Datasets:

debajyotidasgupta

/

dblp_2022_2023

like 0

Title: Personalizing Task-oriented Dialog Systems via Zero-shot Generalizable Reward Function Abstract: ABSTRACTTask-oriented dialog systems enable users to accomplish tasks using natural language. State-of-the-art systems respond to users in the same way regardless of their personalities, although personalizing dialogues can lead to higher levels of adoption and better user experiences. Building personalized dialog systems is an important, yet challenging endeavor, and only a handful of works took on the challenge. Most existing works rely on supervised learning approaches and require laborious and expensive labeled training data for each user profile. Additionally, collecting and labeling data for each user profile is virtually impossible. In this work, we propose a novel framework, P-ToD, to personalize task-oriented dialog systems capable of adapting to a wide range of user profiles in an unsupervised fashion using a zero-shot generalizable reward function. P-ToD uses a pre-trained GPT-2 as a backbone model and works in three phases. Phase one performs task-specific training. Phase two kicks off unsupervised personalization by leveraging the proximal policy optimization algorithm that performs policy gradients guided by the zero-shot generalizable reward function. Our novel reward function can quantify the quality of the generated responses even for unseen profiles. The optional final phase fine-tunes the personalized model using a few labeled training examples. We conduct extensive experimental analysis using the personalized bAbI dialogue benchmark for five tasks and up to 180 diverse user profiles. The experimental results demonstrate that P-ToD, even when it had access to zero labeled examples, outperforms state-of-the-art supervised personalization models and achieves competitive performance on BLEU and ROUGE metrics when compared to a strong fully-supervised GPT-2 baseline.

Title: Hierarchically Fusing Long and Short-Term User Interests for Click-Through Rate Prediction in Product Search Abstract: ABSTRACTEstimating Click-Through Rate (CTR) is a vital yet challenging task in personalized product search. However, existing CTR methods still struggle in the product search settings due to the following three challenges including how to more effectively extract users' short-term interest with respect to multiple aspects, how to extract and fuse users' long-term interest with short-term interest, how to address the entangling characteristic of long and short-term interests. To resolve these challenges, in this paper, we propose a new approach named Hierarchical Interests Fusing Network (HIFN), which consists of four basic modules namely Short-term Interest Extractor (SIE), Long-term Interest Extractor (LIE), Interest Fusion Module (IFM) and Interest Disentanglement Module (IDM). Specifically, SIE is proposed to extract user's short-term interest by integrating three fundamental interest encoders within it namely query-dependent, target-dependent and causal-dependent interest encoder, respectively, followed by delivering the resultant representation to the module LIE, where it can effectively capture user long-term interest by devising an attention mechanism with respect to the short-term interest from SIE module. In IFM, the achieved long and short-term interests are further fused in an adaptive manner, followed by concatenating it with original raw context features for the final prediction result. Last but not least, considering the entangling characteristic of long and short-term interests, IDM further devises a self-supervised framework to disentangle long- and short-term interests. Extensive offline and online evaluations on a real-world e-commerce platform demonstrate the superiority of HIFN over state-of-the-art methods.

Title: Deep Extreme Mixture Model for Time Series Forecasting Abstract: ABSTRACTTime Series Forecasting (TSF) has been a topic of extensive research, which has many real world applications such as weather prediction, stock market value prediction, traffic control etc. Many machine learning models have been developed to address TSF, yet, predicting extreme values remains a challenge to be effectively addressed. Extreme events occur rarely, but tend to cause a huge impact, which makes extreme event prediction important. Assuming light tailed distributions, such as Gaussian distribution, on time series data does not do justice to the modeling of extreme points. To tackle this issue, we develop a novel approach towards improving attention to extreme event prediction. Within our work, we model time series data distribution, as a mixture of Gaussian distribution and Generalized Pareto distribution (GPD). In particular, we develop a novel Deep eXtreme Mixture Model (DXtreMM) for univariate time series forecasting, which addresses extreme events in time series. The model consists of two modules: 1) Variational Disentangled Auto-encoder (VD-AE) based classifier and 2) Multi Layer Perceptron (MLP) based forecaster units combined with Generalized Pareto Distribution (GPD) estimators for lower and upper extreme values separately. VD-AE Classifier model predicts the possibility of occurrence of an extreme event given a time segment, and forecaster module predicts the exact value. Through extensive set of experiments on real-world datasets we have shown that our model performs well for extreme events and is comparable with the existing baseline methods for normal time step forecasting.

Title: Unbiased Learning to Rank with Biased Continuous Feedback Abstract: ABSTRACTIt is a well-known challenge to learn an unbiased ranker with biased feedback. Unbiased learning-to-rank(LTR) algorithms, which are verified to model the relative relevance accurately based on noisy feedback, are appealing candidates and have already been applied in many applications with single categorical labels, such as user click signals. Nevertheless, the existing unbiased LTR methods cannot properly handle continuous feedback, which are essential for many industrial applications, such as content recommender systems. To provide personalized high-quality recommendation results, recommender systems need model both categorical and continuous biased feedback, such as click and dwell time. As unbiased LTR methods could not handle these continuous feedback and pair-wise learning without debiasing often performs worse than point-wise on biased feedback, which is also verified in our experiments, training multiple point-wise rankers to predict the absolute value of multiple objectives and leveraging a distinct shallow tower to estimate and alleviate the impact of position bias has been the mainstream approach in major industrial recommendation applications. However, with such a training paradigm, the optimization target differs a lot from the ranking metrics valuing the relative order of top-ranked items rather than the prediction precision of each item. Moreover, as the existing system tends to recommend more relevant items at higher positions, it is difficult for the shallow tower based methods to precisely attribute the user feedback to the impact of position or relevance. Therefore, there exists an exciting opportunity for us to get enhanced performance if we manage to solve the aforementioned issues. Accordingly, we design a novel unbiased LTR algorithm to tackle the challenges, which innovatively models position bias in the pairwise fashion and introduces the pairwise trust bias to separate the position bias, trust bias, and user relevance explicitly and can work for both continuous and categorical feedback. Experiment results on public benchmark datasets and internal live traffic of a large-scale recommender system at Tencent News show superior results for continuous labels and also competitive performance for categorical labels of the proposed method.

Title: Flow-based Perturbation for Cause-effect Inference Abstract: ABSTRACTA new causal discovery method is introduced to solve the bivariate causal discovery problem. The proposed algorithm leverages the expressive power of flow-based models and tries to learn the complex relationship between two variables. Algorithms have been developed to infer the causal direction according to empirical perturbation errors obtained from an invertible flow-based function. Theoretical results as well as experimental studies are presented to verify the proposed approach. Empirical evaluations demonstrate that our proposed method could outperform baseline methods on both synthetic and real-world datasets.

Title: RSD: A Reinforced Siamese Network with Domain Knowledge for Early Diagnosis Abstract: ABSTRACTThe availability of electronic health record data makes it possible to develop automatic disease diagnosis approaches. In this paper, we study the early diagnosis of diseases. As being a difficult task (even for experienced doctors), early diagnosis of diseases poses several challenges that are not well solved by prior studies, including insufficient training data, dynamic and complex signs of complications and trade-off between earliness and accuracy. To address these challenges, we propose a Reinforced Siamese network with Domain knowledge regularization approach, namely RSD, to achieve high performance for early diagnosis. The RSD approach consists of a diagnosis module and a control module. The diagnosis module adopts any EHR Encoder as a basic framework to extract representations, and introduces two improved training strategies. To overcome the insufficient sample problem, we design a Siamese network architecture to enhance the model learning. Furthermore, we propose a domain knowledge regularization strategy to guide the model learning with domain knowledge. Based on the diagnosis module, our control module learns to automatically determine whether making a disease alert to the patients based on the diagnosis results. Through carefully designed architecture, rewards and policies, it is able to effectively balance earliness and accuracy for diagnosis. Experimental results have demonstrated the effectiveness of our approach on both diagnosis prediction and early diagnosis. We also perform extensive analysis experiments to verify the robustness of the proposed approach.

