File size: 85,977 Bytes
6fa4bc9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 |
{
"paper_id": "N06-1013",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T14:46:29.649132Z"
},
"title": "A Maximum Entropy Approach to Combining Word Alignments",
"authors": [
{
"first": "Necip",
"middle": [],
"last": "Fazil",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Maryland College Park",
"location": {
"postCode": "20742",
"region": "MD"
}
},
"email": ""
},
{
"first": "Bonnie",
"middle": [
"J"
],
"last": "Dorr",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Maryland College Park",
"location": {
"postCode": "20742",
"region": "MD"
}
},
"email": "bonnie@umiacs.umd.edu"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "This paper presents a new approach to combining outputs of existing word alignment systems. Each alignment link is represented with a set of feature functions extracted from linguistic features and input alignments. These features are used as the basis of alignment decisions made by a maximum entropy approach. The learning method has been evaluated on three language pairs, yielding significant improvements over input alignments and three heuristic combination methods. The impact of word alignment on MT quality is investigated, using a phrase-based MT system.",
"pdf_parse": {
"paper_id": "N06-1013",
"_pdf_hash": "",
"abstract": [
{
"text": "This paper presents a new approach to combining outputs of existing word alignment systems. Each alignment link is represented with a set of feature functions extracted from linguistic features and input alignments. These features are used as the basis of alignment decisions made by a maximum entropy approach. The learning method has been evaluated on three language pairs, yielding significant improvements over input alignments and three heuristic combination methods. The impact of word alignment on MT quality is investigated, using a phrase-based MT system.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Word alignment-detection of corresponding words between two sentences that are translations of each other-is usually an intermediate step of statistical machine translation (MT) (Brown et al., 1993; Och and Ney, 2003; Koehn et al., 2003) , but also has been shown useful for other applications such as construction of bilingual lexicons, word-sense disambiguation, projection of resources, and crosslanguage information retrieval.",
"cite_spans": [
{
"start": 178,
"end": 198,
"text": "(Brown et al., 1993;",
"ref_id": "BIBREF2"
},
{
"start": 199,
"end": 217,
"text": "Och and Ney, 2003;",
"ref_id": "BIBREF14"
},
{
"start": 218,
"end": 237,
"text": "Koehn et al., 2003)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Maximum entropy (ME) models have been used in bilingual sense disambiguation, word reordering, and sentence segmentation (Berger et al., 1996) , parsing, POS tagging and PP attachment (Ratnaparkhi, 1998) , machine translation , and FrameNet classification (Fleischman et al., 2003) . They have also been used to solve the word alignment problem (Garcia-Varea et al., 2002; Ittycheriah and Roukos, 2005; Liu et al., 2005 ), but a sentence-level approach to combining knowledge sources is used rather than a word-level approach. This paper describes an approach to combining evidence from alignments generated by existing systems to obtain an alignment that is closer to the true alignment than the individual alignments. The alignment-combination approach (called ACME) operates at the level of alignment links, rather than at the sentence level (as in previous ME approaches). ACME uses ME to decide whether to include/exclude a particular alignment link based on feature functions that are extracted from the input alignments and linguistic features of the words. Since alignment combination relies on evidence from existing alignments, we focus on alignment links that exist in at least one input alignment. An important challenge in this approach is the selection of appropriate links when two aligners make different alignment choices.",
"cite_spans": [
{
"start": 121,
"end": 142,
"text": "(Berger et al., 1996)",
"ref_id": "BIBREF1"
},
{
"start": 184,
"end": 203,
"text": "(Ratnaparkhi, 1998)",
"ref_id": "BIBREF19"
},
{
"start": 256,
"end": 281,
"text": "(Fleischman et al., 2003)",
"ref_id": "BIBREF4"
},
{
"start": 345,
"end": 372,
"text": "(Garcia-Varea et al., 2002;",
"ref_id": "BIBREF5"
},
{
"start": 373,
"end": 402,
"text": "Ittycheriah and Roukos, 2005;",
"ref_id": "BIBREF6"
},
{
"start": 403,
"end": 419,
"text": "Liu et al., 2005",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We show that ACME yields a significant relative error reduction over the input alignment systems and heuristic-based combinations on three different language pairs. Using a higher number of input alignments and partitioning the training data into disjoint subsets yield further error-rate reductions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The next section briefly overviews ME models. Section 3 presents a new ME approach to combining existing word alignment systems. Section 4 describes the evaluation data, input alignments, and evaluation metrics. Section 5 presents experiments on three language pairs, upper bounds for alignment error rate in alignment combination, and MT evaluation on English-Chinese and English-Arabic. Section 6 describes previous work on alignment combination and ME models on word alignment.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In a statistical classification problem, the goal is to estimate the probability of a class y in a given context x, i.e., p(y|x). In an ideal scenario, if the training data contain evidence for all pairs of (y, x), it is trivial to compute the probability distribution p. Unfortunately, due to training-data sparsity, p is generally modeled using only the available evidence.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Maximum Entropy (ME) Models",
"sec_num": "2"
},
{
"text": "Given a collection of facts, ME chooses a model consistent with all the facts, but otherwise as uniform as possible (Berger et al., 1996) . Formally, the evidence is represented as feature functions, i.e., binary valued functions that map a class y and a context x to either 0 or 1, i.e., h m : Y \u00d7 X \u2192 {0, 1}, where Y is the set of all classes and X is the set of all facts. The biggest advantage of maximum entropy models is that they are able to focus on the selection of feature functions rather than on how such functions are used. Any context can be used to define feature functions without concern for the independence of the feature functions from each other or the relevance of the feature functions to the final decision (Ratnaparkhi, 1998) .",
