211.08705v2 | https://arxiv.org/pdf/2211.08705v2.pdf | Resource Allocation of Federated Learning for the Metaverse with Mobile Augmented Reality | The Metaverse has received much attention recently. Metaverse applications via mobile augmented reality (MAR) require rapid and accurate object detection to mix digital data with the real world. Federated learning (FL) is an intriguing distributed machine learning approach due to its privacy-preserving characteristics. Due to privacy concerns and the limited computation resources on mobile devices, we incorporate FL into MAR systems of the Metaverse to train a model cooperatively. Besides, to balance the trade-off between energy, execution latency and model accuracy, thereby accommodating different demands and application scenarios, we formulate an optimization problem to minimize a weighted combination of total energy consumption, completion time and model accuracy. Through decomposing the non-convex optimization problem into two subproblems, we devise a resource allocation algorithm to determine the bandwidth allocation, transmission power, CPU frequency and video frame resolution for each participating device. We further present the convergence analysis and computational complexity of the proposed algorithm. Numerical results show that our proposed algorithm has better performance (in terms of energy consumption, completion time and model accuracy) under different weight parameters compared to existing benchmarks. | ['Jun Zhao', 'Chang Liu', 'Xinyu Zhou'] | 2022-11-16 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [-4.35475539e-03 -3.71052563e-01 -5.46069086e-01 -1.90017983e-01
-8.89658928e-01 -5.39838850e-01 -2.51154561e-04 -8.64232779e-02
-4.11475688e-01 7.17244983e-01 -2.45295390e-01 -3.90923887e-01
-2.62865037e-01 -6.69440031e-01 -6.79650366e-01 -8.28683555e-01
-1.24603540e-01 -2.72604488e-02 -6.64542019e-02 4.78298038e-01
-8.02489072e-02 3.35194111e-01 -1.42742121e+00 -9.69155058e-02
1.07112241e+00 1.59017396e+00 1.53947920e-01 4.61499661e-01
1.58163086e-01 7.19302058e-01 -2.50790328e-01 -8.26945007e-01
4.10418659e-01 -1.64982751e-01 -4.04185027e-01 1.12526424e-01
9.28467289e-02 -5.18313289e-01 -5.06272316e-01 1.28387368e+00
6.36023998e-01 1.70603707e-01 1.99008942e-01 -1.91847777e+00
-2.67673731e-01 3.14297050e-01 -8.11434209e-01 1.26302853e-01
-5.52501678e-02 -2.36696228e-01 5.90054870e-01 -5.55990875e-01
3.44772786e-01 6.61203623e-01 6.08996212e-01 4.44569230e-01
-8.95321310e-01 -7.96207964e-01 1.12558320e-01 4.87862915e-01
-1.45401287e+00 -5.24299502e-01 6.85634255e-01 1.67585202e-02
4.76347148e-01 9.70855296e-01 6.12607658e-01 5.88451147e-01
2.90023327e-01 8.58890891e-01 8.94623637e-01 -2.88363427e-01
6.52355790e-01 3.94489795e-01 -6.70862421e-02 7.32991397e-01
5.71479201e-01 -1.79380357e-01 -6.71671391e-01 -5.89731991e-01
4.56967443e-01 4.23863471e-01 -4.84790027e-01 -6.87311769e-01
-1.09746373e+00 3.66043776e-01 2.38172889e-01 -3.27853382e-01
-4.93904829e-01 3.50462139e-01 4.46089566e-01 5.23817092e-02
3.65333647e-01 -3.45472455e-01 -6.19030058e-01 -2.23807320e-01
-6.93900943e-01 -1.66401863e-01 7.51908123e-01 1.05772209e+00
4.54482585e-01 -2.55944043e-01 6.02526478e-02 6.27339661e-01
5.11924803e-01 5.74453771e-01 3.33617806e-01 -9.45599198e-01
7.89134026e-01 6.02121055e-01 3.51696074e-01 -1.28821838e+00
-1.58705749e-02 -4.07528698e-01 -9.89214778e-01 -2.87851006e-01
3.56190577e-02 -4.58886236e-01 -7.07576349e-02 1.55349791e+00
8.42847943e-01 6.84429169e-01 -1.18128195e-01 1.17307198e+00
2.52501637e-01 8.02891910e-01 1.49185702e-01 -9.30318773e-01
1.31926537e+00 -1.02076769e+00 -9.41415668e-01 8.35966393e-02
6.61753118e-01 -6.55858815e-01 6.60038352e-01 2.62527943e-01
-1.06501174e+00 9.88977551e-02 -1.09782982e+00 1.01883225e-01
-1.82742774e-01 4.20075417e-01 8.41925740e-01 9.40219045e-01
-5.93208849e-01 1.38287902e-01 -9.56947029e-01 -1.92844078e-01
5.98399520e-01 7.50484943e-01 -1.05081052e-01 -1.37197807e-01
-5.94308734e-01 2.95549482e-01 4.73189279e-02 1.54118985e-01
-5.18064737e-01 -7.64737904e-01 -3.20621520e-01 7.64323547e-02
6.71033442e-01 -9.63692427e-01 1.14308083e+00 -6.50887251e-01
-1.51240325e+00 5.91536999e-01 1.69850495e-02 -4.49586838e-01
7.43175507e-01 -6.78795055e-02 -4.05590862e-01 -2.89287895e-01
-3.43834162e-01 9.32714436e-03 5.46066105e-01 -1.04433322e+00
-1.08086276e+00 -7.19142497e-01 4.33521532e-02 3.71626079e-01
-9.50524807e-01 -6.97443113e-02 -7.78019905e-01 -4.53941375e-01
1.18663451e-02 -1.14112604e+00 -2.23517165e-01 2.76960731e-01
-1.39148027e-01 1.22424953e-01 1.21762097e+00 -5.50482810e-01
1.41490757e+00 -2.12116790e+00 -1.10458657e-01 8.65198970e-02
2.94486344e-01 1.78844199e-01 1.15147471e-01 4.11764383e-02
6.25948787e-01 2.42471928e-03 7.15855956e-02 -3.10695142e-01
5.71325645e-02 1.45984404e-02 8.60307831e-03 7.24567413e-01
-5.98485768e-01 7.38244951e-01 -6.65429592e-01 -6.13925338e-01
1.76077068e-01 5.44012129e-01 -5.45593977e-01 2.63614535e-01
-7.97299948e-03 -2.06409376e-02 -7.89585590e-01 9.02138531e-01
1.05378270e+00 -3.56781751e-01 5.30008435e-01 -3.70306879e-01
1.40826955e-01 -2.19846413e-01 -1.43147004e+00 1.52418184e+00
-8.57702196e-01 3.16839874e-01 7.60485888e-01 -7.27720141e-01
6.61122322e-01 2.53452688e-01 8.07793379e-01 -6.06545448e-01
4.41036731e-01 4.43840027e-01 -5.89210331e-01 -6.41332269e-01
4.79829103e-01 3.58742684e-01 8.81863758e-02 5.90403676e-01
-4.27974284e-01 6.21524215e-01 -4.90740120e-01 -6.78973123e-02
1.13944745e+00 -2.46567607e-01 3.23946297e-01 -7.22397044e-02
3.87199908e-01 -1.98879987e-01 7.70416141e-01 2.37841889e-01
-4.43455696e-01 -9.17961299e-02 -2.58984696e-02 -5.66797674e-01
-6.14935875e-01 -5.89256108e-01 2.09767237e-01 1.01583827e+00
7.68785298e-01 -2.35616148e-01 -7.22497463e-01 -6.80891514e-01
1.42932370e-01 3.44817877e-01 -1.58635408e-01 -1.33383274e-01
-4.35358614e-01 -9.69408572e-01 -6.70053298e-03 1.57931224e-01
7.77911305e-01 -3.92974585e-01 -1.10035026e+00 -7.78092295e-02
-3.02908510e-01 -1.11861920e+00 -8.13727736e-01 -1.77601930e-02
-8.88712168e-01 -1.00769949e+00 -5.52771807e-01 -6.68722153e-01
7.42373228e-01 6.93985462e-01 7.66568959e-01 9.30418447e-02
-2.23475039e-01 5.13361394e-01 -6.15415946e-02 -3.33217382e-01
2.35403106e-01 1.44145459e-01 6.01863861e-02 4.07194525e-01
1.15777642e-01 -4.35553879e-01 -7.44659543e-01 5.46679437e-01
-8.91115606e-01 7.80730769e-02 3.49556565e-01 5.67836761e-01
9.76511359e-01 3.97002548e-01 4.10489321e-01 -7.06269979e-01
5.04672766e-01 -7.78281629e-01 -1.00775456e+00 5.44842601e-01
-6.08392656e-01 -5.13975501e-01 3.65302801e-01 -6.11920059e-01
-1.00448263e+00 3.52647752e-01 5.07698774e-01 -6.41641617e-01
5.93581021e-01 2.88990319e-01 -7.93820500e-01 -4.76478159e-01
4.40880023e-02 2.43726805e-01 -1.35156447e-02 -3.08101952e-01
2.18214810e-01 1.04577231e+00 4.88438576e-01 -4.12069887e-01
5.01489222e-01 4.12957549e-01 7.79667054e-04 -3.29056352e-01
-2.78468400e-01 -3.08490664e-01 9.69255641e-02 -3.08012217e-01
3.64284098e-01 -1.08697391e+00 -1.44463885e+00 2.15667069e-01
-1.02155709e+00 2.53360361e-01 -3.92226838e-02 6.28116965e-01
-5.30932128e-01 1.29408315e-01 -1.39378861e-01 -1.31297803e+00
-8.36257219e-01 -1.00583768e+00 8.82461607e-01 3.46961975e-01
4.47283357e-01 -6.29356265e-01 -2.69911528e-01 6.72969639e-01
4.39790279e-01 4.65603560e-01 5.87527752e-01 -2.28917673e-01
-1.03838682e+00 -4.56247568e-01 -3.01061094e-01 -1.23468749e-01
1.50823712e-01 -4.27069068e-01 -6.40296459e-01 -4.92983550e-01
2.00697467e-01 3.21224891e-02 2.50359327e-01 3.01900417e-01
1.66148961e+00 -1.03876793e+00 -6.37199879e-01 8.45214128e-01
1.76986134e+00 2.78035313e-01 2.89626718e-01 3.78575951e-01
6.54951632e-01 2.37267390e-01 8.30886781e-01 7.96070933e-01
5.58298290e-01 8.15364480e-01 8.76716554e-01 1.74661279e-01
3.79436314e-01 -1.44963041e-01 2.92877495e-01 8.17479074e-01
2.29128331e-01 -5.23558855e-01 -6.77777588e-01 5.12494504e-01
-2.27995014e+00 -6.28216803e-01 4.31095138e-02 2.62972927e+00
3.47899348e-01 -3.66361260e-01 2.77217627e-01 1.61403432e-01
9.26846743e-01 -1.15946747e-01 -9.60727751e-01 -1.67569265e-01
1.24973387e-01 -5.67588747e-01 9.53566194e-01 -1.49897113e-01
-9.81424451e-01 2.72360206e-01 5.58508587e+00 9.97513235e-01
-1.23928654e+00 3.97718668e-01 9.34513092e-01 -6.61322773e-01
-2.31196150e-01 -3.83073837e-01 -3.77633005e-01 7.97727048e-01
1.03934956e+00 -5.99071145e-01 7.83764124e-01 1.15798068e+00
2.66851395e-01 1.16696596e-01 -9.97294545e-01 1.58047974e+00
-1.81338385e-01 -1.59422112e+00 -2.09642440e-01 4.04728830e-01
7.81445444e-01 3.90575305e-02 1.17074244e-01 -6.37337267e-02
-5.39741069e-02 -5.68689883e-01 6.16382003e-01 2.78704464e-01
7.32640564e-01 -9.83249187e-01 7.64964461e-01 5.17754376e-01
-1.30106878e+00 -4.95044798e-01 -2.36349151e-01 2.03392446e-01
2.12363200e-03 6.14700019e-01 -2.06039771e-01 5.45399249e-01
7.91730165e-01 2.24810019e-01 -5.99737801e-02 1.25004625e+00
6.61114335e-01 1.88827544e-01 -5.20869374e-01 -2.20682934e-01
-3.82531345e-01 -2.50860721e-01 4.57531214e-01 6.09266579e-01
5.44800818e-01 4.32856083e-01 1.12215757e-01 3.28960776e-01
-7.17449605e-01 4.02113497e-01 -2.17436314e-01 6.85357079e-02
1.05139446e+00 1.46876276e+00 -4.95902389e-01 1.63467862e-02
-4.25144821e-01 8.76897395e-01 2.76678093e-02 1.15978517e-01
-1.04224193e+00 -1.21006541e-01 9.93038595e-01 5.61193414e-02
1.58993840e-01 -5.66059127e-02 -3.90326887e-01 -1.18994915e+00
5.14199257e-01 -7.29996324e-01 4.35894489e-01 -3.96070033e-01
-9.56196785e-01 5.52486777e-01 -4.27703917e-01 -1.42464507e+00
2.83619434e-01 -1.39803544e-01 -3.11224669e-01 4.44820434e-01
-1.47136211e+00 -1.12479985e+00 -4.84782845e-01 7.76787043e-01
3.56746942e-01 -2.04033896e-01 7.12172329e-01 7.31577873e-01
-1.03953063e+00 9.91838396e-01 5.11450469e-01 -4.68373567e-01
1.48937181e-01 -5.47074854e-01 -1.49746194e-01 7.36731887e-01
-1.57631263e-01 4.27567750e-01 4.32696223e-01 -3.94855618e-01
-2.32638168e+00 -1.35265362e+00 5.84462285e-01 1.65460713e-03
4.00202751e-01 -3.70725453e-01 -4.19890732e-01 3.95887911e-01
-2.19657779e-01 8.30508053e-01 8.10108840e-01 -1.83938518e-01
-4.05398533e-02 -5.97308874e-01 -1.59347439e+00 4.66471404e-01
1.04184377e+00 -2.93956250e-01 5.08330047e-01 3.46030563e-01
7.88878500e-01 -6.17225766e-01 -9.49153423e-01 2.31848404e-01
6.20500803e-01 -6.11696541e-01 6.65214241e-01 -2.42202789e-01
-1.08728312e-01 -3.92030388e-01 -4.84153390e-01 -6.27829134e-01
1.26234293e-01 -1.07845509e+00 -9.47625875e-01 1.47188461e+00
9.83368680e-02 -5.85175753e-01 1.36181462e+00 1.22053957e+00
4.57592636e-01 -1.18350542e+00 -1.35713708e+00 -6.95248485e-01
-7.50956953e-01 -3.59044164e-01 1.03658152e+00 9.07081902e-01
-3.26861776e-02 3.71774435e-02 -6.82965577e-01 3.51188570e-01
7.51567781e-01 1.76601842e-01 7.91086078e-01 -1.04461825e+00
-1.83344051e-01 2.18068361e-01 -3.66618842e-01 -8.31379354e-01
-1.28948212e-01 -5.88178515e-01 -3.19713831e-01 -1.09574401e+00
4.06678528e-01 -9.30918276e-01 -5.64597189e-01 3.94052356e-01
3.64030004e-02 7.67334104e-02 4.17561859e-01 5.64654112e-01
-1.03584242e+00 6.37497067e-01 8.21649134e-01 -7.89830685e-02
-3.28327686e-01 2.36682460e-01 -6.48158431e-01 4.27270681e-01
6.27658427e-01 -5.15624225e-01 -6.36633992e-01 -8.05395782e-01
8.49822387e-02 4.81727153e-01 2.05912739e-01 -8.84398043e-01
6.14987969e-01 -3.75097573e-01 8.99610221e-02 -4.27689105e-01
3.98518294e-01 -1.67484856e+00 7.25253880e-01 4.31242228e-01
5.46991965e-03 1.30023122e-01 1.93261188e-02 9.45671797e-01
9.24461931e-02 3.07373285e-01 4.68775570e-01 3.73986155e-01
-4.58873212e-01 6.57177389e-01 -1.79497302e-02 -5.95570624e-01
1.72064972e+00 -2.45181024e-01 -4.38824296e-01 -1.75338611e-01
-2.16490492e-01 5.07129490e-01 5.58519244e-01 4.46154207e-01
3.87856752e-01 -1.52798951e+00 -1.98304608e-01 -9.56633091e-02
-9.91539210e-02 -1.89920247e-01 5.12530208e-01 8.51663470e-01
-4.44327444e-01 4.13723916e-01 -9.10984054e-02 -3.19578975e-01
-1.54390013e+00 6.18960857e-01 3.29019457e-01 -1.80203676e-01
-2.30445981e-01 6.87131643e-01 -3.06272835e-01 -2.59264916e-01
6.40496969e-01 1.52806208e-01 3.35582197e-01 -2.24524900e-01
5.60651064e-01 1.02314436e+00 2.92386502e-01 -4.98322487e-01
-6.13175869e-01 1.56920180e-01 1.29435986e-01 3.79292607e-01
1.30344701e+00 -6.68851554e-01 -8.54647681e-02 -2.78519779e-01
1.47825825e+00 -7.28657767e-02 -1.17738390e+00 -2.06055850e-01
-2.63150811e-01 -8.61913800e-01 6.90791607e-01 -6.58153772e-01
-1.63585842e+00 1.68565020e-01 1.17070341e+00 3.00652836e-03
1.60734785e+00 -5.23075879e-01 1.21253109e+00 -2.85588135e-03
8.67474854e-01 -1.08637607e+00 -3.41057271e-01 -3.33246529e-01
3.50345969e-01 -1.13737965e+00 1.30109057e-01 -7.43466675e-01
-3.26363444e-01 6.73758507e-01 6.70224309e-01 3.12166661e-01
7.17668355e-01 3.30417395e-01 -2.04349890e-01 2.45302945e-01
-8.68175268e-01 6.65572941e-01 -1.02370217e-01 4.83446687e-01
-1.20378107e-01 2.36513272e-01 -5.70235610e-01 1.08068049e+00
1.24534816e-01 4.51228917e-01 2.77505457e-01 1.00892580e+00
-1.66812651e-02 -9.80403244e-01 -5.32197952e-01 6.34291172e-01
-6.00773513e-01 4.29118991e-01 6.22453652e-02 2.54711866e-01
2.90975511e-01 1.07869244e+00 4.12371419e-02 -5.96822619e-01
9.99327078e-02 -4.70642656e-01 1.19156532e-01 1.21184690e-02
-3.93219054e-01 -3.01253349e-02 -1.15160793e-01 -6.75905645e-01
-3.21588576e-01 -3.18009824e-01 -9.35371220e-01 -7.64406621e-01
-7.44537055e-01 2.66379684e-01 1.22494173e+00 7.24305153e-01
8.49782705e-01 3.38476598e-01 9.96278226e-01 -6.39678955e-01
-5.21814048e-01 -2.50438392e-01 -6.14392698e-01 -2.49196589e-01
2.89288551e-01 -3.15602720e-01 -8.24939683e-02 -1.68319821e-01] | [5.908616065979004, 5.931375026702881] |
066fd5b8-3228-4d16-a4fe-f7a16cae739d | unleashing-the-potential-of-unsupervised-deep | 2305.16777 | null | https://arxiv.org/abs/2305.16777v1 | https://arxiv.org/pdf/2305.16777v1.pdf | Unleashing the Potential of Unsupervised Deep Outlier Detection through Automated Training Stopping | Outlier detection (OD) has received continuous research interests due to its wide applications. With the development of deep learning, increasingly deep OD algorithms are proposed. Despite the availability of numerous deep OD models, existing research has reported that the performance of deep models is extremely sensitive to the configuration of hyperparameters (HPs). However, the selection of HPs for deep OD models remains a notoriously difficult task due to the lack of any labels and long list of HPs. In our study. we shed light on an essential factor, training time, that can introduce significant variation in the performance of deep model. Even the performance is stable across other HPs, training time itself can cause a serious HP sensitivity issue. Motivated by this finding, we are dedicated to formulating a strategy to terminate model training at the optimal iteration. Specifically, we propose a novel metric called loss entropy to internally evaluate the model performance during training while an automated training stopping algorithm is devised. To our knowledge, our approach is the first to enable reliable identification of the optimal training iteration during training without requiring any labels. Our experiments on tabular, image datasets show that our approach can be applied to diverse deep models and datasets. It not only enhances the robustness of deep models to their HPs, but also improves the performance and reduces plenty of training time compared to naive training. | ['Xuemin Lin', 'Liping Wang', 'Yuang Zhang', 'Yihong Huang'] | 2023-05-26 | null | null | null | null | ['outlier-detection'] | ['methodology'] | [-2.64773071e-01 -3.53596389e-01 -1.27257824e-01 -3.04364502e-01
-4.06421095e-01 -3.00195813e-01 4.44028854e-01 3.70997250e-01
-5.12773991e-01 5.27199745e-01 -2.54336178e-01 -3.10523391e-01
-3.01530719e-01 -6.03902936e-01 -6.18522942e-01 -7.50849426e-01
-1.15490995e-01 4.11051154e-01 4.37341452e-01 2.13054165e-01
4.35812354e-01 6.78833604e-01 -1.73997307e+00 -1.84401751e-01
1.16553736e+00 1.15156460e+00 -5.22727743e-02 2.24449366e-01
-4.45335731e-02 6.80562794e-01 -8.10048819e-01 -4.49638903e-01
4.02625412e-01 -3.57630402e-01 -5.33574462e-01 -1.04570381e-01
4.06958640e-01 -3.64539862e-01 -1.16471998e-01 1.11689162e+00
8.53057802e-01 9.73622352e-02 5.69988787e-01 -1.43401396e+00
-2.95775473e-01 3.90374780e-01 -2.82499641e-01 3.82050574e-01
-2.15423051e-02 3.01954001e-01 9.94229615e-01 -8.67894590e-01
3.55716467e-01 7.92776823e-01 7.95108736e-01 2.47882515e-01
-1.20568180e+00 -5.45014977e-01 6.77811503e-02 3.93920392e-01
-1.44480884e+00 -2.14990988e-01 7.47898877e-01 -5.59759319e-01
7.60388076e-01 -2.57670339e-02 5.51683366e-01 1.16238320e+00
4.04259920e-01 5.43474019e-01 9.85367239e-01 -4.81776148e-01
6.22364521e-01 8.23114812e-02 3.01967800e-01 7.52715409e-01
6.46168232e-01 1.09800614e-01 -4.72720295e-01 -2.20640913e-01
4.52211618e-01 -1.17941357e-01 -5.94612844e-02 -5.89670479e-01
-6.44589841e-01 8.64404559e-01 2.67933995e-01 3.02344441e-01
-1.81141660e-01 2.04618257e-02 6.35516703e-01 4.24284995e-01
2.61951923e-01 7.50513673e-01 -4.38637942e-01 -2.80653834e-01
-9.67121303e-01 3.29249889e-01 8.75144780e-01 4.19600844e-01
6.45431101e-01 -2.56651901e-02 -1.43467486e-01 9.30158377e-01
6.37226775e-02 -3.57825309e-02 6.76757336e-01 -7.68904746e-01
2.60979950e-01 8.60510588e-01 3.62215303e-02 -1.22330570e+00
-6.01936877e-01 -7.89070010e-01 -7.30843246e-01 1.95334688e-01
5.07881463e-01 -1.49397522e-01 -9.67855275e-01 1.54680073e+00
2.30254844e-01 8.69051591e-02 -1.69981241e-01 7.35234261e-01
5.18636823e-01 1.96991831e-01 9.19788517e-03 -1.59924161e-02
9.57247019e-01 -6.21940851e-01 -5.98438978e-01 -2.33075544e-01
8.54297340e-01 -5.84274888e-01 1.22966719e+00 6.84288263e-01
-7.42027462e-01 -3.47784013e-01 -1.26426375e+00 1.89201936e-01
-3.09964567e-01 6.20011911e-02 5.84493697e-01 7.83483803e-01
-8.45188379e-01 7.77555585e-01 -9.39384282e-01 -3.90892714e-01
3.74150366e-01 6.13199234e-01 -1.25976965e-01 -1.77064925e-01
-1.15697610e+00 9.96533513e-01 6.98952734e-01 2.89126903e-01
-8.59194517e-01 -4.65705156e-01 -5.57956994e-01 -2.03270353e-02
4.01445508e-01 -6.36980057e-01 1.16290879e+00 -8.42305422e-01
-1.38361108e+00 6.95620477e-01 1.51509628e-01 -7.27417171e-01
8.31505060e-01 -2.47076288e-01 -4.27700907e-01 -1.65436834e-01
4.18814830e-02 2.32317805e-01 9.07795310e-01 -1.07940030e+00
-4.35619205e-01 -4.10320073e-01 1.24869158e-03 -2.79787146e-02
-5.01588941e-01 -1.15718268e-01 -5.55109620e-01 -4.48210448e-01
3.42357188e-01 -1.07204449e+00 -1.94300801e-01 -3.17725837e-01
-4.13011521e-01 -2.67072618e-01 4.86937463e-01 -1.82423905e-01
1.61130488e+00 -2.29283738e+00 -1.55860677e-01 2.83488452e-01
1.51247099e-01 4.44211841e-01 2.07855612e-01 4.22571927e-01
8.19120109e-02 5.24083935e-02 -2.64829189e-01 -3.75400722e-01
1.91175684e-01 3.69219542e-01 -2.17365101e-01 5.91529131e-01
1.29618973e-01 5.86432278e-01 -6.71934545e-01 -4.36161071e-01
2.51073360e-01 2.66082495e-01 -6.89989924e-01 2.92569757e-01
-1.82318669e-02 2.72915721e-01 -2.55467743e-01 6.07122540e-01
5.72884917e-01 -3.24796677e-01 -5.23220226e-02 1.09806553e-01
-8.61884281e-02 3.23514432e-01 -1.40526426e+00 1.19453406e+00
-2.96591043e-01 5.64515889e-01 -5.71128070e-01 -1.05645955e+00
1.12943184e+00 2.77559403e-02 5.08004427e-01 -8.61696720e-01
1.48042336e-01 6.17976844e-01 2.87733525e-01 -5.46907544e-01
4.89381671e-01 4.11736891e-02 -2.05946863e-02 2.98430622e-01
-1.23216175e-01 2.76846588e-01 3.73603612e-01 -1.75967559e-01
1.20302284e+00 -1.72655135e-01 3.08421522e-01 -2.42786139e-01
3.70459586e-01 -1.56527475e-01 6.60689354e-01 9.40079570e-01
-4.71453428e-01 7.26821601e-01 8.37961376e-01 -6.61907554e-01
-9.27506208e-01 -1.00961363e+00 -5.69642186e-01 9.69912231e-01
1.20437987e-01 -2.05338478e-01 -6.46613359e-01 -6.46699369e-01
1.96903884e-01 7.08251178e-01 -6.27581239e-01 -5.07310152e-01
-6.10438108e-01 -1.15511644e+00 6.99150562e-01 3.34667116e-01
5.36840558e-01 -1.20027220e+00 -7.28038549e-01 2.53010899e-01
7.50222802e-02 -1.17293406e+00 -5.97789325e-02 5.35773873e-01
-1.03653228e+00 -1.10062504e+00 -5.21244884e-01 -5.01514673e-01
7.11278856e-01 -1.52563080e-01 1.18683290e+00 2.36615777e-01
-8.71362686e-02 -5.37253432e-02 -2.95088023e-01 -4.69120562e-01
-4.46652323e-01 4.42364186e-01 2.06952333e-01 -8.79455507e-02
6.21494055e-01 -4.10025448e-01 -7.19939172e-01 4.31671381e-01
-9.65086997e-01 -4.22546297e-01 6.35632277e-01 6.45108402e-01
4.13131982e-01 3.31518650e-01 5.99274993e-01 -8.89041126e-01
7.88327157e-01 -6.07326984e-01 -6.56740546e-01 -2.08404120e-02
-9.27011549e-01 3.14721972e-01 7.87513435e-01 -1.97663680e-01
-6.32133961e-01 -2.38884613e-01 -1.67950183e-01 -3.74048650e-01
-2.56850123e-01 5.42694867e-01 -7.07860291e-02 7.33575737e-03
8.43146384e-01 -9.30171534e-02 -2.05489889e-01 -5.19019842e-01
-1.81125239e-01 4.15475368e-01 4.21878815e-01 -5.31805634e-01
5.51310122e-01 2.74001837e-01 1.92399286e-02 -7.20363021e-01
-8.95372689e-01 -4.17254865e-01 -5.83348453e-01 -8.21761563e-02
4.53483522e-01 -6.84867680e-01 -4.97245252e-01 6.42924666e-01
-8.91076922e-01 -1.20095707e-01 -1.57713629e-02 3.04355562e-01
-2.99111694e-01 5.14019310e-01 -3.81092519e-01 -5.38532078e-01
-2.25186437e-01 -1.31940746e+00 5.46312869e-01 -5.99432662e-02
-2.55584896e-01 -1.05026388e+00 1.44305214e-01 5.19048274e-02
1.69756547e-01 4.15743709e-01 7.93540418e-01 -7.61017978e-01
-6.12055540e-01 -3.26639265e-01 -1.81651920e-01 5.54349542e-01
-4.73397560e-02 7.81553909e-02 -1.04742897e+00 -3.88294935e-01
1.09408341e-01 -1.22144483e-01 9.19467032e-01 4.17775303e-01
1.32940102e+00 -2.45057434e-01 -1.03755116e-01 7.64076114e-01
1.44348836e+00 7.90080205e-02 7.16306090e-01 1.07462931e+00
6.00188732e-01 3.10794204e-01 6.45387411e-01 5.96163690e-01
7.28173777e-02 7.78297603e-01 7.71997154e-01 -2.46231146e-02
3.13582361e-01 -1.30702764e-01 4.55815315e-01 5.56584120e-01
2.13811412e-01 -1.78789005e-01 -1.15161669e+00 4.93458271e-01
-1.70258021e+00 -6.21188402e-01 -1.00897618e-01 2.48892212e+00
5.87314010e-01 4.89778399e-01 2.34880671e-01 4.49709356e-01
4.63340729e-01 1.63027328e-02 -7.00267375e-01 -5.62366724e-01
7.08124116e-02 -4.49801348e-02 5.00619829e-01 1.44581020e-01
-1.03875542e+00 7.31192112e-01 5.86406708e+00 6.02888703e-01
-1.18502283e+00 -2.09298551e-01 4.37042475e-01 -9.28989574e-02
9.47140157e-02 -2.51010597e-01 -9.58635390e-01 8.25616062e-01
9.74114180e-01 -3.11634820e-02 9.36339200e-02 9.41785634e-01
3.46614420e-01 -1.31834045e-01 -1.14997327e+00 9.20134008e-01
7.80887008e-02 -8.34246397e-01 1.39494259e-02 1.95422068e-01
7.66238689e-01 2.15947270e-01 1.70929283e-01 2.49786317e-01
9.87909213e-02 -9.74127710e-01 6.37253284e-01 1.52181517e-02
3.72039348e-01 -9.90942657e-01 1.07290745e+00 3.64355415e-01
-6.69838130e-01 -4.56882983e-01 -4.44794327e-01 -8.66777003e-02
-2.04227403e-01 1.05867088e+00 -9.42972064e-01 3.98645520e-01
8.06881726e-01 3.99099648e-01 -7.50836134e-01 1.52206790e+00
-2.37513587e-01 5.29568255e-01 -3.62572044e-01 7.80104771e-02
3.02927613e-01 1.74309127e-02 3.96451563e-01 1.09452879e+00
4.30276632e-01 -4.20043886e-01 3.65885571e-02 5.51695168e-01
-6.01409525e-02 2.59278744e-01 -7.91321337e-01 1.33092657e-01
5.21254420e-01 7.97492385e-01 -7.69153059e-01 -8.03179853e-03
-2.80586779e-01 8.09993446e-01 5.45159936e-01 2.24743366e-01
-9.11342263e-01 -1.48725659e-01 7.08388925e-01 2.63624460e-01
3.83280814e-01 -3.34032387e-01 -5.79345465e-01 -9.37525034e-01
2.62265563e-01 -9.13992703e-01 5.26547253e-01 -3.12377256e-03
-1.24056232e+00 5.84555805e-01 -1.62279204e-01 -1.39647388e+00
-1.59560382e-01 -5.70419133e-01 -5.60041606e-01 3.38221580e-01
-1.51634359e+00 -3.85338575e-01 -3.78227055e-01 3.58767629e-01
1.66917011e-01 -1.03557281e-01 5.44496417e-01 6.03179932e-01
-1.09647381e+00 8.40629816e-01 3.36940110e-01 3.82588878e-02
7.43105352e-01 -1.23432362e+00 2.23209679e-01 1.03788626e+00
1.89966530e-01 6.99217796e-01 1.02688789e+00 -3.35280985e-01
-8.60045016e-01 -9.95749116e-01 8.83940756e-01 -4.68143165e-01
5.53420722e-01 -5.38018122e-02 -1.14700639e+00 4.69571412e-01
-2.80505121e-01 -8.39414224e-02 5.98026276e-01 3.01854253e-01
-1.88157499e-01 -2.27698028e-01 -9.54752803e-01 6.05041981e-01
7.97842860e-01 -2.84505188e-01 -4.88947839e-01 3.17680389e-02
3.33356202e-01 -6.39925778e-01 -7.35148668e-01 5.93533874e-01
3.98862511e-01 -1.58357656e+00 8.09228778e-01 -1.70922950e-01
2.83450752e-01 -2.85572231e-01 1.10885523e-01 -1.12206829e+00
-2.74261422e-02 -2.79061794e-01 -2.61774033e-01 1.02901411e+00
2.82174647e-01 -8.13469112e-01 7.65298128e-01 8.42100263e-01
-3.10392231e-01 -1.08949387e+00 -9.16613162e-01 -9.61126089e-01
-1.19410627e-01 -6.54323757e-01 6.19440913e-01 9.10015762e-01
-4.04167861e-01 -1.91547461e-02 -2.27315322e-01 4.50618297e-01
4.95684266e-01 -1.62809249e-02 8.54232907e-01 -1.58494496e+00
-1.42969355e-01 -4.84524667e-01 -5.84079087e-01 -8.01372468e-01
4.67290059e-02 -7.10846961e-01 -3.95933576e-02 -1.33281171e+00
-6.76319972e-02 -6.40198171e-01 -7.13671565e-01 4.91606444e-01
-1.56607017e-01 1.72581062e-01 8.59916359e-02 2.76565552e-01
-6.45247459e-01 5.49907744e-01 9.19727564e-01 2.75985241e-01
-4.92153853e-01 9.12828445e-02 -4.89774376e-01 7.94409275e-01
8.79405677e-01 -8.28209400e-01 -2.81304270e-01 -3.53267133e-01
4.42159384e-01 -5.46063006e-01 4.00622755e-01 -1.43949950e+00
1.73135638e-01 6.42050579e-02 2.29775175e-01 -4.81640160e-01
-7.26437122e-02 -8.26535165e-01 1.72684193e-02 5.79857886e-01
-1.51232660e-01 2.02048108e-01 2.91392297e-01 4.73394096e-01
-2.68025905e-01 -4.90445435e-01 8.65035713e-01 1.14609823e-01
-8.90456319e-01 3.13178211e-01 -3.72272655e-02 1.57243922e-01
1.00005186e+00 -4.63883460e-01 -6.91039711e-02 1.20321505e-01