Title: A Multi-Interest Evolution Story: Applying Psychology in Query-based Recommendation for Inferring Customer Intention Abstract: ABSTRACTThe query-based recommendation now is becoming a basic research topic in the e-commerce scenario. Generally, given a query that a user typed, it aims to provide a set of items that the user may be interested in. In this task, the customer intention (i.e., browsing or purchase) is an important factor to configure the corresponding recommendation strategy for better shopping experiences (i.e., providing diverse items when the user prefers to browse or recommending specific items when detecting the user is willing to purchase). Though necessary, this is usually overlooked in previous works. In addition, the diversity and evolution of user interests also bring challenges to inferring user intentions correctly. In this paper, we propose a predecessor task to infer two important customer intentions, which are purchasing and browsing respectively, and we introduce a novel Psychological Intention Prediction Model (PIPM for short) to address this issue. Inspired by cognitive psychology, we first devise a multi-interest extraction module to adaptively extract interests from the user-item interaction sequence. After this, we design an interest evolution layer to model the evolution of the mined multiple interests. Finally, we aggregate all evolved multiple interests to infer users' intentions in his/her next visit. Extensive experiments are conducted on a large-scale Taobao industrial dataset. The results demonstrate that PIPM gains a significant improvement on AUC and GAUC than state-of-the-art baselines. Notably, PIPM has been deployed on the Taobao e-commerce platform and obtained over 10% improvement on PCTR.

Title: Malicious Repositories Detection with Adversarial Heterogeneous Graph Contrastive Learning Abstract: ABSTRACTGitHub, as the largest social coding platform, has attracted an increasing number of cybercriminals to disseminate malware by posting malicious code repositories. To address the imminent problem, some tools were developed to detect malicious repositories based on the code content. However, most of them ignore the rich relational information among repositories and usually require abundant labeled data to train the model. To this end, one effective way is to exploit unlabeled data to pre-train a model which considers both structural relation and code content of repositories, and further transfer the pre-trained model to the downstream tasks with labeled repository data. In this paper, we propose a novel model adversarial contrastive learning on heterogeneous graph (CLA-HG) to detect malicious repository in GitHub. First of all, CLA-HG builds a heterogeneous graph (HG) to comprehensively model repository data. Afterwards, to exploit unlabeled information in HG, CLA-HG introduces a dual-stream graph contrastive learning mechanism that distinguishes both adversarial subgraph pairs and standard subgraph pairs to pre-train graph neural networks using unlabeled data. Finally, the pre-trained model is fine-tuned to the downstream malicious repository detection task enhanced by a knowledge distillation (KD) module. Extensive experiments on two collected datasets from GitHub demonstrate the effectiveness of CLA-HG in comparison with state-of-the-art methods and popular commercial anti-malware products.

Title: Asymmetrical Context-aware Modulation for Collaborative Filtering Recommendation Abstract: ABSTRACTModern learnable collaborative filtering recommendation models generate user and item representations by deep learning methods (e.g. graph neural networks) for modeling user-item interactions. However, most of them may still have unsatisfied performances due to two issues. Firstly, some models assume that the representations of users or items are fixed when modeling interactions with different objects. However, a user may have different interests in different items, and an item may also have different attractions to different users. Thus the representations of users and items should depend on their contexts to some extent. Secondly, existing models learn representations for user and item by symmetrical dual methods which have identical or similar operations. Symmetrical methods may fail to sufficiently and reasonably extract the features of user and item as their interaction data have diverse semantic properties. To address the above issues, a novel model called Asymmetrical context-awaRe modulation for collaBorative filtering REcommendation (ARBRE) is proposed. It adopts simplified GNNs on collaborative graphs to capture homogeneous user preferences and item attributes, then designs two asymmetrical context-aware modulation models to learn dynamic user interests and item attractions, respectively. The learned representations from user domain and item domain are input pair-wisely into 4 Multi-Layer Perceptrons in different combinations to model user-item interactions. Experimental results on three real-world datasets demonstrate the superiority of ARBRE over various state-of-the-arts.

Title: Rationale Aware Contrastive Learning Based Approach to Classify and Summarize Crisis-Related Microblogs Abstract: ABSTRACTRecent fashion of information propagation on Twitter makes the platform a crucial conduit for tactical data and emergency responses during disasters. However, the real-time information about crises is immersed in a large volume of emotional and irrelevant posts. It brings the necessity to develop an automatic tool to identify disaster-related messages and summarize the information for data consumption and situation planning. Besides, explainability of the methods is crucial in determining their applicability in real-life scenarios. Recent studies also highlight the importance of learning a good latent representation of tweets for several downstream tasks. In this paper, we take advantage of state-of-the-art methods, such as transformers and contrastive learning to build an interpretable classifier. Our proposed model classifies Twitter messages into different humanitarian categories and also extracts rationale snippets as supporting evidence for output decisions. The contrastive learning framework helps to learn better representations of tweets by bringing the related tweets closer in the embedding space. Furthermore, we employ classification labels and rationales to efficiently generate summaries of crisis events. Extensive experiments over different crisis datasets show that (i). our classifier obtains the best performance-interpretability trade-off, (ii). the proposed summarizer shows superior performance (1.4%-22% improvement) with significantly less computation cost than baseline models.

Title: Domain-Agnostic Contrastive Representations for Learning from Label Proportions Abstract: ABSTRACTWe study the weak supervision learning problem of Learning from Label Proportions (LLP) where the goal is to learn an instance-level classifier using proportions of various class labels in a bag -- a collection of input instances that often can be highly correlated. While representation learning for weakly-supervised tasks is found to be effective, they often require domain knowledge. To the best of our knowledge, representation learning for tabular data (unstructured data containing both continuous and categorical features) are not studied. In this paper, we propose to learn diverse representations of instances within the same bags to effectively utilize the weak bag-level supervision. We propose a domain agnostic LLP method, called "Self Contrastive Representation Learning for LLP" (SelfCLR-LLP) that incorporates a novel self--contrastive function as an auxiliary loss to learn representations on tabular data for LLP. We show that diverse representations for instances within the same bags aid efficient usage of the weak bag-level LLP supervision. We evaluate the proposed method through extensive experiments on real-world LLP datasets from e-commerce applications to demonstrate the effectiveness of our proposed SelfCLR-LLP. In this paper, we propose to learn diverse representations of instances within the same bags to effectively utilize the weak bag-level supervision. We propose a domain agnostic LLP method, called "Self Contrastive Representation Learning for LLP" (SelfCLR-LLP) that incorporates a novel self--contrastive function as an auxiliary loss to learn representations on tabular data for LLP. We show that diverse representations for instances within the same bags aid efficient usage of the weak bag-level LLP supervision. We evaluate the proposed method through extensive experiments on real-world LLP datasets from e-commerce applications to demonstrate the effectiveness of our proposed SelfCLR-LLP.

Title: Network Aware Forecasting for eCommerce Supply Planning Abstract: ABSTRACTA real world supply chain planning starts with the demand forecasting as a key input. In most scenarios, especially in fields like e-commerce where demand patterns are complex and are large scale, demand forecasting is done independent of supply chain constraints. There have been a plethora of methods, old and recent, for generating accurate forecasts. However, to the best of our knowledge, none of the methods take supply chain constraints into account during forecasting. In this paper, we are primarily interested in supply chain aware forecasting methods that does not impose any restrictions on demand forecasting process. We assume that the base forecasts follow a distribution from exponential family and are provided as input to supply chain planning by specifying the distribution form and parameters. With this in mind, following are the contributions of our paper. First, we formulate the supply chain aware forecast improvement of a base forecast as finding the game theoretically optimal parameters satisfying the supply chain constraints. Second, for regular distributions from exponential family, we show that this translates to projecting base forecast onto the (convex) set defined by supply constraints, which is at least as accurate as the base forecasts. Third, we note that using off the shelf convex solvers does not scale for large instances of supply chain, which is typical in e-commerce settings. We propose algorithms that scale better with problem size. We propose a general gradient descent based approach that works across different distributions from exponential family. We also propose a network flow based exact algorithm for Laplace distribution (which relates to mean absolute error, which is the most commonly used metric in forecasting). Finally, we substantiate the theoretical results with extensive experiments on a real life e-commerce data set as well as a range of synthetic data sets.