"cite_spans": [
{
"start": 116,
"end": 137,
"text": "(Berger et al., 1996)",
"ref_id": "BIBREF1"
},
{
"start": 731,
"end": 750,
"text": "(Ratnaparkhi, 1998)",
"ref_id": "BIBREF19"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Maximum Entropy (ME) Models",
"sec_num": "2"
},
{
"text": "Each feature function h m is associated with a model parameter \u03bb m . Given a set of M feature functions h 1 , . . . , h M , the probability of class y given a context x is equal to:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Maximum Entropy (ME) Models",
"sec_num": "2"
},
{
"text": "p(y|x) = 1 Z x exp M m=1 \u03bb m h m (y, x)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Maximum Entropy (ME) Models",
"sec_num": "2"
},
{
"text": "where Z x is a normalization constant. The contribution of each feature function to the final decision, i.e., \u03bb m , can be automatically computed using Generalized Iterative Scaling (GIS) algorithm (Darroch and Ratcliff, 1972) . The final classification for a given instance is the class y that maximizes p(y|x).",
"cite_spans": [
{
"start": 198,
"end": 226,
"text": "(Darroch and Ratcliff, 1972)",
"ref_id": "BIBREF3"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Maximum Entropy (ME) Models",
"sec_num": "2"
},
{
"text": "Let e = e 1 , . . . , e I and f = f 1 , . . . , f J be two sentences in two different languages. An alignment link (i, j) corresponds to a translational equivalence between words e i and f j . Let A k be an alignment between sentences e and f , where each element a \u2208 A k is an alignment link (i, j). Let A = {A 1 , . . . , A n } be a set of alignments between e and f . We refer to the true alignment as T , where each a \u2208 T is of the form (i, j). The goal of ACME is to combine the information in A such that the combined alignment A C is closer to T . A straightforward solution is to take the intersection or union of the individual alignments. In this paper, an additional model is learned to combine outputs of A 1 , . . . , A n .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "In our combination framework, first, n different word-alignment systems, A 1 , . . . , A n , generate word alignments between a given English sentence and a foreign-language (FL) sentence. Then a Feature Extractor takes the output of these alignment systems and the parallel corpus (which might be enriched with linguistic features) and extracts a set of feature functions based on linguistic properties of the words and the input alignments. Each feature function h m is associated with a model parameter \u03bb m . Next, an Alignment Combiner decides whether to include or exclude an alignment link based on the extracted feature functions and the model parameters associated with them.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "For each possible alignment link a set of features is extracted from the input alignments and linguistic properties of words. The features that are used for representing an alignment link (i, j) are as follows:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "1. Part-of-speech tags (posE, posF, prevposE, prevposF, nextpostE, nextposF) : POS tags for the previous, current, and the next English and FL words. 2. Outputs of input aligners (out): Whether (i, j) exists in a given input alignment A k . 3. Neighbors (neigh): A neighborhood of an alignment link (i, j)-denoted by N (i, j)consists of 8 possible alignment links in a 3\u00d73 window with (i, j) in the center of the window. Each element of N (i, j) is called a neighboring link of (i, j). Neighbor features include:",
"cite_spans": [
{
"start": 23,
"end": 76,
"text": "(posE, posF, prevposE, prevposF, nextpostE, nextposF)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "(1) Whether a particular neighbor of (i, j) exists in a given input alignment A k ; and (2) Total number of neighbors of (i, j) in a given input alignment A k . 4. Fertilities (fertE, fertF): The number of words that e i (or f j ) is aligned to in a given input alignment A k . 5. Monotonicity (mon): The absolute difference between i and j.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "Our combination approach employs feature functions derived from a subset of the features above. Assuming Y = {yes, no} represents the set of classes, where each class denotes the existence or absence of a link in the combined alignment, and X is the set of features above, we generate various feature functions h (y, x) , where y \u2208 Y and x are instantiations of one or more features in X . Table 1 lists the feature sets with an example feature func- tion for each. 1 For example, the feature function in the fifth row has a value of 1 if there are 2 neighboring links to (i, j) that exist in the input alignment A k and the alignment link",
"cite_spans": [
{
"start": 313,
"end": 319,
"text": "(y, x)",
"ref_id": null
}
],
"ref_spans": [
{
"start": 390,
"end": 397,
"text": "Table 1",
"ref_id": "TABREF0"
}
],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "Features Example Feature Function posE h( yes , i, j) = 1 if (i, j) \u2208 AC and pos(ei) = N oun posF h( no , i, j) = 1 if (i, j) / \u2208 AC and pos(fj) = V erb out h( yes , i, j, k) = 1 if (i, j) \u2208 AC and (i, j) \u2208 A k out, neigh h( yes , i, j, k) = 1 if (i, j) \u2208 AC and (i \u2212 1, j + 1) \u2208 A k h( yes , i, j, k) = 1 if (i, j) \u2208 AC and |N C| = 2 where N C = {n|n \u2208 N (i, j), n \u2208 A k } out, f ertE h( no , i, j, k) = 1 if (i, j) / \u2208 AC and |F T | = 0 where F T = {t|(i, t) \u2208 A k } out, f ertF h( no , i, j, k) = 1 if (i, j) / \u2208 AC and |F T | = 1 where F T = {t|(t, j) \u2208 A k } mon h( yes , i, j) = 1 if (i, j) \u2208 AC and |i \u2212 j| = 2",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "(i, j) exists in A C .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "In combining evidence from different alignments, it is assumed that, when an alignment link is left out by all aligners, that particular link should not be included in the final output. Since the majority of all possible word pairs are unaligned in real data, the inclusion of all possible word pairs in the training data leads to skewed results, where the learning algorithm is biased toward labeling the links as invalid. To offset this problem, our training data includes only alignment links that appear in at least one input alignment.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "Once the feature functions are extracted, we learn the model parameters using the YASMET ME package (Och, 2002) , which is an efficient implementation of the GIS algorithm.",