-4.81455088e-01 2.24087536e-01 7.60203063e-01 5.02734303e-01
2.52715379e-01 -1.09282374e+00 -2.33122632e-01 4.12973851e-01
1.90484658e-01 1.64912701e-01 2.76865084e-02 8.63391936e-01
-6.47311509e-01 4.03820157e-01 -1.84399888e-01 -7.69481301e-01
-1.02469385e+00 3.27000886e-01 4.44928855e-01 -3.94077122e-01
-6.59863591e-01 6.06522083e-01 -2.31341228e-01 -2.14944273e-01
5.77157259e-01 -3.51541013e-01 -1.03944883e-01 1.95554852e-01
1.74065262e-01 5.84392786e-01 5.40312171e-01 -3.10870647e-01
-3.26212108e-01 3.08555245e-01 -2.69773066e-01 3.72560203e-01
1.10828054e+00 -3.90062146e-02 1.69689041e-02 5.59150696e-01
1.15577757e+00 -3.87638867e-01 -1.21217811e+00 2.71536503e-02
3.50708634e-01 -5.81348479e-01 9.95616168e-02 -5.81863344e-01
-9.71891522e-01 8.61721039e-01 7.49366760e-01 3.94858122e-01
1.07464468e+00 -3.57495487e-01 8.36358905e-01 3.75994742e-01
3.23669195e-01 -1.24598622e+00 3.15045297e-01 6.93398416e-01
4.67750400e-01 -1.48311722e+00 2.80737747e-02 1.26188368e-01
-4.92486417e-01 9.44547236e-01 7.35823095e-01 -8.88959542e-02
3.95258486e-01 6.97540864e-02 1.71427384e-01 -2.82111287e-01
-5.71789861e-01 -3.09210494e-02 1.53401047e-01 2.27587730e-01
4.18176383e-01 -2.04110026e-01 -4.20473605e-01 2.54588634e-01
-2.62568831e-01 1.80500194e-01 4.73042011e-01 6.44343019e-01
-5.00232995e-01 -1.14173865e+00 -1.75059259e-01 5.73727667e-01
-7.28338361e-01 1.50055764e-02 -1.32880077e-01 9.00224268e-01
2.17005000e-01 6.24516070e-01 2.97958814e-02 -2.13194489e-01
5.75219691e-01 1.70317367e-01 2.30059311e-01 -5.75775445e-01
-5.31327128e-01 -4.93883848e-01 -1.85236663e-01 -4.88534331e-01
-8.24533701e-02 -6.52259886e-01 -1.29943442e+00 -4.62074637e-01
-4.12424922e-01 1.09561659e-01 5.54486454e-01 1.01586759e+00
4.10476983e-01 2.81173915e-01 6.52290225e-01 -5.24772882e-01
-7.68926442e-01 -6.89325094e-01 -5.83795428e-01 4.74963069e-01
4.97527629e-01 -1.05389512e+00 -7.85387814e-01 -4.73001152e-01] | [7.7150092124938965, 2.6090035438537598] |
5e07f78f-c193-4bb3-a53d-16c189896409 | uniqueness-of-iris-pattern-based-on-ar-model | 2306.12572 | null | https://arxiv.org/abs/2306.12572v1 | https://arxiv.org/pdf/2306.12572v1.pdf | Uniqueness of Iris Pattern Based on AR Model | The assessment of iris uniqueness plays a crucial role in analyzing the capabilities and limitations of iris recognition systems. Among the various methodologies proposed, Daugman's approach to iris uniqueness stands out as one of the most widely accepted. According to Daugman, uniqueness refers to the iris recognition system's ability to enroll an increasing number of classes while maintaining a near-zero probability of collision between new and enrolled classes. Daugman's approach involves creating distinct IrisCode templates for each iris class within the system and evaluating the sustainable population under a fixed Hamming distance between codewords. In our previous work [23], we utilized Rate-Distortion Theory (as it pertains to the limits of error-correction codes) to establish boundaries for the maximum possible population of iris classes supported by Daugman's IrisCode, given the constraint of a fixed Hamming distance between codewords. Building upon that research, we propose a novel methodology to evaluate the scalability of an iris recognition system, while also measuring iris quality. We achieve this by employing a sphere-packing bound for Gaussian codewords and adopting a approach similar to Daugman's, which utilizes relative entropy as a distance measure between iris classes. To demonstrate the efficacy of our methodology, we illustrate its application on two small datasets of iris images. We determine the sustainable maximum population for each dataset based on the quality of the images. By providing these illustrations, we aim to assist researchers in comprehending the limitations inherent in their recognition systems, depending on the quality of their iris databases. | ['Matthew C. Valenti', 'Joseph Skufca', 'Stephanie Schuckers', 'Natalia A. Schmid', 'Priyanka Das', 'Jinyu Zuo', 'Katelyn M. Hampel'] | 2023-06-21 | null | null | null | null | ['iris-recognition'] | ['computer-vision'] | [ 3.75144064e-01 -4.28524576e-02 -2.43745521e-01 -4.17546555e-02
-2.36766145e-01 -6.12467825e-01 2.59057432e-01 3.91285300e-01
-3.34929496e-01 4.11074936e-01 5.02434336e-02 -5.98060548e-01
-7.21465766e-01 -5.50330997e-01 -1.97038531e-01 -7.58846045e-01
-9.12117288e-02 2.91140020e-01 -2.30287880e-01 1.37529880e-01
5.58378339e-01 8.91876400e-01 -2.27857471e+00 -2.96263456e-01
1.09012711e+00 5.97351789e-01 -3.43191981e-01 9.56836700e-01
2.97796279e-01 2.27080435e-01 -7.96570241e-01 -5.24732471e-01
4.64599520e-01 -6.58656538e-01 -7.04715014e-01 3.65633160e-01
6.18534565e-01 -2.74729013e-01 -1.01980433e-01 9.94932890e-01
4.74857837e-01 -2.76379250e-02 6.79921329e-01 -8.74633014e-01
-3.94743443e-01 1.19353347e-01 -3.74328315e-01 3.44747633e-01
5.98594606e-01 2.63891637e-01 7.25380778e-01 -2.19602481e-01
4.04973418e-01 6.07358098e-01 5.05512357e-01 5.61201453e-01
-1.31868184e+00 -4.08448458e-01 -4.38677967e-01 -1.14699639e-01
-1.76400578e+00 -6.51363909e-01 6.92983270e-02 -5.99584937e-01
6.32624686e-01 7.89728701e-01 6.82907820e-01 2.47600138e-01
-5.13477251e-02 3.28359127e-01 1.03507173e+00 -8.34803641e-01
2.66519964e-01 2.14904100e-01 3.63398045e-02 4.92230028e-01
5.70543647e-01 4.96750176e-01 -1.95257097e-01 -3.49436253e-01
6.31377697e-01 -4.13336515e-01 -3.44864666e-01 -4.70101088e-01
-9.29802656e-01 5.89940071e-01 3.76473903e-03 5.89556217e-01
-1.50535256e-03 -5.43426633e-01 4.25556116e-02 7.29402108e-03
1.06131352e-01 5.70816815e-01 1.71302691e-01 -2.15607524e-01
-1.19188869e+00 1.00897275e-01 7.70334959e-01 8.01875651e-01
2.27071971e-01 -4.45129901e-01 -3.48332644e-01 4.89661038e-01
3.30834031e-01 5.40219188e-01 3.57322186e-01 -3.51490080e-01
6.04843497e-02 8.91359508e-01 1.04751281e-01 -8.90992761e-01
-2.06945807e-01 -6.06582940e-01 -6.21094167e-01 2.87681431e-01
8.19240570e-01 -5.36421128e-02 -7.22298205e-01 1.36156321e+00
2.62313575e-01 2.89693981e-01 2.48634875e-01 7.28456497e-01
4.61133301e-01 -1.12297330e-02 -2.94723302e-01 -4.36493605e-01
1.21569741e+00 -1.77740440e-01 -2.66553044e-01 5.57959795e-01
7.18665659e-01 -1.03354764e+00 6.63281977e-01 3.73336107e-01
-8.26629519e-01 -2.81300634e-01 -1.04292035e+00 4.18409914e-01
-1.69101909e-01 3.83270502e-01 3.26808512e-01 1.63415575e+00
-1.09286809e+00 2.01104045e-01 -7.00769961e-01 -5.95791638e-01
1.63133517e-01 7.48094618e-01 -2.00974032e-01 8.29582661e-02
-6.01809382e-01 8.25206816e-01 2.92092383e-01 -9.34626535e-02
-2.31680740e-02 -2.84665525e-01 -6.09445572e-01 6.15075752e-02
1.20989420e-02 -5.40360749e-01 8.22249651e-01 -8.05189013e-01
-1.27541387e+00 1.02877712e+00 -2.71716952e-01 -4.66316164e-01
2.95422673e-01 5.59648275e-01 -5.08273602e-01 2.07435042e-01
-2.34352708e-01 1.16329245e-01 4.52609807e-01 -8.82650673e-01
-7.68902123e-01 -5.66491485e-01 1.89758301e-01 1.64210960e-01
-2.12146655e-01 3.68863642e-01 -2.12480903e-01 -2.63218433e-01
2.00667053e-01 -9.61607635e-01 -5.90149313e-02 -3.82605433e-01
-3.18583190e-01 -1.80176884e-01 1.12994224e-01 -2.18674943e-01
1.67287517e+00 -2.30013227e+00 4.94995713e-02 6.87118948e-01
1.96874410e-01 6.80490375e-01 -4.75310907e-03 2.55276829e-01
-9.31879357e-02 8.73917863e-02 -2.38037333e-01 -1.83495015e-01
-7.50707611e-02 9.20296982e-02 -5.63240834e-02 6.78925037e-01
2.14378536e-02 2.61361390e-01 -8.62972736e-01 -3.10044974e-01
5.76059878e-01 5.37725449e-01 -4.38531190e-01 6.35015145e-02
3.41212153e-01 2.48386309e-01 -1.33185629e-02 8.38070989e-01
7.44887769e-01 -2.06736997e-01 2.43650228e-01 8.75958726e-02
-3.33879322e-01 -1.76818773e-01 -1.56998050e+00 9.27145302e-01
4.01135124e-02 5.65571249e-01 -4.33085144e-01 -7.62788653e-01
9.49352264e-01 4.34006810e-01 5.81094265e-01 -3.13784003e-01
1.72667861e-01 4.92156208e-01 3.05847377e-01 -3.92227173e-01
5.86528063e-01 -1.32145852e-01 4.73698467e-01 4.61758077e-01
-1.43381208e-01 2.79860377e-01 5.46259224e-01 -1.71567529e-01
6.85009003e-01 -3.53492141e-01 5.07594645e-01 -3.52145135e-01
7.56196439e-01 -1.95667073e-01 7.12430403e-02 8.14356685e-01
-3.58651876e-01 6.68596625e-01 4.07853842e-01 -3.55947316e-01
-1.03847265e+00 -9.36710477e-01 -1.01968682e+00 5.71885891e-02
2.56206572e-01 -3.89941305e-01 -8.91203642e-01 -4.09467548e-01
9.48631093e-02 3.83591473e-01 -6.77824855e-01 2.14899316e-01
1.19509902e-02 -1.13573611e+00 6.52235746e-01 -2.45658532e-01
1.77298784e-01 -5.03634572e-01 -9.36830759e-01 -1.44876450e-01
1.90195143e-01 -6.40415728e-01 -3.07851374e-01 -4.40175503e-01
-7.64212728e-01 -1.74573839e+00 -7.27905393e-01 -4.56429750e-01
1.12064564e+00 1.17578998e-01 7.35906661e-01 4.73543316e-01
-7.04689860e-01 6.25300348e-01 -3.57785344e-01 -2.91746855e-01
-6.52238011e-01 -7.77947679e-02 2.54600018e-01 2.31722057e-01
8.16914499e-01 2.74275262e-02 -6.46087766e-01 3.81804913e-01
-1.01881433e+00 -3.95809680e-01 3.11442941e-01 5.45115829e-01
3.71041149e-01 2.89005488e-01 -1.15817323e-01 -4.11772192e-01
5.54437220e-01 -2.96506315e-01 -8.75377357e-01 4.15968925e-01
-9.75471914e-01 -1.00874186e-01 2.17520148e-01 -3.26806635e-01
-2.70798266e-01 1.10154673e-01 1.62574112e-01 5.89574017e-02
-2.73116618e-01 2.92450875e-01 2.84024745e-01 -6.25032365e-01
7.79927552e-01 4.74433273e-01 3.77885789e-01 -2.16426030e-01
6.07610904e-02 1.17706048e+00 4.92555767e-01 -4.05993342e-01
4.24219489e-01 3.04003716e-01 1.65976688e-01 -9.30950820e-01
-2.61027098e-01 -8.35341811e-01 -5.70843220e-01 -2.85791367e-01
4.11634594e-01 -4.08405155e-01 -1.39400935e+00 5.31809568e-01
-6.09727502e-01 4.07353789e-01 -2.71277308e-01 8.47696245e-01
-5.03191769e-01 7.34801829e-01 -1.04135294e-02 -1.21672714e+00
-3.53005566e-02 -1.40459204e+00 7.89462805e-01 5.14686048e-01
-1.46620840e-01 -8.70995700e-01 1.87166810e-01 3.79759133e-01
3.38042706e-01 3.16379994e-01 6.24329090e-01 -4.45664138e-01
-4.94580626e-01 -3.69539022e-01 -2.32677281e-01 2.19786540e-01
3.78919542e-01 3.47251892e-01 -6.23734772e-01 -6.69714928e-01
-1.26937144e-02 2.41927907e-01 4.06257868e-01 5.16158819e-01
7.91188300e-01 -2.85026342e-01 -2.86473364e-01 7.33238041e-01
1.71613753e+00 3.39475870e-01 8.91060710e-01 4.09373611e-01
-2.19712537e-02 6.59604788e-01 3.75219375e-01 6.05608225e-01
8.83346796e-02 7.99818099e-01 1.75275594e-01 -1.67154223e-01
6.42533377e-02 7.30501860e-02 -1.34505391e-01 2.59823412e-01
-4.00654823e-01 -2.43483484e-01 -9.57982719e-01 5.81652820e-01
-1.42819870e+00 -9.24521983e-01 -1.18945912e-03 3.03887892e+00
6.97773159e-01 -1.47267058e-01 4.71217752e-01 3.58961165e-01
8.41438413e-01 -3.80051374e-01 -3.18031371e-01 -4.47930366e-01
-1.58248439e-01 3.07943434e-01 7.00679183e-01 6.90010190e-01
-8.85088146e-01 2.47870117e-01 6.94489193e+00 5.26875198e-01
-1.04292595e+00 -5.38386524e-01 5.42537093e-01 -2.49590769e-01
1.13123223e-01 1.57006472e-01 -1.02157748e+00 7.63890266e-01
1.02780581e+00 -4.22303736e-01 4.11390573e-01 1.45814925e-01
1.34600073e-01 -5.39017498e-01 -8.52561891e-01 9.90701199e-01
3.90174121e-01 -1.17187321e+00 -5.18516116e-02 4.38751429e-01
5.66165686e-01 -3.46473426e-01 2.86100835e-01 -4.25080270e-01
-2.01375365e-01 -1.13980711e+00 1.31087024e-02 7.31269538e-01
1.02276218e+00 -7.37362921e-01 9.15117443e-01 6.39481023e-02
-8.40793133e-01 -1.11733593e-01 -2.92229056e-01 -2.78533138e-02
-3.77793133e-01 4.67460394e-01 -1.01072514e+00 5.88863671e-01
1.81847334e-01 2.30351895e-01 -6.33459687e-01 1.64519632e+00
4.13732082e-01 4.09495056e-01 -5.37255585e-01 -1.48629591e-01
-7.49697164e-02 -4.01957005e-01 6.73274517e-01 8.87272775e-01
4.65354979e-01 2.58625895e-01 -3.09407771e-01 7.28393435e-01
3.99633437e-01 6.30201995e-01 -4.20219719e-01 -1.11501746e-01
5.74563622e-01 5.57964385e-01 -7.37185061e-01 -7.00007454e-02
-3.92203689e-01 5.05609632e-01 -2.70331144e-01 7.96800703e-02
-2.33023956e-01 -5.48548222e-01 7.61159778e-01 1.67593181e-01
-1.55915385e-02 -1.45887453e-02 -4.12305385e-01 -9.62925673e-01
4.91611883e-02 -1.07940531e+00 4.28088367e-01 -3.29957336e-01
-6.76747859e-01 5.03911912e-01 -7.13640302e-02 -1.55396211e+00
-1.85987189e-01 -5.39662540e-01 1.16558122e-02 1.33380091e+00
-1.24571669e+00 -8.94405901e-01 3.47151943e-02 3.44610989e-01
-2.37697497e-01 -5.40125668e-01 1.01865208e+00 1.27192006e-01
-6.32944047e-01 1.01918411e+00 3.94665241e-01 3.94160859e-02
4.21606988e-01 -1.09122443e+00 3.06195877e-02 1.07567084e+00
2.41309360e-01 9.40301538e-01 6.60322845e-01 -4.63705838e-01
-1.17380702e+00 -4.15677816e-01 1.24027944e+00 -4.60695386e-01
2.49628276e-01 2.21652865e-01 -5.68133652e-01 3.14236730e-01
-1.92124665e-01 -2.76953220e-01 1.13602257e+00 3.23190004e-01
-7.46068358e-02 -2.02741139e-02 -1.37050879e+00 4.99088794e-01
5.89186847e-01 -5.41239858e-01 -3.63628685e-01 3.49015415e-01
-6.09384552e-02 -7.30717719e-01 -1.19188797e+00 3.40061128e-01
8.07277024e-01 -1.45554996e+00 6.84006155e-01 -3.99540693e-01
9.17013809e-02 -5.89455664e-01 7.46316686e-02 -7.17090905e-01
-1.24658741e-01 -7.87480533e-01 1.00727215e-01 8.22150230e-01
2.36483619e-01 -8.50373507e-01 9.16614354e-01 8.27028334e-01
5.40538132e-01 -8.84410858e-01 -1.04209220e+00 -8.17979693e-01
-1.86637685e-01 -2.05815714e-02 1.02268338e+00 7.90981054e-01
2.99546212e-01 -5.52977681e-01 -1.19854294e-01 5.64658463e-01
6.34901881e-01 1.28620327e-01 8.39409232e-01 -1.30305922e+00
-2.51040310e-01 -6.88611925e-01 -1.13870347e+00 -5.89781046e-01
-4.05761421e-01 -7.32343078e-01 -4.75803107e-01 -9.12020802e-01
2.58823991e-01 -6.65314138e-01 -2.86087483e-01 2.59793609e-01
-6.57356977e-02 4.89343941e-01 1.77758291e-01 3.19973320e-01
3.55098657e-02 -3.68780017e-01 9.76620018e-01 1.16757080e-01
-4.73300159e-01 4.18386340e-01 -8.24522614e-01 6.94416910e-02
6.11950099e-01 -1.59918040e-01 -3.11096817e-01 1.14461973e-01
1.49331376e-01 9.68783442e-03 1.37985379e-01 -1.18159795e+00
5.31559765e-01 -4.08434272e-02 9.14935097e-02 -2.29641035e-01
-1.75125867e-01 -6.99590147e-01 4.05879647e-01 6.76836669e-01
-2.13493496e-01 -2.44091600e-01 3.06661367e-01 1.57049283e-01
-1.71090052e-01 -5.22815764e-01 9.01316583e-01 3.46529841e-01
-2.74256110e-01 1.21752217e-01 -1.20499596e-01 -2.68033773e-01
1.26850617e+00 -8.79733264e-01 -2.88362920e-01 4.89361286e-02
-7.10750282e-01 -4.74683568e-02 9.93179321e-01 1.77677110e-01
5.32368541e-01 -8.07843626e-01 -8.38385403e-01 9.41347957e-01
5.35133123e-01 -5.07051110e-01 8.91770795e-02 9.20600653e-01
-7.78152645e-01 7.65615702e-01 -1.11292034e-01 -8.07860315e-01
-1.84532785e+00 5.57140946e-01 6.86244845e-01 1.81248844e-01
-3.05146813e-01 6.01874769e-01 -4.78647530e-01 -1.70017555e-02
2.70252794e-01 -2.54095495e-01 -3.06648165e-01 -1.49811476e-01
7.74672687e-01 2.70795166e-01 2.16985106e-01 -7.97612727e-01
-2.04577267e-01 7.90341556e-01 6.93249563e-03 9.53615680e-02
6.57480717e-01 -1.00618273e-01 -2.24731058e-01 -5.82885109e-02
7.35653996e-01 4.01445895e-01 -7.13601172e-01 8.64907652e-02
-6.96515441e-02 -1.08933437e+00 2.59559322e-02 -7.62685478e-01
-6.38663411e-01 4.17129368e-01 1.11773682e+00 4.33818877e-01
1.47639859e+00 -3.60441476e-01 2.00424969e-01 -2.21021935e-01
3.05795401e-01 -7.40857244e-01 -8.28197777e-01 -2.10912116e-02
3.39256734e-01 -1.00795543e+00 2.02792674e-01 -3.35329950e-01
-9.33657959e-02 1.21571660e+00 -4.14071791e-02 3.11207384e-01
5.20703733e-01 1.12268925e-02 1.52973101e-01 2.69763153e-02
-2.18064636e-01 -4.77946132e-01 6.19743288e-01 7.20457435e-01
6.16578698e-01 4.02110577e-01 -9.71473157e-01 -1.99398816e-01
-4.01004016e-01 2.37925738e-01 7.55314887e-01 5.43397129e-01
-2.78483152e-01 -1.45054889e+00 -6.56370819e-01 7.05478132e-01
-3.18424582e-01 -7.00778961e-02 -1.95109800e-01 6.02737188e-01
3.99450362e-01 1.17316341e+00 1.69318765e-01 -3.11388522e-01
3.03181231e-01 -2.04215392e-01 7.72946954e-01 -5.79328597e-01
-6.07932031e-01 -2.29690447e-01 -2.42981330e-01 -7.47495964e-02
-7.48503149e-01 -8.65027428e-01 -6.76611483e-01 -6.75022423e-01
-5.05879939e-01 3.51484627e-01 6.99431121e-01 9.76827323e-01
3.84382576e-01 -3.81244421e-02 7.92324364e-01 -9.62526537e-03
-6.16007745e-01 -6.03368819e-01 -9.77516532e-01 2.33878419e-01
5.63129544e-01 -4.41788703e-01 -6.38092518e-01 -6.01057671e-02] | [3.7472660541534424, -3.624019145965576] |
ca28230e-56d9-470b-bcc5-04e3ea6dc4b7 | pc-dan-point-cloud-based-deep-affinity | 2106.07552 | null | https://arxiv.org/abs/2106.07552v1 | https://arxiv.org/pdf/2106.07552v1.pdf | PC-DAN: Point Cloud based Deep Affinity Network for 3D Multi-Object Tracking (Accepted as an extended abstract in JRDB-ACT Workshop at CVPR21) | In recent times, the scope of LIDAR (Light Detection and Ranging) sensor-based technology has spread across numerous fields. It is popularly used to map terrain and navigation information into reliable 3D point cloud data, potentially revolutionizing the autonomous vehicles and assistive robotic industry. A point cloud is a dense compilation of spatial data in 3D coordinates. It plays a vital role in modeling complex real-world scenes since it preserves structural information and avoids perspective distortion, unlike image data, which is the projection of a 3D structure on a 2D plane. In order to leverage the intrinsic capabilities of the LIDAR data, we propose a PointNet-based approach for 3D Multi-Object Tracking (MOT). | ['Ajmal Mian', 'Mubarak Shah', 'Jyoti Kini', 'Aakash Kumar'] | 2021-06-03 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [ 7.70702735e-02 -4.88406062e-01 -3.22425395e-01 -3.67833018e-01
-6.92231506e-02 -5.29581606e-01 6.81484520e-01 1.13870114e-01
-4.20275360e-01 5.23793936e-01 -4.16281402e-01 -4.80344236e-01
-9.90058035e-02 -1.27589560e+00 -4.17238981e-01 -2.92031258e-01
-2.60259639e-02 9.49415445e-01 9.35459375e-01 -5.19037127e-01
3.57773870e-01 1.31264579e+00 -1.78485155e+00 -3.61142665e-01
8.51925850e-01 1.01542461e+00 3.36699903e-01 2.92431980e-01
-4.17901158e-01 -1.13139361e-01 -1.78741157e-01 -2.04330862e-01
5.59480369e-01 4.05719697e-01 2.27077808e-02 -1.85811058e-01
7.53117740e-01 -2.47147739e-01 -3.58430058e-01 1.09324205e+00
1.51569054e-01 -1.83435470e-01 4.51515317e-01 -1.48475289e+00
-1.53492868e-01 -4.76673752e-01 -8.39115381e-01 -7.13491738e-02
3.75905961e-01 2.65455828e-03 5.26502371e-01 -8.58721375e-01
8.77073169e-01 1.46158385e+00 9.16565180e-01 9.87240002e-02
-9.31767166e-01 -8.40251505e-01 -2.30027914e-01 3.23894173e-01
-1.65152955e+00 -1.58100039e-01 8.31217110e-01 -4.19888377e-01
7.10851371e-01 2.02482119e-01 9.96856689e-01 6.35113060e-01
6.90418124e-01 2.28712395e-01 1.06823087e+00 -2.36125365e-01
1.28439695e-01 -6.11839704e-02 -1.73723355e-01 5.66889822e-01
8.42094183e-01 4.68828350e-01 -6.67805433e-01 -1.30675003e-01
9.67196584e-01 5.43739319e-01 9.53661576e-02 -1.22660017e+00
-9.35361445e-01 7.46549368e-01 8.01694989e-01 -1.31317794e-01
-3.56256813e-01 1.71905547e-01 -1.39780805e-01 7.62544423e-02
3.47317010e-01 -1.18106775e-01 -5.46042919e-02 -1.08420953e-01
-7.60750830e-01 5.26289701e-01 3.50789696e-01 1.36088622e+00
1.16216338e+00 3.02431285e-02 6.45531476e-01 3.35739106e-01
5.43702781e-01 1.28203595e+00 -4.58165482e-02 -1.14864969e+00
3.63544047e-01 1.19653654e+00 1.05778731e-01 -1.04982591e+00
-4.51117575e-01 -1.47051513e-01 -7.45008588e-01 8.61936331e-01
-4.12857607e-02 5.22070050e-01 -6.70878112e-01 1.11061347e+00
7.58057296e-01 4.81413752e-02 -1.69178188e-01 6.62264585e-01
5.61176002e-01 2.54137576e-01 -3.95718455e-01 2.11645231e-01
1.24727929e+00 1.59718283e-02 -4.02891695e-01 -3.66828382e-01
9.72458646e-02 -5.30369043e-01 5.78398108e-01 1.33858144e-01
-7.86522329e-01 -4.54250872e-01 -1.23659897e+00 -2.79854894e-01
-5.27910948e-01 -5.84560812e-01 3.78200203e-01 6.69499099e-01
-1.08455086e+00 1.90002739e-01 -7.62733698e-01 -5.19727707e-01
6.18202925e-01 4.62206095e-01 -5.06261826e-01 -4.81470048e-01
-7.14026392e-01 1.38124573e+00 3.07544440e-01 -1.18046813e-01
-2.74350762e-01 -8.73787999e-01 -8.50958109e-01 -2.93173730e-01
4.63888168e-01 -8.39091241e-01 7.46286154e-01 4.57098007e-01
-9.92870569e-01 9.71059382e-01 -4.62693006e-01 -5.99169493e-01
5.47472000e-01 -3.25868428e-01 -3.43075395e-01 1.77600775e-02
2.04083547e-01 6.52959108e-01 6.27958119e-01 -1.38047886e+00
-1.01633596e+00 -1.08352602e+00 -2.22120017e-01 2.04602793e-01
2.46554106e-01 -3.62549305e-01 -1.30489230e-01 5.53737544e-02
7.78190911e-01 -9.50948715e-01 -2.11974859e-01 6.78726852e-01
-1.55540347e-01 -5.65523170e-02 1.64118934e+00 5.06781191e-02
4.76368159e-01 -1.98414469e+00 -2.46820718e-01 2.58341193e-01
4.18327272e-01 2.22472668e-01 3.11853766e-01 4.68248218e-01
6.22146666e-01 1.32798236e-02 -2.31599644e-01 -2.96474099e-01
-1.28692061e-01 5.45732260e-01 -4.86430645e-01 6.54080391e-01
5.05644269e-02 8.77958834e-01 -7.99378753e-01 -5.43492675e-01
6.68665469e-01 6.59511626e-01 -2.86097705e-01 -3.63975227e-01
-2.14789525e-01 5.09105623e-01 -6.68055415e-01 9.13867533e-01
1.25207424e+00 2.38834038e-01 -2.38404989e-01 7.64975604e-03
-7.71226645e-01 6.18971922e-02 -1.09675634e+00 1.70422351e+00
-3.00840855e-01 6.67860985e-01 9.32801217e-02 -3.69409025e-01
1.37022448e+00 -7.82734305e-02 7.75648236e-01 -9.18738484e-01
3.81828137e-02 3.19586247e-01 -3.12740654e-01 1.82542298e-02
9.21835780e-01 -1.88037798e-01 -1.44120321e-01 2.30772123e-02
-5.16156077e-01 -8.78365934e-01 -3.23074073e-01 8.61211494e-03
1.04748023e+00 1.94670618e-01 5.85435629e-01 -1.04934439e-01
3.95828903e-01 6.57827795e-01 6.51498795e-01 4.30399209e-01
-1.86323240e-01 2.77300239e-01 -3.05369675e-01 -5.29064715e-01
-1.21630430e+00 -1.42785478e+00 -5.05586863e-01 1.22889094e-01
6.59945369e-01 -1.77592039e-01 1.42296463e-01 -5.09392992e-02
8.21352601e-01 3.77524257e-01 -1.10356070e-01 7.81235397e-02
-7.70648539e-01 2.95771230e-02 3.12273234e-01 2.74298191e-01
8.33721340e-01 -4.14353639e-01 -1.25455248e+00 1.60473049e-01
3.34949493e-02 -1.17635477e+00 1.46751255e-02 -1.07755447e-02
-1.18445098e+00 -1.21297884e+00 -2.63169445e-02 -2.68402785e-01
1.74184337e-01 1.13136947e+00 1.02162492e+00 1.93932503e-02
-3.14975053e-01 4.74239349e-01 -1.04232699e-01 -9.78421926e-01
-6.33331388e-02 -2.62189537e-01 4.05545831e-01 -3.89212132e-01
8.76625478e-01 -9.27428961e-01 -3.60538960e-01 5.15448272e-01
-4.23503548e-01 -1.44206345e-01 5.58721840e-01 2.36647293e-01
9.52519715e-01 5.75573295e-02 2.70633772e-02 -4.68773901e-01
9.70396101e-02 -4.15579915e-01 -1.18385637e+00 -2.26682827e-01
-4.19589907e-01 -3.92366678e-01 4.98354435e-04 9.52898338e-03
-6.08344615e-01 3.33053857e-01 2.00975463e-01 -6.00724399e-01
-2.50689715e-01 2.73719758e-01 -1.90668270e-01 -4.80336040e-01
4.44037378e-01 9.95507911e-02 1.92267254e-01 -6.25357151e-01
3.45453858e-01 6.31094277e-01 8.49317014e-01 -2.71747947e-01
1.38644874e+00 1.26725769e+00 9.33769763e-01 -1.20000029e+00
-3.66175622e-01 -8.96150172e-01 -1.44130242e+00 -3.64498973e-01
6.61962092e-01 -8.40664566e-01 -7.68475056e-01 7.99767673e-02
-1.32833421e+00 1.71013430e-01 -4.33424085e-01 4.68851417e-01
-5.29247344e-01 1.71643913e-01 3.06788862e-01 -9.74137187e-01
4.54970114e-02 -8.46926689e-01 1.19211066e+00 2.96525925e-01
-2.35798992e-02 -6.94884717e-01 3.30506444e-01 4.50738892e-02
2.01362267e-01 6.20213389e-01 7.28035927e-01 2.43281499e-02
-1.37442362e+00 -3.84438932e-01 -4.04381752e-01 -3.94198895e-01
1.09079912e-01 -8.93533304e-02 -7.84787834e-01 -7.87952319e-02
1.53454006e-01 2.81934738e-01 5.02033889e-01 2.89603353e-01
3.21136773e-01 3.44958186e-01 -7.56072402e-01 7.43476570e-01
1.75179267e+00 3.04694593e-01 5.36327600e-01 5.75710475e-01
5.85823298e-01 5.52204072e-01 8.73163998e-01 1.41404644e-01
7.35022426e-01 1.00342047e+00 9.82793272e-01 1.34501696e-01
-1.97956726e-01 -5.71861982e-01 -1.51931554e-01 4.38677937e-01
-1.10439219e-01 2.84809858e-01 -1.14191592e+00 3.22715729e-01
-1.70203173e+00 -9.75227118e-01 -6.54560924e-01 2.43169403e+00
1.58885956e-01 1.74970210e-01 6.83875382e-03 9.56967995e-02
5.91132343e-01 5.74617356e-04 -8.43069732e-01 3.84674184e-02
-2.28026494e-01 1.90450132e-01 9.96091306e-01 4.17055607e-01
-6.59475386e-01 1.03655946e+00 6.17013550e+00 2.94748455e-01
-1.13016891e+00 7.69272149e-02 -7.55975783e-01 2.88781285e-01
-2.77897984e-01 1.82677135e-01 -1.18472743e+00 2.21002266e-01
7.72893190e-01 -4.73885864e-01 -1.46699980e-01 8.88441086e-01
4.46555972e-01 -4.92956996e-01 -8.47748518e-01 1.11842394e+00
-1.99605748e-01 -1.49396586e+00 7.30263218e-02 7.96005189e-01
4.38054979e-01 5.33965170e-01 9.36149657e-02 -7.79409409e-02
4.57057744e-01 -7.30257452e-01 8.42780352e-01 3.90642226e-01
9.27520871e-01 -5.31277776e-01 3.79163474e-01 8.65078628e-01
-1.50448418e+00 4.73996587e-02 -7.26913333e-01 -2.09578961e-01
5.34324706e-01 6.65010035e-01 -1.00983095e+00 7.70044386e-01
8.92964363e-01 6.90890133e-01 -4.31091577e-01 1.56748939e+00
3.21879834e-02 -8.12684000e-02 -7.46402025e-01 2.41932705e-01
7.97517672e-02 -6.83741271e-01 9.84088182e-01 5.86290359e-01
4.70550895e-01 -1.24381613e-02 2.94496655e-01 7.88897991e-01
2.81145051e-02 -3.58968854e-01 -1.42670465e+00 6.52588606e-01
9.36553836e-01 9.29567575e-01 -6.29612923e-01 -3.84222679e-02
-3.48875225e-01 2.97220230e-01 -1.79224148e-01 -1.00629248e-01
-4.60053325e-01 -2.76328892e-01 1.02637374e+00 4.49966103e-01
3.60636294e-01 -1.11319983e+00 -6.86480582e-01 -7.30566323e-01
5.48563339e-02 -1.04091868e-01 -1.33056983e-01 -9.83984828e-01
-8.37120056e-01 3.92136365e-01 2.21861586e-01 -1.82435858e+00
-1.42281666e-01 -6.35088921e-01 -4.07032490e-01 1.11838365e+00
-1.95219278e+00 -1.12244332e+00 -7.62298822e-01 6.50571406e-01
2.64279097e-01 7.90403783e-02 6.74472272e-01 1.85623482e-01
2.45219126e-01 -3.71392727e-01 2.62718350e-01 -3.24091077e-01
3.96990716e-01 -9.61369634e-01 6.99629545e-01 8.76597583e-01
1.38448104e-01 6.36487544e-01 7.12566137e-01 -1.15011847e+00