Title: Hierarchical Spatio-Temporal Graph Neural Networks for Pandemic Forecasting Abstract: ABSTRACTThe spread of COVID-19 throughout the world has led to cataclysmic consequences on the global community, which poses an urgent need to accurately understand and predict the trajectories of the pandemic. Existing research has relied on graph-structured human mobility data for the task of pandemic forecasting. To perform pandemic forecasting of COVID-19 in the United States, we curate Large-MG, a large-scale mobility dataset that contains 66 dynamic mobility graphs, with each graph having over 3k nodes and an average of 540k edges. One drawback with existing Graph Neural Networks (GNNs) for pandemic forecasting is that they generally perform information propagation in a flat way and thus ignore the inherent community structure in a mobility graph. To bridge this gap, we propose a Hierarchical Spatio-Temporal Graph Neural Network (HiSTGNN) to perform pandemic forecasting, which learns both spatial and temporal information from a sequence of dynamic mobility graphs. HiSTGNN consists of two network architectures. One is a hierarchical graph neural network (HiGNN) that constructs a two-level neural architecture: county-level and region-level, and performs information propagation in a hierarchical way. The other network architecture is a Transformer-based model that captures the temporal dynamics among the sequence of learned node representations from HiGNN. Additionally, we introduce a joint learning objective to further optimize HiSTGNN. Extensive experiments have demonstrated HiSTGNN's superior predictive power of COVID-19 new case/death counts compared with state-of-the-art baselines.

Title: MORN: Molecular Property Prediction Based on Textual-Topological-Spatial Multi-View Learning Abstract: ABSTRACTPredicting molecular properties has significant implications for the discovery and generation of drugs and further research in the domain of medicinal chemistry. Learning representations of molecules plays a central role in deep learning-driven property prediction. However, the diversity of molecular features (e.g., chemical system languages, structure notations) brings inconsistency in molecular representation. Moreover, the scarcity of labeled molecular data limits the accuracy of the molecular property prediction model. To address the above issues, we proposed a two-stage method, named MORN, for learning molecular representations for molecular property prediction from a multi-view perspective. In the first stage, textual-topological-spatial multi-views were proposed to learn the molecular representations, so as to capture both chemical system language and structure notation features simultaneously. In the second stage, an adaptive strategy was used to fuse molecular representations learned from multi-views to predict molecular properties. To alleviate the limitation of the scarcity of labeled molecular data, the label restriction was introduced in both multi-view representation learning and fusion stages. The performance of MORN was assessed by seven benchmark molecular datasets and one self-built molecular dataset. Experimental results demonstrated that MORN is effective in molecular property prediction.

Title: Knowledge-Sensed Cognitive Diagnosis for Intelligent Education Platforms Abstract: ABSTRACTCognitive diagnosis is a fundamental issue of intelligent education platforms, whose goal is to reveal the mastery of students on knowledge concepts. Recently, certain efforts have been made to improve the diagnosis precision, by designing deep neural networks-based diagnostic functions or incorporating more rich context features to enhance the representation of students and exercises. However, how to interpretably infer the student's mastery over non-interactive knowledge concepts (i.e., knowledge concepts not related to his/her exercising records) still remains challenging, especially when not giving relations between knowledge concepts. To this end, we propose a Knowledge-Sensed Cognitive Diagnosis (KSCD) framework, aiming at learning intrinsic relations among knowledge concepts from student response logs and incorporating them for inferring students' mastery over all knowledge concepts in an end-to-end manner. Specifically, we firstly project students, exercises and knowledge concepts into embedding representation matrices, where the intrinsic relations among knowledge concepts are reflected in the knowledge embedding representation matrix. Then, the knowledge-sensed student knowledge mastery vector and exercise factor vectors are obtained by the multiply product of their embedding representations and the knowledge embedding representation matrix, which make the student's mastery of non-interactive knowledge concepts be interpretably inferred. Finally, we can utilize classical student-exercise interaction functions to predict student's exercising performance and jointly train the model. In additional, we also design a new function to better model the student-exercise interactions. Extensive experimental results on two real-world datasets clearly show the significant performance gain of our KSCD framework, especially in predicting students' mastery over non-interactive knowledge concepts, by comparing to state-of-the-art cognitive diagnosis models (CDMs).

Title: NEST: Simulating Pandemic-like Events for Collaborative Filtering by Modeling User Needs Evolution Abstract: ABSTRACTWe outline a simulation-based study of the effect rapid population-scale concept drifts have on Collaborative Filtering (CF) models. We create a framework for analyzing the effects of macro-trends in population dynamics on the behavior of such models. Our framework characterizes population-scale concept drifts in item preferences and provides a lens to understand the influence events, such as a pandemic, have on CF models. Our experimental results show the initial impact on CF performance at the initial stage of such events, followed by an aggravated population herding effect during the event. The herding introduces a popularity bias that may benefit affected users, but which comes at the expense of a normal user experience. We propose an adaptive ensemble method that can effectively apply optimal algorithms to cope with the change brought about by different stages of the event.

Title: DEMO: Disentangled Molecular Graph Generation via an Invertible Flow Model Abstract: ABSTRACTMolecular graph generation via deep generative models has attracted increasing attention. This is a challenging problem because it requires optimizing a given objective under a huge search space while obeying the chemical valence rules. Although recently developed molecular generation models have achieved promising results on generating novel, valid and unique molecules, few efforts have been made toward interpretable molecular graph generation. In this work, we propose DEMO, a flow-based model for DisEntangled Molecular graph generatiOn in a completely unsupervised manner, which is able to generate molecular graphs w.r.t. the learned disentangled latent factors that are relevant to molecular semantic features and interpretable structural patterns. Specifically, DEMO is composed of a VAE-encoder and a flow-generator. The VAE-encoder focuses on extracting global features of molecular graphs, and the flow-generator aims at disentangling these features to be corresponding to certain types of understandable molecular structure features while learning data distributions. To generate molecular graphs, DEMO simply runs the flow-generator in the reverse order due to the reversibility of the flow-based models. Extensive experimental results on two benchmark datasets demonstrate that DEMO outperforms the state-of-the-art methods in molecular generation, and takes the first step in interpretable molecular graph generation.

Title: Faithful Abstractive Summarization via Fact-aware Consistency-constrained Transformer Abstract: ABSTRACTAbstractive summarization is a classic task in Natural Language Generation (NLG), which aims to produce a concise summary of the original document. Recently, great efforts have been made on sequence-to-sequence neural networks to generate abstractive sum- maries with a high level of fluency. However, prior arts mainly focus on the optimization of token-level likelihood, while the rich semantic information in documents has been largely ignored. In this way, the summarization results could be vulnerable to hallucinations, i.e., the semantic-level inconsistency between a summary and corresponding original document. To deal with this challenge, in this paper, we propose a novel fact-aware abstractive summarization model, named Entity-Relation Pointer Generator Network (ERPGN). Specially, we attempt to formalize the facts in original document as a factual knowledge graph, and then generate the high-quality summary via directly modeling consistency between summary and the factual knowledge graph. To that end, we first leverage two pointer net- work structures to capture the fact in original documents. Then, to enhance the traditional token-level likelihood loss, we design two extra semantic-level losses to measure the disagreement between a summary and facts from its original document. Extensive experi- ments on public datasets demonstrate that our ERPGN framework could outperform both classic abstractive summarization models and the state-of-the-art fact-aware baseline methods, with significant improvement in terms of faithfulness.