"cite_spans": [
{
"start": 100,
"end": 111,
"text": "(Och, 2002)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Alignment Combination: ACME",
"sec_num": "3"
},
{
"text": "The alignment combination techniques are evaluated in this paper using data from three language pairs, as shown in Table 2 . Input alignments are generated using two existing word alignment systems: GIZA++ (Och, 2000) 1 In Table 1 , NC corresponds to the set of (i, j)'s neighbors that exist in the alignment A k , and FT represents the set of words that ei (or fj) is aligned to.",
"cite_spans": [
{
"start": 206,
"end": 217,
"text": "(Och, 2000)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [
{
"start": 115,
"end": 122,
"text": "Table 2",
"ref_id": "TABREF2"
},
{
"start": 223,
"end": 230,
"text": "Table 1",
"ref_id": "TABREF0"
}
],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "2 From (Ayan et al., 2005 ).",
"cite_spans": [
{
"start": 2,
"end": 25,
"text": "From (Ayan et al., 2005",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "3 From (Ittycheriah and Roukos, 2005) . 4 From (Mihalcea and Pedersen, 2003) .",
"cite_spans": [
{
"start": 7,
"end": 37,
"text": "(Ittycheriah and Roukos, 2005)",
"ref_id": "BIBREF6"
},
{
"start": 40,
"end": 41,
"text": "4",
"ref_id": null
},
{
"start": 47,
"end": 76,
"text": "(Mihalcea and Pedersen, 2003)",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "and SAHMM (Lopez and Resnik, 2005) . Both systems are run in two different directions with default configurations. We indicate the two directions using the notation Aligner(en \u2192 f l) and Aligner(f l \u2192 en), where en is English, f l is either Chinese (ch), Arabic (ar), or Romanian (ro).",
"cite_spans": [
{
"start": 10,
"end": 34,
"text": "(Lopez and Resnik, 2005)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "To train both systems, additional data was used for the three language pairs: 107K English-Chinese sentence pairs (4.1M/3.3M English/Chinese words); 44K English-Arabic sentence pairs (1.4M/1M English/Arabic words); 48K English-Romanian sentence pairs (1M/1M English/Romanian words). 5 POS tags were generated using the MXPOST tagger (Ratnaparkhi, 1998) . POS tagger for English was trained on Sections 0-18 of the Penn Treebank Wall Street Journal corpus. On the FL side, we used POS tagger for only Chinese and it was trained on Sections 16-299 of Chinese Treebank.",
"cite_spans": [
{
"start": 283,
"end": 284,
"text": "5",
"ref_id": null
},
{
"start": 333,
"end": 352,
"text": "(Ratnaparkhi, 1998)",
"ref_id": "BIBREF19"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "For comparison purposes, three additional heuristically-induced alignments are generated for each system: (1) Intersection of both directions (Aligner(int)); (2) Union of both directions (Aligner(union)); and (3) The previously bestknown heuristic combination approach called growdiag-final (Koehn et al., 2003 ) (Aligner(gdf)).",
"cite_spans": [
{
"start": 291,
"end": 310,
"text": "(Koehn et al., 2003",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "In our evaluation, we take A to be the set of alignment links for a set of sentences, S to be the set of sure alignment links, and P be the set of probable alignment links (in the gold standard). Precision (P r), recall (Rc) and alignment error rate (AER) are defined as follows:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "6 P r = |A \u2229 P | |A| Rc = |A \u2229 S| |S| AER = 1 \u2212 |A \u2229 S| + |A \u2229 P | |A| + |S|",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "Our gold standard for each language pair is a manually aligned corpus. English-Chinese annotations distinguish between sure and probable alignment links (i.e., S \u2282 P ), but there is no such distinction for the other two language pairs (i.e., P = S).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "Because of the availability of limited manually annotated data, evaluations are performed using 5fold cross validation. Once the alignments are generated for each fold (using one as the test set and the other 4 folds as training set), the results are concatenated to compute precision, recall and error rate on the entire set of sentence pairs for each data set. 7",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment Data, Alignment Inputs, and Metrics",
"sec_num": "4"
},
{
"text": "This section presents several experiments and results comparing AER of ACME to those of standard alignment approaches on English-Chinese data. We also present experiments on additional languages, analyses based on precision and recall, an upperbound oracle analysis, and MT evaluations.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments and Results",
"sec_num": "5"
},
{
"text": "The experiments below test the effects of input alignments, feature set, data partitioning, number of inputs, and size of training data on the performance of ACME. Table 3 shows the AER for GIZA++ and SAHMM (in each direction), three heuristic-based combinations and ACME using 2 uni-directional alignments as input and all features described in Section 3. 8 (We use 'ACME[2]' in this section to refer to ACME applied to two input alignments and ACME[4] in later sections to refer to ACME applied to four input alignments.)",
"cite_spans": [],
"ref_spans": [
{
"start": 164,
"end": 171,
"text": "Table 3",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "English-Chinese Experiments",
"sec_num": "5.1"
},
{
"text": "Using 2 GIZA++ uni-directional alignments as input, ACME yields a 22.0% AER-a relative error reduction of 25.9% over GIZA++(gdf). Similarly, using 2 SAHMM uni-directional alignments as input, ACME produces a 20.6% AER-a relative error reduction of 28.0% and 25.4% over SAHMM(gdf) and SAHMM(int), respectively. 7 Because the NIST MTEval data include sentences that may be related (according to the document in which they appear), the training and test material could potentially be related; however, given the types of features used in our experiments, we do not believe this biases our results.",
"cite_spans": [
{
"start": 310,
"end": 311,
"text": "7",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Input alignments:",