-1.86181045e+00 -1.13028657e+00 7.03427613e-01 -7.79503882e-01
4.92320329e-01 -4.32064891e-01 -8.45175087e-01 6.86858714e-01
-4.17072654e-01 -1.81773603e-02 1.52089179e-01 -1.77418411e-01
-3.80191445e-01 -3.75016421e-01 -1.51266801e+00 5.44421434e-01
1.40787506e+00 -4.67269689e-01 -5.90871334e-01 -1.31831532e-02
6.43197358e-01 -6.66790962e-01 -7.56791115e-01 6.26342595e-01
7.12510645e-01 -1.09685850e+00 1.25407743e+00 -9.44199562e-02
-3.06989133e-01 -6.75858378e-01 -4.71771091e-01 -1.19978356e+00
-1.61509484e-01 -4.17589456e-01 -1.33962810e-01 6.68731689e-01
-2.99446702e-01 -8.56380224e-01 1.21303046e+00 2.19491944e-01
-3.73142600e-01 -2.83620715e-01 -1.51761723e+00 -1.07724941e+00
-2.27422506e-01 -5.93385458e-01 9.42061841e-01 4.36974376e-01
-5.27307272e-01 9.85885412e-02 9.52103361e-02 5.07195473e-01
1.15486550e+00 3.74896884e-01 1.29514778e+00 -2.28863525e+00
7.59962618e-01 -3.15461248e-01 -1.10081458e+00 -1.40989816e+00
-1.52489796e-01 -8.08966100e-01 -4.14929427e-02 -1.52917433e+00
-3.75355512e-01 -1.07009923e+00 4.46873397e-01 2.31094342e-02
5.01508296e-01 5.31630635e-01 2.63095349e-01 7.11967468e-01
-2.35373184e-01 5.99752009e-01 9.46028590e-01 -6.96812151e-03
-1.25693560e-01 3.08790416e-01 -2.46482030e-01 8.08252454e-01
5.91114700e-01 -5.03861427e-01 -1.94522440e-01 -6.50321126e-01
9.69473645e-02 -1.00909047e-01 5.42812228e-01 -1.17850816e+00
5.22368968e-01 -4.43419367e-01 1.77362382e-01 -1.65563154e+00
9.94284928e-01 -1.57033789e+00 3.93885791e-01 5.60575187e-01
7.49542475e-01 4.43798751e-01 1.07797727e-01 7.72053063e-01
-1.40800178e-01 5.02124876e-02 7.82177389e-01 -1.93667486e-01
-9.54569280e-01 6.62886202e-01 3.47233191e-02 -3.82838637e-01
1.22005188e+00 -9.66900170e-01 -5.26583612e-01 -1.21222496e-01
4.21206579e-02 4.10300553e-01 1.26575947e+00 3.74543071e-01
1.03183234e+00 -1.42862093e+00 -4.85595822e-01 4.57674146e-01
3.09673369e-01 5.80526352e-01 -2.92486679e-02 5.87689698e-01
-8.14493895e-01 7.76521206e-01 -5.48341513e-01 -1.29882216e+00
-1.39764082e+00 2.75287628e-01 -1.17186802e-02 3.44755352e-01
-1.16544855e+00 2.38199607e-01 -2.34845534e-01 -4.88389492e-01
-1.37868941e-01 -3.05838734e-01 1.34770554e-02 -4.39078182e-01
4.43865657e-01 6.88732862e-01 1.97104126e-01 -1.08225882e+00
-4.28536892e-01 1.09399652e+00 3.12158704e-01 -2.70057827e-01
1.36696804e+00 -4.06607389e-01 -4.04469520e-02 5.98246336e-01
7.56111264e-01 1.91846400e-01 -1.14487982e+00 -3.83424699e-01
3.47163826e-01 -9.67276990e-01 9.19855684e-02 -1.34907857e-01
-4.45282400e-01 1.02421892e+00 6.42546594e-01 3.62222195e-01
4.88006592e-01 2.89888419e-02 7.43538141e-01 4.10531372e-01
1.35708642e+00 -5.42341173e-01 -3.99708152e-01 6.71654880e-01
6.99444830e-01 -1.05356133e+00 3.86445940e-01 -7.95132875e-01
-1.73722982e-01 1.03856945e+00 3.14256787e-01 -1.47815451e-01
6.68736458e-01 3.18632662e-01 2.66446173e-02 -4.69297945e-01
-2.60423601e-01 -3.48787427e-01 -1.46171665e-02 1.43068147e+00
-3.93158644e-01 1.56216538e-02 2.83483863e-01 -3.15155953e-01
-5.13352990e-01 7.96082541e-02 4.02215928e-01 1.26477349e+00
-1.12666106e+00 -1.32224810e+00 -7.20226526e-01 4.26066041e-01
4.20288503e-01 2.76055694e-01 -2.20835954e-01 1.03494561e+00
2.30117559e-01 7.14816630e-01 4.80741709e-01 -3.00942302e-01
6.83197439e-01 -1.66675732e-01 5.79640985e-01 -6.32402539e-01
1.86950564e-01 -3.99076104e-01 -3.76067519e-01 -7.22414613e-01
-4.34847564e-01 -8.02470088e-01 -1.27812517e+00 -6.28161192e-01
-1.33692369e-01 -9.47940499e-02 1.44643807e+00 7.27233350e-01
4.88922536e-01 -9.01422352e-02 5.34790039e-01 -1.21180737e+00
-2.69323140e-01 -5.45460463e-01 -6.75912321e-01 -6.29298985e-02
3.35749775e-01 -1.30099881e+00 1.06309570e-01 -3.16916317e-01] | [7.641634941101074, -2.4727327823638916] |
a7548e38-8968-4fc1-b204-cb8a95620334 | unified-2d-and-3d-pre-training-of-molecular | 2207.08806 | null | https://arxiv.org/abs/2207.08806v1 | https://arxiv.org/pdf/2207.08806v1.pdf | Unified 2D and 3D Pre-Training of Molecular Representations | Molecular representation learning has attracted much attention recently. A molecule can be viewed as a 2D graph with nodes/atoms connected by edges/bonds, and can also be represented by a 3D conformation with 3-dimensional coordinates of all atoms. We note that most previous work handles 2D and 3D information separately, while jointly leveraging these two sources may foster a more informative representation. In this work, we explore this appealing idea and propose a new representation learning method based on a unified 2D and 3D pre-training. Atom coordinates and interatomic distances are encoded and then fused with atomic representations through graph neural networks. The model is pre-trained on three tasks: reconstruction of masked atoms and coordinates, 3D conformation generation conditioned on 2D graph, and 2D graph generation conditioned on 3D conformation. We evaluate our method on 11 downstream molecular property prediction tasks: 7 with 2D information only and 4 with both 2D and 3D information. Our method achieves state-of-the-art results on 10 tasks, and the average improvement on 2D-only tasks is 8.3%. Our method also achieves significant improvement on two 3D conformation generation tasks. | ['Tie-Yan Liu', 'Houqiang Li', 'Wengang Zhou', 'Tao Qin', 'Shufang Xie', 'Lijun Wu', 'Yingce Xia', 'Jinhua Zhu'] | 2022-07-14 | null | null | null | null | ['molecular-property-prediction'] | ['miscellaneous'] | [ 4.63907182e-01 3.61845076e-01 -5.12845516e-01 -3.18518937e-01
-8.16881120e-01 -5.79409420e-01 6.83688164e-01 5.28572619e-01
-9.48863402e-02 8.83409441e-01 3.87672782e-01 -6.58224046e-01
1.78710997e-01 -9.90623772e-01 -9.66782391e-01 -9.65804875e-01
-3.56096774e-01 5.64204633e-01 -1.38745874e-01 -1.70532018e-01
3.64938200e-01 9.47565794e-01 -9.48475957e-01 3.29367697e-01
7.03523874e-01 8.21041346e-01 6.20783418e-02 5.42584538e-01
-2.23072767e-01 7.48829603e-01 -3.78175288e-01 -3.04859638e-01
1.94224596e-01 -5.14755070e-01 -7.84720004e-01 -3.24845314e-01
4.42891330e-01 2.06119698e-02 -5.89462996e-01 7.72463858e-01
7.13542998e-01 2.83630759e-01 1.04548764e+00 -5.96836448e-01
-1.04798949e+00 4.55781132e-01 -7.43550181e-01 -2.45229617e-01
7.41749883e-01 2.61929810e-01 1.17934811e+00 -8.30010831e-01
9.32006061e-01 1.34516299e+00 4.59242433e-01 5.04711926e-01
-1.67211890e+00 -4.88220543e-01 3.96143258e-01 -9.31164622e-02
-1.24987388e+00 -2.24092916e-01 9.57177222e-01 -5.55685997e-01
1.73596883e+00 -1.16279744e-01 4.62229371e-01 1.19364154e+00
4.91959095e-01 3.83445382e-01 6.40950620e-01 -2.93698817e-01
1.72196627e-01 -5.89029729e-01 2.06351355e-01 9.78416920e-01
4.92342144e-01 9.71498266e-02 -4.28926229e-01 -5.78898549e-01
7.01958418e-01 3.34825784e-01 -1.86309427e-01 -5.18415272e-01
-1.05897737e+00 1.04361975e+00 9.91990626e-01 -1.01699606e-01
-6.12649083e-01 3.02472144e-01 1.06279969e-01 2.07000598e-01
4.74465311e-01 8.24582279e-01 -2.70668983e-01 1.18614547e-01
-4.74406451e-01 3.98624957e-01 7.88092613e-01 1.01175213e+00
8.97358775e-01 -1.42203301e-01 -7.11745545e-02 6.31465375e-01
4.04870182e-01 1.49622202e-01 2.84655124e-01 -4.82973874e-01
5.13886034e-01 7.22412348e-01 -4.74724881e-02 -8.29987764e-01
-7.02244520e-01 -2.17051864e-01 -9.38604236e-01 1.14726059e-01
1.38478711e-01 5.96341491e-02 -1.26415884e+00 1.78303909e+00
2.71968544e-01 -8.34186673e-02 2.05956653e-01 6.87700450e-01
1.26833665e+00 9.80744481e-01 1.66187048e-01 -1.96570948e-01
1.04556763e+00 -1.04336333e+00 -4.59829152e-01 1.15230531e-01
9.72121894e-01 -5.56039572e-01 3.01612049e-01 2.84189969e-01
-1.20282340e+00 -6.32665515e-01 -1.13285339e+00 -6.30025148e-01
-5.68014979e-01 -1.35465741e-01 1.19273007e+00 3.71413887e-01
-7.31048346e-01 1.08847725e+00 -7.20451176e-01 1.81081980e-01
4.43495095e-01 6.56215847e-01 -7.05003798e-01 -2.88825691e-01
-1.17055058e+00 7.93756902e-01 4.66169357e-01 -1.88977644e-01
-6.58111155e-01 -7.13536322e-01 -1.04535961e+00 -1.60532609e-01
2.80710250e-01 -1.02331901e+00 8.27184319e-01 -7.87125751e-02
-1.34896231e+00 7.36631513e-01 -5.55090427e-01 -3.17029893e-01
1.00605749e-01 8.67827237e-02 -2.25550979e-01 1.35408193e-01
-4.76830751e-02 6.37569666e-01 5.24838567e-01 -1.08533788e+00
1.30782679e-01 -6.41845703e-01 1.55627459e-01 2.03514680e-01
3.59015167e-01 -4.91605103e-01 -3.31756979e-01 -7.16067493e-01
4.50020909e-01 -9.03231680e-01 -6.06790245e-01 -1.35845587e-01
-7.59248972e-01 -3.30523282e-01 2.53809750e-01 -5.39806545e-01
1.05862904e+00 -1.60724545e+00 7.21978486e-01 4.31454122e-01
5.55647075e-01 8.11495259e-02 -3.13703090e-01 7.16387510e-01
-5.33708811e-01 3.50796968e-01 -3.71433973e-01 -4.43947136e-01
-4.46348563e-02 4.55758563e-04 -2.74299592e-01 3.83782625e-01
5.02462804e-01 1.03636301e+00 -1.05878031e+00 8.49133059e-02
-9.61169414e-03 8.27068269e-01 -7.29227602e-01 3.52682799e-01
-4.86441940e-01 4.93416578e-01 -7.44781256e-01 7.90741742e-01
8.42162013e-01 -6.49665773e-01 6.23226047e-01 -5.14837742e-01
2.79499125e-02 8.22540343e-01 -7.54384339e-01 2.05212712e+00
-9.15281698e-02 -8.76846164e-02 -5.83755195e-01 -9.39439476e-01
1.20463037e+00 1.96464226e-01 5.79898655e-01 -6.61262691e-01
-3.04318219e-01 1.28377870e-01 1.59112662e-01 -1.76115990e-01
4.57388490e-01 -9.36135650e-02 1.61163300e-01 5.98027468e-01
9.73654836e-02 -3.52390677e-01 2.11646948e-02 4.35526639e-01
1.08442950e+00 5.48336148e-01 5.76527596e-01 5.63390031e-02
2.01127633e-01 -2.43578449e-01 1.97160065e-01 6.67256892e-01
5.30979455e-01 5.91555774e-01 1.02150106e+00 -7.68491507e-01
-1.03546536e+00 -8.33143413e-01 1.40539676e-01 7.75134325e-01
2.83910185e-02 -9.95160818e-01 -2.52985030e-01 -8.97801876e-01
3.85177702e-01 5.21591008e-01 -8.59918177e-01 -4.76125926e-01
-4.58716989e-01 -1.03508985e+00 1.52558520e-01 3.07616860e-01
-1.36607096e-01 -8.32547545e-01 1.24417000e-01 4.77735430e-01
3.93631369e-01 -5.33349991e-01 -5.05411744e-01 5.78527689e-01
-1.12326908e+00 -1.26092172e+00 -5.36057055e-01 -5.05347371e-01
8.52496684e-01 6.01545513e-01 1.19050121e+00 1.18203998e-01
-4.05923814e-01 -2.15317965e-01 -2.75993258e-01 -2.59987175e-01
-3.42805266e-01 2.27738798e-01 -3.01866736e-02 -1.44937873e-01
1.98469341e-01 -8.62098396e-01 -7.22869158e-01 -1.45418495e-01
-6.29850268e-01 2.85175562e-01 5.97102225e-01 7.59354651e-01
1.05431259e+00 -5.87898195e-01 5.32160163e-01 -1.37252414e+00
4.76651609e-01 -5.43970287e-01 -2.28316605e-01 1.33927777e-01
-5.77910602e-01 5.68478227e-01 6.91251338e-01 -5.96033037e-02
-6.97442770e-01 3.37912172e-01 -3.05482388e-01 -4.54592109e-01
-5.11751100e-02 7.79356897e-01 -2.02078521e-01 -1.52663648e-01
7.25996196e-01 8.87852162e-02 8.05446655e-02 -7.98677325e-01
6.26608789e-01 3.91143352e-01 2.36880481e-02 -9.84019399e-01
5.25601625e-01 2.53964067e-01 7.18515813e-01 -7.23939419e-01
-8.83965909e-01 -1.96504563e-01 -1.05122101e+00 5.11759460e-01
6.77616715e-01 -1.12093771e+00 -8.86632621e-01 5.32609560e-02
-1.37409353e+00 -1.31901324e-01 -3.03000063e-02 5.00697553e-01
-4.77417260e-01 7.89545238e-01 -4.84242707e-01 -5.35251498e-01
-4.84625518e-01 -1.55109370e+00 1.38727164e+00 -2.74687201e-01
-2.41565332e-01 -9.50459123e-01 2.60003895e-01 1.79585382e-01
1.49309263e-01 7.16862381e-01 1.45953774e+00 -6.84417307e-01
-6.21867657e-01 -2.62962788e-01 -2.17222124e-01 -1.05363533e-01
3.52092803e-01 -1.76284850e-01 -8.24589074e-01 -4.66836274e-01
-6.50141716e-01 -5.86399198e-01 1.43352711e+00 2.99344271e-01
1.40244925e+00 -1.99696139e-01 -5.73785961e-01 8.31489921e-01
1.01208746e+00 2.14179173e-01 4.93295372e-01 -3.21283668e-01
1.18425727e+00 4.14959699e-01 1.52399495e-01 3.58117193e-01
2.91672081e-01 8.43898237e-01 5.95966041e-01 -1.36248186e-01
-1.99288681e-01 -6.82547152e-01 1.78611562e-01 5.85553527e-01
-7.94876516e-01 -4.08228606e-01 -4.99735087e-01 -2.14760929e-01
-1.81010735e+00 -8.84446979e-01 -1.16656251e-01 2.23292303e+00
8.82459700e-01 1.15900062e-01 1.94713101e-01 -2.68872499e-01
4.19769675e-01 5.36806405e-01 -1.03572571e+00 -2.94925064e-01
-5.85632473e-02 6.02028131e-01 3.71288717e-01 6.91299498e-01
-8.65128458e-01 9.86616433e-01 6.51988935e+00 6.44131422e-01
-9.00832474e-01 -5.14345944e-01 7.39104450e-01 9.58712306e-04
-7.10788369e-01 3.23186293e-02 -8.64480495e-01 3.16533953e-01
7.51150608e-01 1.43645734e-01 2.80568540e-01 7.32502818e-01
9.18226019e-02 3.81243199e-01 -1.33096302e+00 1.19188559e+00
4.39214818e-02 -2.00889587e+00 7.90575862e-01 5.04606843e-01
6.69807255e-01 2.81093456e-02 -6.40173405e-02 1.16071850e-01
2.53077745e-01 -1.53830194e+00 3.19428265e-01 6.73166931e-01
1.10541105e+00 -7.54210413e-01 3.45186412e-01 -9.27123874e-02
-1.30035913e+00 6.19334579e-01 -5.38533092e-01 -1.33494034e-01
4.83585075e-02 5.06937325e-01 -7.50124216e-01 1.09783518e+00
-8.89514685e-02 1.44513988e+00 -3.04938614e-01 6.29655182e-01
-3.30689430e-01 3.21449131e-01 -7.06370622e-02 -9.02628899e-02
3.14168662e-01 -7.76729465e-01 3.01013798e-01 1.06696105e+00
2.51021355e-01 3.13189596e-01 2.88357794e-01 9.19215500e-01
-7.08380163e-01 -6.57409951e-02 -9.58797336e-01 -3.27259332e-01
5.05098045e-01 9.49370444e-01 -3.53857636e-01 -3.35264236e-01
-4.72677886e-01 9.02828336e-01 6.83332682e-01 4.80341583e-01
-6.32585466e-01 -5.11107266e-01 9.97093856e-01 -3.92748863e-02
3.39955568e-01 -5.37902594e-01 2.39943311e-01 -1.13049209e+00
-3.57399434e-01 -6.96064830e-01 3.58211786e-01 -6.67234659e-01
-1.45173776e+00 5.26300907e-01 -2.99164802e-01 -6.45963132e-01
-1.59253061e-01 -1.17517829e+00 -5.47497809e-01 1.19415224e+00
-1.71559763e+00 -8.86539876e-01 6.54870197e-02 1.85271353e-01
4.99593504e-02 -2.33600125e-01 1.19517851e+00 -5.94938844e-02
-7.35341072e-01 5.52157581e-01 1.72006682e-01 -1.37303755e-01
7.35608935e-01 -1.40365946e+00 1.02787793e+00 8.86977762e-02
1.78735301e-01 9.73533750e-01 2.04692349e-01 -8.15800905e-01
-2.05123687e+00 -1.25234771e+00 8.55224848e-01 -5.64877331e-01
2.35045463e-01 -4.92613256e-01 -9.50153053e-01 6.29386127e-01
-2.98555732e-01 1.73132177e-02 1.12393665e+00 3.01941365e-01
-7.96955824e-01 4.91002858e-01 -9.21708882e-01 5.20718873e-01
1.64559674e+00 -5.68478882e-01 -3.78183842e-01 8.38835895e-01
1.14552283e+00 -5.28039396e-01 -1.23574138e+00 3.01522613e-01
3.30934435e-01 -6.48526907e-01 1.36904144e+00 -1.34081709e+00
3.45398366e-01 -2.45834455e-01 -3.33163917e-01 -1.13952863e+00
-5.75661242e-01 -1.04354894e+00 -6.02798164e-01 3.12101990e-01
7.75744557e-01 -6.86855257e-01 8.36331487e-01 5.01621604e-01
-4.77936059e-01 -1.08084965e+00 -7.21587360e-01 -5.39625347e-01
1.32979915e-01 -1.89072490e-01 8.10037911e-01 8.28859925e-01
-2.73685418e-02 9.18649197e-01 -4.12764192e-01 -2.66845934e-02
3.78262162e-01 7.04009652e-01 8.06270540e-01 -1.30995309e+00
-4.27372962e-01 -4.18600291e-01 -3.71401727e-01 -1.59253538e+00
1.65394798e-01 -1.61119545e+00 -6.12959981e-01 -1.63893080e+00
3.65725160e-01 -1.57234564e-01 -2.91151226e-01 5.90850472e-01
-1.20001696e-01 -9.43627954e-02 3.38410735e-02 2.44731650e-01
-5.24173498e-01 8.08263719e-01 1.48916793e+00 -2.84967840e-01
-2.51870036e-01 -2.89343059e-01 -9.43647265e-01 1.41131848e-01
5.76507747e-01 -2.43132740e-01 -2.59548217e-01 -2.23358989e-01
3.67496789e-01 1.71013561e-03 2.09945887e-02 -4.90407854e-01
-3.99792567e-02 -1.59454629e-01 7.19828486e-01 -7.57887185e-01
9.08810139e-01 -3.53963822e-01 1.37517542e-01 4.32905883e-01
-3.84301126e-01 -9.73613560e-02 -5.81159703e-02 9.98524964e-01
5.08614555e-02 -2.52238363e-02 6.19702756e-01 -4.20800239e-01
-1.79446071e-01 9.16126966e-01 9.30455700e-03 -3.43991727e-01
7.52745390e-01 -1.88146204e-01 -3.30601722e-01 -2.65387535e-01
-8.56531382e-01 4.49288748e-02 6.20523453e-01 7.26172850e-02
9.29127693e-01 -1.32588792e+00 -5.12654722e-01 3.68530631e-01
7.84689635e-02 2.27126077e-01 4.76426771e-03 3.74906093e-01
-4.85146046e-01 7.08182096e-01 3.76749132e-03 -2.25250721e-01
-1.07823336e+00 8.43855500e-01 3.10620755e-01 -4.08288479e-01
-4.76315647e-01 8.00135374e-01 2.82420844e-01 -5.13463974e-01
1.33019209e-01 -3.37011397e-01 -5.37434034e-02 6.91233650e-02
5.00328720e-01 3.85294519e-02 6.64420128e-02 -6.02615058e-01
-2.73682415e-01 8.18733871e-01 -5.82243204e-01 6.55732512e-01
1.59821272e+00 5.36817253e-01 -1.83724597e-01 1.73860490e-01
1.54792488e+00 -3.09642497e-02 -1.21223938e+00 -3.52720320e-01
-2.65862405e-01 -2.74117559e-01 -1.49090216e-01 -6.43695772e-01
-8.23847294e-01 1.04887116e+00 -1.12230899e-02 -2.50611436e-02
4.76479441e-01 9.12863985e-02 8.31333160e-01 9.25936282e-01
3.00725132e-01 -3.57529253e-01 3.15199912e-01 6.05192959e-01
9.45053458e-01 -1.22626746e+00 4.85927522e-01 -5.37866950e-01
-4.69359815e-01 1.36884058e+00 1.73597991e-01 -1.10581912e-01
4.84359324e-01 -4.44770157e-01 -5.95119596e-01 -6.23841524e-01
-7.78647184e-01 -1.53076693e-01 6.03762984e-01 5.72810292e-01
1.03331041e+00 1.65931925e-01 1.72415063e-01 5.50802708e-01
1.18346043e-01 -4.98660624e-01 1.74266309e-01 1.01014435e+00
-4.92543638e-01 -1.62410438e+00 3.27452123e-02 5.40558815e-01
-1.41159028e-01 -3.00310165e-01 -9.42707062e-01 5.60489953e-01
-1.98004812e-01 6.66606486e-01 -1.55030653e-01 -4.96970564e-01
2.97175109e-01 1.75342098e-01 8.49986315e-01 -9.94541109e-01
-1.95616916e-01 -7.42341131e-02 9.19169933e-02 -6.27552330e-01
-4.15566415e-01 -2.73491055e-01 -1.44692135e+00 -4.12570149e-01
-1.27751753e-01 4.89444375e-01 4.95886058e-01 4.30220991e-01
1.07795882e+00 4.98445660e-01 7.10890234e-01 -1.19018340e+00
-4.02169377e-01 -9.00642335e-01 -3.76720130e-01 3.98372620e-01
4.51794654e-01 -7.60994911e-01 -7.58995488e-02 -2.22308710e-01] | [5.094351768493652, 5.78941011428833] |
5f4f7c81-f593-4619-9736-3ca8f6dcdd33 | efficient-distributed-framework-for | 2205.05248 | null | https://arxiv.org/abs/2205.05248v1 | https://arxiv.org/pdf/2205.05248v1.pdf | Efficient Distributed Framework for Collaborative Multi-Agent Reinforcement Learning | Multi-agent reinforcement learning for incomplete information environments has attracted extensive attention from researchers. However, due to the slow sample collection and poor sample exploration, there are still some problems in multi-agent reinforcement learning, such as unstable model iteration and low training efficiency. Moreover, most of the existing distributed framework are proposed for single-agent reinforcement learning and not suitable for multi-agent. In this paper, we design an distributed MARL framework based on the actor-work-learner architecture. In this framework, multiple asynchronous environment interaction modules can be deployed simultaneously, which greatly improves the sample collection speed and sample diversity. Meanwhile, to make full use of computing resources, we decouple the model iteration from environment interaction, and thus accelerate the policy iteration. Finally, we verified the effectiveness of propose framework in MaCA military simulation environment and the SMAC 3D realtime strategy gaming environment with imcomplete information characteristics. | ['Jing Xiao', 'Xuan Wang', 'Jiajia Zhang', 'Xiaohan Hou', 'Shuhao Zhang', 'Shuhan Qi'] | 2022-05-11 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-4.46991473e-01 -5.07447243e-01 -2.72392392e-01 1.50999442e-01
-5.37324786e-01 -3.04955512e-01 1.83289289e-01 1.64587691e-01
-8.96070600e-01 1.07177031e+00 -1.37424812e-01 -1.77662671e-01
-2.88003504e-01 -8.35036814e-01 -1.98473722e-01 -8.89187813e-01
-2.63794780e-01 6.56546891e-01 2.50645936e-01 -3.61615509e-01
1.48713619e-01 1.79081619e-01 -1.45936477e+00 -3.31467062e-01
1.00679493e+00 5.70197642e-01 6.32544816e-01 6.75781608e-01
-8.10972303e-02 8.17485332e-01 -8.23660851e-01 3.35429609e-01
2.09256798e-01 -5.73981404e-01 -4.11612749e-01 -2.48004645e-02
-6.18056059e-01 -9.89879549e-01 -2.53154516e-01 9.43587899e-01
1.07452643e+00 3.12811106e-01 2.19610021e-01 -1.51073813e+00
2.47841090e-01 5.73277235e-01 -9.06037807e-01 1.86820418e-01
2.46797249e-01 4.01092410e-01 4.84189421e-01 -3.71385604e-01
4.12317276e-01 1.44552255e+00 2.20361903e-01 5.25533199e-01
-5.14222264e-01 -8.29922318e-01 5.62933981e-01 4.57527190e-01
-1.11995578e+00 -1.21752903e-01 6.75002277e-01 1.06787391e-01
6.05372727e-01 1.04895279e-01 1.09146237e+00 8.07434440e-01
1.79975495e-01 1.22980487e+00 1.17350066e+00 -1.99588820e-01
6.57130897e-01 -1.34227052e-01 -3.30475658e-01 6.85270071e-01
1.22710951e-01 3.60685110e-01 -1.93081006e-01 -2.13996261e-01
1.09551740e+00 2.87615806e-01 1.86918274e-01 -1.81690782e-01
-1.14767647e+00 7.52841175e-01 3.81316282e-02 -1.34557858e-01
-5.26650548e-01 1.20951913e-01 6.76095426e-01 6.66892171e-01
2.11128533e-01 -2.54801620e-04 -3.02950412e-01 -6.97548985e-01
-4.57383037e-01 6.55507743e-01 6.54550493e-01 8.53988051e-01
8.70883405e-01 4.49136078e-01 1.99540645e-01 7.55285800e-01
4.60395515e-01 8.41461003e-01 6.79733217e-01 -1.02222657e+00
4.49821949e-01 4.47838068e-01 3.05442810e-01 -9.54876482e-01
-6.48573041e-01 -4.05341417e-01 -9.22418058e-01 7.36176550e-01
1.45581573e-01 -8.89728367e-01 -3.22871238e-01 1.43009448e+00
8.06986868e-01 3.23960900e-01 3.92271101e-01 1.08867431e+00
5.47879040e-01 6.46737516e-01 6.52397946e-02 -6.42962277e-01
9.95795965e-01 -9.59121466e-01 -6.91923618e-01 -1.80233806e-01
4.85369027e-01 -6.02102935e-01 8.23786378e-01 4.67001557e-01
-1.11492968e+00 -5.83925664e-01 -1.14974618e+00 7.69225419e-01
-3.76943611e-02 -3.31273861e-02 7.35006094e-01 5.22540808e-01
-5.76828480e-01 2.77187169e-01 -9.16777074e-01 5.34075685e-02
1.93546310e-01 4.19502378e-01 1.51984289e-01 -3.30290198e-02
-1.21277630e+00 6.63140953e-01 6.07470930e-01 1.27377316e-01
-1.40575278e+00 -3.07621032e-01 -3.73691589e-01 -1.73344463e-02
7.22547054e-01 -3.76064062e-01 1.29665279e+00 -8.94344270e-01
-2.06595278e+00 -1.97820947e-01 2.37292722e-01 -1.08106673e-01
7.09662616e-01 5.61955906e-02 -3.45404088e-01 1.50780287e-02
-1.59416854e-01 2.68077672e-01 6.64094448e-01 -1.34164822e+00
-1.09230602e+00 -3.72719705e-01 3.07580173e-01 1.03017962e+00
-3.45294476e-01 -1.99237317e-01 -9.90909860e-02 -1.13698393e-01
-4.07913387e-01 -7.03074634e-01 -8.29135060e-01 -5.50239444e-01
3.32157731e-01 -2.10527450e-01 8.73445451e-01 -2.96429425e-01
1.27717638e+00 -2.10245824e+00 2.62761921e-01 2.96437651e-01
2.26808622e-01 2.12136075e-01 -2.71652579e-01 5.64366519e-01
4.92876589e-01 -2.03083351e-01 2.15200201e-01 -7.52538368e-02
8.27023853e-03 3.37026685e-01 1.09066926e-01 3.42678547e-01
-3.99450690e-01 3.49650800e-01 -1.17772293e+00 -8.57035995e-01
2.26262555e-01 -1.99970305e-01 -6.29093230e-01 3.36701155e-01
-3.19201589e-01 5.75957298e-01 -1.09854162e+00 6.74207091e-01
9.87041235e-01 9.94238704e-02 3.23176712e-01 2.62535512e-01
-4.91001874e-01 -4.31627989e-01 -1.91820443e+00 1.84037423e+00
-6.33057415e-01 -1.12637714e-01 5.87209642e-01 -9.99480844e-01
8.72207820e-01 3.00356418e-01 8.48303676e-01 -9.68232870e-01
2.03893572e-01 2.88182318e-01 1.57732248e-01 -6.86061800e-01
5.86070120e-01 1.56823844e-01 1.14635952e-01 7.38998830e-01
-2.68054426e-01 -3.12545747e-02 3.92800093e-01 5.69105744e-02
1.00955260e+00 2.74132699e-01 1.92027584e-01 1.83456596e-02
5.17115235e-01 1.08443499e-01 8.73154640e-01 8.31157088e-01
-4.48538989e-01 -1.34970129e-01 2.32216582e-01 -4.93537724e-01
-8.63917649e-01 -7.44790912e-01 4.25315738e-01 1.04777026e+00
7.87586510e-01 -4.12879407e-01 -4.57119465e-01 -5.76235235e-01
-1.30629003e-01 1.17024839e-01 -4.83134799e-02 6.92910552e-02
-6.83391869e-01 -8.54826510e-01 2.98203111e-01 2.13156175e-02
9.24953043e-01 -1.39746714e+00 -9.97633398e-01 7.43510544e-01
5.14291935e-02 -5.85308492e-01 -1.34614721e-01 -3.21279883e-01
-7.72920072e-01 -1.05690992e+00 -5.93254864e-01 -7.59658635e-01
3.10952306e-01 3.84514540e-01 7.45722115e-01 3.85800958e-01
-1.77294672e-01 4.28173393e-01 -5.04981697e-01 -5.99170744e-01
-3.45053375e-01 2.41118625e-01 2.16147110e-01 -2.43481919e-01
-9.92968306e-02 -4.41931695e-01 -6.56060755e-01 3.18348914e-01
-9.97430027e-01 2.17754304e-01 7.08326578e-01 1.02521157e+00
1.87743112e-01 2.90070117e-01 1.09897208e+00 -3.90932202e-01
9.87588286e-01 -5.31725466e-01 -8.46690238e-01 2.16377005e-01
-5.46557963e-01 -2.37337559e-01 9.70207095e-01 -5.47273040e-01
-1.38474739e+00 -1.47018746e-01 -3.30643237e-01 -1.06190518e-01
-3.09592746e-02 7.24384308e-01 -5.61591312e-02 -4.21028808e-02
2.73699641e-01 3.68113488e-01 5.80909610e-01 -1.17621087e-01
-1.03265077e-01 8.99181962e-01 -2.28652433e-01 -7.73793101e-01
5.35748363e-01 1.14192985e-01 4.68936302e-02 -6.67195857e-01
-3.51067670e-02 -1.97115436e-01 1.50494337e-01 -5.97776711e-01
4.24430460e-01 -1.19522595e+00 -1.49038458e+00 6.81231260e-01
-9.41879332e-01 -5.38617849e-01 5.17237484e-02 9.07637358e-01
-5.73005199e-01 3.95597607e-01 -5.01668155e-01 -1.05912530e+00
-3.38274837e-01 -1.37920415e+00 5.60769022e-01 8.07280540e-01
5.26979566e-01 -8.95836234e-01 3.60297769e-01 -1.82308003e-01
6.02132916e-01 1.49295613e-01 5.78225255e-01 -1.98048547e-01
-5.91481328e-01 4.00043018e-02 2.60151535e-01 -1.88866228e-01
-3.56343836e-02 -2.91354150e-01 -4.06837881e-01 -8.02511454e-01
-1.71956480e-01 -7.25330651e-01 1.00257672e-01 2.18993977e-01
9.75492895e-01 -2.68794477e-01 -2.45465204e-01 2.82247484e-01
1.53033316e+00 5.71003556e-01 4.25757438e-01 6.59022510e-01
3.72969389e-01 1.55577376e-01 1.22402513e+00 1.36440992e+00
3.57753009e-01 3.56656373e-01 6.71682060e-01 -2.11685881e-01
4.39511836e-01 -1.20936483e-01 4.11973566e-01 1.12551558e+00
-2.44838402e-01 -9.98713449e-02 -6.54742599e-01 2.61650801e-01
-2.35491538e+00 -1.16039360e+00 2.51980603e-01 2.02114558e+00
8.51738513e-01 3.69462557e-02 4.37860310e-01 -2.93812394e-01
4.84194428e-01 1.57325193e-01 -8.54542911e-01 -4.57902253e-02
2.39046998e-02 -1.35095283e-01 4.46622103e-01 5.90467930e-01
-6.74412668e-01 8.28615069e-01 5.74620485e+00 1.22852933e+00
-8.90325487e-01 1.23708904e-01 4.78070855e-01 -2.37603337e-01