Title: Dual-Task Learning for Multi-Behavior Sequential Recommendation Abstract: ABSTRACTRecently, sequential recommendation has become a research hotspot while multi-behavior sequential recommendation (MBSR) that exploits users' heterogeneous interactions in sequences has received relatively little attention. Existing works often overlook the complementary effect of different perspectives when addressing the MBSR problem. In addition, there are two specific challenges remained to be addressed. One is the heterogeneity of a user's intention and the context information, the other one is the sparsity of the interactions of target behavior. To release the potential of multi-behavior interaction sequences, we propose a novel framework named NextIP that adopts a dual-task learning strategy to convert the problem to two specific tasks, i.e., next-item prediction and purchase prediction. For next-item prediction, we design a target-behavior aware context aggregator (TBCG), which utilizes the next behavior to guide all kinds of behavior-specific item sub-sequences to jointly predict the next item. For purchase prediction, we design a behavior-aware self-attention (BSA) mechanism to extract a user's behavior-specific interests and treat them as negative samples to learn the user's purchase preferences. Extensive experimental results on two public datasets show that our NextIP performs significantly better than the state-of-the-art methods.

Title: Learning Chinese Word Embeddings By Discovering Inherent Semantic Relevance in Sub-characters Abstract: ABSTRACTLearning Chinese word embeddings is important in many tasks of Chinese language information processing, such as entity linking, entity extraction, and knowledge graph. A Chinese word consists of Chinese characters, which can be decomposed into sub-characters (radical, component, stroke, etc). Similar to roots in English words, sub-characters also indicate the origins and basic semantics of Chinese characters. So, many researches follow the approaches designed for learning embeddings of English words to improve Chinese word embeddings. However, some Chinese characters sharing the same sub-characters have different meanings. Furthermore, with more cultural interaction and the popularization of the Internet and web, many neologisms, such as transliterated loanwords and network terms, are emerging, which are only close to the pronunciation of their characters, but far from their semantics. Here, a tripartite weighted graph is proposed to model the semantic relationship among words, characters, and sub-characters, in which the semantic relationship is evaluated according to the Chinese linguistic information. So, the semantic relevance hidden in lower components (sub-characters, characters) can be used to further distinguish the semantics of corresponding higher components (characters, words). Then, the tripartite weighted graph is fed into our Chinese word embedding modelinsideCC to reveal the semantic relationship among different language components, and learn the embeddings of words. Extensive experimental results on multiple corpora and datasets verify that our proposed methods outperform the state-of-the-art counterparts by a significant margin.

Title: Social Graph Transformer Networks for Pedestrian Trajectory Prediction in Complex Social Scenarios Abstract: ABSTRACTPedestrian trajectory prediction is essential for many modern applications, such as abnormal motion analysis and collision avoidance for improved traffic safety. Previous studies still face challenges in embracing high social interaction, dynamics, and multi-modality for achieving high accuracy with long-time predictions. We propose Social Graph Transformer Networks for multi-modal prediction of pedestrian trajectories, where we combine Graph Convolutional Network and Transformer Network by generating stable resolution pseudo-images from Spatio-temporal graphs through a designed stacking and interception method. Specifically, we adopt adjacency matrices to obtain Spatio-temporal features and Transformer for long-time trajectory predictions. As such, we retrain the advantages of both, i.e., the ability to aggregate information over an arbitrary number of neighbors and to conduct complex time-dependent data processing. Our experimental results show that our model reduces the final displacement error and achieves state-of-the-art in multiple metrics. The module's effectiveness is demonstrated through ablation experiments.

Title: HeGA: Heterogeneous Graph Aggregation Network for Trajectory Prediction in High-Density Traffic Abstract: ABSTRACTTrajectory prediction enables the fast and accurate response of autonomous driving navigation in complex and dense traffics. In this paper, we present a novel trajectory prediction network called Heterogeneous Graph Aggregation (HeGA) for high-density heterogeneous traffic, where the traffic agents of various categories interact densely with each other. To predict the trajectory of a target agent, HeGA first automatically selects neighbors that interact with it by our proposed adaptive neighbor selector, and then aggregates their interactions based on a novel two-phase aggregation transformer block. At last, the historical residual connection LSTM enhances the historical information awareness and decodes the spatial coordinates as the prediction results. Extensive experiments on real data demonstrate that the proposed network significantly outperforms the existing state-of-the-art competitors by over 27% on average displacement error (ADE) and over 31% on final displacement error (FDE). We also deploy HeGA in a state-of-the-art framework for autonomous driving, demonstrating its superior applicability based on three simulated environments with different densities and complexities.

Title: Efficient Learning with Pseudo Labels for Query Cost Estimation Abstract: ABSTRACTQuery cost estimation, which is to estimate the query plan cost and query execution cost, is of utmost importance to query optimizers. Query plan cost estimation heavily relies on accurate cardinality estimation, and query execution cost estimation gives good hints on query latency, both of which are challenging in database management systems. Despite decades of research, existing studies either over-simplify the models only using histograms and polynomial calculation that leads to inaccurate estimates, or over-complicate them by using cumbersome neural networks with the requirements for large amounts of training data hence poor computational efficiency. Besides, most of the studies ignore the diversity of query plan structures. In this work, we propose a plan-based query cost estimation framework, called Saturn, which can eStimate cardinality and latency accurately and efficiently, for any query plan structures. Saturn first encodes each query plan tree into a compressed vector by using a traversal-based query plan autoencoder to cope with diverse plan structures. The compressed vectors can be leveraged to distinguish different query types, which is highly useful for downstream tasks. Then a pseudo label generator is designed to acquire all cardinality and latency labels with the execution part of the query plans in the training workload, which can significantly reduce the overhead of collecting the real cardinality and latency labels. Finally, a chain-wise transfer learning module is proposed to estimate the cardinality and latency of the query plan in a pipeline paradigm, which further enhances the efficiency. An extensive empirical study on benchmark data offers evidence that Saturn outperforms the state-of-the-art proposals in terms of accuracy, efficiency, and generalizability for query cost estimation.

Title: Unsupervised Hierarchical Graph Pooling via Substructure-Sensitive Mutual Information Maximization Abstract: ABSTRACTGraph pooling plays a vital role in learning graph embeddings. Due to the lack of label information, unsupervised graph pooling has received much attention, primarily via mutual information (MI). However, most existing MI-based pooling methods only preserve node features while overlooking the hierarchical substructural information. In this paper, we propose SMIP, a novel unsupervised hierarchical graph pooling method based on substructure-sensitive MI maximization. SMIP reconstructs a hard-style substructure encoder based on cluster-based pooling paradigm, and trains it with two substructure-sensitive MI-based objectives, i.e., node-substructure MI and node-node MI. The node-substructure MI guides to transfer maximum node feature information into corresponded substructures and the node-node MI guarantees a more accurate node allocation. Moreover, to avoid extra computation of augmented graphs and prevent noise information during MI estimation, we propose a local-scope contrastive MI estimation method, making SMIP more potent in capturing intrinsic features of the input graph. Experiments on six benchmark graph classification datasets demonstrate that our hierarchical deep learning approach outperforms all state-of-the-art unsupervised GNN-based methods and even surpasses the performance of nine supervised ones. Generalization study shows that the proposed substructure-sensitive MI objective can be successfully embedded into other cluster-based pooling methods to improve their performance.

Title: Task Assignment with Federated Preference Learning in Spatial Crowdsourcing Abstract: ABSTRACTSpatial Crowdsourcing (SC) is ubiquitous in the online world today. As we have transitioned from crowdsourcing applications (e.g., Wikipedia) to SC applications (e.g., Uber), there is a substantial precedent that SC systems have a responsibility not only to effective task assignment but also to privacy protection. To address these often-conflicting responsibilities, we propose a framework, Task Assignment with Federated Preference Learning, which performs task assignment based on worker preferences while keeping the data decentralized and private in each platform center (e.g., each delivery center of an SC company). The framework includes two phases, i.e., a federated preference learning and a task assignment phase. Specifically, in the first phase, we design a local preference model for each platform center based on historical data. Meanwhile, the horizontal federated learning with a client-server structure is introduced to collaboratively train these local preference models under the orchestration of a central server. The task assignment phase aims to achieve effective and efficient task assignment by considering workers' preferences. Extensive evaluations over real data show the effectiveness and efficiency of the paper's proposals.