"sec_num": "2"
},
{
"text": "8 For ease of readability, in the rest of this paper, we will report precision, recall, and AER in percentages. Table 4 presents the AER for ACME[2], using either two GIZA++ alignments or two SAHMM alignments, on English-Chinese data. Without any partitioning, ACME achieves an AER of 22.0 (GIZA++) and 20.6 (SAHMM). Using English POS tags for data partitioning results in a significant reduction in AER: 19.8% (GIZA++) and 18.0% (SAHMM). Interestingly, using foreign-language (FL) tags on their own or together with English POS tags does not provide any improvement. Overall when ACME[2] is applied to partitioned data (using posE for partitioning) a relative error reduction of 33-37% over GIZA++(gdf) and SAHMM(gdf) is achieved. Table 5 presents the English-Chinese AER for ACME[1] (using either GIZA++ or SAHMM in only one direction), ACME[2] (using either GIZA++ or SAHMM in two directions) and ACME[4] (using GIZA++ and SAHMM, each in two directions).",
"cite_spans": [],
"ref_spans": [
{
"start": 112,
"end": 119,
"text": "Table 4",
"ref_id": "TABREF5"
},
{
"start": 732,
"end": 739,
"text": "Table 5",
"ref_id": "TABREF7"
}
],
"eq_spans": [],
"section": "Input alignments:",
"sec_num": "2"
},
{
"text": "Regardless of the number of inputs, partitioning the data (using English POS tags) yields lower AER than no partitioning. Using one GIZA++ alignment as input, ACME Size of Training Data to Obtain Input Alignments: In general, statistical alignment systems improve as the size of the training data increases. We present the AER for GIZA++ and ACME[2] using GIZA++ alignments as input, where GIZA++ is trained on different sizes of data. We started with 20K sentence pairs of FBIS data and increased it to all available FBIS data (241K sentence pairs). Figure 1 compares the alignment performance of: (1) uni-directional GIZA++ (each direction);",
"cite_spans": [],
"ref_spans": [
{
"start": 551,
"end": 559,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Number of Input Alignments:",
"sec_num": null
},
{
"text": "(2) GIZA++(gdf); and (3) ACME[2] with all fea- We should emphasize that ACME[2] on only 20K sentence pairs yields a lower AER than those of all GIZA++ alignments obtained on 241K sentence pairs. Overall ACME[2] achieves a relative error reduction of 31-38% over the input alignments, and a relative error reduction of 31-34% over GIZA++(gdf) for different sizes of training data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Number of Input Alignments:",
"sec_num": null
},
{
"text": "We also investigated the applicability of ACME to additional language pairs. Table 6 presents the AER for GIZA++ and SAHMM (in each direction), three combination heuristics (gdf, int and union), and ACME[2] and ACME[4] on English-Arabic and English-Romanian data. We should emphasize that no POS tagger on the FL side was used for these experiments.",
"cite_spans": [],
"ref_spans": [
{
"start": 77,
"end": 84,
"text": "Table 6",
"ref_id": null
}
],
"eq_spans": [],
"section": "Expanding to Additional Languages",
"sec_num": "5.2"
},
{
"text": "On English-Arabic data, ACME[2] (with POS partitioning and including all features) yields 21.4% (20.7%) AER-a relative error reduction of 24.6% (13.0%) over the best combination heuristic with GIZA++ (SAHMM) alignments. ACME[4] reduces the AER to 18.1%-a relative error reduction of 36.3% and 23.9% over GIZA++(int) Table 6 : AER for Input Alignments, Heuristic-based Alignments, and ACME Using 2 and 4 Input Alignments (on English-Arabic and English-Romanian).",
"cite_spans": [
{
"start": 304,
"end": 315,
"text": "GIZA++(int)",
"ref_id": null
}
],
"ref_spans": [
{
"start": 316,
"end": 323,
"text": "Table 6",
"ref_id": null
}
],
"eq_spans": [],
"section": "Expanding to Additional Languages",
"sec_num": "5.2"
},
{
"text": "duces the AER to 22.3%-a relative error reduction of 22.6% and 23.9% over GIZA++(int) and SAHMM(int), respectively.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expanding to Additional Languages",
"sec_num": "5.2"
},
{
"text": "We now turn to a precision vs. recall analysis of different alignments to elucidate the nature of the differences between two alignments. Figure 2 presents precision and recall values for three combined alignments using GIZA++ (int, union, gdf) as well as results for ACME[2] and ACME[4] on three different language pairs. For all three pairs, the ranking of the combined alignments is the same with respect to precision and recall. GIZA++(int) yields the highest precision (nearly 95%) but the lowest recall (53-57%). Both union and gdf methods achieve low precision (56-68%) but high recall (75-83%), and gdf is better than union. By contrast, ACME[2] yields significantly higher precision (nearly 87%) but lower recall (67-75%) with respect to union and gdf. ACME[4] has higher precision and recall than ACME[2]-an absolute increase of 2-3% and 4%, respectively.",
"cite_spans": [],
"ref_spans": [
{
"start": 138,
"end": 146,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Precision, Recall and Upper-Bound Analysis",
"sec_num": "5.3"
},
{
"text": "Next we compute an oracle upper-bound in AER where mismatched input alignments are assumed to be resolved perfectly within the alignment combination framework (i.e., an oracle chooses the correct output in cases where the input aligners make different choices). 9 Table 7 presents the upper bounds using a generic alignment combiner (denoted Oracle) with 2 and 4 input alignments on three language pairs, assuming a perfect resolution of mismatched input alignments. For English-Chinese, the upper bound is 9.4% (us- [4] ). The English-Arabic data exhibits a slightly higher upper bound of 5.5% for Oracle [4] . The upper bounds for AER on English-Romanian data are even higher (up to 17.7%), which indicates that the input alignments are significantly worse than others. This may be one of the main contributing factors to the lower improvement of ACME on English-Romanian in comparison to the other two language pairs.",
"cite_spans": [
{
"start": 606,
"end": 609,
"text": "[4]",
"ref_id": null
}
],
"ref_spans": [
{
"start": 264,
"end": 271,
"text": "Table 7",
"ref_id": "TABREF11"
},
{
"start": 517,
"end": 520,
"text": "[4]",
"ref_id": null
}
],
"eq_spans": [],
"section": "Precision, Recall and Upper-Bound Analysis",
"sec_num": "5.3"
},
{