-1.82356775e-01 2.52533592e-02 -7.19832420e-01 6.70525670e-01
6.21750295e-01 -3.89129966e-01 9.33584452e-01 9.52735364e-01
5.28088093e-01 -3.84035528e-01 -3.63611788e-01 1.18539846e+00
-2.69053638e-01 -1.20026886e+00 -3.28994930e-01 8.39758590e-02
6.04724288e-01 4.97088023e-02 -2.85296112e-01 7.61165738e-01
6.01206601e-01 -6.71127319e-01 2.16528311e-01 2.91257411e-01
3.86955887e-01 -1.30902326e+00 6.49692118e-01 8.35226417e-01
-1.31380761e+00 -3.82417113e-01 -7.42629826e-01 -4.43003625e-01
5.89431860e-02 2.29410693e-01 -5.34342468e-01 1.01627219e+00
4.65876877e-01 5.86006522e-01 -2.29168430e-01 1.23504734e+00
2.77044713e-01 3.39834720e-01 -4.69951957e-01 -5.97213686e-01
4.00222033e-01 -5.45640826e-01 5.55906534e-01 5.45897245e-01
3.19789767e-01 1.00957289e-01 1.09757924e+00 1.59556240e-01
4.03389961e-01 3.13279688e-01 -4.52030927e-01 4.36185926e-01
8.29405963e-01 1.45147741e+00 -4.56586331e-01 -3.76426280e-01
-3.59085381e-01 5.26750565e-01 3.25986177e-01 4.47831899e-01
-8.70370924e-01 -5.34382939e-01 5.55890560e-01 -5.37688434e-01
-1.20383754e-01 -4.29419160e-01 1.84491113e-01 -1.07437861e+00
-2.12429911e-01 -1.27151489e+00 2.75687486e-01 -2.08187342e-01
-9.24369037e-01 3.78237814e-01 -5.59757836e-03 -1.50206411e+00
-5.11563957e-01 -2.53953695e-01 -8.07382464e-01 6.33471310e-01
-1.51523924e+00 -7.82565117e-01 -4.67744261e-01 7.44860590e-01
7.35193312e-01 -7.52833426e-01 8.40119362e-01 3.98276806e-01
-7.31108189e-01 3.79173070e-01 6.58060372e-01 -1.85872585e-01
5.52840114e-01 -1.05016685e+00 -2.45701507e-01 4.30602431e-01
-6.37657404e-01 1.57432765e-01 6.08718991e-01 -6.54880822e-01
-1.82712162e+00 -7.21806288e-01 -2.96585292e-01 4.52928245e-01
4.35856342e-01 -4.68190871e-02 -4.61594969e-01 1.19710997e-01
5.28835595e-01 -2.11140290e-01 3.20885241e-01 2.04507664e-01
5.58664739e-01 -4.73977655e-01 -1.01221752e+00 8.53451788e-01
6.46766901e-01 2.59790033e-01 1.51496470e-01 2.81034678e-01
3.23973805e-01 -6.45436108e-01 -8.90021145e-01 5.08196168e-02
5.49405515e-01 -7.61134505e-01 7.70192444e-01 -4.78948176e-01
2.98940446e-02 -5.11795998e-01 2.00077906e-01 -1.75475836e+00
-3.06141078e-01 -7.84592092e-01 -8.22370350e-02 1.07557666e+00
3.05550974e-02 -6.83928370e-01 1.05232716e+00 1.42888382e-01
-8.17287713e-02 -7.22392738e-01 -1.13377321e+00 -6.74547553e-01
-2.09340211e-02 9.04356912e-02 9.14368331e-01 5.50423324e-01
1.99361786e-01 4.46921051e-01 -8.21138084e-01 3.56816575e-02
6.68649435e-01 1.83323890e-01 1.12395394e+00 -9.03709173e-01
-7.11017251e-01 -2.08677560e-01 9.26337615e-02 -1.14371610e+00
1.10606861e-03 -2.50448734e-01 1.02989383e-01 -1.41333425e+00
1.48526251e-01 -9.46934462e-01 -2.72930622e-01 2.27641255e-01
-3.42063397e-01 -5.16591847e-01 2.46549666e-01 2.28369653e-01
-1.24680007e+00 9.04302955e-01 1.61490798e+00 -1.27565742e-01
-4.18472886e-01 6.18290566e-02 -1.56507984e-01 4.69174713e-01
1.13748193e+00 -2.96891749e-01 -7.59936094e-01 -6.12272680e-01
1.80579364e-01 8.67426693e-01 -5.00164516e-02 -1.20087397e+00
3.22753787e-01 -7.44865775e-01 3.89598191e-01 -4.17560905e-01
2.44351640e-01 -8.98672938e-01 -4.16786708e-02 8.60552609e-01
-4.55505168e-03 5.80828071e-01 9.09398347e-02 7.31416821e-01
-2.33019367e-01 -4.19886649e-01 4.64128315e-01 -5.09000182e-01
-7.22581029e-01 6.20748758e-01 -7.93442130e-01 1.42106012e-01
1.25556922e+00 1.71610430e-01 -2.94763535e-01 -4.59909856e-01
-2.68979967e-01 9.29025829e-01 2.27494717e-01 1.63135052e-01
8.27362716e-01 -1.42735100e+00 -8.33495438e-01 2.30717659e-01
-2.56994247e-01 7.76106343e-02 6.94260478e-01 6.78943336e-01
-6.15606487e-01 -2.78004944e-01 -5.39142191e-01 -3.85388464e-01
-1.20328665e+00 3.98407966e-01 2.36107469e-01 -5.85633874e-01
-3.72574449e-01 6.09143712e-02 -2.97191828e-01 -6.22753143e-01
3.10632914e-01 2.24024162e-01 -2.70993322e-01 -5.42308912e-02
6.77901983e-01 6.19030356e-01 -4.14943874e-01 9.91945639e-02
-2.25909263e-01 2.04302415e-01 -1.95247144e-01 -3.90146792e-01
1.24471962e+00 -3.27966720e-01 2.20147759e-01 2.30874792e-01
7.50793099e-01 -1.11258872e-01 -1.40250361e+00 -6.45299479e-02
-4.79309052e-01 -5.82054436e-01 1.03690855e-01 -6.58956587e-01
-9.09456432e-01 5.12978792e-01 9.12116885e-01 2.43613496e-01
9.77233708e-01 -7.75669158e-01 7.00909793e-01 6.62317455e-01
7.58522749e-01 -1.59173369e+00 2.14613140e-01 6.56199276e-01
5.00107586e-01 -1.44759202e+00 2.18472302e-01 2.25869492e-01
-8.93214285e-01 1.15472484e+00 1.12285900e+00 -2.55490303e-01
5.62453568e-01 2.92078167e-01 2.13268384e-01 8.88922065e-02
-9.13759708e-01 -1.67519316e-01 -8.12280476e-01 6.90518737e-01
-4.73230593e-02 -2.89893858e-02 -4.79671806e-01 4.48013693e-01
2.65516579e-01 1.16960369e-01 7.43452549e-01 1.19625664e+00
-6.13917887e-01 -1.27558458e+00 -3.75294626e-01 4.21002388e-01
-1.41954049e-01 2.96164632e-01 4.98759180e-01 9.04921710e-01
-9.53154936e-02 1.09165776e+00 -1.22765198e-01 -3.33349794e-01
1.70109794e-01 -5.90662420e-01 4.18236166e-01 -8.63871947e-02
-6.30532324e-01 3.99464548e-01 4.27563637e-02 -3.37363362e-01
-3.27237725e-01 -7.05397487e-01 -1.41746378e+00 -4.39058512e-01
-2.19394624e-01 5.95507026e-01 5.48789382e-01 8.69054496e-01
3.60874057e-01 6.81290865e-01 1.10549223e+00 -6.46031857e-01
-9.56039846e-01 -1.03992724e+00 -7.87050188e-01 -5.07608801e-02
1.87917128e-01 -5.98839879e-01 7.40679502e-02 -6.72714651e-01] | [3.7989656925201416, 2.026935577392578] |
d982d8c6-fbdb-4cc2-80d6-85306b547cf6 | asymmetric-modality-translation-for-face | 2110.09108 | null | https://arxiv.org/abs/2110.09108v2 | https://arxiv.org/pdf/2110.09108v2.pdf | Asymmetric Modality Translation For Face Presentation Attack Detection | Face presentation attack detection (PAD) is an essential measure to protect face recognition systems from being spoofed by malicious users and has attracted great attention from both academia and industry. Although most of the existing methods can achieve desired performance to some extent, the generalization issue of face presentation attack detection under cross-domain settings (e.g., the setting of unseen attacks and varying illumination) remains to be solved. In this paper, we propose a novel framework based on asymmetric modality translation for face presentation attack detection in bi-modality scenarios. Under the framework, we establish connections between two modality images of genuine faces. Specifically, a novel modality fusion scheme is presented that the image of one modality is translated to the other one through an asymmetric modality translator, then fused with its corresponding paired image. The fusion result is fed as the input to a discriminator for inference. The training of the translator is supervised by an asymmetric modality translation loss. Besides, an illumination normalization module based on Pattern of Local Gravitational Force (PLGF) representation is used to reduce the impact of illumination variation. We conduct extensive experiments on three public datasets, which validate that our method is effective in detecting various types of attacks and achieves state-of-the-art performance under different evaluation protocols. | ['Alex C. Kot', 'Kwok-Yan Lam', 'Yongjian Hu', 'Xin Luo', 'Haoliang Li', 'Zhi Li'] | 2021-10-18 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 7.24935472e-01 -3.61874819e-01 -1.48043588e-01 -2.60984242e-01
-6.38804317e-01 -6.95763111e-01 7.79064298e-01 -3.82094204e-01
-5.10696881e-02 3.74347955e-01 -1.96010530e-01 -2.88819671e-01
-1.61274746e-02 -7.00595975e-01 -6.63966119e-01 -1.21847200e+00
2.28557944e-01 -1.43709868e-01 1.23383805e-01 -2.60525018e-01
1.96194932e-01 5.12917042e-01 -1.28243077e+00 3.37022185e-01
6.13238692e-01 1.20325935e+00 -3.03246796e-01 4.09426630e-01
1.33199006e-01 2.80184090e-01 -6.56776547e-01 -9.80273843e-01
4.79295015e-01 -4.78823423e-01 -4.03797030e-01 1.27654582e-01
4.86219764e-01 -2.30961844e-01 -4.73152757e-01 1.33480597e+00
7.90457487e-01 -1.39742315e-01 5.11405587e-01 -1.69447768e+00
-6.38735414e-01 1.92745268e-01 -8.61463070e-01 1.90988660e-01
5.75015426e-01 2.23605871e-01 4.48150903e-01 -8.22804451e-01
3.73438984e-01 1.55684423e+00 3.41347516e-01 7.44994164e-01
-1.13313293e+00 -1.05378461e+00 1.55589692e-02 3.47098708e-01
-1.37569475e+00 -5.22269249e-01 1.04996371e+00 -1.62106037e-01
1.69668287e-01 1.72801644e-01 -1.12549178e-02 1.51802993e+00
6.28798902e-02 4.83692110e-01 1.27403307e+00 -3.02323550e-01
-1.48223057e-01 4.21860129e-01 -2.54645318e-01 6.36932969e-01
1.76836550e-01 1.46654204e-01 -5.62917829e-01 -4.62618321e-01
4.74751949e-01 -3.13395262e-02 -4.28396046e-01 -1.72627822e-01
-7.64597118e-01 7.91137636e-01 2.77556121e-01 1.74333185e-01
-1.93562210e-01 -4.18093234e-01 3.64781618e-01 4.07135963e-01
3.36244076e-01 -1.84120089e-01 -2.33060829e-02 3.49375278e-01
-7.19054639e-01 1.02018259e-01 6.35073900e-01 4.27951068e-01
4.90405977e-01 -5.10916635e-02 -1.43157631e-01 7.41692066e-01
5.08632898e-01 9.70835567e-01 3.22486371e-01 -3.89924437e-01
7.51014292e-01 4.11561877e-01 -3.00220013e-01 -1.40732276e+00
-5.79459174e-03 -2.67378330e-01 -8.07322800e-01 1.40595466e-01
3.20615530e-01 -1.62399411e-01 -5.92941999e-01 1.99042559e+00
5.75735152e-01 5.60023010e-01 1.88400626e-01 9.47295904e-01
6.13408267e-01 5.16623616e-01 1.62386239e-01 -3.18720639e-01
1.63400555e+00 -5.13262153e-01 -6.52219415e-01 -1.42358188e-02
2.18866676e-01 -1.04285979e+00 7.17483044e-01 2.92646587e-01
-7.27958679e-01 -3.43850195e-01 -9.99437094e-01 3.30027521e-01
-3.57829034e-01 -7.09657222e-02 1.69107318e-01 1.36096120e+00
-7.36984491e-01 9.89280418e-02 -3.44384730e-01 -3.44334781e-01
5.87514043e-01 4.15322959e-01 -7.23230004e-01 -2.20822483e-01
-1.45305848e+00 5.42137146e-01 1.80385694e-01 1.21360093e-01
-8.32317352e-01 -4.35715914e-01 -7.33709753e-01 -9.24402773e-02
3.50018799e-01 -3.61911297e-01 7.52859116e-01 -1.01691866e+00
-1.46853018e+00 9.07013416e-01 1.77297667e-02 -1.49883747e-01
5.08130729e-01 1.75025538e-01 -8.95352542e-01 3.98302495e-01
-1.60540137e-02 2.18482420e-01 1.48182893e+00 -1.24262238e+00
-4.38713312e-01 -7.69176602e-01 8.44427496e-02 2.30931640e-02
-6.94266558e-01 5.58405757e-01 -4.13261890e-01 -6.35507107e-01
5.10328077e-02 -9.69591618e-01 4.97170419e-01 8.43800008e-02
-3.51405084e-01 -7.67703354e-02 1.46433640e+00 -8.87697339e-01
1.03302479e+00 -2.34945917e+00 1.43319309e-01 5.08252382e-01
-8.39435235e-02 6.45115972e-01 -2.32673526e-01 4.03487116e-01
-2.67343581e-01 2.58323122e-02 -3.32723737e-01 -4.62831318e-01
-2.53025610e-02 1.23940436e-02 -4.85824734e-01 7.97538996e-01
4.21083629e-01 3.61241162e-01 -6.26884937e-01 -5.98446667e-01
1.85447603e-01 8.66526663e-01 -3.47467184e-01 2.93582171e-01
2.04549655e-01 6.46371663e-01 -4.35007393e-01 9.04121757e-01
1.40294504e+00 9.20664966e-02 2.07811818e-01 -2.83020586e-01
3.26919407e-01 -1.55910790e-01 -1.26542115e+00 1.15810132e+00
-2.06179112e-01 3.02486241e-01 2.92529494e-01 -9.45010662e-01
9.35138285e-01 6.51590645e-01 1.97412685e-01 -6.21985734e-01
4.73132432e-01 1.22925118e-01 -7.26383030e-02 -5.91748118e-01
-1.09542474e-01 -2.00128704e-01 5.06098233e-02 4.96526599e-01
8.83882195e-02 3.12059611e-01 2.06849705e-02 1.41208142e-01
7.56046534e-01 -1.49342328e-01 -5.24631292e-02 1.10405371e-01
1.20140862e+00 -7.11255312e-01 4.72085059e-01 4.22671914e-01
-4.74967420e-01 4.09944534e-01 5.80151618e-01 -7.87064061e-02
-6.77406728e-01 -1.12217069e+00 -2.18979865e-01 1.14328694e+00
3.67016196e-01 -1.14209384e-01 -9.16621089e-01 -1.05388260e+00
1.71907358e-02 9.44087952e-02 -5.50525665e-01 -3.78001630e-01
-4.80247557e-01 -8.53755474e-01 1.08690631e+00 2.10856661e-01
9.33021188e-01 -8.56929302e-01 -7.66090304e-02 -3.95366669e-01
-3.97254288e-01 -1.42813241e+00 -3.65197659e-01 -5.44495702e-01
-3.58986199e-01 -1.20742226e+00 -5.77165246e-01 -8.39380920e-01
6.82307661e-01 6.01294160e-01 3.35441291e-01 1.53248027e-01
-2.11917117e-01 3.73987317e-01 -1.99649215e-01 -2.52065241e-01
-4.51309681e-01 -2.78074384e-01 2.64581323e-01 8.60274792e-01
3.87684733e-01 -3.55689764e-01 -5.58235407e-01 5.33500314e-01
-1.29119861e+00 -3.95522177e-01 4.67485368e-01 8.05840790e-01
-8.05279985e-03 2.32449442e-01 4.55256641e-01 -7.13659286e-01
5.59643030e-01 -6.40518188e-01 -4.09805447e-01 4.17362720e-01
-1.02341212e-01 -1.02498069e-01 5.02201080e-01 -7.62405992e-01
-1.43825984e+00 -1.26322716e-01 -2.55916640e-02 -5.34120440e-01
-1.58577308e-01 1.09794632e-01 -7.86904573e-01 -5.85734487e-01
4.59974229e-01 2.81558096e-01 2.94447690e-01 -2.75022745e-01
1.83396518e-01 8.40441883e-01 5.45681000e-01 -5.14373004e-01
1.32184839e+00 5.98762274e-01 1.72852114e-01 -8.86450529e-01
-2.34783366e-01 -1.59043655e-01 -2.35081077e-01 -2.64643520e-01
6.32523298e-01 -6.99556112e-01 -8.94555211e-01 8.81289661e-01
-1.07946563e+00 5.23040831e-01 6.78376675e-01 3.44104826e-01
-8.35340619e-02 7.88441598e-01 -5.98367214e-01 -8.82459223e-01
-2.53016293e-01 -1.35703897e+00 1.10690129e+00 4.30777073e-01
5.16554594e-01 -9.78139341e-01 -1.69032842e-01 6.98925138e-01
4.47373152e-01 2.51813650e-01 7.32401848e-01 -7.40988612e-01
-2.16117442e-01 -3.98881614e-01 -3.97416204e-01 6.48416638e-01
2.51909584e-01 -9.58986022e-03 -1.35283244e+00 -5.17873764e-01
3.68339330e-01 -2.53020316e-01 6.71084762e-01 -2.76547492e-01
1.04553211e+00 -2.67779231e-01 -3.10661674e-01 5.63497663e-01
1.16062784e+00 9.62122381e-02 7.41862535e-01 4.23285179e-02
6.35533929e-01 7.61674643e-01 3.96347344e-01 4.33787435e-01
4.87754717e-02 7.65472770e-01 5.61126649e-01 -7.15956092e-04
6.77751079e-02 -9.69410837e-02 5.96241117e-01 1.06516384e-01
1.16806775e-01 -3.80538046e-01 -5.74215829e-01 8.12675804e-04
-1.51450896e+00 -1.15947580e+00 2.66752481e-01 2.29349041e+00
5.55773318e-01 -2.07636449e-02 1.11972153e-01 4.56618965e-01
1.11526120e+00 2.29780287e-01 -2.29091316e-01 -1.94705442e-01
-3.56141508e-01 8.39866847e-02 2.61364341e-01 2.67464757e-01
-1.34957850e+00 7.73540735e-01 4.94554567e+00 9.48790729e-01
-1.56873727e+00 1.61338523e-01 5.09448409e-01 3.34743708e-01
3.69182900e-02 -1.35402843e-01 -8.02120686e-01 7.50209451e-01
5.72890282e-01 7.37438649e-02 4.54268217e-01 4.02323484e-01
-6.45356700e-02 2.82018960e-01 -6.22230887e-01 1.15929890e+00
6.26359820e-01 -5.76943099e-01 8.54737386e-02 2.29865015e-01
4.48200047e-01 -7.08705306e-01 7.10358381e-01 7.06995353e-02
-3.28679651e-01 -7.63202190e-01 2.75172114e-01 1.20771267e-01
7.93994308e-01 -7.66331732e-01 7.64861703e-01 1.11138783e-01
-1.17722988e+00 -2.85256863e-01 -1.83545679e-01 4.08270210e-01
6.74823858e-03 2.94337839e-01 -7.95335293e-01 8.99715662e-01
3.24711472e-01 2.64800012e-01 -5.95839024e-01 7.02344716e-01
-2.22302258e-01 6.32109284e-01 -4.13717747e-01 4.40516174e-01
-1.57469437e-02 -2.11346522e-02 7.28122950e-01 1.09455037e+00
2.26797074e-01 -1.45420492e-01 1.39662251e-01 4.64367211e-01
-2.17083380e-01 1.87336668e-01 -7.83091486e-01 -7.20274299e-02
5.51842332e-01 1.40752506e+00 -3.78957093e-01 -4.63918261e-02
-4.78584886e-01 9.22349811e-01 -4.77956273e-02 4.11022604e-01
-1.08110487e+00 -3.07574779e-01 6.24035299e-01 -6.26300871e-02
2.03833118e-01 2.61568606e-01 5.06094918e-02 -1.25186622e+00
2.33386010e-01 -1.06122577e+00 7.12806821e-01 -3.10290128e-01
-1.44061899e+00 5.23971081e-01 -1.50896320e-02 -1.22475076e+00
-1.13869406e-01 -7.61298299e-01 -7.23112881e-01 9.19107258e-01
-1.62056494e+00 -1.49202895e+00 -2.89215028e-01 1.08169150e+00
2.92797945e-02 -4.70450640e-01 6.83191657e-01 7.27791309e-01
-8.17087889e-01 1.17484117e+00 -2.53311127e-01 3.47876698e-01
9.72768605e-01 -5.49475789e-01 -8.26946944e-02 1.06341815e+00
-1.80880904e-01 7.36755788e-01 5.70672154e-01 -5.79208732e-01
-1.59744835e+00 -9.58555639e-01 6.92888021e-01 -1.81264162e-01
5.21454215e-01 -3.25313985e-01 -9.08769965e-01 3.16567838e-01
1.39075980e-01 2.52637267e-01 8.64941001e-01 -4.74424779e-01
-1.00474072e+00 -2.78234184e-01 -1.64798760e+00 4.45173860e-01
7.64684618e-01 -8.42485845e-01 -3.45933795e-01 2.87057877e-01
3.98450404e-01 -1.60809696e-01 -6.93905950e-01 5.07086277e-01
6.66202366e-01 -7.87062228e-01 1.10985184e+00 -5.40708601e-01
1.15899354e-01 -2.96785146e-01 -3.81144583e-01 -8.53700876e-01
9.09903497e-02 -6.50017023e-01 -4.28965800e-02 1.58602297e+00
-8.05773884e-02 -8.39398444e-01 4.79914784e-01 1.42960548e-01
3.95093203e-01 -3.54457617e-01 -1.00422621e+00 -7.82711387e-01
-1.93663299e-01 -8.04496706e-02 7.80610383e-01 9.64437664e-01
-1.13239691e-01 8.21326971e-02 -5.42335153e-01 6.47166967e-01
9.61255550e-01 -1.74370125e-01 7.27263331e-01 -9.50321019e-01
-2.45479941e-01 -3.02991033e-01 -6.69149280e-01 -5.82594514e-01
3.80490601e-01 -7.08037615e-01 -2.94726789e-01 -5.29307246e-01
2.19552517e-01 -6.71909899e-02 -4.24454749e-01 3.40282083e-01
-2.92924076e-01 8.23505878e-01 2.84472674e-01 6.05405159e-02
-2.68652558e-01 3.75952750e-01 9.77660000e-01 -1.27467006e-01
1.48523510e-01 1.79680839e-01 -6.87932849e-01 4.88249630e-01
7.80844331e-01 -3.93502235e-01 -3.15695703e-01 -2.06499010e-01
-2.95388043e-01 1.06273025e-01 6.09102726e-01 -9.89730060e-01
1.30521536e-01 -7.75901824e-02 4.46699321e-01 -1.40565753e-01
4.34092343e-01 -9.93718803e-01 -1.05596766e-01 4.16696429e-01
-1.23469442e-01 -8.58971924e-02 1.34947717e-01 5.72620451e-01
-1.29847094e-01 -2.34196614e-03 1.00829649e+00 3.26655805e-01
-2.99382985e-01 4.61955011e-01 1.09568559e-01 -1.72590092e-01
1.12419224e+00 -1.95877790e-01 -7.39133000e-01 -1.99962705e-01
-3.90948057e-01 -1.06160939e-01 2.71907389e-01 6.28818393e-01
7.33135462e-01 -1.49031174e+00 -8.03076386e-01 7.10697353e-01
2.53403783e-01 -8.91704142e-01 3.22473407e-01 7.76274145e-01
-8.41566250e-02 1.50958896e-01 -2.66191304e-01 -5.61875343e-01
-1.78938842e+00 6.42832816e-01 2.61503607e-01 -6.82862625e-02
-9.40646231e-02 6.35326028e-01 3.33284885e-01 -2.90249676e-01
1.98016226e-01 6.46046042e-01 -1.70117557e-01 2.14385465e-02
9.50970829e-01 2.18328506e-01 1.04369000e-01 -1.28630352e+00
-5.35961092e-01 5.58224082e-01 -1.26580879e-01 -2.16448337e-01
8.26379180e-01 -1.23069867e-01 -2.10760325e-01 -3.26392174e-01
1.50344288e+00 4.82498072e-02 -9.86273408e-01 -4.67926592e-01
-3.33356261e-01 -9.73345876e-01 -2.07100719e-01 -5.91013968e-01
-1.25271714e+00 9.88315105e-01 1.00463796e+00 1.72698677e-01
1.29951906e+00 -1.23388454e-01 7.69044876e-01 7.10749403e-02
2.07429856e-01 -6.21328473e-01 3.51434529e-01 1.70651928e-01
6.37375653e-01 -1.33641744e+00 -3.70106488e-01 -6.49688840e-01
-5.42200923e-01 9.21464741e-01 5.41876495e-01 1.50176138e-01
6.12597704e-01 7.73559231e-03 -1.89153254e-02 1.72622111e-02
-2.62715131e-01 9.09216851e-02 2.66256452e-01 8.35532904e-01
8.21081363e-03 -7.17069507e-02 -1.94448978e-01 3.90471667e-01
8.74686241e-02 -3.54846954e-01 5.04584871e-02 9.86521780e-01
-1.29973426e-01 -1.31693316e+00 -8.49070430e-01 -3.67722400e-02
-8.12649250e-01 2.11695701e-01 -4.40994322e-01 4.62867975e-01
3.01653028e-01 1.25032830e+00 -2.73659199e-01 -6.56800807e-01
1.23395622e-01 1.75136536e-01 5.09163618e-01 -2.00392872e-01
-5.91436565e-01 -8.58111493e-03 -3.54242623e-01 -3.49231482e-01
-4.57549721e-01 -5.76010942e-01 -7.58942366e-01 -6.11138999e-01
-3.44086230e-01 -3.51141356e-02 7.44204164e-01 9.49557424e-01
2.16116697e-01 4.49305587e-02 1.25560451e+00 -6.00384235e-01
-6.87751770e-01 -6.97605729e-01 -4.62544769e-01 8.74239624e-01
4.66088831e-01 -8.96154583e-01 -6.46534562e-01 3.67525592e-03] | [13.044256210327148, 1.1708290576934814] |
0504310b-f0d4-4f33-a256-13e6d46990b3 | synthesis-of-opacity-enforcing-winning | 2304.01286 | null | https://arxiv.org/abs/2304.01286v1 | https://arxiv.org/pdf/2304.01286v1.pdf | Synthesis of Opacity-Enforcing Winning Strategies Against Colluded Opponent | This paper studies a language-based opacity enforcement in a two-player, zero-sum game on a graph. In this game, player 1 (P1) wins if it can achieve a secret temporal goal described by the language of a finite automaton, no matter what strategy the opponent player 2 (P2) selects. In addition, P1 aims to win while making its goal opaque to a passive observer with imperfect information. However, P2 colludes with the observer to reveal P1's secret whenever P2 cannot prevent P1 from achieving its goal, and therefore, opacity must be enforced against P2. We show that a winning and opacity-enforcing strategy for P1 can be computed by reducing the problem to solving a reachability game augmented with observer's belief states. Furthermore, if such a strategy does not exist, winning for P1 must entail the price of revealing his secret to the observer. We demonstrate our game-theoretic solution of opacity-enforcement control through a small illustrative example and in a robot motion planning problem. | ['Jie Fu', 'Hazhar Rahmani', 'Abhishek N. Kulkarni', 'Chongyang Shi'] | 2023-04-03 | null | null | null | null | ['motion-planning'] | ['robots'] | [ 4.11234051e-01 1.22597146e+00 4.63492796e-02 4.32712376e-01
-6.41750693e-01 -9.12773252e-01 4.30611551e-01 -8.91495775e-03
-5.89601040e-01 8.34384680e-01 -3.92650329e-02 -6.60003304e-01
2.35987082e-01 -1.17144334e+00 -6.19023740e-01 -7.36638784e-01
-3.52566570e-01 4.05554295e-01 8.51712644e-01 -1.57576978e-01
4.44867695e-03 2.11022988e-01 -1.07940400e+00 -2.93076038e-01
-1.56196197e-02 4.38407898e-01 -1.04429387e-01 9.82984304e-01
2.74675906e-01 1.52924860e+00 -6.72988176e-01 -1.83437809e-01
4.94983047e-01 -5.23858547e-01 -1.24348092e+00 1.92286834e-01
-1.72416285e-01 -3.29624325e-01 -3.23263347e-01 1.61245441e+00
-2.95361102e-01 1.37736633e-01 4.15201932e-01 -1.69917202e+00
-4.50936735e-01 4.36547518e-01 -3.35767716e-01 -1.96854934e-01
3.58849078e-01 7.29264259e-01 1.11362708e+00 3.68798465e-01
8.20082188e-01 8.57056975e-01 -8.64469782e-02 1.01012433e+00
-1.26195133e+00 -1.89975813e-01 2.20348254e-01 -3.07219088e-01
-1.42732668e+00 -2.08607703e-01 6.21078193e-01 -2.51200646e-01
1.00836945e+00 8.24942589e-01 7.07351089e-01 3.80164295e-01
4.36357170e-01 3.00953537e-01 1.45700419e+00 -3.43482554e-01
4.84935969e-01 1.60709471e-01 -2.06683099e-01 9.31210876e-01
4.50073481e-01 6.86689198e-01 -5.35607457e-01 -5.60105324e-01
6.61043286e-01 -5.29825807e-01 -4.20701355e-01 -4.76950407e-01
-1.00076175e+00 6.78114891e-01 7.83186555e-02 1.66210294e-01
-2.95668185e-01 6.66212142e-01 -1.65469777e-02 9.27832544e-01
-7.77685419e-02 6.80400789e-01 -4.16185334e-03 -1.26978204e-01
-1.05511248e-01 2.99450606e-01 1.11604178e+00 7.35167384e-01
6.80796802e-01 2.28018790e-01 -1.15713859e-02 -5.17658293e-01
4.83097434e-01 5.63669145e-01 -4.09420222e-01 -1.59853697e+00
1.96231715e-02 5.04286230e-01 7.24730015e-01 -7.89641023e-01
-2.29570314e-01 2.39764422e-01 -3.35851699e-01 1.32253015e+00
6.45985305e-01 -1.62316501e-01 -5.36790788e-01 2.22617674e+00
2.51580000e-01 -5.05051762e-02 5.49167752e-01 1.16081762e+00
2.81458944e-01 9.12333965e-01 -1.66040674e-01 -5.74570656e-01
1.51041949e+00 -7.68550038e-01 -3.74390811e-01 -3.41102570e-01
6.95526242e-01 -1.55558512e-01 6.56055510e-01 3.57919931e-01
-1.43875515e+00 4.14191544e-01 -1.02671444e+00 1.17009349e-01
1.51897252e-01 -8.16632271e-01 3.21649075e-01 8.11649203e-01
-1.35365784e+00 1.84549615e-01 -7.77965188e-01 1.09736826e-02
-5.76290190e-02 6.81598425e-01 -5.78413904e-01 1.72628582e-01
-1.25675619e+00 8.85977626e-01 4.78188097e-01 -4.84294176e-01
-1.44907331e+00 2.77319074e-01 -8.36851656e-01 2.31332421e-01
1.31093609e+00 -7.75741458e-01 1.55113590e+00 -1.25964272e+00
-1.87503767e+00 1.06793046e+00 6.71040341e-02 -6.33660495e-01
5.82952857e-01 5.38061082e-01 -4.70719598e-02 2.68789262e-01
2.60988891e-01 2.15872452e-01 8.03564847e-01 -1.61318231e+00
-7.05922425e-01 -4.07558262e-01 1.29301584e+00 4.07417953e-01
2.97700644e-01 2.44375154e-01 -3.64773601e-01 2.49490768e-01
1.17128849e-01 -1.35281515e+00 -5.26465058e-01 1.75731733e-01
-5.98404646e-01 2.05099896e-01 5.99197507e-01 3.37060913e-02
7.63390362e-01 -2.22044468e+00 1.80404708e-02 4.32393909e-01
9.74219441e-01 5.29016145e-02 -9.54045877e-02 2.64898270e-01
3.23688567e-01 4.21381027e-01 -3.32969010e-01 -2.53941298e-01
2.42518440e-01 4.40814674e-01 -2.89604366e-01 8.76311779e-01
-1.34257987e-01 5.50470710e-01 -8.72673333e-01 -4.13780093e-01
-2.01315641e-01 -1.10964902e-01 -8.83097172e-01 2.00004593e-01
-6.12322390e-01 5.14946103e-01 -6.52151287e-01 1.32839888e-01
4.42221314e-01 -3.60290736e-01 7.02662826e-01 8.49598646e-01
-3.94810200e-01 2.21278891e-01 -1.20649409e+00 8.03153396e-01
2.56847385e-02 3.62884730e-01 7.38768101e-01 -1.63469106e-01
2.00867966e-01 6.59771323e-01 1.55234486e-02 -3.90034348e-01
6.87879324e-01 -1.25815282e-02 1.97597682e-01 9.16944370e-02
7.04173684e-01 -5.86552739e-01 -4.58936512e-01 9.87879038e-01
-6.39314532e-01 -4.16794240e-01 -2.94449627e-01 5.72648764e-01
1.46575630e+00 -1.79222360e-01 4.71578032e-01 -1.53331473e-01
7.68907845e-01 1.45364955e-01 6.78036094e-01 1.27158296e+00
-4.60792482e-01 1.06650047e-01 1.24279928e+00 -4.99985516e-02
-6.82653606e-01 -8.07904541e-01 1.01428652e+00 6.73022866e-01
6.38114095e-01 -4.86904144e-01 -7.72053421e-01 -3.62962157e-01
-3.40990305e-01 7.91716814e-01 -6.54307485e-01 -2.76607692e-01
-3.67706150e-01 3.93677324e-01 4.96998638e-01 -1.54403433e-01
8.25698435e-01 -1.11012387e+00 -1.17774248e+00 5.90220951e-02
-1.58120394e-01 -8.82285297e-01 -6.63222432e-01 8.07946920e-02
-4.61998373e-01 -1.34996295e+00 1.84047118e-01 -4.11812752e-01
8.77723932e-01 4.74263668e-01 6.64183974e-01 7.38085210e-01
6.31588399e-01 6.59104764e-01 1.25719324e-01 -4.25191253e-01
-9.12352562e-01 -2.76692808e-01 -1.53513983e-01 -2.20859766e-01
-9.33633819e-02 1.69549882e-02 -5.66888638e-02 1.19431145e-01
-9.58667815e-01 1.12499274e-01 -1.79324776e-01 3.44618887e-01
5.83172262e-01 5.50188303e-01 -6.20576680e-01 -9.71086085e-01
5.69926143e-01 -2.01925084e-01 -1.52437043e+00 3.23066637e-02
-1.71362549e-01 5.58154099e-02 6.57077312e-01 -1.60133600e-01