Title: Predicting Intraoperative Hypoxemia with Hybrid Inference Sequence Autoencoder Networks Abstract: ABSTRACTWe present an end-to-end model using streaming physiological time series to predict near-term risk for hypoxemia, a rare, but life-threatening condition known to cause serious patient harm during surgery. Inspired by the fact that a hypoxemia event is defined based on a future sequence of low SpO2 (i.e., blood oxygen saturation) instances, we propose the hybrid inference network (hiNet) that makes hybrid inference on both future low SpO2 instances and hypoxemia outcomes. hiNet integrates 1) a joint sequence autoencoder that simultaneously optimizes a discriminative decoder for label prediction, and 2) two auxiliary decoders trained for data reconstruction and forecast, which seamlessly learn contextual latent representations that capture the transition from present states to future states. All decoders share a memory-based encoder that helps capture the global dynamics of patient measurement. For a large surgical cohort of 72,081 surgeries at a major academic medical center, our model outperforms strong baselines including the model used by the state-of-the-art hypoxemia prediction system. With its capability to make real-time predictions of near-term hypoxemic at clinically acceptable alarm rates, hiNet shows promise in improving clinical decision making and easing burden of perioperative care.

Title: Cascade Variational Auto-Encoder for Hierarchical Disentanglement Abstract: ABSTRACTWhile deep generative models pave the way for many emerging applications, decreased interpretability for larger model sizes and complexities hinders their generalizability to wide domains such as economy, security, healthcare, etc. Considering this obstacle, a common practice is to learn interpretable representations through latent feature disentanglement, aiming for exposing a set of mutually independent factors of data variations. However, existing methods either fail to catch the trade-off between the synthetic data quality and model interpretability, or consider the first-order feature disentangling only, overlooking the fact that a subset of salient features can carry decomposable semantic meanings and hence be of high-order in nature. Hence, we in this paper propose a novel generative modeling paradigm by introducing a Bayesian network-based regularize on a cascade Variational Auto-Encoder (VAE). Specifically, this regularizer guides the learner to discover a representation space that comprises both first-order disentangled features and high-order salient features, with the feature interplay captured by the Bayesian structure. Experiments demonstrate that this regularizer gives us free control over the representation space and can guide the learner to discover decomposable semantic meanings by capturing the interplay among independent factors. Meanwhile, we benchmark extensive experiments on six widely-used vision datasets, and the results exhibit that our approach outperforms the state-of-the-art VAE competitors in terms of the trade-off between the synthetic data quality and model interpretability. Although our design is framed in the VAE regime, it in effect is generic and can be better amenable to both GANs and VAEs in terms of letting them concurrently enjoy both high model interpretability and high synthesis quality.

Title: TrajFormer: Efficient Trajectory Classification with Transformers Abstract: ABSTRACTTransformers have been an efficient alternative to recurrent neural networks in many sequential learning tasks. When adapting transformers to modeling trajectories, we encounter two major issues. First, being originally designed for language modeling, transformers assume regular intervals between input tokens, which contradicts the irregularity of trajectories. Second, transformers often suffer high computational costs, especially for long trajectories. In this paper, we address these challenges by presenting a novel transformer architecture entitled TrajFormer. Our model first generates continuous point embeddings by jointly considering the input features and the information of spatio-temporal intervals, and then adopts a squeeze function to speed up the representation learning. Moreover, we introduce an auxiliary loss to ease the training of transformers using the supervision signals provided by all output tokens. Extensive experiments verify that our TrajFormer achieves a preferable speed-accuracy balance compared to existing approaches.

Title: Dynamic Network Embedding via Temporal Path Adjacency Matrix Factorization Abstract: ABSTRACTNetwork embedding has been widely investigated to learn low dimensional nodes representation of networks, and serves for many downstream machine learning tasks. Previous network embedding studies mainly focus on static networks, and cannot adapt well to the characteristics of dynamic networks which are evolving over time. Some works on dynamic network embedding have tried to improve the computation efficiency for incremental updates of embedding vectors, while others have made efforts to utilize temporal information to enhance the quality of embedding vectors. However, few existing works can fulfill both efficiency and quality requirements. In this article, a novel dynamic network embedding model named TPANE (Temporal Path Adjacency Matrix based Network Embedding) is proposed. It employs a new network proximity measure: Temporal Path Adjacency, which is capable of capturing the temporal dependency between edges as well as being incrementally computed in an efficient way. It evaluates the similarity between nodes via the count of temporal paths between them, rather than making random sampling approximation, and adopts matrix factorization to obtain embedding vectors. Link prediction experiments on various real-world dynamic networks have been conducted to show the superior performance of TPANE against other state-of-the-art methods. Time consumption analysis also shows that TPANE is more efficient in incremental updates.

Title: Spatiotemporal-aware Session-based Recommendation with Graph Neural Networks Abstract: ABSTRACTSession-based recommendation (SBR) aims to recommend items based on user behaviors in a session. For the online life service platforms, such as Meituan, both the user's location and the current time primarily cause the different patterns and intents in user behaviors. Hence, spatiotemporal context plays a significant role in the recommendation on those platforms, which motivates an important problem of spatiotemporal-aware session-based recommendation (STSBR). Since the spatiotemporal context is introduced, there are two critical challenges: 1) how to capture session-level relations of spatiotemporal context (inter-session view), and 2) how to model the complex user decision-making process at a specific location and time (intra-session view). To address them, we propose a novel solution named STAGE in this paper. Specifically, STAGE first constructs a global information graph to model the multi-level relations among all sessions, and a session decision graph to capture the complex user decision process for each session. STAGE then performs inter-session and intra-session embedding propagation on the constructed graphs with the proposed graph attentive convolution (GAC) to learn representations from the above two perspectives. Finally, the learned representations are combined with spatiotemporal-aware soft-attention for final recommendation. Extensive experiments on two datasets from Meituan demonstrate the superiority of STAGE over state-of-the-art methods. Further studies also verify that each component is effective.

Title: Gromov-Wasserstein Guided Representation Learning for Cross-Domain Recommendation Abstract: ABSTRACTCross-Domain Recommendation (CDR) has attracted increasing attention in recent years as a solution to the data sparsity issue. The fundamental paradigm of prior efforts is to train a mapping function based on the overlapping users/items and then apply it to the knowledge transfer. However, due to the commercial privacy policy and the sensitivity of user data, it is unrealistic to explicitly share the user mapping relations and behavior data. Therefore, in this paper, we consider a more practical cross-domain scenario, where there is no explicit overlap between the source and target domains in terms of users/items. Since the user sets of both domains are drawn from the entire population, there may be commonalities between their user characteristics, resulting in comparable user preference distributions. Thus, without the mapping relations at user level, it is feasible to model this distribution-level relation to transfer knowledge between domains. To this end, we propose a novel framework that improves the effect of representation learning on the target domain by aligning the representation distributions between the source and target domains. In addition, GWCDR can be easily integrated with existing single-domain collaborative filtering methods to achieve cross-domain recommendation. Extensive experiments on two pairs of public bidirectional datasets demonstrate the effectiveness of our proposed framework in enhancing the recommendation performance.

Title: ℘-MinHash Algorithm for Continuous Probability Measures: Theory and Application to Machine Learning Abstract: ABSTRACTThis paper studies the scale-invariant "probability Jaccard'' (ProbJ), noted as ℐ℘, which is another variant of weighted Jaccard similarity. The standard and commonly used Jaccard index is not invariant of data scaling. Thus, the probability Jaccard can be a potentially useful extension to probability distributions. Before our paper, the problem of hashing the ℐ℘ for continuous probability measures is an open problem, where rigorous definitions and analysis are still absent in literature. In our work, we solve this problem systematically and completely. Specifically, we formalize the definition of ℐ℘ in continuous measure space, and propose a general ℘-MinHash sampling algorithm which generates samples following any target distribution, and preserves ℐ℘ between two distributions by the hash collision. In addition, a refined early stopping rule is proposed under a practical boundedness assumption. We validate the theory through simulation and experiments, and demonstrate the application of our method in machine learning problems.