"text": "To determine the contribution of improved alignment in an external application, we examined the improvement in an off-the-shelf phrase-based MT system Pharaoh (Koehn, 2004) on both Chinese and Arabic data. In these experiments, all components of the MT system were kept the same except for the component that generates a phrase table from a given alignment. The input alignments were generated using GIZA++ and SAHMM on 107K (44K) sentence pairs for Chinese (Arabic). ACME (with English POS partitioning) combines alignments using model parameters learned from the corresponding manually aligned data. MT output is evaluated using the standard MT evaluation metric BLEU (Papineni et al., 2002) . 10 Table 8 presents the BLEU scores on Figure 2 : Precision and Recall Scores for GIZA++ and ACME Using 2 and 4 Input Alignments.",
"cite_spans": [
{
"start": 159,
"end": 172,
"text": "(Koehn, 2004)",
"ref_id": "BIBREF8"
},
{
"start": 670,
"end": 693,
"text": "(Papineni et al., 2002)",
"ref_id": "BIBREF18"
},
{
"start": 696,
"end": 698,
"text": "10",
"ref_id": null
}
],
"ref_spans": [
{
"start": 699,
"end": 706,
"text": "Table 8",
"ref_id": "TABREF13"
},
{
"start": 735,
"end": 743,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "MT Evaluation",
"sec_num": "5.4"
},
{
"text": "MTEval'03 data for 5 different Pharaoh runs, one for each alignment. The parameters of the MT system were optimized on MTEval'02 data using minimum error rate training (Och, 2003) .",
"cite_spans": [
{
"start": 168,
"end": 179,
"text": "(Och, 2003)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "MT Evaluation",
"sec_num": "5.4"
},
{
"text": "For the language model, the SRI Language Modeling Toolkit was used to train a trigram model with modified Kneser-Ney smoothing on 155M words of English newswire text, mostly from the Xinhua portion of the Gigaword corpus. During decoding, the number of English phrases per FL phrase was limited to 100 and the distortion of phrases was limited by 4. Based on the observations in (Koehn et al., 2003) , we also limited the phrase length to 3 for computational reasons. ACME[4] , for instance, yields the BLEU scores of 25.59% for Chinese and 45.54% for Arabic-an absolute 1.6-1.7% BLEU point increase over the best of the other three alignment combinations. The differences between the BLEU scores for ACME and the other three BLEU scores are statistically significant, using a significance test with bootstrap resampling (Zhang et al., 2004) .",
"cite_spans": [
{
"start": 379,
"end": 399,
"text": "(Koehn et al., 2003)",
"ref_id": "BIBREF7"
},
{
"start": 468,
"end": 475,
"text": "ACME[4]",
"ref_id": null
},
{
"start": 821,
"end": 841,
"text": "(Zhang et al., 2004)",
"ref_id": "BIBREF22"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "MT Evaluation",
"sec_num": "5.4"
},
{
"text": "ME models have been previously applied to several NLP problems, including word alignments. For intings: case-insensitive matching of n-grams up to n = 4, and the shortest reference sentence for the brevity penalty. stance, the IBM models (Brown et al., 1993) can be improved by adding more context dependencies into the translation model using a ME framework rather than using only p(f j |e i ) (Garcia-Varea et al., 2002) . In a later study, Och and Ney (2003) present a loglinear combination of the HMM and IBM Model 4 that produces better alignments than either of those. The major advantage of these two methods is that they do not require manually annotated data.",
"cite_spans": [
{
"start": 238,
"end": 258,
"text": "(Brown et al., 1993)",
"ref_id": "BIBREF2"
},
{
"start": 395,
"end": 422,
"text": "(Garcia-Varea et al., 2002)",
"ref_id": "BIBREF5"
},
{
"start": 443,
"end": 461,
"text": "Och and Ney (2003)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "6"
},
{
"text": "The alignment process can be modeled as a product of a transition model and an observation model, where ME models the observations (Ittycheriah and Roukos, 2005) . Significant improvements are reported using this approach but the need for large manually aligned data is a bottleneck. An alternative ME approach models alignment directly as a log-linear combination of feature functions (Liu et al., 2005) . Moore (2005) and Taskar et al. (2005) represent alignments with several feature functions that are then combined in a weighted sum to model word alignments. Once a confidence score is assigned to all links, a non-trivial search is invoked to find the best alignment using the scores associated with the links. The major difference between these approaches and that of ACME is that we use the ME model to predict the correct class for each alignment link independently using outputs of existing alignment systems, instead of generating them from scratch at the level of the whole sentence, thus eliminating the need for an exhaustive search over all possible alignments, i.e., previous approaches work globally while ACME is a localized model. A discussion of these two contrasting approaches can be found in (Tillmann and Zhang, 2005) .",
"cite_spans": [
{
"start": 131,
"end": 161,
"text": "(Ittycheriah and Roukos, 2005)",
"ref_id": "BIBREF6"
},
{
"start": 386,
"end": 404,
"text": "(Liu et al., 2005)",
"ref_id": "BIBREF9"
},
{
"start": 407,
"end": 419,
"text": "Moore (2005)",
"ref_id": "BIBREF12"
},
{
"start": 424,
"end": 444,
"text": "Taskar et al. (2005)",
"ref_id": "BIBREF20"
},
{
"start": 1215,
"end": 1241,
"text": "(Tillmann and Zhang, 2005)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "6"
},
{
"text": "A recent attempt to combine outputs of different alignments views the combination problem as a classifier ensemble in the neural network framework (Ayan et al., 2005) . However, this method is subject to the unpredictability of random network initialization, whereas ACME is guaranteed to find the model that maximizes the likelihood of training data.",
"cite_spans": [
{
"start": 147,
"end": 166,
"text": "(Ayan et al., 2005)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "6"
},
{
"text": "We presented a new approach, ACME, to combining the outputs of different word alignment systems by reducing the combination problem to the level of alignment links and using a maximum entropy model to learn whether a particular alignment link is included in the final alignment.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions",
"sec_num": "7"
},
{