-9.12779033e-01 -5.48353978e-02 6.68944895e-01 -8.65913406e-02
1.31688491e-01 4.71527517e-01 -5.20303905e-01 -3.82334828e-01
4.68888849e-01 1.84897885e-01 1.94248453e-01 2.81273276e-01
2.78613746e-01 -3.95229794e-02 6.69052482e-01 -8.93702805e-01
1.00508678e+00 6.92268670e-01 6.19033515e-01 -5.25027096e-01
-2.35509619e-01 1.46357328e-01 7.10048229e-02 -2.95479029e-01
8.75750482e-01 -5.84067047e-01 -1.75778973e+00 5.99644065e-01
-1.34738445e+00 -5.48812568e-01 -5.79605341e-01 3.00053833e-03
-8.97362232e-01 5.29470444e-01 -2.28184357e-01 -1.59603190e+00
3.31098437e-01 -1.36807692e+00 4.11772519e-01 3.46403904e-02
-2.42825091e-01 -5.12082517e-01 1.49445400e-01 2.09057644e-01
2.73334771e-01 1.56169772e-01 7.29505479e-01 -4.57172245e-01
-1.47498274e+00 -3.37789416e-01 1.04102962e-01 -5.96359707e-02
2.75535788e-03 3.00213099e-02 -7.90824294e-01 -4.12429065e-01
2.79965937e-01 -2.07194582e-01 8.44153985e-02 -3.38697359e-02
5.65538108e-01 -8.73602808e-01 -1.19297907e-01 3.81860025e-02
1.42322564e+00 4.29455459e-01 5.94597936e-01 5.07216096e-01
3.00038904e-01 3.49683642e-01 3.34258735e-01 9.62175578e-02
4.88323510e-01 6.27425730e-01 9.79221821e-01 2.34949782e-01
4.36606795e-01 -1.33220524e-01 4.02890503e-01 -2.33095348e-01
-6.32693768e-01 -3.99936110e-01 -5.48979461e-01 3.39783967e-01
-1.92443931e+00 -1.12070107e+00 -3.14439135e-03 2.41460633e+00
7.45545864e-01 5.58389843e-01 5.46509922e-02 -7.32601807e-02
3.86335909e-01 4.60708410e-01 -3.64366651e-01 -7.00064003e-01
-1.46362588e-01 -7.32145011e-02 9.50550139e-01 1.30708098e+00
-6.23676598e-01 1.18383920e+00 6.09434271e+00 3.48759919e-01
-8.21430087e-01 5.10746360e-01 2.45597288e-01 -1.37144119e-01
-6.67153716e-01 7.88090050e-01 -1.85153097e-01 1.49333149e-01
8.62489879e-01 -7.14439094e-01 8.56030107e-01 6.96346700e-01
-9.38339159e-03 -8.81969512e-01 -9.19514835e-01 2.76179194e-01
-1.98670194e-01 -1.35746813e+00 -4.54428613e-01 4.51523215e-01
5.82611024e-01 -2.82747328e-01 -1.54746342e-02 1.26416668e-01
1.16304135e+00 -8.21923614e-01 1.26772702e+00 5.57720028e-02
4.78693575e-01 -8.87234092e-01 3.10114235e-01 8.05456996e-01
-1.04178786e+00 -1.32569715e-01 1.07309125e-01 -6.72687232e-01
2.41911396e-01 -3.44514310e-01 -4.37442303e-01 3.89338583e-01
1.98909745e-01 -6.95448518e-01 4.75729614e-01 5.84862173e-01
-7.66414702e-01 2.44788930e-01 -4.39142197e-01 4.59078811e-02
5.54372907e-01 -5.32557845e-01 1.01670933e+00 2.81744331e-01
2.52863280e-02 1.16062963e+00 4.61046457e-01 9.79446650e-01
-1.01907551e-01 -3.89981687e-01 -9.36237276e-01 -2.62204498e-01
2.85582423e-01 6.27396047e-01 -8.29789698e-01 -5.41435003e-01
-4.35377151e-01 1.01356936e+00 2.13987201e-01 4.52329516e-01
-7.64194191e-01 4.76923771e-02 1.12371039e+00 1.03991352e-01
-1.28192067e-01 -5.01094401e-01 -9.60570723e-02 -7.70318687e-01
-2.55823702e-01 -9.74608719e-01 4.88996118e-01 -1.08634639e+00
-2.86520630e-01 6.11282647e-01 -1.71454370e-01 -7.40931034e-01
-9.62031707e-02 -3.46741527e-01 -6.84810281e-01 1.01983988e+00
-1.32228923e+00 -8.10487449e-01 3.09207290e-01 7.93797910e-01
-9.76234451e-02 9.86104310e-02 9.69366372e-01 -6.63468242e-01
-1.37441173e-01 -7.17694461e-02 -7.36995578e-01 -8.98254141e-02
-1.63494214e-01 -1.08897281e+00 3.18694949e-01 1.45834184e+00
-1.54658943e-01 6.33069932e-01 9.52253044e-01 -8.70474160e-01
-1.78245544e+00 -7.14683533e-01 1.12236178e+00 -5.26297033e-01
9.69741106e-01 4.52957153e-02 -6.65439248e-01 1.14654291e+00
2.25442305e-01 -2.30155841e-01 2.45415345e-01 -4.05578911e-01
-4.11900222e-01 5.15929163e-01 -1.40843415e+00 1.13014352e+00
8.82542312e-01 -1.00901258e+00 -4.30833161e-01 2.09218144e-01
8.28848660e-01 -8.05234671e-01 -1.48868188e-01 -1.47429869e-01
9.37756598e-02 -9.28193092e-01 3.50541085e-01 -7.59570956e-01
6.18522204e-02 -1.05170846e+00 -1.12570293e-01 -8.10457647e-01
-5.02307773e-01 -1.31950653e+00 1.15564048e-01 2.66345143e-01
2.45918423e-01 -1.04082632e+00 9.95371997e-01 1.33486366e+00
1.93895668e-01 4.28026430e-02 -1.59108746e+00 -9.42480922e-01
1.05463929e-01 -7.09340394e-01 6.24730647e-01 4.90585834e-01
4.02496755e-01 8.10390115e-02 -6.42554104e-01 1.00140166e+00
4.91077483e-01 7.53855472e-03 9.33350325e-01 -8.72068465e-01
-6.51000142e-01 -2.95163780e-01 -1.42297387e-01 -8.07345986e-01
2.87151843e-01 -5.46644270e-01 4.37184036e-01 -1.22992587e+00
1.25939012e-01 -2.08340451e-01 1.56311214e-01 7.83982277e-01
2.11484835e-01 1.74160644e-01 7.63962388e-01 2.83979833e-01
-7.42238045e-01 1.40579179e-01 1.50772583e+00 -4.70116735e-01
-1.10590339e-01 2.12642342e-01 -8.66167724e-01 7.52241313e-01
6.82581425e-01 -6.28008723e-01 -5.65594792e-01 -1.61045920e-02
6.18876755e-01 1.00007641e+00 4.70310450e-01 -4.97273237e-01
4.31238115e-01 -8.71525109e-01 -9.36566353e-01 2.18044564e-01
5.75379252e-01 -1.07057130e+00 7.36562192e-01 1.03412068e+00
-1.14412107e-01 -1.38830647e-01 1.38008920e-02 5.80728590e-01
1.43128127e-01 -3.66707176e-01 7.72209466e-01 -3.58339995e-01
-7.59261787e-01 2.44907364e-01 -1.35497606e+00 -1.17031962e-01
1.15146780e+00 -3.15646321e-01 -5.29435158e-01 -8.30721557e-01
-9.79646564e-01 3.27568293e-01 1.33214521e+00 -2.86478311e-01
5.10608017e-01 -8.45652282e-01 -2.25336537e-01 1.24496795e-01
1.28059372e-01 -1.27617344e-01 1.48771405e-01 5.36047757e-01
-4.03800040e-01 1.38701499e-01 -6.74152747e-02 1.58444390e-01
-1.82371199e+00 8.17152321e-01 6.04516208e-01 -4.42647457e-01
-3.79487962e-01 8.74124348e-01 5.59651375e-01 1.12712421e-01
-1.52605385e-01 -2.10377097e-01 1.00981854e-01 -8.29412878e-01
5.51690102e-01 -7.80705512e-02 -5.55146873e-01 -8.40206623e-01
-4.69927311e-01 2.11760625e-01 2.17422634e-01 -8.93466175e-01
5.98329425e-01 -2.60489404e-01 -5.28880239e-01 -9.55829863e-03
4.12557989e-01 4.37115133e-01 -1.08597577e+00 -2.89825112e-01
-3.60222012e-01 -6.83519483e-01 -8.76411572e-02 -4.06034499e-01
-1.21266806e+00 1.39408156e-01 -1.06986396e-01 6.92835510e-01
8.00456226e-01 1.32067472e-01 1.70299172e-01 6.60519660e-01
1.33479428e+00 -8.38386476e-01 -3.85174379e-02 7.78668523e-01
5.83748460e-01 -8.10675144e-01 -1.74537450e-01 -7.02044845e-01
-8.77763331e-01 5.07382691e-01 5.97263217e-01 2.96823289e-02
1.90964669e-01 5.28772175e-01 3.36812511e-02 -4.98828948e-01
-8.66289556e-01 -2.50500470e-01 -4.13927615e-01 7.75630891e-01
-5.23512185e-01 4.72428083e-01 -4.73137721e-02 3.02707851e-01
-1.89940959e-01 1.09776944e-01 1.32163703e+00 1.37613285e+00
-8.46436977e-01 -1.23326230e+00 -5.57244122e-01 -2.76727587e-01
-3.27809900e-01 -7.56972134e-02 -6.12388015e-01 8.91119778e-01
-5.11602052e-02 9.86415207e-01 -7.08821788e-02 -1.29095033e-01
6.98959306e-02 -4.07285154e-01 5.32239974e-01 -1.02571225e+00
-4.56131041e-01 -1.92336932e-01 3.96639526e-01 -7.01094449e-01
-3.99451405e-01 -3.13746840e-01 -1.66476643e+00 -8.55106831e-01
-1.70808002e-01 4.18080896e-01 2.92113405e-02 8.53671730e-01
-1.44335806e-01 1.85451075e-01 5.24303854e-01 3.72150764e-02
-2.09655911e-01 3.40701155e-02 -9.17073965e-01 -2.46992797e-01
5.11499822e-01 -2.80186981e-01 -6.63562119e-01 2.02174727e-02] | [4.402095794677734, 2.091395616531372] |
351e5f1c-5543-454d-82ac-89a8fea3afcb | self-governing-neural-networks-for-on-device | null | null | https://aclanthology.org/D18-1105 | https://aclanthology.org/D18-1105.pdf | Self-Governing Neural Networks for On-Device Short Text Classification | Deep neural networks reach state-of-the-art performance for wide range of natural language processing, computer vision and speech applications. Yet, one of the biggest challenges is running these complex networks on devices such as mobile phones or smart watches with tiny memory footprint and low computational capacity. We propose on-device Self-Governing Neural Networks (SGNNs), which learn compact projection vectors with local sensitive hashing. The key advantage of SGNNs over existing work is that they surmount the need for pre-trained word embeddings and complex networks with huge parameters. We conduct extensive evaluation on dialog act classification and show significant improvement over state-of-the-art results. Our findings show that SGNNs are effective at capturing low-dimensional semantic text representations, while maintaining high accuracy. | ['Zornitsa Kozareva', 'Sujith Ravi'] | 2018-10-01 | self-governing-neural-networks-for-on-device-1 | https://aclanthology.org/D18-1092 | https://aclanthology.org/D18-1092.pdf | emnlp-2018-10 | ['dialog-act-classification', 'dialogue-act-classification'] | ['natural-language-processing', 'natural-language-processing'] | [-1.86267406e-01 1.26554817e-01 -3.23463500e-01 -5.89871228e-01
-4.05333847e-01 -3.30384493e-01 6.79933131e-01 3.19652483e-02
-9.18083489e-01 4.02313054e-01 6.34050012e-01 -2.88705766e-01
1.95286304e-01 -7.54881144e-01 -2.00523794e-01 -4.68310535e-01
2.71713406e-01 5.33681452e-01 2.81873822e-01 -4.34732407e-01
1.99552272e-02 2.70519078e-01 -1.00771928e+00 2.08787605e-01
2.31366530e-01 1.16727316e+00 -1.06287740e-01 6.81419790e-01
-1.02492198e-01 4.61130500e-01 -4.08937663e-01 -5.90143621e-01
2.24365041e-01 2.64082432e-01 -5.01447082e-01 -4.33405191e-01
4.41961735e-01 -7.89329767e-01 -1.01014173e+00 8.87807131e-01
8.42809379e-01 2.86609083e-01 4.12544727e-01 -1.18094492e+00
-9.75911200e-01 5.77479064e-01 -1.41924217e-01 1.80703774e-01
-4.07622382e-02 3.08937699e-01 1.22254109e+00 -9.82521355e-01
2.13535428e-02 1.47073293e+00 9.80409622e-01 9.38194096e-01
-8.96097600e-01 -6.85236573e-01 -7.61924684e-02 5.58957681e-02
-1.17463410e+00 -7.53216386e-01 4.35093403e-01 9.25077423e-02
1.58590627e+00 -2.17550114e-01 2.45532826e-01 1.24593174e+00
2.13515848e-01 7.91347742e-01 6.73247933e-01 -1.31034970e-01
3.86200935e-01 2.64468610e-01 4.52420592e-01 7.85744786e-01
2.71009892e-01 -4.02749360e-01 -6.15367532e-01 -3.83649141e-01
5.93998849e-01 6.55037045e-01 -3.15705277e-02 -1.61804166e-02
-7.38206089e-01 1.23344684e+00 5.29703498e-01 2.21435085e-01
-4.34557319e-01 1.97570890e-01 7.41002619e-01 -1.32890791e-02
2.82692462e-01 -1.08163934e-02 -5.57022274e-01 -5.69692850e-01
-4.99189824e-01 -1.20458230e-02 9.30596471e-01 5.77114880e-01
6.40260994e-01 1.81014851e-01 -1.84387919e-02 1.08974719e+00
3.39476883e-01 5.48866212e-01 1.05640876e+00 -6.06197834e-01
6.20544732e-01 7.45606422e-01 -3.24631095e-01 -9.12903011e-01
-3.82283956e-01 2.62745246e-02 -1.16339445e+00 -1.51877239e-01
6.54392838e-02 -2.87042648e-01 -8.40048969e-01 1.61638832e+00
8.71069580e-02 -7.32690543e-02 3.10689867e-01 6.64505184e-01
8.54105234e-01 7.83355534e-01 2.67186135e-01 3.75757396e-01
1.66398335e+00 -1.16095352e+00 -7.25281119e-01 -5.93966722e-01
5.66459537e-01 -2.88548648e-01 1.42826355e+00 7.91126862e-02
-8.73921573e-01 -4.73495781e-01 -1.13841271e+00 -5.47558248e-01
-6.20564699e-01 1.81662947e-01 1.03947818e+00 8.62885654e-01
-1.02558148e+00 4.85226244e-01 -9.83839214e-01 -4.72140163e-01
7.60842443e-01 8.44292581e-01 -4.31096613e-01 9.46542062e-03
-1.26655400e+00 7.87382066e-01 3.45791012e-01 6.41430775e-03
-6.54854476e-01 -5.20422757e-01 -9.68239903e-01 5.59426248e-01
9.84095313e-05 -2.33119920e-01 1.28584981e+00 -2.00156689e-01
-1.82731402e+00 7.39180923e-01 2.27958225e-02 -8.67576778e-01
-1.75497010e-01 -5.81582189e-01 -2.79899359e-01 1.72580630e-01
-5.23344815e-01 9.46508884e-01 6.82772756e-01 -3.93461436e-01
-8.27985629e-02 -4.50093478e-01 8.87356624e-02 2.51096666e-01
-1.50821447e+00 -8.55264999e-03 -2.87629366e-01 -3.03381860e-01
-1.64888918e-01 -1.11268675e+00 -9.88499820e-02 1.63872764e-01
-2.95080513e-01 -6.70991898e-01 1.42118573e+00 -4.45069134e-01
1.12038410e+00 -2.26781082e+00 -2.37768650e-01 -4.18842077e-01
3.12500477e-01 9.55802798e-01 2.97630671e-02 4.12042409e-01
4.21100289e-01 -1.02378227e-01 1.88250262e-02 -6.32405460e-01
4.15660024e-01 4.66016829e-01 -5.19101679e-01 4.25553322e-01
-5.72389513e-02 1.22707355e+00 -2.61806101e-01 -2.01592103e-01
4.73848909e-01 8.43594134e-01 -6.33136928e-01 1.53130099e-01
-1.86790749e-01 -4.38976347e-01 -2.71319985e-01 4.29764092e-01
5.29058993e-01 -4.59744275e-01 1.17236726e-01 -3.17046583e-01
3.82577896e-01 7.73304582e-01 -8.45669210e-01 1.71843517e+00
-5.94296277e-01 6.55459821e-01 2.52764791e-01 -9.96676266e-01
8.54027867e-01 4.37955976e-01 8.42396840e-02 -5.56073725e-01
4.95795637e-01 -1.44505769e-01 -2.51603901e-01 -3.15226465e-01
6.11228824e-01 -3.09135169e-01 -2.24682584e-01 7.40531564e-01
3.74058247e-01 3.20887953e-01 -3.69516641e-01 2.71903604e-01
1.10401249e+00 -6.99057341e-01 3.29791367e-01 -5.09656221e-02
4.78484809e-01 -5.27745187e-01 3.43102813e-01 4.42458630e-01
-5.51424742e-01 3.75360459e-01 2.22194090e-01 -7.14143932e-01
-8.83280337e-01 -9.88248110e-01 1.16344281e-01 1.69447386e+00
-2.96585917e-01 -5.35710871e-01 -7.91623712e-01 -6.32999539e-01
3.37736532e-02 1.70956686e-01 -3.59375447e-01 -3.46826732e-01
-6.25384212e-01 -7.49898612e-01 1.07929647e+00 9.09180284e-01
1.03183472e+00 -1.06416845e+00 -5.92401803e-01 -4.52182032e-02
1.10704735e-01 -1.40037668e+00 -4.99386370e-01 1.41081318e-01
-9.95006144e-01 -4.54344392e-01 -6.16281450e-01 -9.90211189e-01
4.90176916e-01 4.47483391e-01 7.59431839e-01 -3.86325605e-02
-2.63970286e-01 1.02108784e-01 4.50423658e-02 -1.99951544e-01
4.89510149e-02 4.64567155e-01 5.29436827e-01 -1.05824023e-01
9.70356643e-01 -7.24998176e-01 -7.24649727e-01 1.47102550e-01
-1.00757039e+00 -4.63925421e-01 7.03175902e-01 9.99644816e-01
4.56759185e-02 7.11598843e-02 5.07551193e-01 -7.65107572e-01
1.03835106e+00 -2.45559111e-01 -3.31127495e-01 1.13992736e-01
-7.14084744e-01 3.23735654e-01 8.65487576e-01 -6.34034932e-01
-9.98482883e-01 -5.65429293e-02 -3.85202914e-01 -2.45576739e-01
-1.11715458e-01 1.92524493e-01 -1.12601265e-01 1.58410162e-01
7.06708908e-01 1.60943314e-01 8.80546346e-02 -4.90744144e-01
4.59482789e-01 1.31026518e+00 4.37433302e-01 -3.37954253e-01
5.05734503e-01 4.89220530e-01 -2.49343261e-01 -9.27432477e-01
-6.73307955e-01 -5.82367599e-01 -4.71048534e-01 4.43391055e-01
9.25939143e-01 -1.10228038e+00 -1.12656665e+00 5.84800482e-01
-1.04434669e+00 -2.04877436e-01 -2.93803215e-03 4.06564265e-01
-1.07222132e-01 2.14299887e-01 -1.11793244e+00 -8.52627993e-01
-9.98147607e-01 -8.94342184e-01 1.05816805e+00 5.26845515e-01
-1.91352010e-01 -1.17555523e+00 5.56919053e-02 5.40060282e-01
8.93605113e-01 -5.40485740e-01 9.73297894e-01 -1.02946210e+00
-5.46661735e-01 -3.66044730e-01 -4.05734062e-01 6.57425106e-01
1.32341728e-01 -6.47779226e-01 -1.36458278e+00 -4.13891286e-01
8.53944048e-02 -8.79078329e-01 8.90441716e-01 2.35870942e-01
8.83779526e-01 -6.96758330e-01 -2.89574355e-01 4.20622945e-01
1.35338318e+00 3.74470912e-02 5.10895193e-01 -5.15064150e-02
8.28907132e-01 2.41632193e-01 -7.32795969e-02 4.86456752e-01
4.08273667e-01 5.71842074e-01 1.28033504e-01 1.90807849e-01
1.07899748e-01 -3.18762273e-01 4.33266371e-01 1.09666133e+00
5.14394999e-01 -2.64999539e-01 -9.11892831e-01 6.54642344e-01
-1.90988624e+00 -7.29551852e-01 5.92071176e-01 1.91935551e+00
8.15040767e-01 3.03637981e-01 1.39378995e-01 4.47809473e-02
4.95088458e-01 4.77966219e-01 -7.61529744e-01 -5.61087668e-01
9.35134292e-02 4.53524590e-01 5.17787695e-01 5.17556190e-01
-1.10815299e+00 1.15195704e+00 6.26289177e+00 6.81077421e-01
-1.30304492e+00 3.96634310e-01 7.26038933e-01 -3.27791393e-01
1.27915397e-01 -4.38985139e-01 -1.14779449e+00 1.67696580e-01
1.53224313e+00 2.33930051e-01 4.48089689e-01 1.26756072e+00
-3.14429551e-01 2.56716758e-01 -8.67258966e-01 1.16482639e+00
2.87026882e-01 -1.44485450e+00 3.98986116e-02 2.86336452e-01
3.63482147e-01 4.31180626e-01 3.60222787e-01 5.77684402e-01
5.11288643e-01 -1.30149543e+00 -1.25352070e-01 -2.17057481e-01
6.87696815e-01 -6.20675385e-01 8.49126101e-01 2.16074452e-01
-9.45670605e-01 -2.46868297e-01 -8.58709991e-01 -3.66780609e-01
2.06695914e-01 4.75822449e-01 -9.55754340e-01 -4.34042245e-01
6.52390599e-01 5.70395172e-01 -3.54177445e-01 2.19535485e-01
1.68992013e-01 7.00705588e-01 -6.11409724e-01 -5.85220993e-01
4.60075289e-01 1.20850444e-01 -9.33636799e-02 1.24814785e+00
5.59338331e-02 1.52494088e-01 -3.42765711e-02 4.10146654e-01
-5.66850662e-01 -1.36284962e-01 -8.17441702e-01 -4.61659431e-01
7.21355379e-01 1.25234640e+00 -5.29522181e-01 -4.05972540e-01
-4.65074807e-01 1.13768053e+00 5.45618117e-01 -2.78592557e-02
-6.01966619e-01 -6.24939322e-01 1.22145331e+00 -2.00508356e-01
4.76044655e-01 -4.79109019e-01 -4.58275765e-01 -1.08077574e+00
1.21488065e-01 -6.09479308e-01 2.37874001e-01 -3.79989177e-01
-1.25125182e+00 6.81365609e-01 -3.63845646e-01 -6.64178967e-01
-2.01705843e-01 -1.09231949e+00 -5.76903284e-01 4.56079662e-01
-1.14384186e+00 -1.13103032e+00 -1.91510230e-01 7.54004180e-01
6.24795556e-01 -5.41030109e-01 1.28871942e+00 2.87441671e-01
-6.43142998e-01 8.51758718e-01 1.67545035e-01 5.53807974e-01
6.08396530e-01 -9.61241722e-01 7.84596503e-01 3.62139076e-01
2.33400032e-01 1.08453834e+00 2.38925353e-01 -2.36268386e-01
-1.65715587e+00 -8.43851507e-01 9.40564334e-01 -3.60780358e-01
5.75164139e-01 -9.80224788e-01 -8.46574783e-01 5.76506197e-01
4.89095002e-01 1.85080692e-01 9.51026618e-01 2.98874080e-01
-7.07409620e-01 -2.53770739e-01 -1.21022666e+00 6.46651566e-01
7.89206982e-01 -1.20872462e+00 -5.76401949e-01 4.69228864e-01
9.89262640e-01 -2.52476066e-01 -6.71081781e-01 1.84177719e-02
7.37447739e-01 -8.49235654e-01 1.22531247e+00 -6.33496165e-01
3.60877514e-01 3.27213079e-01 -4.46536869e-01 -9.41813231e-01
-6.00283556e-02 -8.09076309e-01 -2.67054528e-01 9.93471682e-01
2.86167473e-01 -7.16339588e-01 1.29680860e+00 1.08093810e+00
1.37836918e-01 -8.74438107e-01 -9.52525556e-01 -5.53084850e-01
9.12912562e-02 -1.32949844e-01 4.31317091e-01 8.90601754e-01
3.37144017e-01 9.90809739e-01 -6.01857901e-01 -1.99809164e-01
3.00534159e-01 -4.80725646e-01 5.98365009e-01 -1.23253953e+00
-2.46378496e-01 -8.59424770e-02 -5.93911350e-01 -1.44787180e+00
3.48166347e-01 -4.12264079e-01 8.88342932e-02 -1.42974222e+00
1.83383167e-01 -2.24844590e-01 -3.72648835e-01 7.61678696e-01
-1.63552240e-02 3.38710129e-01 1.76190555e-01 -7.63917435e-03
-8.86758149e-01 9.56014693e-01 4.32013899e-01 -2.40334958e-01
-9.28461552e-02 -3.40367436e-01 -8.58202279e-01 6.26787186e-01
9.82716441e-01 -4.91125286e-01 -6.58897519e-01 -5.93787551e-01
-4.30611856e-02 -2.96969056e-01 1.86874792e-01 -1.08242011e+00
5.87340832e-01 3.22370708e-01 3.09936479e-02 -4.06785011e-01
9.65127289e-01 -7.81635880e-01 -6.80808306e-01 5.26349485e-01
-5.45162439e-01 2.99441993e-01 3.20018202e-01 5.90487719e-01
-1.23879567e-01 -2.12820023e-01 7.32867301e-01 -7.80020133e-02
-4.02358919e-01 2.37689495e-01 -3.58392537e-01 4.60871309e-02
5.56656003e-01 -4.15076725e-02 -7.69371212e-01 -7.11487770e-01
-1.51615620e-01 -1.72700956e-01 1.02405764e-01 5.35954535e-01
7.06981719e-01 -1.34475946e+00 -1.44728571e-01 5.00612259e-01
-6.23337105e-02 6.19104458e-03 3.10257822e-01 3.34545076e-01
-3.61390114e-01 1.12702405e+00 -2.14503422e-01 -3.38890940e-01
-1.24265516e+00 2.15801775e-01 -2.42157467e-02 -4.07838941e-01
-6.54515862e-01 1.07224977e+00 1.12966321e-01 -8.24511588e-01
6.25813246e-01 -4.67904985e-01 5.82056530e-02 -4.37311620e-01
8.02756369e-01 1.94443300e-01 4.17731628e-02 -4.48035955e-01
-3.30554694e-01 2.26458639e-01 -5.89152694e-01 -1.79570675e-01
1.45463121e+00 1.58930108e-01 1.06000587e-01 2.38804057e-01
1.55624509e+00 -3.63630712e-01 -1.26671994e+00 -6.53242409e-01
-4.29989874e-01 -5.51560633e-02 3.60478699e-01 -4.80193317e-01
-1.07984543e+00 1.32589245e+00 9.02589142e-01 -3.15268748e-02
7.72860885e-01 -1.77361906e-01 1.71792567e+00 1.27638316e+00
9.36419964e-02 -1.29320443e+00 4.71173435e-01 8.97195041e-01
3.30478340e-01 -1.40073860e+00 -4.93012257e-02 1.37077808e-01
-8.70151639e-01 8.92281055e-01 5.98918319e-01 -1.64719254e-01
1.00811172e+00 3.45892519e-01 1.50532037e-01 -3.39197576e-01
-8.72132897e-01 2.44600058e-01 -4.57044840e-02 5.57144642e-01
2.95942515e-01 -4.87740058e-03 1.70493409e-01 6.06660843e-01
-2.91973531e-01 -1.44730493e-01 7.25991875e-02 9.67066884e-01
-6.60991728e-01 -1.09014308e+00 1.25060648e-01 4.63768780e-01
-6.16350293e-01 -4.93633121e-01 -3.80409837e-01 3.48189980e-01
-5.82975507e-01 7.89722443e-01 1.10549107e-01 -6.98465586e-01
2.78970804e-02 4.84904289e-01 -6.83372021e-02 -5.86663961e-01
-7.21905708e-01 -5.60295761e-01 7.55885392e-02 -4.15620685e-01
-1.20548934e-01 -1.86643764e-01 -1.26972151e+00 -5.97641945e-01
-2.88018852e-01 -7.60852396e-02 9.10977423e-01 9.84428704e-01
7.99825847e-01 3.78020220e-02 2.60868549e-01 -6.61583304e-01
-1.05524552e+00 -1.23778403e+00 -2.24816263e-01 4.38857079e-01
3.13028932e-01 -3.70723009e-01 -2.25787997e-01 -2.22839236e-01] | [10.798141479492188, 7.905714511871338] |
fc97a451-9a19-4726-b60c-f531cc992d26 | deep-learning-and-image-super-resolution | 2305.13929 | null | https://arxiv.org/abs/2305.13929v1 | https://arxiv.org/pdf/2305.13929v1.pdf | Deep Learning and Image Super-Resolution-Guided Beam and Power Allocation for mmWave Networks | In this paper, we develop a deep learning (DL)-guided hybrid beam and power allocation approach for multiuser millimeter-wave (mmWave) networks, which facilitates swift beamforming at the base station (BS). The following persisting challenges motivated our research: (i) User and vehicular mobility, as well as redundant beam-reselections in mmWave networks, degrade the efficiency; (ii) Due to the large beamforming dimension at the BS, the beamforming weights predicted by the cutting-edge DL-based methods often do not suit the channel distributions; (iii) Co-located user devices may cause a severe beam conflict, thus deteriorating system performance. To address the aforementioned challenges, we exploit the synergy of supervised learning and super-resolution technology to enable low-overhead beam- and power allocation. In the first step, we propose a method for beam-quality prediction. It is based on deep learning and explores the relationship between high- and low-resolution beam images (energy). Afterward, we develop a DL-based allocation approach, which enables high-accuracy beam and power allocation with only a portion of the available time-sequential low-resolution images. Theoretical and numerical results verify the effectiveness of our proposed | ['Tony Q. S. Quek', 'Setareh Maghsudi', 'Tomoaki Ohtsuki', 'Yuwen Cao'] | 2023-05-08 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [-1.31159902e-01 -1.86242864e-01 -1.58820584e-01 -9.20426399e-02
-7.07946777e-01 -2.73979604e-01 5.44611663e-02 -4.43658888e-01
-8.35853964e-02 9.40523446e-01 3.36564869e-01 -5.50722301e-01
-6.94765449e-01 -1.11589742e+00 -2.82610297e-01 -1.15569484e+00
-2.73629993e-01 2.28172243e-02 -1.57506526e-01 -1.39102727e-01
-1.07530013e-01 6.37538612e-01 -8.45406771e-01 -8.36953223e-02
8.67268920e-01 1.14949679e+00 5.25127590e-01 5.36098540e-01
3.08079291e-02 5.14388919e-01 -3.23704600e-01 -1.57696202e-01
3.18126023e-01 -3.64097863e-01 1.46490438e-02 -5.29329740e-02
9.30589959e-02 -4.63731557e-01 -6.93594754e-01 9.02734220e-01
1.15821719e+00 -3.60160410e-01 4.78494018e-01 -1.05870581e+00
-1.92738950e-01 5.70660412e-01 -6.78359866e-01 3.23145807e-01
1.42396586e-02 -8.07375554e-03 6.39681816e-01 -9.32937145e-01
2.55807161e-01 7.38859236e-01 8.11164558e-01 3.01716685e-01
-5.96179545e-01 -1.16956639e+00 -1.81774609e-02 3.41021448e-01
-1.52126682e+00 -6.74837530e-01 9.01017845e-01 -1.70590580e-01
1.77961081e-01 3.44259739e-01 8.78178716e-01 6.54252052e-01
5.07111013e-01 4.18386638e-01 6.82177603e-01 -4.03549939e-01
3.02788943e-01 -3.07711542e-01 -5.45366824e-01 9.22312975e-01
5.23807108e-01 2.58976728e-01 -3.35869074e-01 -1.24227814e-01
1.11659038e+00 2.07768455e-02 -5.38766265e-01 -5.20396233e-01
-1.02047884e+00 6.14183724e-01 6.62237346e-01 5.90995371e-01
-4.34662402e-01 8.13661367e-02 -3.75376105e-01 2.99855143e-01
2.33083621e-01 3.46479177e-01 -2.92408556e-01 3.16361070e-01
-1.13953876e+00 4.04531485e-04 5.00142515e-01 1.09740245e+00
5.22970200e-01 2.75114745e-01 -4.00040895e-01 6.04890108e-01
7.41663694e-01 8.06257188e-01 -5.93316481e-02 -6.30817771e-01
7.43029416e-01 -8.03663135e-02 3.38296413e-01 -1.07316124e+00
-1.00072062e+00 -1.62086260e+00 -1.63689458e+00 -1.15251672e-02
2.53898293e-01 -7.00039446e-01 -7.86279142e-01 1.70215678e+00
1.37742255e-02 3.65855485e-01 3.84262539e-02 9.12547469e-01
8.83644283e-01 6.15621030e-01 -1.82768643e-01 -1.00930893e+00
8.93619776e-01 -5.75586498e-01 -8.25975716e-01 -2.21760139e-01
4.89800870e-01 -5.24331689e-01 4.73351270e-01 3.67108524e-01
-1.25232422e+00 -2.43204832e-01 -1.33387661e+00 8.01159263e-01
1.92329347e-01 3.08484398e-02 5.44223487e-01 9.40433621e-01
-8.68344963e-01 -5.20185940e-02 -5.05318403e-01 -2.57083774e-02
7.56261647e-01 3.59969258e-01 4.61886227e-01 -4.28906858e-01
-1.04382908e+00 3.96528929e-01 -2.71398425e-01 2.08761856e-01
-4.42422092e-01 -1.05271089e+00 -3.52916867e-01 1.41490027e-01
1.46873847e-01 -8.95818472e-01 9.35896695e-01 -2.75858462e-01
-1.43390787e+00 -2.15982020e-01 -1.06497526e-01 -1.87313035e-01
5.99259138e-01 1.19307540e-01 -5.72511137e-01 2.29653344e-02
-1.88591242e-01 -6.85327426e-02 5.96969187e-01 -1.70053065e+00
-1.01789439e+00 -4.88949865e-01 3.13122161e-02 2.40926489e-01
-5.51799834e-01 -5.68553686e-01 -4.30841744e-01 -7.60530233e-01
6.71910882e-01 -5.81285775e-01 -5.60130060e-01 -3.33540052e-01
-3.71886551e-01 2.87304938e-01 4.88311857e-01 -2.69356459e-01
1.59505749e+00 -1.93729591e+00 2.19597161e-01 4.18363154e-01
4.28036183e-01 2.28536911e-02 -9.75653082e-02 2.90503383e-01
2.23261505e-01 -3.95121276e-02 1.75880924e-01 -6.99084476e-02
-3.15950841e-01 -1.43137351e-02 -8.35469440e-02 5.98273635e-01
-7.09634721e-01 7.24395633e-01 -9.01078224e-01 -3.18402827e-01
2.43230626e-01 3.01334411e-01 -6.37556076e-01 2.70949274e-01
3.00466925e-01 8.62185657e-01 -6.95586026e-01 7.62860239e-01