Title: Heterogeneous Graph Attention Network for Drug-Target Interaction Prediction Abstract: ABSTRACTIdentification of drug-target interactions (DTIs) is crucial for drug discovery and drug repositioning. Existing graph neural network (GNN) based methods only aggregate information from directly connected nodes restricted in a drug-related or a target-related network, and are incapable of capturing long-range dependencies in the biological heterogeneous graph. In this paper, we propose the heterogeneous graph attention network (HGAN) to capture the complex structures and rich semantics in the biological heterogeneous graph for DTI prediction. HGAN enhances heterogeneous graph structure learning from both the intra-layer perspective and the inter-layer perspective. Concretely, we develop an enhanced graph attention diffusion layer (EGADL), which efficiently builds connections between node pairs that may not be directly connected, enabling information passing from important nodes multiple hops away. By stacking multiple EGADLs, we further enlarge the receptive field from the inter-layer perspective. HGAN advances 15 state-of-the-art methods on two heterogeneous biological datasets, achieving the results near to 1 in terms of AUC and AUPR. We also find that enlarging receptive fields from the inter-layer perspective (stacking layers) is more effective than that from the intra-layer perspective (attention diffusion) for HGAN to achieve promising DTI prediction performances. The code is available at https://github.com/Zora-LM/HGAN-DTI.

Title: AdaDebunk: An Efficient and Reliable Deep State Space Model for Adaptive Fake News Early Detection Abstract: ABSTRACTAutomatically detecting fake news as early as possible becomes increasingly necessary. Conventional approaches of fake news early detection (FNED) verify news' veracity with a predefined and indiscriminate detection position, which depends on domain experience and leads to unstable performance. More advanced methods address this problem with a proposed concept of adaptive detection position (ADP), i.e. the position where the veracity of the news record can be concluded. Yet these methods either lack theoretical reliability or weaken complex dependencies among multi-aspect clues, thus failing to provide practical and reasonable detection. This work focuses on the adaptive FNED problem and proposes a novel efficient and reliable deep state space model, namely AdaDebunk, which models the complex probabilistic dependencies. Specifically, a Bayes' theorem-based dynamic inference algorithm is designed to infer the ADPs and veracity, supporting the accumulation of multi-aspect clues. Besides, a training mechanism with hybrid loss is also designed to solve the over-/under-fitting problems, which further trades off the performance and generalization ability. Experiments on two real-world fake news datasets are conducted to evaluate the effectiveness and superiority of AdaDebunk. Compared with the state-of-the-art baselines, AdaDebunk achieves a 10% increase in F1 performance. Meanwhile, a case study is provided to demonstrate the reliability of AdaDebunk as well as our research motivation.

Title: Frequent Itemset Mining with Local Differential Privacy Abstract: ABSTRACTWith the development of the Internet, a large amount of transaction data (e.g., shopping records, web browsing history), which represents user data, has been generated. By collecting user transaction data and learning specific patterns and association rules from it, service providers can provide better services. However, because of the increasing privacy awareness and the formulation of laws on data protection, collecting data directly from users will raise privacy concerns. The concept of local differential privacy (LDP), which provides strict data privacy protection on the user side and allows effective statistical analysis on the server side, is able to protect user privacy and perform statistics on sensitive issues at the same time. This paper adopts padding-and-sampling-based frequent oracle (PSFO), combined with an interactive query-response method satisfying local differential privacy, to identify frequent itemsets in an efficient and accurate way. Therefore, this paper proposes FIML, an improved algorithm for finding frequent itemsets in the LDP setting of transaction data. The data collector generates frequent candidate sets based on the results of the previous stage and uses them for querying, and users randomize their responses in a reduced domain to achieve local differential privacy. Extensive experiments on real-world and synthetic datasets show that the FIML algorithm can find frequent itemsets more efficiently with the same privacy protection and computational cost.

Title: Multi-agent Transformer Networks for Multimodal Human Activity Recognition Abstract: ABSTRACTHuman activity recognition has become an important challenge yet to resolve while also having promising benefits in various applications for years. Existing approaches have made great progress by applying deep-learning and attention-based methods. However, the deep learning-based approaches may not fully exploit the features to resolve multimodal human activity recognition tasks. Also, the potential of attention-based methods still has not been fully explored to better extract the multimodal spatial-temporal relationship and produce robust results. In this work, we propose Multi-agent Transformer Network (MATN), a multi-agent attention-based deep learning algorithm, to address the above issues in multimodal human activity recognition. We first design a unified representation learning layer to encode the multimodal data, which preprocesses the data in a generalized and efficient way. Then we develop a multimodal spatial-temporal transformer module that applies the attention mechanism to extract the salient spatial-temporal features. Finally, we use a multi-agent training module to collaboratively select the informative modalities and predict the activity labels. We have extensively conducted experiments to evaluate MATN's performance on two public multimodal human activity recognition datasets. The results show that our model has achieved competitive performance compared to the state-of-the-art approaches, which also demonstrates scalability, effectiveness, and robustness.

Title: SPOT: Knowledge-Enhanced Language Representations for Information Extraction Abstract: ABSTRACTKnowledge-enhanced pre-trained models for language representation have been shown to be more effective in knowledge base construction tasks (i.e.,~relation extraction) than language models such as BERT. These knowledge-enhanced language models incorporate knowledge into pre-training to generate representations of entities or relationships. However, existing methods typically represent each entity with a separate embedding. As a result, these methods struggle to represent out-of-vocabulary entities and a large amount of parameters, on top of their underlying token models (i.e., the transformer), must be used and the number of entities that can be handled is limited in practice due to memory constraints. Moreover, existing models still struggle to represent entities and relationships simultaneously. To address these problems, we propose a new pre-trained model that learns representations of both entities and relationships from token spans and span pairs in the text respectively. By encoding spans efficiently with span modules, our model can represent both entities and their relationships but requires fewer parameters than existing models. We pre-trained our model with the knowledge graph extracted from Wikipedia and test it on a broad range of supervised and unsupervised information extraction tasks. Results show that our model learns better representations for both entities and relationships than baselines, while in supervised settings, fine-tuning our model outperforms RoBERTa consistently and achieves competitive results on information extraction tasks.

Title: CoPatE: A Novel Contrastive Learning Framework for Patent Embeddings Abstract: ABSTRACTPatents are legal rights issued to inventors to protect their inventions for a certain period and play an important role in today's artificial innovation. With the ever-increasing number of patents each year, an effective and efficient patent management and search system is indispensable for determining how different an invention is from prior works from the vast amount of patent data. However, the chnologists are using now is still based on the strategy of traditional keyword-based Boolean, which requires complex bool expressions. This type of strategy leads to poor performance and costs too much labor power to filter in post-processing. To address these issues, we proposed CoPatE: a novel Contrastive Learning Framework for Patent Embeddings to capture the high-level semantics of the large-scale patents, where a patent semantic compression module learns the informative claims to reduce the computational complexity, and a tags auxiliary learning module is to enhance the semantics of a patent from the structure to learn the high-quality patent embeddings. The CoPatE is trained with the patents from USPTO from 2013 to 2020 and tested by the patents from 2021 with the CPC scheme. The experimental results demonstrate that our model achieves a 17.7% increase at [email protected] compared to the second-best method on the patent retrieval task and achieves 64.5% at Micro-F1 in the patent classification task.