"text": "Our results indicate that ACME yields significant relative error reduction over the input alignments and their heuristic-based combinations on three different language pairs. Moreover, ACME provides similar relative improvements for different sizes of training data for the input alignment systems. We have also shown that using a higher number of input alignments, and partitioning the training data into disjoint subsets and learning a different model for each partition yield further improvements.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions",
"sec_num": "7"
},
{
"text": "We have tested impact of the reduced AER on MT and have shown that alignments generated by ACME yield statistically significant improvements in BLEU scores in two different languages, even if we don't employ a POS tagger on the FL side. However, additional studies are needed to investigate why huge improvements in AER result in relatively smaller improvements in BLEU scores.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions",
"sec_num": "7"
},
{
"text": "Because ACME is a supervised learning approach, it requires annotated data; however, our experiments have shown that significant improvements can be obtained using a small set of annotated data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions",
"sec_num": "7"
},
{
"text": "Note that both GIZA++ and SAHMM are unsupervised learning systems. Sentence-aligned parallel texts are the only required input.6 Note that AER= 1 -F-score when there is no distinction between probable and sure alignment links.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "If the input aligners agree on a particular link, that decision is taken as the final output in computing the upper bound.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We used the NIST script (version 11a) with its default set-",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "Acknowledgments This work has been supported, in part, under ONR MURI Contract FCPO.810548265 and the GALE program of the Defense Advanced Research Projects Agency, Contracts No. HR0011-06-2-0001. We also thank anonymous reviewers for their helpful comments.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "acknowledgement",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Neuralign: Combining word alignments using neural networks",
"authors": [
{
"first": "F",
"middle": [],
"last": "Necip",
"suffix": ""
},
{
"first": "Bonnie",
"middle": [
"J"
],
"last": "Ayan",
"suffix": ""
},
{
"first": "Christof",
"middle": [],
"last": "Dorr",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Monz",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of EMNLP'2005",
"volume": "",
"issue": "",
"pages": "65--72",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Necip F. Ayan, Bonnie J. Dorr, and Christof Monz. 2005. Neu- ralign: Combining word alignments using neural networks. In Proceedings of EMNLP'2005, pages 65-72.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "A maximum entropy approach to natural language processing",
"authors": [
{
"first": "Adam",
"middle": [
"L"
],
"last": "Berger",
"suffix": ""
},
{
"first": "Stephan",
"middle": [
"A"
],
"last": "Della-Pietra",
"suffix": ""
},
{
"first": "Vincent",
"middle": [
"J"
],
"last": "Della-Pietra",
"suffix": ""
}
],
"year": 1996,
"venue": "Computational Linguistics",
"volume": "",
"issue": "1",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Adam L. Berger, Stephan A. Della-Pietra, and Vincent J. Della- Pietra. 1996. A maximum entropy approach to natural lan- guage processing. Computational Linguistics, 22(1).",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "The mathematics of statistical machine translation: Parameter estimation",
"authors": [
{
"first": "F",
"middle": [],
"last": "Peter",
"suffix": ""
},
{
"first": "Stephan",
"middle": [
"A"
],
"last": "Brown",
"suffix": ""
},
{
"first": "Robert",
"middle": [
"L"
],
"last": "Della Pietra",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Mercer",
"suffix": ""
}
],
"year": 1993,
"venue": "Computational Linguistics",
"volume": "19",
"issue": "2",
"pages": "263--311",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Peter F. Brown, Stephan A. Della Pietra, and Robert L. Mer- cer. 1993. The mathematics of statistical machine trans- lation: Parameter estimation. Computational Linguistics, 19(2):263-311.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Generalized iterative scaling for log-linear models",
"authors": [
{
"first": "J",
"middle": [
"N"
],
"last": "Darroch",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Ratcliff",
"suffix": ""
}
],
"year": 1972,
"venue": "Annals of Mathematical Statistics",
"volume": "43",
"issue": "",
"pages": "1470--1480",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. N. Darroch and D. Ratcliff. 1972. Generalized iterative scal- ing for log-linear models. Annals of Mathematical Statistics, 43:1470-1480.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Maximum entropy models for framenet classification",
"authors": [
{
"first": "Michael",
"middle": [],
"last": "Fleischman",
"suffix": ""
},
{
"first": "Namhee",
"middle": [],
"last": "Kwon",
"suffix": ""
},
{
"first": "Eduard",
"middle": [],
"last": "Hovy",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Michael Fleischman, Namhee Kwon, and Eduard Hovy. 2003. Maximum entropy models for framenet classification. In Proceedings of EMNLP'2003.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Improving alignment quality in statistical machine translation using context-dependent maximum entropy models",
"authors": [
{
"first": "Ismael",
"middle": [],
"last": "Garcia-Varea",
"suffix": ""
},
{
"first": "Franz",
"middle": [
"Josef"
],
"last": "Och",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Ney",
"suffix": ""
},
{
"first": "Francisco",
"middle": [],
"last": "Casacuberta",
"suffix": ""
}
],
"year": 2002,
"venue": "Proceedings of COLING",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ismael Garcia-Varea, Franz Josef Och, Hermann Ney, and Fran- cisco Casacuberta. 2002. Improving alignment quality in statistical machine translation using context-dependent max- imum entropy models. In Proceedings of COLING'2002.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "A maximum entropy word aligner for arabic-english machine translation",
"authors": [
{
"first": "Abraham",
"middle": [],
"last": "Ittycheriah",
"suffix": ""
},
{
"first": "Salim",
"middle": [],
"last": "Roukos",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of EMNLP'2005",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Abraham Ittycheriah and Salim Roukos. 2005. A maximum entropy word aligner for arabic-english machine translation. In Proceedings of EMNLP'2005.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Statistical phrase-based translation",