1.12433934e+00 -1.38720065e-01 2.83071339e-01 -9.26403224e-01
-3.05997342e-01 -3.03406090e-01 -1.25488937e+00 1.55774403e+00
-9.39696550e-01 3.06706250e-01 4.40603316e-01 -8.96309555e-01
5.23717880e-01 2.66440868e-01 9.54583645e-01 -1.23770857e+00
1.34852931e-01 2.14723110e-01 5.00255898e-02 -5.95030785e-01
-7.89950937e-02 -2.35371724e-01 -5.04665151e-02 3.65129977e-01
-5.45857707e-03 5.11991419e-02 -1.83588862e-01 -2.34231763e-02
1.20711720e+00 -4.76242602e-01 1.02118008e-01 -3.25679421e-01
5.37236929e-01 -6.96408510e-01 7.32304037e-01 8.76063228e-01
1.78596005e-02 1.78616911e-01 -2.02654451e-01 -2.17006192e-01
-7.71312475e-01 -1.19013989e+00 -1.91570237e-01 9.45718169e-01
5.79088926e-01 -2.58624498e-02 -3.05758536e-01 -2.62327343e-01
-2.05199569e-01 5.78522086e-01 6.38646306e-03 -7.79127181e-02
-7.24128366e-01 -1.37445521e+00 -1.68157481e-02 3.12622666e-01
6.10478222e-01 -1.66542038e-01 -4.52373832e-01 3.87080193e-01
-2.01668933e-01 -1.08069646e+00 -8.89084786e-02 1.72357798e-01
-6.22132123e-01 -7.50752568e-01 -7.78612792e-01 -6.94478989e-01
6.86579823e-01 5.84969163e-01 9.20077860e-01 9.54338722e-03
1.27268419e-01 3.75922620e-01 -4.05711681e-01 -4.20452088e-01
1.42922387e-01 6.92921132e-02 2.10088521e-01 8.11995044e-02
-4.01183665e-01 -1.06523287e+00 -9.58818316e-01 2.95235813e-01
-2.56574571e-01 1.99396551e-01 1.14004660e+00 7.07066059e-01
3.90220821e-01 6.30796790e-01 9.06044960e-01 -5.18593490e-01
3.95308226e-01 -7.07497954e-01 -7.61686027e-01 1.77928805e-01
-5.84966063e-01 -2.66164571e-01 7.31564820e-01 -6.90546213e-03
-1.00883937e+00 -1.39362842e-01 -4.98725414e-01 1.66823268e-02
3.76276761e-01 5.26862919e-01 -6.49965346e-01 -6.43170536e-01
3.52345258e-01 3.50800276e-01 -6.89136326e-01 -1.79131493e-01
3.76553953e-01 7.29372025e-01 3.85824233e-01 -4.95952256e-02
1.36209714e+00 4.99325067e-01 4.40289855e-01 -8.55869412e-01
-8.97165656e-01 -1.61502391e-01 -4.34384763e-01 -5.35871923e-01
4.30598646e-01 -1.19216824e+00 -4.47476685e-01 1.67309985e-01
-8.77627671e-01 -2.13605672e-01 2.94906020e-01 6.76731706e-01
-3.92502964e-01 1.73058391e-01 -2.40935937e-01 -1.05139697e+00
-3.34793597e-01 -8.49785268e-01 6.56880438e-01 3.62081349e-01
3.37289959e-01 -6.29408598e-01 -2.81044185e-01 2.02303752e-01
8.16782117e-01 -3.72251421e-02 1.06124139e+00 1.04004212e-01
-9.50503767e-01 2.04001237e-02 -3.57699692e-01 -5.05337417e-01
2.43849665e-01 -8.94913733e-01 -5.02119780e-01 -5.76016605e-01
7.63932522e-03 1.41075641e-01 5.59085965e-01 9.00140584e-01
1.32994103e+00 -2.81064481e-01 -7.29295135e-01 1.21775365e+00
1.67592967e+00 3.06202829e-01 4.76497740e-01 5.46357334e-02
5.97186923e-01 -1.16200939e-01 5.17417431e-01 9.31667447e-01
5.68146765e-01 7.44552493e-01 8.15521181e-01 -2.95149237e-01
-1.64773345e-01 5.16178682e-02 -2.62808323e-01 7.81838953e-01
-1.74676761e-01 -6.65716410e-01 -5.09050429e-01 2.83188790e-01
-1.72191274e+00 -9.30944443e-01 -7.44277164e-02 2.15621400e+00
4.42916453e-01 2.15212807e-01 -1.81354433e-01 3.75271216e-02
2.50779837e-01 3.65287244e-01 -5.03040850e-01 5.55422246e-01
-1.18179493e-01 1.09627955e-02 9.44271743e-01 5.32082915e-01
-1.05484951e+00 2.19822437e-01 6.03171253e+00 8.84880602e-01
-1.04455805e+00 4.20249075e-01 4.13638145e-01 -4.03915465e-01
-5.76773584e-01 -7.59579062e-01 -6.60246253e-01 6.09281778e-01
6.24156356e-01 -9.81257781e-02 3.44466418e-01 3.89938861e-01
5.56760609e-01 1.27970845e-01 -6.92588151e-01 1.43075132e+00
-9.72283185e-02 -1.75637186e+00 -2.40036488e-01 2.83745915e-01
6.91638887e-01 -1.26902957e-03 1.78959500e-02 1.82597741e-01
9.50655043e-02 -6.77541733e-01 6.06284142e-01 7.37166762e-01
7.63782442e-01 -1.10169697e+00 8.87343287e-01 5.19868076e-01
-1.38473272e+00 -7.07631826e-01 -3.30113113e-01 1.30201325e-01
3.51765603e-01 1.27158773e+00 -2.72664160e-01 1.03327394e+00
6.05523825e-01 4.03904468e-01 -1.58489328e-02 1.37417972e+00
-6.25679493e-02 5.70452034e-01 -2.46898785e-01 -4.11365442e-02
-1.02501616e-01 -1.16657622e-01 6.68291807e-01 1.07345259e+00
9.27407861e-01 6.40672326e-01 2.72578537e-01 4.08049434e-01
-8.05985853e-02 -2.07571566e-01 -3.53315324e-01 6.12715960e-01
8.88318241e-01 1.29361570e+00 -4.29476976e-01 2.08241850e-01
-7.08536804e-01 4.11037713e-01 -1.77299008e-01 6.38710499e-01
-7.22009897e-01 -2.39172936e-04 7.29294538e-01 2.52800316e-01
3.14137936e-01 -6.79677904e-01 -7.40109622e-01 -8.62214983e-01
-1.82559162e-01 -1.79936320e-01 1.22072794e-01 -4.37095731e-01
-1.15218651e+00 3.84172559e-01 -4.16271687e-01 -1.60046113e+00
-6.55354261e-02 -5.62852509e-02 -5.95420063e-01 6.58057630e-01
-1.95381308e+00 -1.17416179e+00 -4.75864857e-01 3.56643677e-01
4.84193861e-01 -5.15860677e-01 6.53755784e-01 8.90989602e-01
-5.57547331e-01 8.02404404e-01 3.23116153e-01 5.59097379e-02
2.39560753e-01 -7.41399050e-01 -4.08341318e-01 9.21751142e-01
-5.01348078e-02 3.82565595e-02 8.21984947e-01 -3.68456244e-01
-1.94839108e+00 -1.31292760e+00 2.99118400e-01 6.87977076e-02
3.53157163e-01 -3.54604870e-01 -2.12055489e-01 1.49518266e-01
1.38025135e-01 1.35551393e-01 9.30290222e-01 -1.17444098e-02
2.84960091e-01 -5.00006974e-01 -1.10174263e+00 7.38859594e-01
1.46997499e+00 1.03448428e-01 5.19348010e-02 6.00164592e-01
2.36604944e-01 -4.54808056e-01 -6.59708261e-01 8.24024737e-01
6.29545689e-01 -9.65000808e-01 1.35238647e+00 3.95403095e-02
1.26117215e-01 -2.94936389e-01 -4.95644897e-01 -1.45117891e+00
-1.11031318e+00 -4.69498873e-01 -6.08496904e-01 9.77297604e-01
3.22955817e-01 -2.81963795e-01 1.09217572e+00 -1.70421541e-01
-1.98310331e-01 -1.19087446e+00 -1.21539640e+00 -6.07300639e-01
-1.10808000e-01 -5.31198382e-01 7.01704979e-01 6.31042898e-01
-2.70435065e-01 3.00203621e-01 -5.41730821e-01 1.06534660e+00
1.10078073e+00 2.81968325e-01 5.78786373e-01 -1.22718728e+00
-3.54583472e-01 -4.99847531e-01 -3.43156047e-02 -1.46131694e+00
-2.74075300e-01 -5.88470399e-01 8.74107629e-02 -2.03139806e+00
-1.09923556e-02 -8.95561516e-01 -5.38216054e-01 1.49837017e-01
1.26303971e-01 4.09080595e-01 -1.85789075e-02 1.39388129e-01
-5.81992686e-01 6.82848632e-01 1.25494003e+00 -3.02910656e-01
-2.42918238e-01 6.04459107e-01 -9.37189996e-01 7.60494649e-01
5.23427665e-01 -9.94079188e-02 -4.65375185e-01 -7.18583763e-01
4.41631377e-01 4.65830535e-01 3.83641981e-02 -1.56344187e+00
5.29768407e-01 -5.36971390e-01 8.19242060e-01 -7.95554459e-01
2.54916728e-01 -1.32051051e+00 -5.65134510e-02 6.25910640e-01
3.59816641e-01 -5.95532000e-01 -8.33148807e-02 6.70849919e-01
2.58524001e-01 1.60749897e-01 9.30430055e-01 1.07933760e-01
-5.00749290e-01 6.88929141e-01 -4.19142038e-01 -5.07597685e-01
9.16491926e-01 -1.10685319e-01 -8.67742598e-02 -8.69217634e-01
-4.42017794e-01 4.63849843e-01 -2.08163485e-01 2.00868487e-01
5.56206048e-01 -1.59162378e+00 -8.15798461e-01 2.78125584e-01
-3.57934795e-02 -2.06675202e-01 5.06492257e-01 9.19793725e-01
-4.70996611e-02 5.43998659e-01 9.07643139e-03 -4.59104121e-01
-8.65309060e-01 -4.83621061e-02 5.63980222e-01 -7.53037333e-02
-5.14697015e-01 1.07587004e+00 1.59373596e-01 -1.90464687e-02
1.15768366e-01 1.02454253e-01 -3.90724361e-01 2.20060498e-02
5.26221931e-01 1.56468317e-01 1.28193095e-01 -3.67236286e-01
-2.99055994e-01 8.32243383e-01 3.93258810e-01 1.24075592e-01
1.68343604e+00 -6.60543442e-01 1.83076113e-01 -2.03818247e-01
7.37322330e-01 5.91285169e-01 -1.08738422e+00 -2.74702936e-01
-5.15868545e-01 -7.09237099e-01 6.51181579e-01 -9.07540739e-01
-1.39888692e+00 3.14812630e-01 1.05373132e+00 3.64875495e-01
1.55922031e+00 -1.79196578e-02 1.03737450e+00 2.78219551e-01
1.03405106e+00 -8.83822739e-01 2.31972337e-02 3.16630363e-01
5.57790041e-01 -8.03636491e-01 2.81579852e-01 -6.16659164e-01
2.89516777e-01 1.00535429e+00 5.05446017e-01 3.74460697e-01
1.04603541e+00 5.76704741e-01 -8.02803263e-02 -1.12216525e-01
-3.88221622e-01 -7.38203451e-02 4.76453453e-02 7.26432800e-01
2.18815550e-01 3.69373672e-02 -3.75215411e-01 1.08920670e+00
-2.20711857e-01 -1.55916698e-02 4.62992668e-01 7.61197746e-01
-8.30936074e-01 -9.52434123e-01 -4.46764886e-01 8.55290771e-01
-9.89823937e-02 1.76835936e-02 4.66740161e-01 2.78126657e-01
4.47466642e-01 1.15647066e+00 -1.66593596e-01 -6.01834834e-01
4.11932200e-01 -5.99365711e-01 6.57814741e-01 -3.85284960e-01
3.97984803e-01 3.59104186e-01 -1.15707852e-01 -4.84957159e-01
-1.72520116e-01 -3.57038021e-01 -1.08410263e+00 1.57175343e-02
-2.23371357e-01 9.62110385e-02 4.88895208e-01 1.36977375e+00
4.95081902e-01 1.10515344e+00 1.11475849e+00 -1.01308060e+00
-2.31119320e-01 -7.24699140e-01 -7.28782594e-01 -4.43492830e-01
4.86622810e-01 -7.76512623e-01 -2.12919399e-01 -5.28237820e-01] | [6.23029899597168, 1.3396475315093994] |
1a0073e8-a149-4438-aa8c-357d65ce0b34 | lepus-prompt-based-unsupervised-multi-hop | 2205.1265 | null | https://arxiv.org/abs/2205.12650v3 | https://arxiv.org/pdf/2205.12650v3.pdf | Few-shot Reranking for Multi-hop QA via Language Model Prompting | We study few-shot reranking for multi-hop QA with open-domain questions. To alleviate the need for a large number of labeled question-document pairs for retriever training, we propose PromptRank, which relies on large language models prompting for multi-hop path reranking. PromptRank first constructs an instruction-based prompt that includes a candidate document path and then computes the relevance score between a given question and the path based on the conditional likelihood of the question given the path prompt according to a language model. PromptRank yields strong retrieval performance on HotpotQA with only 128 training examples compared to state-of-the-art methods trained on thousands of examples -- 73.6 recall@10 by PromptRank vs. 77.8 by PathRetriever and 77.5 by multi-hop dense retrieval. Code available at https://github.com/mukhal/PromptRank | ['Lu Wang', 'Honglak Lee', 'Moontae Lee', 'Lajanugen Logeswaran', 'Muhammad Khalifa'] | 2022-05-25 | null | null | null | null | ['passage-re-ranking'] | ['natural-language-processing'] | [-2.35429525e-01 5.12691401e-02 -2.14446008e-01 -2.89607525e-01
-1.99569893e+00 -7.42790878e-01 8.17499459e-01 5.07108331e-01
-7.08176851e-01 6.78834736e-01 6.91443682e-01 -4.95776892e-01
-4.68946576e-01 -5.94568789e-01 -5.10082066e-01 -1.00326248e-01
2.10994959e-01 1.37328792e+00 7.41930544e-01 -8.74358535e-01
4.69153225e-01 2.69515216e-02 -1.22084105e+00 5.66424429e-01
1.10215294e+00 5.00250995e-01 2.06075847e-01 1.11814356e+00
-4.22430485e-01 9.46246207e-01 -6.51877642e-01 -4.68204856e-01
-1.71771962e-02 -3.62132490e-01 -1.59307873e+00 -7.07181871e-01
9.64245498e-01 -8.31989050e-01 -7.33283579e-01 6.80886626e-01
5.83820045e-01 6.74391568e-01 7.35119939e-01 -8.11655819e-01
-8.14564824e-01 5.05959749e-01 -2.97478825e-01 6.93659782e-01
8.46900940e-01 5.59115857e-02 1.74216247e+00 -8.96879971e-01
8.99416566e-01 1.24003184e+00 -6.80731535e-02 4.44003046e-01
-7.50209391e-01 -2.43181020e-01 -3.77190650e-01 6.54334188e-01
-1.13701892e+00 -3.85144264e-01 8.14867094e-02 -1.07070886e-01
1.07535291e+00 2.55608410e-01 2.55194068e-01 5.51844656e-01
4.18622456e-02 8.99229467e-01 8.94904912e-01 -6.29586279e-01
-1.12369560e-01 -2.48946130e-01 1.09288931e+00 9.21854436e-01
-2.39636153e-01 -1.47822872e-01 -5.99282503e-01 -7.33052850e-01
1.95705205e-01 -2.10513920e-01 -2.57869244e-01 2.13840410e-01
-9.82225478e-01 1.00768352e+00 4.28827703e-01 2.63852984e-01
-3.73717964e-01 1.65834218e-01 1.74865827e-01 7.07381725e-01
2.49516144e-01 1.02186239e+00 -4.81478870e-01 -2.64143407e-01
-9.85414147e-01 5.68161726e-01 8.47163200e-01 7.01606572e-01
1.02835822e+00 -6.55188680e-01 -1.02999437e+00 1.00231862e+00
4.14258599e-01 7.53413558e-01 4.19031233e-01 -1.22140956e+00
4.27135795e-01 4.90645945e-01 2.70105958e-01 -5.01238942e-01
-4.98882420e-02 -1.80985630e-01 2.75270231e-02 -6.03030205e-01
4.95816499e-01 1.71327144e-01 -1.00319123e+00 1.20663285e+00
4.67891365e-01 -1.38425425e-01 1.65617689e-01 8.93956780e-01
1.16653764e+00 1.05606520e+00 9.34354216e-02 1.06580585e-01
1.72833610e+00 -1.48562539e+00 -5.30023634e-01 -7.45979995e-02
1.18414974e+00 -1.01065636e+00 1.56725252e+00 1.20937862e-01
-9.35404003e-01 -1.72696173e-01 -5.44521749e-01 -7.17671037e-01
-1.42317027e-01 -6.40438944e-02 2.55155534e-01 6.13099299e-02
-1.35805666e+00 3.81677419e-01 -2.15668723e-01 -5.14755309e-01
-1.89850062e-01 -6.32003397e-02 -2.67779324e-02 -8.25975776e-01
-1.63142657e+00 9.58475292e-01 -4.44103032e-02 -7.04752147e-01
-1.30305982e+00 -7.85916567e-01 -3.80194128e-01 3.19565475e-01
4.25581276e-01 -8.46657455e-01 1.93470049e+00 -7.03027174e-02
-1.29059684e+00 5.92052698e-01 -5.12799501e-01 -3.70378345e-01
1.25736818e-01 -8.15450311e-01 -3.72059911e-01 8.90316486e-01
4.73435938e-01 8.61404419e-01 7.31709182e-01 -8.58188152e-01
-5.71424723e-01 -6.81547746e-02 5.30975163e-01 5.23286879e-01
-2.32821614e-01 3.22225451e-01 -6.64888978e-01 -1.10891904e-03
-1.73154160e-01 -6.32511556e-01 -2.41990462e-01 -3.97543728e-01
-1.62420034e-01 -1.00319695e+00 5.81485987e-01 -8.96411717e-01
1.23945236e+00 -1.73511875e+00 -3.40427071e-01 -3.51058692e-02
3.44134241e-01 3.89640719e-01 -8.54618132e-01 8.21602106e-01
4.31421489e-01 -7.38908798e-02 8.82199705e-02 1.91948012e-01
-7.79750012e-03 -7.03762323e-02 -5.83009303e-01 1.04899015e-02
-2.27555454e-01 1.08234298e+00 -1.39373505e+00 -7.66432762e-01
-3.56638074e-01 9.05392021e-02 -5.52363396e-01 5.85328519e-01
-7.25537181e-01 8.75955522e-02 -6.77546501e-01 5.34258425e-01
9.81363505e-02 -7.45411396e-01 -1.95185423e-01 9.75035429e-02
4.86468732e-01 9.66637254e-01 -5.50012708e-01 1.79685390e+00
-3.04367572e-01 4.48168904e-01 -2.73854464e-01 -1.75751671e-01
6.84548616e-01 4.83275384e-01 2.62362510e-01 -1.06923962e+00
-8.96252543e-02 3.25076699e-01 -3.75794113e-01 -4.97271478e-01
1.23196113e+00 1.78065971e-01 -9.67288092e-02 1.11530006e+00
3.72387052e-01 -3.11060786e-01 5.68542778e-01 1.28772545e+00
1.57235253e+00 -3.01943660e-01 -1.98176373e-02 -1.39941216e-01
4.39120352e-01 4.79455739e-01 -1.61617234e-01 1.13576186e+00
-3.87576781e-02 2.33776331e-01 1.56946182e-01 -3.98342572e-02
-6.49025679e-01 -1.00636768e+00 2.96913892e-01 1.70926297e+00
2.12301880e-01 -7.28801012e-01 -4.67524707e-01 -9.48670387e-01
-9.02079120e-02 1.10275745e+00 -9.36161429e-02 -1.31527036e-01
-5.08105636e-01 -1.37411961e-02 6.77130342e-01 -4.61639976e-03
1.94730759e-01 -9.13388431e-01 -1.84464917e-01 6.77644759e-02
-6.88462138e-01 -9.17176127e-01 -7.74306357e-01 -2.51127332e-01
-8.09582651e-01 -1.17992902e+00 -9.09238279e-01 -6.64498031e-01
4.68812585e-01 7.98343420e-01 1.69612443e+00 5.88484228e-01
-7.89352953e-02 1.07174289e+00 -8.46866667e-01 1.60367414e-01
-4.28531885e-01 2.96775907e-01 -2.17653155e-01 -7.29767025e-01
3.75541359e-01 -7.41931284e-03 -8.64640415e-01 2.88497746e-01
-8.85271788e-01 -3.34444284e-01 4.74903673e-01 9.00619566e-01
4.30157453e-01 -6.67696595e-01 6.75336778e-01 -7.46001244e-01
1.07971835e+00 -6.75530791e-01 -5.56342125e-01 9.02123570e-01
-8.10272872e-01 2.39661798e-01 2.14605734e-01 -2.75754511e-01
-1.17975307e+00 -6.85900927e-01 -2.37748206e-01 6.52501285e-02
1.33894756e-01 6.37214422e-01 6.78436100e-01 9.98698920e-02
1.00645697e+00 6.02538958e-02 -4.39185679e-01 -5.96947610e-01
1.12139082e+00 7.69165337e-01 2.30277359e-01 -8.45228791e-01
7.28036821e-01 1.36780754e-01 -5.53373396e-01 -6.11091316e-01
-1.18307531e+00 -1.29278982e+00 -2.69061416e-01 -2.66276777e-01
7.23660827e-01 -9.06473041e-01 -3.30679119e-01 -5.86401820e-02
-1.37342882e+00 -5.40074348e-01 -4.24492180e-01 4.35312361e-01
-1.66483745e-01 5.51643550e-01 -1.06340456e+00 -4.15817052e-01
-9.78441119e-01 -7.48000920e-01 1.10149264e+00 3.07016909e-01
-3.89373213e-01 -8.12467396e-01 6.02943718e-01 1.03138649e+00
3.54549259e-01 -9.76699352e-01 1.20884657e+00 -1.07776582e+00
-8.72688115e-01 -3.81654710e-01 -3.30306023e-01 2.19414569e-02
-1.67943969e-01 -2.87684292e-01 -7.49142170e-01 -3.46460760e-01
-3.37729812e-01 -1.03259873e+00 8.72189760e-01 -1.37547195e-01
4.46723640e-01 -4.50626224e-01 -2.45416149e-01 -2.89985776e-01
1.10214460e+00 -1.67258665e-01 4.69663054e-01 1.90948859e-01
4.94818419e-01 5.53734660e-01 1.16476536e+00 2.24111885e-01
6.46899343e-01 4.33237672e-01 -6.26584841e-03 3.06648046e-01
-3.41577321e-01 -4.66781616e-01 3.42244983e-01 1.16505015e+00
5.25872052e-01 -5.30374110e-01 -1.13278604e+00 8.69234622e-01
-1.40758657e+00 -6.51541293e-01 -4.08885807e-01 2.04602885e+00
1.01421714e+00 -2.85042882e-01 -1.33352429e-01 -5.73947072e-01
3.52357566e-01 3.67908031e-01 -2.60122180e-01 -1.87653929e-01
1.74951121e-01 4.31036025e-01 2.06985012e-01 1.02859533e+00
-3.19919437e-01 1.26394188e+00 6.09938717e+00 1.09607387e+00
-5.25495350e-01 4.11096931e-01 3.31625700e-01 -3.22237834e-02
-7.76254237e-01 3.49263787e-01 -1.06813705e+00 -1.16572164e-01
1.21276867e+00 -4.88887995e-01 3.76372397e-01 6.25225782e-01
-3.51259261e-01 -4.36060548e-01 -8.10134411e-01 5.14129281e-01
2.84411281e-01 -1.24466145e+00 4.06726748e-01 -3.46798152e-01
8.02944899e-01 4.74695116e-01 -2.03530639e-01 7.45588601e-01
9.62118447e-01 -6.30651176e-01 1.14508428e-01 3.32156718e-01
6.11525178e-01 -2.45393530e-01 5.00796974e-01 5.28918803e-01
-7.88793504e-01 1.16146922e-01 -4.36189085e-01 2.07701966e-01
3.80686402e-01 5.14218986e-01 -1.35489798e+00 6.26253188e-01
6.73306584e-01 1.61366299e-01 -7.07741618e-01 1.05412328e+00
-5.96691906e-01 8.93317044e-01 -2.50301421e-01 -3.26085240e-01
4.52688217e-01 1.08374350e-01 6.03426099e-01 9.90428925e-01
2.71659702e-01 5.46398878e-01 1.02339007e-01 1.81316167e-01
-4.19130087e-01 4.79862541e-01 -1.63963497e-01 -1.98155671e-01
6.79404914e-01 1.33812189e+00 -2.40261629e-01 -7.77412593e-01
-2.84463048e-01 1.06133974e+00 6.50910020e-01 4.69511390e-01
-3.90002877e-01 -4.78863686e-01 1.84682339e-01 1.02909170e-01
-7.46534765e-02 -1.13224193e-01 4.68250841e-01 -9.55755293e-01
-1.13609992e-01 -1.06427646e+00 1.14891112e+00 -1.19469094e+00
-1.39059639e+00 6.36566639e-01 9.88923758e-02 -8.26156676e-01
-3.67981851e-01 -1.92281842e-01 -5.62768161e-01 7.82444835e-01
-1.86103833e+00 -7.51519382e-01 -1.50844529e-01 6.88015938e-01
6.02138460e-01 3.53850946e-02 9.67628658e-01 3.02040696e-01
1.99745327e-01 4.16367114e-01 1.94851130e-01 -1.00089259e-01
1.29910457e+00 -1.20460880e+00 4.00271833e-01 6.12992525e-01
5.40027678e-01 7.82150090e-01 4.92030680e-01 -6.93329811e-01
-1.31490421e+00 -6.19612575e-01 1.57260525e+00 -8.29090178e-01
8.32181633e-01 2.74302661e-01 -1.16875911e+00 4.87348765e-01
6.53880477e-01 -2.17690811e-01 6.85647249e-01 1.85816497e-01
-7.06419408e-01 -2.00191997e-02 -9.02580440e-01 7.14479387e-01
6.58269227e-01 -9.14361358e-01 -1.20919728e+00 9.32376146e-01
1.23651946e+00 -2.99479097e-01 -8.07355046e-01 1.38427973e-01
2.10149810e-01 -3.91973794e-01 1.01614201e+00 -8.06394160e-01
2.70356238e-01 -1.31344348e-01 -7.28193745e-02 -1.12914097e+00
-3.86953920e-01 -6.54288113e-01 -4.06754643e-01 9.68974352e-01
6.49407029e-01 -4.62324262e-01 5.42376161e-01 4.36321914e-01
-1.60788551e-01 -6.39204919e-01 -9.41097736e-01 -6.93191051e-01
1.96861029e-01 -6.37603700e-02 4.53778356e-01 7.46843517e-01
1.70558780e-01 8.83727729e-01 -7.87199587e-02 4.51167449e-02
3.52996409e-01 1.94034129e-01 6.77011013e-01 -1.00803745e+00
-4.20009941e-01 6.00860566e-02 4.69025582e-01 -1.76761222e+00
7.17523620e-02 -8.53498876e-01 2.99637645e-01 -2.14292502e+00
3.96917820e-01 -5.80248177e-01 -3.47713709e-01 4.29846942e-01
-6.73669219e-01 -1.76839810e-02 6.99388981e-03 4.68320727e-01
-1.50583827e+00 7.85428643e-01 1.41881645e+00 -3.09876084e-01
-2.42432356e-02 -3.43363672e-01 -5.93957067e-01 1.85078636e-01
7.69491971e-01 -8.61320794e-01 -6.43522441e-01 -7.56873429e-01
4.01829004e-01 6.69801533e-01 1.54040739e-01 -6.18528366e-01
5.75790226e-01 -1.12391608e-02 -3.89039874e-01 -7.62710273e-01
4.80508268e-01 -2.61859447e-01 -6.25312567e-01 4.43406641e-01
-8.35943341e-01 4.15468574e-01 -1.36160776e-01 5.92693031e-01
-2.66034186e-01 -7.63202429e-01 3.76924306e-01 -3.38638574e-01
-8.59887779e-01 2.28041202e-01 -3.78079027e-01 7.99905598e-01
3.26040268e-01 5.45274854e-01 -1.16057611e+00 -8.08153033e-01
-2.55241245e-01 6.93544269e-01 1.38776004e-01 4.41987067e-01
7.37870872e-01 -1.07568407e+00 -9.52569544e-01 -5.38628161e-01
4.37589347e-01 -2.41895661e-01 3.90613794e-01 5.78935742e-01
-4.08283651e-01 9.17492986e-01 2.57144541e-01 -3.14920962e-01
-1.51246464e+00 1.22211367e-01 -9.81275588e-02 -1.05903292e+00
-3.63455057e-01 1.08369994e+00 -1.64967939e-01 -6.24106169e-01
-3.98788527e-02 1.49764553e-01 -3.25143814e-01 8.13873336e-02
8.18297207e-01 6.01578474e-01 2.42515355e-01 -3.48941594e-01
-6.83159381e-02 3.51397455e-01 -6.55127883e-01 -7.00291574e-01
8.76650691e-01 -1.47507936e-01 -3.58031571e-01 1.28435150e-01
1.29436862e+00 4.49582413e-02 -6.39382482e-01 -7.73764849e-01
2.72930741e-01 -4.83792245e-01 -2.87614930e-02 -1.15722692e+00
-3.17138165e-01 8.37036729e-01 4.27690923e-01 -1.60922557e-01
6.51939452e-01 4.73113924e-01 1.35322785e+00 1.13676190e+00
5.00339866e-01 -9.13737655e-01 9.68386650e-01 1.12067115e+00
8.33161473e-01 -1.12153578e+00 4.18213420e-02 -1.35780526e-02
-6.01886928e-01 6.37376666e-01 7.17094421e-01 1.09907337e-01
2.92665362e-01 -6.68937922e-01 5.05130947e-01 -7.28400111e-01
-1.02089560e+00 -3.68647188e-01 6.30103767e-01 2.03449413e-01
3.94346684e-01 -1.34237945e-01 -3.89317542e-01 1.65763363e-01
-1.41938210e-01 -1.23865075e-01 3.47646475e-01 9.66078222e-01
-9.29147899e-01 -1.13273013e+00 -2.41992265e-01 7.76858866e-01
-3.82965326e-01 -6.96038246e-01 -2.98706084e-01 2.59154677e-01
-8.27787042e-01 1.40444458e+00 -1.18993372e-01 -1.06939562e-01
6.20677881e-02 3.58756483e-01 3.95325035e-01 -1.00381207e+00
-7.58791029e-01 -2.79583454e-01 5.01622200e-01 -5.13143420e-01
3.29785310e-02 -2.51142859e-01 -1.27675414e+00 -1.69875830e-01
-5.49250066e-01 8.92814219e-01 2.27305248e-01 9.11731958e-01
5.89113533e-01 1.15119815e-01 3.49094629e-01 7.75068551e-02
-9.18513834e-01 -1.20102417e+00 -2.02955797e-01 3.01213235e-01
3.71832132e-01 -2.19916686e-01 -2.74419099e-01 -3.61872047e-01] | [11.436332702636719, 7.751377105712891] |
8c15c917-1863-4705-8610-b621a6750189 | improving-action-localization-by-progressive-1 | 1905.11575 | null | https://arxiv.org/abs/1905.11575v1 | https://arxiv.org/pdf/1905.11575v1.pdf | Improving Action Localization by Progressive Cross-stream Cooperation | Spatio-temporal action localization consists of three levels of tasks: spatial localization, action classification, and temporal segmentation. In this work, we propose a new Progressive Cross-stream Cooperation (PCSC) framework to use both region proposals and features from one stream (i.e. Flow/RGB) to help another stream (i.e. RGB/Flow) to iteratively improve action localization results and generate better bounding boxes in an iterative fashion. Specifically, we first generate a larger set of region proposals by combining the latest region proposals from both streams, from which we can readily obtain a larger set of labelled training samples to help learn better action detection models. Second, we also propose a new message passing approach to pass information from one stream to another stream in order to learn better representations, which also leads to better action detection models. As a result, our iterative framework progressively improves action localization results at the frame level. To improve action localization results at the video level, we additionally propose a new strategy to train class-specific actionness detectors for better temporal segmentation, which can be readily learnt by focusing on "confusing" samples from the same action class. Comprehensive experiments on two benchmark datasets UCF-101-24 and J-HMDB demonstrate the effectiveness of our newly proposed approaches for spatio-temporal action localization in realistic scenarios. | ['Rui Su', 'Wanli Ouyang', 'Luping Zhou', 'Dong Xu'] | 2019-05-28 | improving-action-localization-by-progressive | http://openaccess.thecvf.com/content_CVPR_2019/html/Su_Improving_Action_Localization_by_Progressive_Cross-Stream_Cooperation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Su_Improving_Action_Localization_by_Progressive_Cross-Stream_Cooperation_CVPR_2019_paper.pdf | cvpr-2019-6 | ['spatio-temporal-action-localization'] | ['computer-vision'] | [ 3.84073615e-01 -3.99437338e-01 -4.93622392e-01 -1.64202869e-01
-9.22640681e-01 -4.14599627e-01 3.29492211e-01 2.55580395e-01
-5.57660162e-01 6.53440475e-01 1.36216715e-01 1.37891948e-01
1.84178818e-02 -8.06140602e-01 -6.40615165e-01 -8.59080076e-01
-8.72733667e-02 9.90355760e-02 8.95837963e-01 9.29720029e-02
3.19319725e-01 5.52725494e-01 -1.23872685e+00 6.88565850e-01
7.28324413e-01 1.15710068e+00 2.73408502e-01 5.23752153e-01
-5.08600883e-02 1.11404347e+00 -3.16191345e-01 1.65223137e-01
1.42154664e-01 -5.59981883e-01 -9.49852884e-01 5.76672673e-01
-2.28201468e-02 -5.76723993e-01 -1.59613818e-01 7.90587723e-01
2.65234709e-01 4.67317313e-01 2.74327248e-01 -1.13083255e+00
-2.89422832e-02 3.74721885e-01 -8.18785787e-01 3.94633204e-01
4.68986571e-01 4.66145694e-01 7.91855395e-01 -5.96568525e-01
5.94176769e-01 1.44567108e+00 2.20010579e-01 4.56806988e-01
-9.22644436e-01 -6.36870325e-01 6.93257511e-01 4.58109587e-01
-1.28992975e+00 -3.89685661e-01 8.23718488e-01 -8.47310275e-02
5.71730912e-01 9.59720463e-02 8.53855789e-01 8.01757038e-01
-8.80886316e-02 1.38661289e+00 8.23293209e-01 -1.19229056e-01
3.77544791e-01 -2.27371618e-01 -2.36345962e-01 7.81545401e-01
-4.25481707e-01 -1.69427499e-01 -6.23648345e-01 8.81961584e-02
9.85205352e-01 2.95677066e-01 -4.30585206e-01 -2.82612026e-01
-1.38933206e+00 6.15158856e-01 7.24759698e-01 5.13929784e-01
-6.59629166e-01 4.22634035e-01 4.69969660e-01 -1.31474227e-01
4.16912138e-01 2.17445102e-02 -4.13838565e-01 -2.81833887e-01