Title: MDGCF: Multi-Dependency Graph Collaborative Filtering with Neighborhood- and Homogeneous-level Dependencies Abstract: ABSTRACTDue to the success of graph convolutional networks (GCNs) in effectively extracting features in non-Euclidean spaces, GCNs has become the rising star in implicit collaborative filtering. Existing works, while encouraging, typically adopt simple aggregation operation on the user-item bipartite graph to model user and item representations, but neglect to mine the sufficient dependencies between nodes, e.g., the relationships between users/items and their neighbors (or congeners), resulting in inadequate graph representation learning. To address these problems, we propose a novel Multi-Dependency Graph Collaborative Filtering (MDGCF) model, which mines the neighborhood- and homogeneous-level dependencies to enhance the representation power of graph-based CF models. Specifically, for neighborhood-level dependencies, we explicitly consider both popularity score and preference correlation by designing a joint neighborhood-level dependency weight, based on which we construct a neighborhood-level dependencies graph to capture higher-order interaction features. Besides, by adaptively mining the homogeneous-level dependencies among users and items, we construct two homogeneous graphs, based on which we further aggregate features from homogeneous users and items to supplement their representations, respectively. Extensive experiments on three real-world benchmark datasets demonstrate the effectiveness of the proposed MDGCF. Further experiments reveal that our model can capture rich dependencies between nodes for explaining user behaviors.

Title: Automated Spatio-Temporal Synchronous Modeling with Multiple Graphs for Traffic Prediction Abstract: ABSTRACTTraffic prediction plays an important role in many intelligent transportation systems. Many existing works design static neural network architecture to capture complex spatio-temporal correlations, which is hard to adapt to different datasets. Although recent neural architecture search approaches have addressed this problem, it still adopts a coarse-grained search with pre-defined and fixed components in the search space for spatio-temporal modeling. In this paper, we propose a novel neural architecture search framework, entitled AutoSTS, for automated spatio-temporal synchronous modeling in traffic prediction. To be specific, we design a graph neural network (GNN) based architecture search module to capture localized spatio-temporal correlations, where multiple graphs built from different perspectives are jointly utilized to find a better message passing way for mining such correlations. Further, we propose a convolutional neural network (CNN) based architecture search module to capture temporal dependencies with various ranges, where gated temporal convolutions with different kernel sizes and convolution types are designed in search space. Extensive experiments on six public datasets demonstrate that our model can achieve 4%-10% improvements compared with other methods.

Title: Parallel Skyline Processing Using Space Pruning on GPU Abstract: ABSTRACTSkyline computation is an essential database operation that has many applications in multi-criteria decision making scenarios such as recommender systems. Existing algorithms have focused on checking point domination, which lack efficiency over large datasets. We propose a grid-based structure that enables grid cell domination checks. We show that only a small constant number of cells need to be checked which is independent from the number of data points. Our structure also enables parallel processing. We thus obtain a highly efficient parallel skyline algorithm named SkyCell, taking advantage of the parallelization power of graphics processing units. Experimental results confirm the effectiveness and efficiency of SkyCell -- it outperforms state-of-the-art algorithms consistently and by up to over two orders of magnitude in the computation time.

Title: Maximum Norm Minimization: A Single-Policy Multi-Objective Reinforcement Learning to Expansion of the Pareto Front Abstract: ABSTRACTIn this paper, we propose Maximum Norm Minimization (MNM), a single-policy Multi-Objective Reinforcement Learning (MORL) algorithm to solve the multi-objective RL problem. The main objective of our MNM is to provide the Pareto optimal points constituting the Pareto front in the multi-objective space. First, MNM measures distances among the Pareto optimal points in the current Pareto front and then normalizes the distances based on maximum and minimum reward values for each objective in the multi-objective space. Second, MNM identifies the maximum norm, i.e., the maximum value of the normalized Pareto optimal distances. Then MNM seeks to find a new Pareto optimal point, which corresponds to the middle of the two Pareto optimal points constituting the maximum norm. By iterating these two processes, MNM is able to expand and densify the Pareto front with increasing summation of the Pareto front volumes and decreasing mean-squared distance of the Pareto optimal points. To validate the performance of MNM, we provide the experimental results of five complex robotic multi-objective environments. In particular, we compare the performance of MNM with those of other state-of-the-art methods in terms of the summation of volumes and the mean-squared distance of the Pareto optimal points.

Title: Sliding Cross Entropy for Self-Knowledge Distillation Abstract: ABSTRACTKnowledge distillation (KD) is a powerful technique for improving the performance of a small model by leveraging the knowledge of a larger model. Despite its remarkable performance boost, KD has a drawback with the substantial computational cost of pre-training larger models in advance. Recently, a method called self-knowledge distillation has emerged to improve the model's performance without any supervision. In this paper, we present a novel plug-in approach called Sliding Cross Entropy (SCE) method, which can be combined with existing self-knowledge distillation to significantly improve the performance. Specifically, to minimize the difference between the output of the model and the soft target obtained by self-distillation, we split each softmax representation by a certain window size, and reduce the distance between sliced parts. Through this approach, the model evenly considers all the inter-class relationships of a soft target during optimization. The extensive experiments show that our approach is effective in various tasks, including classification, object detection, and semantic segmentation. We also demonstrate SCE consistently outperforms existing baseline methods.

Title: Accelerating CNN via Dynamic Pattern-based Pruning Network Abstract: ABSTRACTRecently, dynamic pruning methods have been actively researched, as they have shown very effective and remarkable performance in reducing computation complexity of deep neural networks. Nevertheless, most dynamic pruning methods fail to achieve actual acceleration due to the extra overheads caused by indexing and weight-copying to implement the dynamic sparse patterns for every input sample. To address this issue, we propose Dynamic Pattern-based Pruning Network (DPPNet), which preserves the advantages of both static and dynamic networks. First, our method statically prunes the weight kernel into various sparse patterns. Then, the dynamic convolution kernel is generated via aggregating input-dependent attention weights and static kernels. Unlike previous dynamic pruning methods, our novel method dynamically fuses static kernel patterns, enhancing the kernel's representational power without additional overhead. Moreover, our dynamic sparse pattern enables an efficient process using BLAS libraries, accomplishing actual acceleration. We demonstrate the effectiveness of the proposed DPPNet on CIFAR and ImageNet, outperforming the state-of-the-art methods achieving better accuracy with lower computational cost. For example, on ImageNet classification, ResNet34 utilizing DPP module achieves state-of-the-art performance with 65.6% FLOPs reduction and the inference speed increased by 35.9% without loss in accuracy. Code is available at https://github.com/lee-gwang/DPPNet.

Title: Legal Charge Prediction via Bilinear Attention Network Abstract: ABSTRACTThe legal charge prediction task aims to judge appropriate charges according to the given fact description in cases. Most existing methods formulate it as a multi-class text classification problem and have achieved tremendous progress. However, the performance on low-frequency charges is still unsatisfactory. Previous studies indicate leveraging the charge label information can facilitate this task, but the approaches to utilizing the label information are not fully explored. In this paper, inspired by the vision-language information fusion techniques in the multi-modal field, we propose a novel model (denoted as LeapBank) by fusing the representations of text and labels to enhance the legal charge prediction task. Specifically, we devise a representation fusion block based on the bilinear attention network to interact the labels and text tokens seamlessly. Extensive experiments are conducted on three real-world datasets to compare our proposed method with state-of-the-art models. Experimental results show that LeapBank obtains up to 8.5% Macro-F1 improvements on the low-frequency charges, demonstrating our model's superiority and competitiveness.