"authors": [
{
"first": "Philipp",
"middle": [],
"last": "Koehn",
"suffix": ""
},
{
"first": "Franz",
"middle": [
"J"
],
"last": "Och",
"suffix": ""
},
{
"first": "Daniel",
"middle": [],
"last": "Marcu",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of HLT-NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Philipp Koehn, Franz J. Och, and Daniel Marcu. 2003. Sta- tistical phrase-based translation. In Proceedings of HLT- NAACL'2003.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Pharaoh: A beam search decoder for phrase-based statistical machine translation",
"authors": [
{
"first": "Philipp",
"middle": [],
"last": "Koehn",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of AMTA",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Philipp Koehn. 2004. Pharaoh: A beam search decoder for phrase-based statistical machine translation. In Proceedings of AMTA'2004.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Log-linear models for word alignment",
"authors": [
{
"first": "Yang",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Qun",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Shouxun",
"middle": [],
"last": "Lin",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of ACL'2005",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yang Liu, Qun Liu, and Shouxun Lin. 2005. Log-linear models for word alignment. In Proceedings of ACL'2005.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Improved HMM alignment models for languages with scarce resources",
"authors": [
{
"first": "Adam",
"middle": [],
"last": "Lopez",
"suffix": ""
},
{
"first": "Philip",
"middle": [],
"last": "Resnik",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of the ACL'2005 Workshop on Building and Using Parallel Texts: Data Driven Machine Translation and Beyond",
"volume": "",
"issue": "",
"pages": "83--86",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Adam Lopez and Philip Resnik. 2005. Improved HMM align- ment models for languages with scarce resources. In Pro- ceedings of the ACL'2005 Workshop on Building and Using Parallel Texts: Data Driven Machine Translation and Be- yond, pages 83-86.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "An evaluation exercise for word alignment",
"authors": [
{
"first": "Rada",
"middle": [],
"last": "Mihalcea",
"suffix": ""
},
{
"first": "Ted",
"middle": [],
"last": "Pedersen",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of the HLT-NAACL'2003 Workshop: Building and Using Parallel Texts: Data Driven Machine Translation and Beyond",
"volume": "",
"issue": "",
"pages": "1--10",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rada Mihalcea and Ted Pedersen. 2003. An evaluation ex- ercise for word alignment. In Proceedings of the HLT- NAACL'2003 Workshop: Building and Using Parallel Texts: Data Driven Machine Translation and Beyond, pages 1-10.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "A discriminative framework for bilingual word alignment",
"authors": [
{
"first": "Robert",
"middle": [
"C"
],
"last": "Moore",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of EMNLP'2005",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Robert C. Moore. 2005. A discriminative framework for bilin- gual word alignment. In Proceedings of EMNLP'2005.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Discriminative training and maximum entropy models for statistical machine translation",
"authors": [
{
"first": "J",
"middle": [],
"last": "Franz",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Och",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2002,
"venue": "Proceedings of ACL'2002",
"volume": "",
"issue": "",
"pages": "295--302",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Franz J. Och and Hermann Ney. 2002. Discriminative training and maximum entropy models for statistical machine trans- lation. In Proceedings of ACL'2002, pages 295-302.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "A systematic comparison of various statistical alignment models",
"authors": [
{
"first": "J",
"middle": [],
"last": "Franz",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Och",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2003,
"venue": "Computational Linguistics",
"volume": "29",
"issue": "1",
"pages": "9--51",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Franz J. Och and Hermann Ney. 2003. A systematic compari- son of various statistical alignment models. Computational Linguistics, 29(1):9-51, March.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "GIZA++: Training of statistical translation models",
"authors": [
{
"first": "J",
"middle": [],
"last": "Franz",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Och",
"suffix": ""
}
],
"year": 2000,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Franz J. Och. 2000. GIZA++: Training of statistical transla- tion models. Technical report, RWTH Aachen, University of Technology.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Yet another maxent toolkit: YASMET",
"authors": [
{
"first": "Franz",
"middle": [
"J"
],
"last": "Och",
"suffix": ""
}
],
"year": 2002,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Franz J. Och. 2002. Yet another maxent toolkit: YASMET. Available at http://www.fjoch.com/YASMET.html.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Minimum error rate training in statistical machine translation",
"authors": [
{
"first": "Franz",
"middle": [
"J"
],
"last": "Och",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of ACL'2003",
"volume": "",
"issue": "",
"pages": "160--167",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Franz J. Och. 2003. Minimum error rate training in statisti- cal machine translation. In Proceedings of ACL'2003, pages 160-167.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "BLEU: A method for automatic evaluation of machine translation",
"authors": [
{
"first": "Kishore",
"middle": [],
"last": "Papineni",
"suffix": ""
},
{
"first": "Salim",
"middle": [],
"last": "Roukos",
"suffix": ""
},
{
"first": "Todd",
"middle": [],
"last": "Ward",
"suffix": ""
},
{
"first": "Wei-Jing",
"middle": [],
"last": "Zhu",
"suffix": ""
}
],
"year": 2002,
"venue": "Proceedings of ACL'2002",
"volume": "",
"issue": "",
"pages": "311--318",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: A method for automatic evaluation of machine translation. In Proceedings of ACL'2002, pages 311-318.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Maximum Entropy Models for Natural Language Ambiguity Resolution",