-9.03209269e-01 -5.05349748e-02 1.76171437e-01 7.08129764e-01
8.95260632e-01 -2.28277847e-01 -5.54830670e-01 8.15911591e-01
3.36380333e-01 3.14771384e-01 4.82791454e-01 -1.19792771e+00
8.75846207e-01 6.90904498e-01 3.02929878e-01 -9.84196246e-01
-2.19129398e-01 -3.30248982e-01 -6.26953006e-01 8.72614980e-03
5.11028528e-01 3.88990752e-02 -8.67289662e-01 1.49566877e+00
5.74060559e-01 7.10855067e-01 6.30442053e-02 9.47898507e-01
2.82389730e-01 8.75126541e-01 2.20223442e-01 -2.74390608e-01
1.22944403e+00 -1.20332456e+00 -3.97475868e-01 -1.67281911e-01
1.04737508e+00 -4.63451743e-01 6.86288655e-01 4.10666466e-01
-1.12936139e+00 -8.26422632e-01 -7.41637170e-01 1.79490536e-01
-7.11983517e-02 3.94186437e-01 4.27579969e-01 1.93266541e-01
-7.66242921e-01 6.78259671e-01 -1.20349050e+00 -2.25178957e-01
8.80490720e-01 2.09982008e-01 -2.07077876e-01 -3.82069856e-01
-8.82332087e-01 1.98037252e-01 6.53108776e-01 4.66243140e-02
-1.09228384e+00 -4.50058728e-01 -8.50267768e-01 -1.22197822e-01
6.82475746e-01 -3.61230433e-01 1.10912323e+00 -1.24458790e+00
-1.43133318e+00 1.95211634e-01 -4.26423997e-01 -5.23157537e-01
6.02483869e-01 -1.00463212e-01 -1.08529076e-01 6.98504627e-01
2.00145170e-01 1.07954264e+00 7.87831187e-01 -1.14071035e+00
-1.36962259e+00 -2.62317777e-01 3.32842767e-01 3.75342280e-01
-3.33857030e-01 -2.37319414e-02 -7.73600042e-01 -5.14378309e-01
2.70773113e-01 -8.13889205e-01 -4.14963931e-01 2.45738268e-01
-1.84448943e-01 -4.24542159e-01 8.46083581e-01 -5.82940757e-01
1.15628672e+00 -2.17568159e+00 1.98120713e-01 1.20173357e-01
2.46503986e-02 4.13826138e-01 -1.48577154e-01 6.39078990e-02
-2.02137064e-02 1.97975393e-02 -2.33646542e-01 -4.22825128e-01
-5.52967846e-01 3.18000376e-01 4.34051529e-02 4.40131426e-01
4.37013656e-01 8.84179473e-01 -1.23534822e+00 -8.16260815e-01
4.96525109e-01 2.65597165e-01 -4.94186968e-01 -6.88134432e-02
-2.40856633e-01 8.91900599e-01 -8.79972279e-01 6.40957475e-01
5.34114420e-01 -2.56171435e-01 -1.68386921e-01 -2.05273643e-01
-1.36685431e-01 1.14576466e-01 -1.33668697e+00 1.94441748e+00
-6.20678902e-01 1.93358138e-01 -2.09383741e-01 -1.14474082e+00
7.02420950e-01 2.78494775e-01 8.42199087e-01 -8.21969330e-01
2.13165600e-02 -1.06863156e-02 -2.36750767e-01 -4.40504551e-01
1.22262113e-01 1.33684874e-01 4.91270609e-02 4.35723037e-01
-7.36948773e-02 5.33015907e-01 4.17763501e-01 1.63540676e-01
1.22488630e+00 4.10686940e-01 1.36291504e-01 3.09636891e-01
1.08141971e+00 -9.58727077e-02 8.10383916e-01 5.62559009e-01
-5.39763927e-01 4.88261998e-01 4.67053950e-01 -3.70602012e-01
-5.06942272e-01 -8.16318095e-01 2.42897496e-01 1.12240028e+00
4.04462159e-01 -3.61845344e-01 -6.17405593e-01 -1.09423780e+00
-3.05285096e-01 4.23737437e-01 -6.43926263e-01 -1.81156561e-01
-8.56363893e-01 -3.99055094e-01 2.54747123e-01 1.00521779e+00
8.98571551e-01 -1.13447189e+00 -7.96479762e-01 4.89221841e-01
-3.64907384e-01 -1.11324382e+00 -5.29910147e-01 5.22088446e-02
-1.13008344e+00 -9.80704904e-01 -9.71802235e-01 -5.69976926e-01
7.47943282e-01 4.81155097e-01 4.58051950e-01 2.10028604e-01
-3.22109871e-02 3.86504322e-01 -7.95502722e-01 2.43921757e-01
-2.92241216e-01 -1.69447392e-01 -3.95144671e-01 5.70207179e-01
-7.22657144e-02 -2.95582980e-01 -9.34740067e-01 5.31514764e-01
-9.59028304e-01 1.04414240e-01 6.92039490e-01 4.99361128e-01
8.18527699e-01 2.83893198e-01 4.73791599e-01 -4.81053263e-01
-3.03631765e-03 -4.41557169e-01 -3.63880664e-01 3.48315150e-01
6.89275563e-02 -7.36676976e-02 6.67963207e-01 -4.34272259e-01
-1.25941420e+00 4.77562964e-01 -4.32681255e-02 -6.16706192e-01
-3.86097699e-01 2.07009390e-01 -1.98509127e-01 8.85703564e-02
4.03809279e-01 6.44749343e-01 -9.18075517e-02 -3.78784746e-01
5.29942870e-01 4.81858253e-01 4.00496244e-01 -4.36130315e-01
5.05189359e-01 7.27103949e-01 -2.54790962e-01 -4.40251291e-01
-4.43078488e-01 -8.27915847e-01 -8.37139487e-01 -4.89504576e-01
1.15368998e+00 -9.19128239e-01 -6.50467277e-01 4.95894134e-01
-1.12679708e+00 -4.18656260e-01 -1.89890131e-01 6.07566178e-01
-6.19047642e-01 5.45636773e-01 -5.46075702e-01 -8.44057143e-01
1.43432692e-02 -1.29907203e+00 1.45291936e+00 4.23153728e-01
1.43365026e-01 -8.68597150e-01 -2.34934092e-01 3.91750455e-01
-6.20622225e-02 1.22167103e-01 3.80067557e-01 -4.53090608e-01
-9.87941027e-01 -2.13423401e-01 -3.74691397e-01 3.84295493e-01
3.10427397e-01 -1.25162244e-01 -5.88624775e-01 -2.12144226e-01
-2.48517513e-01 -1.74173966e-01 1.06823409e+00 2.50091434e-01
1.34767544e+00 -5.53918667e-02 -6.89812720e-01 4.13004011e-01
1.17344189e+00 5.98500252e-01 7.38194764e-01 1.35447755e-01
6.75872862e-01 4.04911369e-01 1.19369960e+00 6.47943079e-01
2.86803901e-01 8.30898821e-01 3.23874086e-01 -1.26621366e-01
-1.31325036e-01 -2.05756843e-01 4.56377149e-01 3.25218171e-01
-1.56730890e-01 -2.44998813e-01 -5.59590459e-01 5.01140118e-01
-2.16529155e+00 -9.20667291e-01 8.87136441e-03 2.06383181e+00
6.21709704e-01 1.43169895e-01 4.80714142e-01 3.06843072e-01
6.97921813e-01 2.07271919e-01 -4.98253196e-01 1.94401577e-01
3.49405706e-01 7.53984451e-02 3.83163691e-01 1.48463428e-01
-1.28253043e+00 1.02393484e+00 4.20823193e+00 1.16746795e+00
-1.02747726e+00 1.68579414e-01 8.65567446e-01 -1.11363567e-01
8.04146379e-03 9.70614702e-02 -7.17427254e-01 7.09946692e-01
5.17526865e-01 1.59846470e-01 2.57482797e-01 7.66482294e-01
6.23228371e-01 -6.68038368e-01 -9.99512851e-01 8.45567286e-01
-1.04197405e-01 -1.34676301e+00 -4.51493636e-02 -1.87946752e-01
5.39614081e-01 -2.59212613e-01 -3.23720813e-01 2.94325024e-01
1.45369679e-01 -4.68496680e-01 6.84360564e-01 6.33601069e-01
2.66877413e-01 -8.47774029e-01 5.20982683e-01 5.61332703e-01
-1.70277679e+00 -3.02236795e-01 -3.39032888e-01 2.78415650e-01
3.91704530e-01 3.09711128e-01 -5.77781439e-01 8.59407306e-01
6.22779071e-01 1.26663506e+00 -5.91978848e-01 1.25020134e+00
-3.47070783e-01 4.53626573e-01 -3.09323311e-01 2.64989380e-02
5.10959804e-01 1.85389128e-02 2.65056789e-01 1.01338315e+00
1.73028126e-01 3.88687462e-01 6.25680208e-01 5.90201259e-01
2.31756985e-01 1.34537995e-01 -2.05815107e-01 1.86065301e-01
3.16246331e-01 1.24251974e+00 -1.31189525e+00 -6.09258950e-01
-4.71895009e-01 1.26480532e+00 2.38589719e-01 4.74427611e-01
-1.14468396e+00 -1.12736605e-01 3.55649024e-01 1.50590762e-02
5.60274601e-01 -2.01535210e-01 2.05804244e-01 -1.03104281e+00
-6.69761151e-02 -5.18119395e-01 6.27830207e-01 -7.14312196e-01
-7.87731767e-01 2.52671242e-01 5.62045164e-02 -1.58522713e+00
-3.58196013e-02 -2.01014578e-01 -6.38358235e-01 5.79595983e-01
-1.47895026e+00 -1.02545238e+00 -4.27809238e-01 8.89968157e-01
8.92763138e-01 2.66296715e-01 2.06371158e-01 3.98216486e-01
-6.78084850e-01 3.87819767e-01 -2.86732376e-01 2.74539828e-01
4.61542457e-01 -9.94674325e-01 -1.10900616e-02 8.97270560e-01
2.67498732e-01 8.44293460e-02 -6.06676303e-02 -7.59191930e-01
-9.39801872e-01 -1.44984376e+00 2.38580421e-01 -8.57068077e-02
4.64733124e-01 8.90242085e-02 -8.66492450e-01 3.76619965e-01
-2.96630561e-01 1.92965746e-01 3.29136968e-01 -4.44362193e-01
7.57783875e-02 -3.65903586e-01 -9.09588039e-01 4.27655637e-01
1.00125885e+00 -2.48174354e-01 -2.27225915e-01 5.04675448e-01
6.16804361e-01 -2.12763458e-01 -6.29774451e-01 4.18182582e-01
3.91555041e-01 -9.99864697e-01 9.63939607e-01 -4.01798517e-01
4.19437617e-01 -6.40345931e-01 -5.76477051e-02 -9.72619116e-01
-1.39945552e-01 -3.78531247e-01 1.41948517e-02 1.22477531e+00
6.50625676e-02 -4.14747447e-01 8.85384738e-01 1.87605992e-01
-1.43262476e-01 -1.08884776e+00 -1.01428461e+00 -8.10490966e-01
-2.40390345e-01 -7.51466215e-01 3.54723692e-01 5.47707200e-01
-1.86445698e-01 -6.55587390e-02 -3.61138135e-01 1.00390539e-01
2.51984030e-01 2.39440456e-01 7.58617342e-01 -5.79594195e-01
-3.94587398e-01 -4.08167034e-01 -5.95030963e-01 -1.67120588e+00
-7.05077276e-02 -6.93233788e-01 3.37485313e-01 -1.59357584e+00
1.70224100e-01 -5.70878983e-01 -5.43324709e-01 7.79124320e-01
-3.46583068e-01 3.97715032e-01 3.88225615e-01 1.83987930e-01
-1.14936984e+00 6.40173376e-01 1.48288190e+00 1.90585963e-02
-4.36821461e-01 3.24507862e-01 -3.45997483e-01 7.59889007e-01
6.29463851e-01 -6.07469797e-01 -5.35259783e-01 -1.43870085e-01
-4.41791147e-01 4.34135020e-01 5.86966217e-01 -1.34899652e+00
2.35604897e-01 -1.86656907e-01 6.15474701e-01 -6.29725873e-01
4.23028976e-01 -6.92548513e-01 -3.24671715e-01 5.54273605e-01
-3.43950301e-01 -4.51878041e-01 3.75886522e-02 1.02376819e+00
-1.96321160e-01 -1.20310068e-01 9.07675683e-01 -2.52152622e-01
-9.75386083e-01 4.96593297e-01 -3.53226304e-01 -1.23270944e-01
1.52312601e+00 -3.07238579e-01 -1.19314464e-02 -2.81512886e-01
-9.54134345e-01 5.68746746e-01 2.36123279e-01 4.38242614e-01
7.00721622e-01 -1.37162900e+00 -5.37519455e-01 1.21789023e-01
8.75849351e-02 -1.10926200e-03 4.07061100e-01 1.37855935e+00
-3.10650855e-01 3.30115646e-01 -1.12201452e-01 -9.71395791e-01
-1.25409913e+00 5.47966301e-01 2.11962774e-01 -3.68708700e-01
-5.46904445e-01 9.70544398e-01 4.34243262e-01 2.19284028e-01
2.13733107e-01 -5.89707017e-01 -3.66946757e-01 1.41507909e-01
6.48842335e-01 4.19353992e-01 -2.32082933e-01 -7.06097424e-01
-5.70569158e-01 5.56551695e-01 -6.77493736e-02 -1.57348230e-01
1.01444304e+00 -2.85171628e-01 2.32525960e-01 2.15145901e-01
1.25092554e+00 -3.39583993e-01 -1.64092481e+00 -4.07985181e-01
-8.44360366e-02 -8.37248921e-01 -3.11749652e-02 -6.36950970e-01
-1.45204568e+00 8.22353184e-01 6.46975815e-01 -5.79974568e-03
1.58510566e+00 2.80069888e-01 9.60871994e-01 3.77578102e-02
4.00524467e-01 -1.01935911e+00 7.40688860e-01 1.54195338e-01
6.01206064e-01 -1.05055010e+00 -1.56152159e-01 -4.38267529e-01
-8.36850584e-01 1.16962171e+00 5.89933276e-01 7.99059346e-02
3.62448931e-01 -3.80668715e-02 -2.61237800e-01 1.21036671e-01
-5.32285690e-01 -6.36841476e-01 1.65586993e-01 4.34702069e-01
1.11360647e-01 -1.74498081e-01 -2.74747342e-01 5.09113252e-01
7.36360610e-01 2.48379290e-01 2.03286245e-01 1.09711277e+00
-6.64675117e-01 -1.33545852e+00 -2.62618750e-01 3.71741802e-01
-1.36084557e-01 2.58865029e-01 -1.16211131e-01 6.18305087e-01
4.55221653e-01 1.00550973e+00 5.70437824e-03 -3.99210811e-01
2.07191318e-01 -9.63098779e-02 3.68858397e-01 -6.91317737e-01
-2.40288779e-01 4.59363580e-01 -2.92518530e-02 -1.07562757e+00
-7.11881161e-01 -9.18431163e-01 -1.67389333e+00 1.79059908e-01
-1.91982776e-01 1.96780656e-02 2.40340099e-01 1.18442822e+00
3.11269790e-01 7.25092411e-01 7.30121076e-01 -9.66638684e-01
-7.71700293e-02 -6.86885655e-01 -3.08866531e-01 3.12334657e-01
9.72821414e-02 -7.32684016e-01 -1.30727261e-01 1.90001860e-01] | [8.470512390136719, 0.6310554146766663] |
b859e83b-f667-4ec7-8161-697800d7857c | from-abstractions-to-natural-languages-for | 1905.00517 | null | https://arxiv.org/abs/1905.00517v2 | https://arxiv.org/pdf/1905.00517v2.pdf | From Abstractions to "Natural Languages" for Coordinating Planning Agents | Despite significant advancements in developing autonomous agents, communication between them often relies on a set of pre-specified symbols for a given domain. In this paper, we investigate the automatic construction of these symbols from abstractions to form "natural languages" for such agents. The focus of this initial investigation is on a task planning setting where one agent (the speaker) directly communicates a "plan sketch" to another agent (the listener) to achieve coordination. In contrast to prior work, we view language formation as a fundamental requirement for resolving miscoordination. This view enables us to "compute" a language from the ground up by mapping physical states to symbols, thus reverse engineering the function of languages. Languages that arise from this process are only approximately expressive of the actual plans, meaning that they specify abstractions over the plan space, which is not only theoretically appealing as it provides the desired flexibility to the listener to choose its plan during execution, but also practically useful since it both reduces the communication cost of the speaker and computational cost of the listener. We formulate this language construction problem and show that it is NEXP-complete. An approximate solution is then developed to relate this problem to task planning problems that have efficient off-the-shelf solutions. Finally, we discuss a multi-agent path-finding domain in our evaluation to provide a comprehensive set of results to illustrate the benefits of the constructed languages and their applications. | ['Li Wang', 'Yu Zhang'] | 2019-05-01 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 3.25467348e-01 6.11924291e-01 1.48352727e-01 -1.90939739e-01
-6.11972570e-01 -8.19241405e-01 8.87279391e-01 2.60882676e-01
-3.66934866e-01 7.99939036e-01 2.91226268e-01 -5.51472247e-01
-1.54080078e-01 -1.05824935e+00 -6.45784199e-01 -5.44661403e-01
-3.77295822e-01 9.48057353e-01 3.91667038e-01 -4.91036952e-01
1.34581670e-01 5.43026984e-01 -1.41529632e+00 -6.42210320e-02
5.91956198e-01 5.35479665e-01 4.84732509e-01 6.28668725e-01
-2.72823960e-01 8.00530732e-01 -5.76218605e-01 2.98090249e-01
4.19127166e-01 -5.82249284e-01 -1.37584686e+00 4.78141308e-01
-4.75334316e-01 -6.53560311e-02 8.60329419e-02 8.85210931e-01
-1.72873382e-02 2.11961627e-01 3.33581030e-01 -1.62514663e+00
4.38710004e-02 8.63760054e-01 -1.65329054e-01 -4.04592812e-01
7.76467502e-01 2.45746180e-01 9.73236322e-01 -1.33071542e-01
7.50288129e-01 1.12974477e+00 2.56115854e-01 5.73807538e-01
-1.32734942e+00 -1.95779324e-01 3.44652236e-01 -2.40956500e-01
-1.44359994e+00 -6.07217312e-01 5.15973270e-01 -5.52867830e-01
1.17617857e+00 3.73634547e-01 5.90821326e-01 4.78079855e-01
2.61658728e-01 3.13858360e-01 1.01347744e+00 -7.77658939e-01
4.47670043e-01 1.93448722e-01 1.36343062e-01 7.35604346e-01
1.47961482e-01 1.51391655e-01 -4.40508276e-01 -4.73445177e-01
7.75857091e-01 -3.62619311e-01 -3.67914796e-01 -6.77378774e-01
-1.23842835e+00 6.78011358e-01 7.96450004e-02 2.53917128e-01
-4.33605313e-01 3.18579942e-01 3.40003073e-01 5.95082104e-01
-1.34794593e-01 5.81683457e-01 -2.14167774e-01 -1.94587141e-01
-2.89516896e-01 4.76804823e-01 1.35151768e+00 1.04755628e+00
8.68593574e-01 -1.85995668e-01 2.84408510e-01 2.54952908e-01
4.28235233e-01 1.98053539e-01 2.18829513e-02 -1.31733632e+00
2.68101841e-01 6.09841943e-01 5.79350770e-01 -6.94053888e-01
-5.95215082e-01 -1.14990018e-01 -2.62947977e-01 6.66159809e-01
6.09019160e-01 -1.59876823e-01 -4.05678064e-01 1.96279657e+00
4.95064050e-01 -2.64444202e-01 4.67225283e-01 6.75963819e-01
-7.68827798e-04 8.73257995e-01 -1.69224828e-01 -4.83602017e-01
1.51240575e+00 -8.42089176e-01 -4.65849072e-01 -2.93534636e-01
9.68045831e-01 -3.74925673e-01 9.59519506e-01 3.34469885e-01
-1.21581101e+00 2.14705206e-02 -9.10460472e-01 1.42678261e-01
-1.61156226e-02 -4.47041035e-01 7.65253842e-01 4.38591391e-01
-1.24748743e+00 2.26846606e-01 -1.01449740e+00 -4.75878835e-01
-1.56475425e-01 4.93484288e-01 -2.65965998e-01 3.25557217e-02
-9.99350607e-01 8.63653123e-01 4.71454650e-01 -1.92841455e-01
-1.09569156e+00 -5.45346513e-02 -9.45145905e-01 9.40763652e-02
9.05629098e-01 -7.99345016e-01 1.69755554e+00 -9.61389720e-01
-1.81617177e+00 5.18769622e-01 -2.97255844e-01 -5.92946887e-01
3.90614957e-01 4.66544509e-01 5.74773923e-02 1.43066533e-02
4.27624404e-01 4.92632151e-01 2.13098973e-01 -1.41445565e+00
-8.50917995e-01 -2.10173219e-01 9.42520261e-01 3.99031579e-01
3.25130343e-01 2.17868567e-01 -2.62733996e-01 5.88873737e-02
2.41156250e-01 -1.22427821e+00 -5.30276299e-01 -2.25955188e-01
-4.28818375e-01 -3.67421359e-01 3.45120102e-01 1.33777885e-02
8.79461527e-01 -2.03819346e+00 2.49732435e-01 4.83297855e-01
2.20797300e-01 -4.42733169e-01 -2.04959080e-01 1.06080854e+00
1.55253023e-01 -6.87899590e-02 -4.44695592e-01 -2.29541332e-01
2.44925648e-01 5.32822907e-01 -5.01403391e-01 5.41551828e-01
-1.53471515e-01 5.08831620e-01 -8.85407269e-01 -3.70855153e-01
-1.28614251e-02 -9.30856839e-02 -7.37903893e-01 1.73317775e-01
-7.48662353e-01 4.75587636e-01 -7.46406376e-01 1.72197565e-01
2.23978326e-01 -1.69028625e-01 7.04806983e-01 5.87675750e-01
-5.14427006e-01 7.36961901e-01 -1.43084610e+00 1.64922392e+00
-6.37392461e-01 2.44918898e-01 5.38837254e-01 -8.29682469e-01
7.40141630e-01 4.17556763e-01 5.43639004e-01 -2.78057754e-01
-2.96532223e-03 1.15660124e-01 1.51495382e-01 -1.94954172e-01
4.36504751e-01 -3.86989981e-01 -5.98210394e-01 1.03273511e+00
-6.17762566e-01 -3.88736546e-01 3.85905594e-01 2.79593885e-01
1.24873698e+00 1.66661561e-01 5.58416426e-01 -4.45483595e-01
6.65448487e-01 4.60949868e-01 5.58808327e-01 7.39745736e-01
7.73275718e-02 -2.91963089e-02 9.11135912e-01 -4.88499463e-01
-6.70485079e-01 -7.17667997e-01 3.67777675e-01 9.73063886e-01
3.00828248e-01 -7.58414745e-01 -8.86973739e-01 -3.45359147e-01
-4.57361609e-01 9.98075485e-01 -2.41318971e-01 1.96894661e-01
-7.00053632e-01 5.88630792e-04 4.24520969e-01 1.26721710e-01
2.49178156e-01 -1.19484603e+00 -1.50388074e+00 4.70518112e-01
-4.42393333e-01 -1.12395167e+00 -4.38593119e-01 3.23266208e-01
-4.57134604e-01 -1.10036361e+00 -1.13014337e-02 -5.81962943e-01
8.24924052e-01 4.49072570e-01 9.18523431e-01 1.66682854e-01
2.57894993e-01 6.41872346e-01 -2.82604218e-01 -5.21369874e-01
-9.59483325e-01 4.48343121e-02 1.01333171e-01 -2.37065405e-01
-2.77817696e-02 -5.93650877e-01 3.28069702e-02 2.18803585e-01
-9.70413327e-01 4.60980773e-01 1.46676868e-01 4.28181291e-01
3.85608792e-01 6.05205178e-01 1.10624753e-01 -6.43022120e-01
9.12724197e-01 -2.07039222e-01 -1.11588848e+00 1.08797424e-01
-3.59209031e-01 5.12159944e-01 7.19499052e-01 -1.93366278e-02
-1.02056575e+00 4.73320603e-01 3.89215529e-01 4.01216149e-01
-3.17142069e-01 6.89949811e-01 -3.01983953e-01 9.19977874e-02
5.54000318e-01 3.16827029e-01 1.43144384e-01 -1.23580262e-01
3.38059366e-01 3.64871353e-01 4.49012697e-01 -1.17311943e+00
6.53975904e-01 5.63640773e-01 2.06947565e-01 -8.80415857e-01
-2.72020519e-01 -1.43472672e-01 -2.87467360e-01 2.00708378e-02
5.41282654e-01 -5.39499104e-01 -1.14153552e+00 4.79961522e-02
-1.51196432e+00 -7.57454157e-01 -4.37970847e-01 1.65894106e-01
-1.12395966e+00 3.87334436e-01 -2.95359462e-01 -1.04224265e+00
2.58749336e-01 -1.48543334e+00 9.39973176e-01 -1.10451192e-01
-5.79007208e-01 -7.38850296e-01 2.66394615e-01 3.09393648e-02
1.87822327e-01 2.08414644e-01 9.40859079e-01 -4.63926196e-01
-8.98982465e-01 1.09754086e-01 7.65649751e-02 -3.38727027e-01
1.70690745e-01 -3.69204968e-01 -5.08482397e-01 -1.57753944e-01
8.03543404e-02 -9.77489725e-03 9.85103995e-02 2.89795995e-01
4.87534583e-01 -4.76007432e-01 -4.67824548e-01 8.03392604e-02
1.22369504e+00 5.13171434e-01 5.35631895e-01 5.78778923e-01
1.13055274e-01 1.04353464e+00 4.39070255e-01 5.74388504e-01
7.96310842e-01 9.55853224e-01 3.55535120e-01 3.00853163e-01
2.25033313e-01 -1.44100294e-01 5.72139859e-01 1.95619583e-01
-1.37326300e-01 -1.33721009e-01 -1.22061598e+00 5.46453595e-01
-2.01683593e+00 -8.50501120e-01 -9.55876149e-03 2.12135530e+00
8.63304853e-01 1.02291010e-01 3.43541235e-01 2.70002671e-02
4.52809274e-01 -2.03419924e-02 -1.80382937e-01 -5.67584634e-01
3.02594423e-01 -4.67319004e-02 4.47146952e-01 1.11452794e+00
-6.48745835e-01 1.05754232e+00 6.08246136e+00 2.31394067e-01
-8.25789988e-01 6.84182299e-03 -4.46057282e-02 2.91893393e-01
-3.71670336e-01 6.27207160e-01 -4.86167014e-01 2.93810293e-02
1.00642812e+00 -4.11323458e-01 1.03237116e+00 7.11980283e-01
5.31719506e-01 -5.04419804e-01 -1.57003129e+00 4.24062192e-01
-3.77366483e-01 -1.21511912e+00 -2.26777345e-01 4.52909470e-01
2.60868162e-01 -9.06412154e-02 -5.33740520e-01 -3.87489870e-02
8.90281439e-01 -9.11219895e-01 1.12743235e+00 2.11636692e-01
4.50147837e-01 -7.64625669e-01 2.82081336e-01 8.48900735e-01
-1.38444436e+00 -1.17123768e-01 -1.39436089e-02 -7.06902325e-01
6.30859256e-01 1.91143245e-01 -1.01442182e+00 6.21810377e-01
1.72462702e-01 -1.94147453e-01 2.93867081e-01 6.24688327e-01
-3.50343019e-01 3.99159603e-02 -6.59932494e-01 -4.64557251e-03
4.94570971e-01 -2.32569307e-01 6.87102616e-01 8.98898244e-01
2.31836975e-01 4.65644419e-01 6.83301628e-01 1.02156615e+00
4.70131129e-01 -2.80763954e-01 -9.07360077e-01 -7.62228146e-02
5.58399558e-01 8.16637456e-01 -8.96355271e-01 -3.26535970e-01
-3.04111362e-01 5.28791606e-01 1.87821388e-01 3.39097977e-01
-6.16078198e-01 -2.86885828e-01 8.10069501e-01 2.05224305e-01
-8.75739828e-02 -7.60773301e-01 -8.07074755e-02 -7.18092680e-01
2.20158752e-02 -1.07808769e+00 1.85693011e-01 -6.31110251e-01
-5.41424215e-01 6.97361171e-01 3.90140921e-01 -9.93641675e-01
-7.44681358e-01 -2.10547969e-01 -4.44826275e-01 8.44691694e-01
-1.37425923e+00 -8.29112768e-01 -1.55808225e-01 4.86800194e-01
4.77767557e-01 1.15540713e-01 1.12266612e+00 -2.46476069e-01
-2.10210443e-01 -1.31227031e-01 -6.63267255e-01 -1.91682637e-01
-1.70928966e-02 -9.83937502e-01 3.26539993e-01 9.96502101e-01
7.08847120e-02 8.16768587e-01 1.03657377e+00 -5.52564740e-01
-1.74427962e+00 -6.51523411e-01 1.11757874e+00 -2.97543138e-01
5.85842192e-01 -2.56876230e-01 -5.68053365e-01 1.08077443e+00
2.50233203e-01 -4.67529505e-01 2.75707275e-01 -1.21896707e-01
-6.17230423e-02 1.96610749e-01 -8.42412651e-01 8.62815857e-01
1.07612896e+00 -3.67047876e-01 -5.97312570e-01 3.50492388e-01
7.52908826e-01 -4.07659411e-01 -4.07238126e-01 -1.33120775e-01
2.97969997e-01 -1.07480824e+00 5.63691080e-01 -4.45830226e-01
1.02327922e-02 -9.18789327e-01 -2.44872063e-01 -1.22630382e+00
-6.85886666e-02 -1.15850067e+00 4.18069899e-01 8.85924041e-01
5.18572211e-01 -1.06787264e+00 6.27276659e-01 1.04114223e+00
-2.85429507e-01 -3.17815930e-01 -9.64436233e-01 -8.45995605e-01
-1.46809235e-01 -7.51556635e-01 8.28552485e-01 7.30770767e-01
6.41625583e-01 4.74975437e-01 -1.92476995e-02 4.96824294e-01
4.16720152e-01 3.57484847e-01 1.04636216e+00 -1.15176988e+00
-6.25765026e-01 -4.09606248e-01 8.56861994e-02 -1.28376210e+00
3.56410265e-01 -8.62483263e-01 5.07893205e-01 -1.60792601e+00
-2.36796379e-01 -8.94380808e-01 4.01491880e-01 7.92253077e-01
7.21481025e-01 -6.78617597e-01 3.40527624e-01 2.39115968e-01
-4.60266858e-01 3.12535465e-01 1.05466938e+00 -3.75913680e-02
-6.06833816e-01 1.46503881e-01 -8.63485456e-01 8.21757495e-01
7.94322670e-01 -5.20821452e-01 -7.88667560e-01 -5.10842383e-01
3.62443924e-01 6.67040765e-01 2.26303369e-01 -7.36022711e-01
4.76118177e-01 -7.39939868e-01 -8.56144249e-01 -1.46210575e-02
3.95024598e-01 -1.08621991e+00 5.33755362e-01 6.49366021e-01
-4.34918076e-01 2.45124072e-01 6.11579083e-02 2.92773664e-01
2.64803022e-02 -5.51237881e-01 4.36684549e-01 -4.51469213e-01
-5.79409301e-01 -1.93674475e-01 -9.09717917e-01 -2.31697991e-01
1.23983645e+00 -2.66910851e-01 -1.11401767e-01 -7.22693086e-01
-5.69019794e-01 4.36609924e-01 7.43487716e-01 -1.51467144e-01
3.36325854e-01 -8.36961210e-01 -5.20369828e-01 9.83929634e-02
1.09081334e-02 3.87074947e-01 -2.38660589e-01 7.78818786e-01
-5.40732741e-01 5.17929494e-01 -1.85815185e-01 -3.15817982e-01
-1.09622097e+00 4.67069387e-01 3.64982009e-01 -4.06337589e-01
-7.55334258e-01 3.39664519e-01 5.93167484e-01 -3.87561619e-01
1.63650438e-01 -6.32731497e-01 4.91182283e-02 -3.49767089e-01
5.77038050e-01 1.06782116e-01 -3.64421934e-01 -4.92660046e-01
-5.81306696e-01 4.33569014e-01 2.71077335e-01 -6.92307174e-01
1.23237228e+00 -3.13157141e-01 -4.87599880e-01 1.99226186e-01
4.10726458e-01 1.97161257e-01 -1.00418854e+00 -2.39518434e-01
2.28911281e-01 -1.40627027e-01 -2.98119456e-01 -3.66692394e-01
-4.53211099e-01 4.11888063e-01 -2.81336874e-01 8.27282488e-01
9.09788251e-01 3.08594137e-01 5.50178826e-01 7.60383129e-01
1.07959557e+00 -9.10501420e-01 -1.44659847e-01 7.64337301e-01
9.54865992e-01 -5.98171890e-01 -1.01483151e-01 -7.77639925e-01
-7.27979243e-01 1.06933439e+00 3.19747120e-01 6.34040162e-02
7.33424351e-02 7.38639593e-01 -9.13001224e-02 -3.42786223e-01
-9.34651732e-01 -3.26082170e-01 -6.30672634e-01 8.38328838e-01
-9.48776724e-04 2.80811906e-01 -2.75924206e-01 4.28941727e-01
-4.88545448e-01 4.14768383e-02 1.13683140e+00 1.35834908e+00
-7.39089429e-01 -1.52595210e+00 -5.10535598e-01 -2.53490239e-01
-2.25476129e-03 2.43136853e-01 -4.82239395e-01 9.37469244e-01
-3.70192416e-02 1.28157914e+00 4.95096948e-03 -2.11796816e-02
3.91007096e-01 1.07785231e-02 5.03502071e-01 -1.06097078e+00
-2.70221680e-01 9.80740972e-03 7.10559666e-01 -5.09083807e-01
-5.17076612e-01 -6.38660848e-01 -1.86951840e+00 -3.14840198e-01
-7.56555572e-02 4.65547562e-01 4.87016201e-01 1.07204735e+00
2.19972461e-01 5.47652423e-01 4.10781264e-01 -8.75256002e-01
-5.74630737e-01 -2.82862097e-01 -4.80294287e-01 -1.19658984e-01
2.15115830e-01 -5.68073094e-01 -9.44681559e-03 1.24091357e-01] | [4.440442085266113, 1.1946226358413696] |
870cd6af-f8d1-460e-b93d-3f1fde007268 | conditional-and-residual-methods-in-scalable | 2305.02562 | null | https://arxiv.org/abs/2305.02562v2 | https://arxiv.org/pdf/2305.02562v2.pdf | Conditional and Residual Methods in Scalable Coding for Humans and Machines | We present methods for conditional and residual coding in the context of scalable coding for humans and machines. Our focus is on optimizing the rate-distortion performance of the reconstruction task using the information available in the computer vision task. We include an information analysis of both approaches to provide baselines and also propose an entropy model suitable for conditional coding with increased modelling capacity and similar tractability as previous work. We apply these methods to image reconstruction, using, in one instance, representations created for semantic segmentation on the Cityscapes dataset, and in another instance, representations created for object detection on the COCO dataset. In both experiments, we obtain similar performance between the conditional and residual methods, with the resulting rate-distortion curves contained within our baselines. | ['Ivan V. Bajić', 'Yalda Foroutan', 'Alon Harell', 'Anderson de Andrade'] | 2023-05-04 | null | null | null | null | ['image-reconstruction'] | ['computer-vision'] | [ 6.67498112e-01 7.84333110e-01 1.22094765e-01 -4.15115923e-01