"authors": [
{
"first": "Adwait",
"middle": [],
"last": "Ratnaparkhi",
"suffix": ""
}
],
"year": 1998,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Adwait Ratnaparkhi. 1998. Maximum Entropy Models for Nat- ural Language Ambiguity Resolution. Ph.D. thesis, Univer- sity of Pennsylvania, Philadelphia, PA.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "A discriminative matching approach to word alignment",
"authors": [
{
"first": "Ben",
"middle": [],
"last": "Taskar",
"suffix": ""
},
{
"first": "Simon",
"middle": [],
"last": "Lacoste-Julien",
"suffix": ""
},
{
"first": "Dan",
"middle": [],
"last": "Klein",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ben Taskar, Simon Lacoste-Julien, and Dan Klein. 2005. A discriminative matching approach to word alignment. In Proceedings of EMNLP'2005.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "A localized prediction model for statistical machine translation",
"authors": [
{
"first": "Christoph",
"middle": [],
"last": "Tillmann",
"suffix": ""
},
{
"first": "Tong",
"middle": [],
"last": "Zhang",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christoph Tillmann and Tong Zhang. 2005. A localized predic- tion model for statistical machine translation. In Proceed- ings of ACL'2005.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Interpreting BLEU/NIST scores: How much improvement do we need to have a better system?",
"authors": [
{
"first": "Ying",
"middle": [],
"last": "Zhang",
"suffix": ""
},
{
"first": "Stephan",
"middle": [],
"last": "Vogel",
"suffix": ""
},
{
"first": "Alex",
"middle": [],
"last": "Waibel",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of LREC'2004",
"volume": "",
"issue": "",
"pages": "2051--2054",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ying Zhang, Stephan Vogel, and Alex Waibel. 2004. Inter- preting BLEU/NIST scores: How much improvement do we need to have a better system? In Proceedings of LREC'2004, pages 2051-2054.",
"links": null
}
},
"ref_entries": {
"TABREF0": {
"type_str": "table",
"text": "Feature Functions.",
"html": null,
"content": "<table/>",
"num": null
},
"TABREF2": {
"type_str": "table",
"text": "",
"html": null,
"content": "<table/>",
"num": null
},
"TABREF4": {
"type_str": "table",
"text": "Comparison of GIZA++ and SAHMM to ACME[2] (on English-Chinese).",
"html": null,
"content": "<table><tr><td colspan=\"2\">Feature Set: To examine the effects of each fea-ture on the performance of ACME, we compute the</td></tr><tr><td colspan=\"2\">AER under a variety of conditions, removing each</td></tr><tr><td colspan=\"2\">feature one at a time. ACME is evaluated using</td></tr><tr><td colspan=\"2\">2 uni-directional GIZA++ alignments as input on</td></tr><tr><td colspan=\"2\">English-Chinese data. Using all features, the AER</td></tr><tr><td colspan=\"2\">is 22.0%. Our experiments show that there is no sig-</td></tr><tr><td colspan=\"2\">nificant increase in AER for the removal of features</td></tr><tr><td colspan=\"2\">corresponding to monotonicity (22.1%), neighbors</td></tr><tr><td colspan=\"2\">(22.8%), POS on English side (22.9%), POS on</td></tr><tr><td colspan=\"2\">foreign-language side (22.9%). On the other hand,</td></tr><tr><td colspan=\"2\">deleting POS tags on both sides yields an AER of</td></tr><tr><td colspan=\"2\">25.2% and deleting the fertility features increases</td></tr><tr><td colspan=\"2\">the AER to 25.9%. This indicates that both POS</td></tr><tr><td colspan=\"2\">tags (or fertilities) contribute heavily toward the de-</td></tr><tr><td colspan=\"2\">cision as to whether a particular alignment should be</td></tr><tr><td>included/excluded.</td><td/></tr><tr><td colspan=\"2\">Partitioning Data: Previous work showed that partitioning the data into disjoint subsets and learn-</td></tr><tr><td colspan=\"2\">ing a different model for each partition improves</td></tr><tr><td colspan=\"2\">the performance of the alignment systems (Ayan et</td></tr><tr><td colspan=\"2\">al., 2005). To test whether this same principle ap-</td></tr><tr><td colspan=\"2\">plies to alignment combination with maximum en-</td></tr><tr><td colspan=\"2\">tropy modeling, the training data was partitioned us-</td></tr><tr><td colspan=\"2\">ing POS tags for English and the FL, and different</td></tr><tr><td colspan=\"2\">weights were learned for each partition.</td></tr><tr><td>Alignments ACME[2] ACME[2]-Part[posE] ACME[2]-Part[posF ] ACME[2]-Part[posE, posF ]</td><td>GIZA++ SAHMM 22.0 20.6 19.8 18.0 20.0 18.1 20.0 18.4</td></tr></table>",
"num": null
},
"TABREF5": {
"type_str": "table",
"text": "",
"html": null,
"content": "<table><tr><td>: Application of ACME[2] on Partitioned</td></tr><tr><td>Data (on English-Chinese).</td></tr></table>",
"num": null
},
"TABREF6": {
"type_str": "table",
"text": "[1] with partitioning improves the AER to 26.9% and 25.5% for each direction, respectively. Similarly, using one SAHMM alignment as input, ACME[1] with partitioning reduces the AER to 22.9% and 24.7%. ACME[2] with partitioning reduces the AER to 19.8% and 18.0% for GIZA++ and SAHMM, respectively. Finally, using all four input alignments, ACME[4] with partitioning yields a 15.6% AER-a relative error reduction of 21.2% and 13.3% over each ACME[2] case.",
"html": null,
"content": "<table><tr><td>Alignments ACME[1](en \u2192 f l) ACME[1]-Part[posE](en \u2192 f l) ACME[1](f l \u2192 en) ACME[1]-Part[posE](f l \u2192 en) ACME[2] ACME[2]-Part[posE]</td><td>GIZA++ SAHMM 28.1 24.4 26.9 22.9 26.6 26.9 25.5 24.7 22.0 20.6 19.8 18.0</td></tr><tr><td>ACME[4] ACME[4]-Part[posE]</td><td>17.8 15.6</td></tr></table>",
"num": null
},
"TABREF7": {
"type_str": "table",
"text": "",
"html": null,
"content": "<table><tr><td>: Application of ACME to 1, 2 and 4 Input</td></tr><tr><td>Alignments (on English-Chinese).</td></tr></table>",
"num": null
},
"TABREF11": {
"type_str": "table",
"text": "",
"html": null,
"content": "<table><tr><td>: Oracle Upper Bounds on AER for Align-</td></tr><tr><td>ment Combination</td></tr><tr><td>ing Oracle[2]) and 4.7% (using Oracle</td></tr></table>",
"num": null
},
"TABREF13": {
"type_str": "table",
"text": "",
"html": null,
"content": "<table><tr><td>: Evaluation of Pharaoh with Different Initial</td></tr><tr><td>Alignments using BLEU (in percentages)</td></tr><tr><td>For both languages, ACME[2] and ACME[4]</td></tr><tr><td>outperform the other three alignment combination</td></tr><tr><td>techniques.</td></tr></table>",
"num": null
}
}
}
} |