-1.20367825e+00 -5.65037131e-01 9.70759988e-01 1.25771701e-01
-5.41473567e-01 4.03699547e-01 6.18852973e-01 -2.82487392e-01
1.72896370e-01 -4.40631777e-01 -6.97654545e-01 -5.43290377e-01
4.93871383e-02 4.51191336e-01 2.95631826e-01 2.88285881e-01
4.90223616e-01 2.05098793e-01 -1.55505824e+00 4.37030464e-01
8.73393655e-01 1.14192116e+00 5.09379745e-01 1.01055562e+00
3.53623450e-01 1.01638913e+00 -3.03937882e-01 -8.42386544e-01
3.55691135e-01 -4.38703299e-01 -1.09684896e+00 4.60287422e-01
4.59361643e-01 -3.32602799e-01 -4.61572021e-01 9.23788786e-01
2.35137269e-01 1.92431808e-01 1.07723522e+00 -8.50696921e-01
-5.42360783e-01 5.24707139e-01 -3.17470402e-01 4.27647009e-02
5.30085087e-01 1.71991765e-01 1.10234880e+00 -6.35256052e-01
8.97571206e-01 1.51839566e+00 4.94652361e-01 4.51279610e-01
-1.61540461e+00 -1.51731178e-01 -6.77871928e-02 -1.21503919e-01
-1.09794617e+00 -7.54718006e-01 8.61240327e-02 -6.60778284e-01
1.29667187e+00 1.46877602e-01 4.33019042e-01 8.66362274e-01
-4.34671566e-02 9.74918425e-01 8.33908916e-01 -5.04854143e-01
3.02266717e-01 2.76150554e-01 -1.74609393e-01 5.96432447e-01
1.38005048e-01 3.51304114e-01 -9.99632031e-02 6.24499880e-02
7.02796519e-01 -3.25627774e-01 -2.46252939e-01 -3.39298368e-01
-1.19404781e+00 9.90346193e-01 5.50858796e-01 9.90589336e-02
-1.76367685e-01 4.29364324e-01 3.86538863e-01 -2.88943164e-02
7.27645218e-01 3.83752406e-01 -3.06447119e-01 -7.19714984e-02
-1.34818327e+00 2.82920778e-01 8.49555194e-01 1.27759469e+00
6.10197842e-01 2.08847541e-02 -2.98687100e-01 7.22370982e-01
2.38627955e-01 2.20199898e-01 4.37896937e-01 -1.54511464e+00
7.60003388e-01 6.88550398e-02 6.21517375e-02 -7.66169786e-01
-1.09709568e-01 -1.58701628e-01 -6.25678539e-01 5.56933135e-02
2.22816870e-01 1.01447716e-01 -1.37889814e+00 1.65427864e+00
-1.10691905e-01 8.83005485e-02 2.93259352e-01 5.10806024e-01
4.71553266e-01 6.68872833e-01 9.43388566e-02 -2.39284381e-01
1.20083570e+00 -1.16029680e+00 -4.72225815e-01 -2.72068799e-01
8.30224693e-01 -6.35163784e-01 5.22753000e-01 4.02298629e-01
-1.34488630e+00 -5.19986272e-01 -1.01134574e+00 -5.26413083e-01
-1.49595402e-02 1.41803131e-01 6.33619905e-01 4.98598963e-01
-1.42388737e+00 6.83547139e-01 -1.01852441e+00 -3.29649687e-01
4.50212300e-01 -1.18047530e-02 -1.10472776e-01 -2.04437822e-01
-7.87922859e-01 1.11572301e+00 5.60835540e-01 -5.73193848e-01
-1.14299679e+00 -4.20900494e-01 -1.16593671e+00 2.99214542e-01
-9.04015526e-02 -4.64380562e-01 1.58832765e+00 -1.10137391e+00
-1.23876441e+00 1.05250025e+00 -1.96162295e-02 -1.07840705e+00
7.70311832e-01 2.65148096e-02 1.46653682e-01 3.72090399e-01
1.78263262e-01 1.23753679e+00 6.64962888e-01 -1.18383169e+00
-3.68871450e-01 -1.34312600e-01 1.21506840e-01 3.10056686e-01
2.58907229e-01 -4.58033271e-02 -8.00184190e-01 -8.27139795e-01
2.72952020e-01 -1.13400936e+00 -5.97325802e-01 1.03767395e-01
-3.15627128e-01 3.17976847e-02 2.07291663e-01 -1.03822565e+00
8.86529505e-01 -2.45318508e+00 4.04870421e-01 3.09947710e-02
6.18812814e-03 -9.48841423e-02 -1.24323390e-01 4.18330841e-02
-1.79134533e-01 4.24325317e-01 -9.10402954e-01 -7.72755504e-01
-8.31614360e-02 2.15729371e-01 -2.41640359e-01 4.61269051e-01
3.19259018e-01 7.22770751e-01 -5.36348522e-01 -6.36098683e-01
2.05731928e-01 5.90412855e-01 -1.06261182e+00 2.63933510e-01
-2.45140612e-01 2.96891659e-01 -1.20336726e-01 2.83978701e-01
5.87597728e-01 -2.05572814e-01 2.17158809e-01 1.71254978e-01
1.63553253e-01 5.72358906e-01 -9.79693413e-01 1.99674916e+00
-6.47065938e-01 9.04141605e-01 1.75654188e-01 -9.57046390e-01
5.41195095e-01 4.74078327e-01 3.40994090e-01 -7.02934921e-01
-8.49255249e-02 1.39007315e-01 -2.57169276e-01 -1.32909134e-01
7.04442143e-01 -1.82276219e-01 3.59570011e-02 3.53459537e-01
4.30614084e-01 -7.24925697e-01 2.62399048e-01 6.83585942e-01
1.02464831e+00 1.59196332e-01 3.46035898e-01 -1.57099009e-01
-4.43888269e-03 -5.66743799e-02 2.09617481e-01 6.21971667e-01
-8.16639513e-02 1.23995805e+00 4.57888007e-01 -3.08294166e-02
-1.41895878e+00 -1.06917822e+00 -4.58717346e-01 6.55020893e-01
-1.61684945e-01 -6.47608995e-01 -1.00664306e+00 -5.03731906e-01
-2.56043851e-01 9.90023136e-01 -5.72130680e-01 -1.08432723e-02
-3.09796184e-01 -3.96110803e-01 5.31800568e-01 5.02645195e-01
6.83635592e-01 -1.09921408e+00 -9.28257465e-01 6.28435146e-03
-5.47239900e-01 -1.39613795e+00 -2.59034425e-01 3.44323397e-01
-1.09577096e+00 -8.67720306e-01 -7.56355345e-01 -7.35901415e-01
4.64714825e-01 4.77504656e-02 1.33013260e+00 1.76951915e-01
-3.04040045e-01 5.19733489e-01 -3.95483941e-01 -2.30237052e-01
-5.33874393e-01 -9.08816457e-02 -4.05940920e-01 -3.34916651e-01
-2.71506608e-01 -3.16795081e-01 -5.61669052e-01 -4.68771085e-02
-1.07875729e+00 2.43319720e-01 4.48039532e-01 5.10567784e-01
4.88676697e-01 -2.07494557e-01 -2.24382564e-01 -9.09494102e-01
2.87471503e-01 -6.68997109e-01 -6.74561977e-01 6.58887019e-03
-7.25644231e-01 3.20947737e-01 1.67785078e-01 -4.90157232e-02
-1.32064390e+00 4.29629743e-01 -2.66154826e-01 -2.14051828e-01
-1.02833807e-01 4.61263299e-01 2.46183917e-01 2.86494315e-01
8.30032229e-01 2.08562732e-01 -3.03450197e-01 -4.01726246e-01
7.15467811e-01 6.06006086e-01 7.00694442e-01 -3.88827801e-01
4.31764275e-01 4.37663227e-01 -2.90026426e-01 -6.42271876e-01
-5.92118144e-01 -4.49368894e-01 -6.47346020e-01 2.65280724e-01
1.26022041e+00 -1.31662452e+00 5.51629104e-02 9.66611356e-02
-1.39175856e+00 -3.36384892e-01 -5.39171994e-01 4.71419781e-01
-1.30382967e+00 4.38297391e-01 -6.95096374e-01 -8.80768359e-01
7.39123672e-02 -1.52392316e+00 1.33657241e+00 -1.83967516e-01
-7.36427307e-02 -9.62016284e-01 1.07888699e-01 3.29093635e-01
2.57461578e-01 2.71166652e-01 7.15052903e-01 -4.41787809e-01
-7.20899224e-01 -9.64532271e-02 -4.94624615e-01 4.63797748e-01
-4.41998363e-01 -9.13809463e-02 -1.19683671e+00 -3.01998377e-01
1.66931614e-01 -6.00255489e-01 1.57954800e+00 5.80919504e-01
1.14794457e+00 -4.05571669e-01 -2.30997637e-01 8.17958176e-01
1.60332608e+00 -9.43230279e-03 1.12626278e+00 -1.44693512e-03
4.49075490e-01 7.89598584e-01 2.71039099e-01 4.15667176e-01
4.28298533e-01 7.71242678e-01 3.80366027e-01 3.93540636e-02
-3.77118051e-01 -2.85604119e-01 2.95155168e-01 6.77634716e-01
-9.17755067e-02 -2.62822777e-01 -7.52252996e-01 7.88925648e-01
-1.77698410e+00 -9.13815081e-01 2.83184629e-02 2.12192917e+00
7.16313839e-01 1.12059802e-01 -1.36334360e-01 -2.33235341e-02
5.60765028e-01 2.52493441e-01 -2.83629507e-01 -7.09250748e-01
1.12436831e-01 1.92620009e-01 6.93577588e-01 6.18882477e-01
-1.19540584e+00 8.76126051e-01 7.98731756e+00 6.89276278e-01
-3.96067113e-01 2.09819496e-01 9.08277810e-01 -9.38185379e-02
-3.09293509e-01 2.90939838e-01 -3.23505372e-01 3.21424454e-01
1.53930092e+00 1.16095066e-01 6.51732981e-01 9.82991278e-01
-3.00314128e-01 -3.97498727e-01 -1.22550619e+00 9.42517579e-01
1.44258246e-01 -1.32161880e+00 -6.38455003e-02 1.89569995e-01
8.85268033e-01 2.85469711e-01 1.81648076e-01 2.10460067e-01
8.35626364e-01 -1.03423333e+00 1.01524258e+00 2.59679347e-01
1.01816344e+00 -5.25342464e-01 4.74736631e-01 2.23506793e-01
-1.16015625e+00 -1.00922756e-01 -4.29332405e-01 9.23612416e-02
1.87298745e-01 4.92582500e-01 -5.44217944e-01 4.05540019e-01
6.16923213e-01 7.13038683e-01 -7.00730741e-01 1.11632955e+00
-3.43617678e-01 6.48324430e-01 -3.35538030e-01 5.37478864e-01
2.97469050e-01 -1.22556970e-01 3.96135628e-01 1.57685208e+00
2.40369558e-01 9.23590586e-02 9.22149569e-02 9.52134490e-01
-2.35040799e-01 -1.76715165e-01 -8.18064570e-01 1.48853347e-01
2.62865275e-01 8.49578738e-01 -8.42689335e-01 -6.32821739e-01
-3.03619236e-01 1.13428736e+00 4.78447229e-01 3.83371294e-01
-7.40020633e-01 -1.26810759e-01 4.10576165e-01 -3.78902219e-02
6.28048301e-01 -2.41771981e-01 -5.01867175e-01 -1.30155194e+00
-1.73536822e-01 -6.08326674e-01 4.22062546e-01 -9.47114289e-01
-7.80367255e-01 4.41228062e-01 3.46606195e-01 -9.20773387e-01
-6.77998245e-01 -3.35057944e-01 -2.17442304e-01 7.23627925e-01
-1.28859997e+00 -7.01636493e-01 4.54094000e-02 3.43665302e-01
9.28727627e-01 1.61196530e-01 7.60748684e-01 -8.01798403e-02
-5.93477562e-02 3.66557032e-01 2.35336512e-01 1.82147212e-02
1.72532395e-01 -1.28187716e+00 8.15953493e-01 1.14106166e+00
5.10930300e-01 3.68102521e-01 6.35319948e-01 -4.61922705e-01
-8.90958846e-01 -1.09963250e+00 7.07251847e-01 -5.92194259e-01
1.44272804e-01 -4.59230959e-01 -6.15613461e-01 1.07978404e+00
2.84185290e-01 -2.36343071e-01 2.66938478e-01 -8.97885486e-02
-6.14495039e-01 7.26115227e-01 -1.45609558e+00 3.54601264e-01
9.94415164e-01 -7.38117814e-01 -5.89994848e-01 4.86899704e-01
1.20941079e+00 -4.50784564e-01 -5.94347119e-01 7.48496950e-02
4.10949200e-01 -1.24113977e+00 9.66717601e-01 -2.14347020e-01
7.84970045e-01 1.04193710e-01 -5.68822861e-01 -1.18015134e+00
-4.56105441e-01 -3.17738414e-01 2.26676300e-01 9.53662038e-01
6.08094513e-01 -3.43289047e-01 4.79649961e-01 8.99443328e-01
-2.12747812e-01 -2.65146583e-01 -1.06676424e+00 -6.46333814e-01
1.97992459e-01 -7.92980313e-01 1.21072471e-01 5.17577946e-01
-1.34454221e-01 2.01948032e-01 -2.05088258e-01 -2.01836675e-01
5.42223811e-01 -8.45694318e-02 4.02204067e-01 -1.03053010e+00
-5.01118779e-01 -3.50000024e-01 -4.81393456e-01 -1.23835433e+00
6.45010918e-02 -1.23088634e+00 4.37491447e-01 -1.66322315e+00
4.68575805e-01 -1.82835579e-01 1.04320362e-01 2.31932729e-01
1.18233085e-01 3.94173473e-01 7.04129696e-01 4.18377876e-01
-6.52061522e-01 5.29437363e-01 8.89490545e-01 -2.22006962e-01
5.00889681e-02 -3.24140102e-01 -6.96893215e-01 7.35855639e-01
4.73897398e-01 -6.24976516e-01 -5.92989445e-01 -6.87881529e-01
1.35778010e-01 1.92933634e-01 3.82125080e-01 -1.31512904e+00
3.45301395e-03 1.02106847e-01 2.49651715e-01 -3.58938694e-01
4.10321712e-01 -8.18132043e-01 2.67224340e-03 5.43049216e-01
-8.02426100e-01 -6.00939058e-02 8.87563750e-02 8.32289755e-01
-2.30402946e-01 -4.25558448e-01 1.01874268e+00 -5.43576419e-01
-4.57928151e-01 1.02795072e-01 -5.43337226e-01 2.13788778e-01
7.56095350e-01 -2.82341987e-01 -1.66922610e-03 -6.59767509e-01
-9.50729907e-01 7.38220885e-02 6.17839456e-01 9.41470489e-02
7.72767007e-01 -1.18550730e+00 -8.26331556e-01 1.08995043e-01
1.00037634e-01 2.22703800e-01 -1.03713147e-01 5.14231026e-01
-6.07608974e-01 6.93888247e-01 -4.71307859e-02 -5.98120749e-01
-8.48340154e-01 7.39729524e-01 1.74412832e-01 -4.31530684e-01
-4.71576065e-01 9.16105747e-01 4.28368300e-01 -3.70338589e-01
2.89700955e-01 -5.06291509e-01 -1.31367482e-02 -2.36884132e-02
3.44791174e-01 1.63772583e-01 -2.57992931e-02 -7.42377460e-01
-1.37189150e-01 2.18538865e-01 1.59990098e-02 -7.39517391e-01
1.20671558e+00 -3.02095830e-01 1.22928157e-01 1.89897984e-01
1.44984925e+00 -5.19551814e-01 -1.57954133e+00 -5.75445928e-02
1.87578686e-02 -5.05988121e-01 1.72147900e-01 -7.09936261e-01
-1.05156016e+00 9.89904642e-01 6.85314894e-01 2.06959218e-01
1.13284266e+00 1.38185561e-01 3.46995294e-01 2.95249194e-01
3.90942246e-01 -1.05157316e+00 -7.46388286e-02 5.01129806e-01
9.33389843e-01 -1.15740740e+00 2.18806788e-01 -3.56057912e-01
-9.10806119e-01 8.74379098e-01 8.59355256e-02 -2.96673775e-01
5.94993353e-01 1.59291327e-01 -5.82899511e-01 -5.23098819e-02
-9.27518487e-01 -4.00309443e-01 1.95950285e-01 7.70626128e-01
4.90625978e-01 1.57638237e-01 -2.10531324e-01 -5.35943061e-02
-1.75452232e-01 -1.51656523e-01 5.51637232e-01 8.69420588e-01
-3.82781267e-01 -6.98610246e-01 -4.64367010e-02 4.76603657e-01
-4.38758999e-01 -4.82328415e-01 -3.17296386e-01 5.34033060e-01
-6.95834681e-02 1.14375114e+00 3.06147993e-01 -2.57164270e-01
1.14386901e-02 8.46975669e-02 4.53451276e-01 -8.01618636e-01
-3.97920907e-01 1.52298614e-01 3.11509013e-01 -8.12615812e-01
-4.57980990e-01 -9.12628591e-01 -1.19941139e+00 -5.03028231e-03
-3.58310491e-01 -8.96474645e-02 9.30242956e-01 8.55912149e-01
3.08008939e-01 3.10513496e-01 3.21411967e-01 -1.13108861e+00
-3.74134719e-01 -1.13120115e+00 -3.87111127e-01 5.62424481e-01
3.30123156e-01 -5.23315728e-01 -4.68026787e-01 4.07005459e-01] | [10.917855262756348, 0.10810727626085281] |
c0bd3c0e-d056-4d11-8101-aa9a8035fd6f | joint-span-segmentation-and-rhetorical-role | 2302.06448 | null | https://arxiv.org/abs/2302.06448v1 | https://arxiv.org/pdf/2302.06448v1.pdf | Joint Span Segmentation and Rhetorical Role Labeling with Data Augmentation for Legal Documents | Segmentation and Rhetorical Role Labeling of legal judgements play a crucial role in retrieval and adjacent tasks, including case summarization, semantic search, argument mining etc. Previous approaches have formulated this task either as independent classification or sequence labeling of sentences. In this work, we reformulate the task at span level as identifying spans of multiple consecutive sentences that share the same rhetorical role label to be assigned via classification. We employ semi-Markov Conditional Random Fields (CRF) to jointly learn span segmentation and span label assignment. We further explore three data augmentation strategies to mitigate the data scarcity in the specialized domain of law where individual documents tend to be very long and annotation cost is high. Our experiments demonstrate improvement of span-level prediction metrics with a semi-Markov CRF model over a CRF baseline. This benefit is contingent on the presence of multi sentence spans in the document. | ['Matthias Grabmair', 'Philipp Bock', 'T. Y. S. S. Santosh'] | 2023-02-13 | null | null | null | null | ['argument-mining'] | ['natural-language-processing'] | [ 7.48991966e-01 5.01586556e-01 -7.65428603e-01 -4.91291642e-01
-1.34687078e+00 -8.21746290e-01 7.32525229e-01 5.79392433e-01
-5.41689038e-01 1.24010932e+00 9.35923398e-01 -6.26048446e-01
-1.88898876e-01 -3.33854735e-01 -3.23346138e-01 -2.67153651e-01
2.42921770e-01 5.53232908e-01 2.84460992e-01 -1.36954814e-01
8.74222159e-01 2.20462903e-01 -9.05184150e-01 6.50896728e-01
1.11982942e+00 4.36124325e-01 5.88359572e-02 7.16554463e-01
-5.22252262e-01 1.59821653e+00 -9.83687222e-01 -6.16972268e-01
-8.00003633e-02 -2.63752431e-01 -1.65427601e+00 6.41049147e-02
4.45589006e-01 -1.10486530e-01 -3.14440466e-02 6.33977532e-01
3.88794392e-01 5.88577464e-02 1.02148557e+00 -8.74730825e-01
-2.33497486e-01 8.61530483e-01 -1.01929271e+00 6.41632915e-01
4.47747111e-01 -3.94528329e-01 1.20567107e+00 6.58054836e-03
1.01530802e+00 1.26583791e+00 6.69025481e-01 2.93979973e-01
-1.06437147e+00 -1.82829708e-01 1.47905588e-01 2.37087935e-01
-6.80291176e-01 -4.98849511e-01 9.10321712e-01 -7.50762343e-01
1.05779004e+00 4.89280112e-02 1.68348700e-01 9.31611001e-01
2.30879873e-01 8.58332276e-01 9.44023728e-01 -8.73280406e-01
1.63147017e-01 -2.81439841e-01 5.71598411e-01 4.15591806e-01
1.10834949e-01 -7.34204888e-01 -4.41703588e-01 -5.67376733e-01
1.89625651e-01 -3.67819250e-01 3.28474849e-01 3.99419963e-01
-7.21712470e-01 1.14978957e+00 -2.90091544e-01 5.01084983e-01
-5.13192773e-01 4.01047431e-02 1.00572801e+00 6.53778464e-02
5.56163728e-01 4.58076179e-01 -4.72666919e-01 -3.53827149e-01
-1.29585862e+00 3.24313790e-01 7.52900958e-01 9.12971973e-01
3.90966147e-01 -5.30510426e-01 -9.24326360e-01 9.54461932e-01
2.72306874e-02 4.13353443e-02 2.73002565e-01 -1.20719004e+00
1.05584908e+00 6.50166571e-01 1.24249123e-01 -5.94291449e-01
-1.70485869e-01 -7.04136789e-02 -2.85097182e-01 -4.01320875e-01
3.01929504e-01 -2.54807919e-01 -9.26535785e-01 1.62393105e+00
2.92105436e-01 -3.56907219e-01 9.22497585e-02 1.21724784e-01
6.42747223e-01 4.43466574e-01 7.32693434e-01 -6.75481617e-01
1.59641373e+00 -8.40421796e-01 -8.41942608e-01 -1.12080634e-01
9.25704002e-01 -1.01124203e+00 7.39632368e-01 -1.47411749e-01
-1.00193667e+00 -2.13053197e-01 -6.60728335e-01 -3.64457339e-01
9.29434747e-02 1.05905324e-01 6.62515640e-01 3.65256041e-01
-3.42243433e-01 3.65949512e-01 -4.20294225e-01 -5.40558919e-02
8.84465992e-01 -4.58693132e-02 -1.34075254e-01 1.44566640e-01
-1.19275451e+00 1.06567967e+00 5.82736552e-01 -3.84642631e-01
-4.70793873e-01 -6.12749994e-01 -9.72906590e-01 5.38319116e-03
5.69140017e-01 -4.03364271e-01 1.52055085e+00 -3.84576917e-01
-8.15334857e-01 1.26501894e+00 -5.08329034e-01 -7.53105223e-01
4.42503691e-01 -3.83810043e-01 -2.41926908e-01 4.28888440e-01
8.98464680e-01 4.92844909e-01 6.63290799e-01 -9.35777366e-01
-5.92300832e-01 -3.77214938e-01 2.41022810e-01 1.42050192e-01
-1.71098411e-02 6.16256714e-01 5.75815551e-02 -5.34762502e-01
-2.41705611e-01 -6.27492964e-01 -5.11887491e-01 -8.73471498e-01
-7.74268687e-01 -9.41719294e-01 6.13544226e-01 -1.11030412e+00
1.55209434e+00 -1.60793376e+00 -2.17058077e-01 -2.27547616e-01
2.20768582e-02 1.80246383e-02 1.96986541e-01 7.91968107e-01
-2.02179234e-02 5.13261616e-01 -4.29454118e-01 -2.68447131e-01
-1.12622567e-01 3.28021467e-01 -5.58741331e-01 3.73535335e-01
1.61274686e-01 9.13052976e-01 -6.61839783e-01 -1.52117777e+00
-3.25758040e-01 -2.17198461e-01 -1.62602618e-01 -5.36293611e-02
-4.52788919e-01 7.29883611e-01 -6.53469145e-01 5.40605307e-01
4.64433849e-01 -1.08856551e-01 2.85720706e-01 7.44675547e-02
-1.30680948e-01 8.07922006e-01 -5.22124290e-01 1.90636098e+00
-3.22207063e-01 5.90949595e-01 -7.53101259e-02 -1.24986696e+00
6.07031524e-01 4.81673449e-01 6.12073600e-01 -5.97996414e-01
2.16197241e-02 6.04258925e-02 -3.86650860e-03 -8.78984392e-01
7.21956968e-01 -2.99234599e-01 -6.29199982e-01 5.92165232e-01
-1.45062074e-01 5.70135601e-02 8.38542581e-01 5.38956583e-01
1.24745452e+00 8.52308199e-02 5.45639336e-01 -1.52592272e-01
5.81382334e-01 3.44949037e-01 8.12151313e-01 7.88484275e-01
-1.69708744e-01 2.35414982e-01 1.04589152e+00 -1.53296916e-02
-1.11053813e+00 -5.33462107e-01 -3.93294156e-01 1.04818499e+00
-2.53410131e-01 -3.47068489e-01 -7.92627394e-01 -1.14395618e+00
-2.67704785e-01 8.17951262e-01 -5.30331552e-01 3.43039006e-01
-1.00202000e+00 -5.14490604e-01 6.72091186e-01 4.97353256e-01
3.65620196e-01 -1.34276581e+00 -9.51187372e-01 4.42408174e-01
-4.11095053e-01 -1.40559304e+00 -3.77082020e-01 1.29943132e-01
-6.71490729e-01 -1.22671294e+00 -3.55730534e-01 -8.58900189e-01
3.73024732e-01 -3.20292294e-01 1.03981674e+00 -8.24721530e-02
-3.38958293e-01 5.37949875e-02 -4.92978036e-01 -6.13560081e-01
-5.71545482e-01 4.67135847e-01 -5.07574618e-01 -4.41015601e-01
3.45255911e-01 -6.78814948e-01 -3.63151282e-01 -3.58800143e-01
-9.00767505e-01 1.76532939e-02 5.03985882e-01 7.22414374e-01
2.47049540e-01 -1.65573478e-01 9.83817756e-01 -1.58519268e+00
1.12020862e+00 -4.29421574e-01 -2.08859906e-01 6.38666213e-01
-3.94816399e-01 2.34591350e-01 2.81447709e-01 -2.32730761e-01
-1.39098895e+00 -1.64138824e-01 -1.28560662e-01 2.50384629e-01
-3.30662251e-01 6.74107790e-01 -3.56335044e-02 7.48411834e-01
5.96417427e-01 -8.73853713e-02 -2.32997760e-02 -5.37714660e-01
4.05301362e-01 9.40667808e-01 3.34530801e-01 -6.68015599e-01
3.46862674e-01 2.02680320e-01 4.59863469e-02 -4.82672989e-01
-1.48660207e+00 -8.11824024e-01 -9.06170070e-01 -7.74738789e-02
9.84162986e-01 -7.10768878e-01 -4.11501586e-01 -2.09411219e-01
-1.62480879e+00 -1.08169824e-01 -2.89025694e-01 2.20429793e-01
-4.93466645e-01 8.67225528e-01 -8.04940522e-01 -9.63428140e-01
-6.68096125e-01 -6.12503827e-01 1.21415949e+00 2.95885086e-01
-6.60868287e-01 -8.60569358e-01 2.71739662e-01 1.00820518e+00
-9.17803049e-02 6.38044059e-01 1.22930622e+00 -8.24457765e-01
1.88634202e-01 -3.01587880e-01 -1.32190421e-01 1.23124093e-01
-5.14041744e-02 -1.81983054e-01 -7.20369220e-01 2.78745830e-01
2.26223264e-02 -5.40906787e-01 1.01542521e+00 4.84740198e-01
9.13481772e-01 -6.04560733e-01 -3.72616857e-01 -4.21955585e-01
1.40153825e+00 3.10518086e-01 6.87053978e-01 2.09580481e-01
4.83920932e-01 1.07216167e+00 8.30299437e-01 2.93194413e-01
1.58433244e-01 4.49673682e-01 -1.65765554e-01 1.59732997e-01
-2.21936062e-01 -2.82906860e-01 1.03652269e-01 4.49484974e-01
5.58661520e-02 -3.34776610e-01 -1.03157783e+00 9.44052458e-01
-1.96525049e+00 -1.33960545e+00 -4.51309651e-01 1.56757510e+00
1.14170718e+00 4.38508928e-01 2.54808158e-01 3.65227312e-01
9.08861518e-01 3.66314650e-01 -1.52373582e-01 -7.78856933e-01
1.02602819e-03 1.43869743e-01 4.06071484e-01 5.46640456e-01
-1.24048615e+00 1.04526567e+00 6.00668478e+00 1.29178452e+00
-4.84175056e-01 4.39943492e-01 8.08241844e-01 1.91971004e-01
-2.96249330e-01 3.78579289e-01 -8.78693342e-01 6.32727146e-01
9.03096616e-01 -3.56313027e-02 -4.25053954e-01 6.95982039e-01
3.90329093e-01 -6.80203915e-01 -7.41823733e-01 4.29449260e-01
2.02248860e-02 -1.58624935e+00 -1.55279472e-01 2.74031579e-01
9.48446095e-01 -2.60767758e-01 -3.78840327e-01 3.29631150e-01
3.60982895e-01 -9.94844139e-01 5.86437285e-01 3.97994369e-01
8.32169712e-01 -5.21572888e-01 8.80550861e-01 5.07265449e-01
-8.11823845e-01 -1.29005641e-01 -1.08040318e-01 -1.53117791e-01
7.69963443e-01 6.38249874e-01 -1.04447985e+00 6.06749415e-01
4.20283675e-02 4.09727156e-01 -4.47365791e-01 9.60548162e-01
-4.48602319e-01 1.16001904e+00 5.25917625e-03 7.25102276e-02
3.39893222e-01 6.50143325e-02 4.79204744e-01 1.28453255e+00
-3.99940670e-01 2.15635449e-01 4.41422790e-01 4.53093022e-01
-2.62471974e-01 3.88720959e-01 -5.05380630e-01 2.94323992e-02
5.66304445e-01 1.06176162e+00 -9.89857912e-01 -4.54245597e-01
-2.04540655e-01 6.23648524e-01 4.90973055e-01 4.61870991e-02
-8.00085366e-01 -2.13223502e-01 -1.97774991e-01 1.81443736e-01
2.23116398e-01 3.42512801e-02 -7.26525307e-01 -6.68878973e-01
1.74560189e-01 -3.67683887e-01 7.09494948e-01 -4.03261125e-01
-1.19854558e+00 4.19917256e-01 2.85277009e-01 -1.02764416e+00
-3.82064551e-01 -6.02615215e-02 -9.27504718e-01 5.68228900e-01
-1.44705117e+00 -1.29323578e+00 4.67885047e-01 1.19959988e-01
9.77818131e-01 -5.31647727e-02 5.92729926e-01 4.26818095e-02
-4.07419026e-01 2.84659237e-01 -2.25839823e-01 3.98619622e-01
5.34233093e-01 -1.24389017e+00 -1.30074427e-01 8.17570746e-01
4.13977616e-02 4.89730626e-01 6.34633720e-01 -9.98743057e-01
-2.88681746e-01 -8.88585806e-01 1.56027699e+00 -4.97590512e-01
4.97535139e-01 -4.44304720e-02 -6.07790649e-01 4.48632747e-01
5.62073529e-01 -5.21758497e-01 9.89188910e-01 3.18476230e-01
-1.59110010e-01 2.65052497e-01 -1.13058257e+00 1.68377355e-01
1.04832172e+00 -6.65832162e-01 -1.15262425e+00 7.37204432e-01
7.51985490e-01 2.99162939e-02 -9.53790188e-01 2.47793913e-01
1.81710973e-01 -5.32423973e-01 6.34624660e-01 -9.89926398e-01
9.43708479e-01 4.47453186e-02 1.60583168e-01 -4.03953254e-01
1.78337784e-03 -6.31059706e-01 -1.40515924e-01 1.65804470e+00
6.94164991e-01 -9.02885124e-02 8.35270107e-01 6.06595755e-01
-1.12915725e-01 -8.80711138e-01 -1.13412285e+00 -6.76920891e-01
2.74680912e-01 -2.46184289e-01 2.46667281e-01 7.12430120e-01
2.93666959e-01 1.12291574e+00 -3.48389655e-01 -3.05349350e-01
4.47674692e-01 5.98263443e-01 3.51305991e-01 -1.44361830e+00
-3.18684839e-02 -2.43128344e-01 5.22724837e-02 -7.40891457e-01
5.69349825e-01 -8.41325343e-01 6.63149729e-02 -2.01304078e+00
6.41662836e-01 -3.48888367e-01 2.44189575e-01 5.22945046e-01
-3.13456088e-01 -1.93586841e-01 -2.20250785e-02 4.97301489e-01
-1.12275004e+00 2.86663085e-01 1.16020882e+00 -1.85156822e-01
-2.45769918e-01 1.83949634e-01 -8.48098934e-01 6.76500082e-01
9.36480582e-01 -7.29049087e-01 -4.03833807e-01 -2.45123312e-01
1.40310079e-02 6.14182591e-01 -1.49408191e-01 -5.01815796e-01
2.63606608e-01 -3.65850270e-01 5.69147393e-02 -8.62792790e-01
-1.31035298e-01 -1.91817269e-01 -3.53896290e-01 3.18566501e-01
-9.34903860e-01 -1.34821311e-01 -9.20488834e-02 6.56299770e-01
-2.53738016e-01 -9.77673292e-01 3.20363820e-01 -3.72313470e-01
-4.34249610e-01 -1.19043760e-01 -6.83918417e-01 5.00847816e-01
1.00473845e+00 -1.53375983e-01 -3.80690992e-01 -2.18586341e-01
-7.03337491e-01 2.83691764e-01 -2.69428547e-02 8.31110552e-02
1.48582563e-01 -7.30509102e-01 -1.07501066e+00 -7.81125307e-01
-4.85389456e-02 2.97324270e-01 2.44025692e-01 9.23816144e-01
-3.58840972e-01 7.49785125e-01 2.10507527e-01 -1.91587791e-01
-1.56509721e+00 3.49819064e-01 -3.80329609e-01 -1.33872259e+00
-4.00075585e-01 6.23188376e-01 -2.70947486e-01 -4.88508753e-02
6.32006899e-02 1.07807666e-01 -9.73720133e-01 4.22678560e-01
2.89277583e-01 4.14582849e-01 -5.83985187e-02 -8.14900041e-01
-2.00572237e-01 2.86612570e-01 -4.61857229e-01 -1.88934222e-01
1.41224074e+00 -1.10835381e-01 -5.19106209e-01 2.11445034e-01
1.11092377e+00 3.35456461e-01 -8.08560491e-01 -1.62729457e-01
7.96671689e-01 -1.90350451e-02 -2.56243050e-01 -8.18675458e-01
-2.75017798e-01 5.76161981e-01 -9.68308002e-02 3.32704216e-01
7.29447186e-01 4.58868086e-01 7.79573143e-01 -8.00035000e-02
2.34721415e-02 -1.44906056e+00 1.01892248e-01 5.36363840e-01
8.90760481e-01 -9.84236181e-01 2.77033895e-01 -7.67580390e-01
-8.84054542e-01 9.15616691e-01 5.01255512e-01 1.17827006e-01
2.16901526e-01 1.75075784e-01 -1.86408460e-01 -5.48717558e-01
-5.94851196e-01 -3.10632765e-01 2.15667859e-01 3.24624121e-01
5.18241823e-01 -1.33065686e-01 -1.27814448e+00 5.15316904e-01
-2.86540300e-01 -1.86469793e-01 4.25544173e-01 1.31464589e+00
-4.91850913e-01 -1.23609769e+00 -2.92837232e-01 8.32105160e-01
-1.13030803e+00 -2.91308105e-01 -3.70373666e-01 2.79816180e-01
2.58223772e-01 1.28488863e+00 -2.57002525e-02 2.72794336e-01
-3.42739448e-02 4.26087886e-01 3.98295492e-01 -9.26158488e-01
-6.46457553e-01 1.70093387e-01 8.60524058e-01 1.22462921e-01
-8.87740493e-01 -9.89798069e-01 -1.41233981e+00 2.09308639e-01
-4.52018619e-01 6.83010101e-01 3.07296842e-01 1.46953881e+00
1.39009818e-01 5.07105947e-01 3.30925494e-01 -1.23717681e-01
-6.86463058e-01 -1.50129509e+00 -2.90715754e-01 4.74096656e-01
7.46987090e-02 -4.22471762e-01 7.54581913e-02 3.70410591e-01] | [10.873517036437988, 9.439335823059082] |