Title: Loyalty-based Task Assignment in Spatial Crowdsourcing Abstract: ABSTRACTWith the fast-paced development of mobile networks and the widespread usage of mobile devices, Spatial Crowdsourcing (SC) has drawn increasing attention in recent years. SC has the potential for collecting information for a broad range of applications such as on-demand local delivery and on-demand transportation. One of the critical issues in SC is task assignment that allocates location-based tasks (e.g., delivering food and packages) to appropriate moving workers (i.e., intelligent device carriers). In this paper, we study a loyalty-based task assignment problem, which aims to maximize the overall rewards of workers while considering worker loyalty. We propose a two-phase framework to solve the problem, including a worker loyalty prediction and a task assignment phase. In the first phase, we use a model based on an efficient time series prediction method called Prophet and an Entropy Weighting method to extract workers' short-term and long-term loyalty and then predict workers' current loyalty scores. In the task assignment phase, we design a Kuhn-Munkras-based algorithm that achieves the optimal task assignment and an efficient Degree-Reduction-based algorithm with minority first scheme. Extensive experiments offer insight into the effectiveness and efficiency of the proposed solutions.

Title: Semorph: A Morphology Semantic Enhanced Pre-trained Model for Chinese Spam Text Detection Abstract: ABSTRACTChinese spam text detection is essential for social media since these texts affect the user experience of Chinese speakers and pollute the community. The underlying text classification method is employed to explore the unique combinations of characters that represent clues of spam information from annotated or further augmented data. However, based on the diversity of Chinese characters in glyphs, the spammers frequently wrap the spam content in another visually close text to fool the model but make sure people understand. This paper proposes to adopt the essence of human cognition of these adversarial texts into spam text detection models, by designing a pre-trained model to learn the morphology semantics of Chinese characters and represent their contextual meanings from scratch. The model pre-trains on self-supervised Chinese corpus and fine-tunes on spam-annotated community texts. Besides, cooperating with the pre-trained model that can capture the morphological features of Chinese, a new data perturbation method is introduced to guide the optimization towards the direction of recognizing the actual meaning of a text after spammers tamper with partial characters by visually close ones. The experimental results have shown that our proposed methodology can notably improve the performance of spam text detection as well as maintain robustness against adversarial samples.

Title: MARIO: Modality-Aware Attention and Modality-Preserving Decoders for Multimedia Recommendation Abstract: ABSTRACTWe address the multimedia recommendation problem, which utilizes items' multimodal features, such as visual and textual modalities, in addition to interaction information. While a number of existing multimedia recommender systems have been developed for this problem, we point out that none of these methods individually capture the influence of each modality at the interaction level. More importantly, we experimentally observe that the learning procedures of existing works fail to preserve the intrinsic modality-specific properties of items. To address above limitations, we propose an accurate multimedia recommendation framework, named MARIO, based on modality-aware attention and modality-preserving decoders. MARIO predicts users' preferences by considering the individual influence of each modality on each interaction while obtaining item embeddings that preserve the intrinsic modality-specific properties. The experiments on four real-life datasets demonstrate that MARIO consistently and significantly outperforms seven competitors in terms of the recommendation accuracy: MARIO yields up to 14.61% higher accuracy, compared to the best competitor.

Title: SWAG-Net: Semantic Word-Aware Graph Network for Temporal Video Grounding Abstract: ABSTRACTIn this paper, to effectively capture non-sequential dependencies among semantic words for temporal video grounding, we propose a novel framework called Semantic Word-Aware Graph Network (SWAG-Net), which adopts graph-guided semantic word embedding in an end-to-end manner. Specifically, we define semantic word features as node features of semantic word-aware graphs and word-to-word correlations as three edge types (i.e., intrinsic, extrinsic, and relative edges) for diverse graph structures. We then apply Semantic Word-aware Graph Convolutional Networks (SW-GCNs) to the graphs for semantic word embedding. For modality fusion and context modeling, the embedded features and video segment features are merged into bi-modal features, and the bi-modal features are aggregated by incorporating local and global contextual information. Leveraging the aggregated features, the proposed method effectively finds a temporal boundary semantically corresponding to a sentence query in an untrimmed video. We verify that our SWAG-Net outperforms state-of-the-art methods on Charades-STA and ActivityNet Captions datasets.

Title: Can Adversarial Training benefit Trajectory Representation?: An Investigation on Robustness for Trajectory Similarity Computation Abstract: ABSTRACTTrajectory similarity computation as the fundamental problem for various downstream analytic tasks, such as trajectory classification and clustering, has been extensively studied in recent years. However, how to infer an accurate and robust similarity over two trajectories is difficult due to the some trajectory characteristics in practice, e.g. non-uniform sampling rate, nonmalignant fluctuation, and noise points, etc. To circumvent such challenges, we in this paper introduce the adversarial training idea into the trajectory representation learning for the first time to enhance the robustness and accuracy. Specifically, our proposed method AdvTraj2Vec has two novelties: i) it perturbs the weight parameters of embedding layers to learn a robust model to infer an accurate pairwise similarity over each two trajectories; and ii) it employs the GAN momentum to harness the perturbation extent to which an appropriate trajectory representation can be learned for the similarity computation. Extensive experiments using two real-world trajectory datasets Porto and Beijing validate our proposed AdvTraj2Vec on the robustness and accuracy aspects. The multi-facet results show that our AdvTraj2Vec significantly outperforms the stat-of-the-art methods in terms of different distortions, such as trajectory-point addition, deletion, disturbance, and outlier injection.

Title: Extracting Drug-drug Interactions from Biomedical Texts using Knowledge Graph Embeddings and Multi-focal Loss Abstract: ABSTRACTThe field of Drug-drug interaction (DDI) aims to detect descriptions of interactions between drugs from biomedical texts. Currently, researchers have extracted DDIs using pre-trained language models such as BERT, which often misclassify two kinds of DDI types, "Effect" and "Int", on the DDIExtraction 2013 corpus because of highly similar expressions. The use of knowledge graphs can alleviate this problem by incorporating different relationships for each, thus allowing them to be distinguished. Thus, we propose a novel framework to integrate the neural network with a knowledge graph, where the features from these components are complementary. Specifically, we take text features at different levels into account in the neural network part. This is done by firstly obtaining a word-level position feature using PubMedBERT together with a convolution neural network, secondly, getting a phrase-level key path feature using a dependency parsing tree, thirdly, using PubMedBERT with an attention mechanism to obtain a sentence-level language feature, and finally, fusing these three kinds of representation into a synthesized feature. We also extract a knowledge feature from a drug knowledge graph which takes just a few minutes to construct, then concatenate the synthesized feature with the knowledge feature, feed the result into a multi-layer perceptron and obtain the result by a softmax classifier. In order to achieve a good integration of the synthesized feature and the knowledge feature, we train the model using a novel multifocal loss function, KGE-MFL, which is based on a knowledge graph embedding. Finally we attain state-of-the-art results on the DDIExtraction 2013 dataset (micro F-score 86.24%) and on the ChemProt dataset (micro F-score 77.75%), which proves our framework to be effective for biomedical relation extraction tasks. In particular, we fill the performance gap (more than 5.57%) between methods that rely on and do not rely on knowledge graph embedding on the DDIExtraction 2013 corpus, when predicting the "Int" type. The implementation code is available at https://github.com/NWU-IPMI/DDIE-KGE-MFL.

Title: Estimating Causal Effects on Networked Observational Data via Representation Learning Abstract: ABSTRACTIn this paper, we study the causal effects estimation problem on networked observational data. We theoretically prove that standard graph machine learning (ML) models, e.g., graph neural networks (GNNs), fail in estimating the causal effects on networks. We show that graph ML models exhibit two distribution mismatches of their objective functions compared to causal effects estimation, leading to the failure of traditional ML models. Motivated by this, we first formulate the networked causal effects estimation as a data-driven multi-task learning problem, and then propose a novel framework NetEst to conduct causal inference in the network setting. NetEst uses GNNs to learn representations for confounders, which are from both a unit's own characteristics and the network effects. The embeddings are then used to sufficiently bridge the distribution gaps via adversarial learning and estimate the observed outcomes simultaneously. Extensive experimental studies on two real-world networks with semi-synthetic data demonstrate the effectiveness of NetEst. We also provide analyses on why and when NetEst works.

Title: Diverse Effective Relationship Exploration for Cooperative Multi-Agent Reinforcement Learning Abstract: ABSTRACTIn some complex multi-agent environments, the types of relationships between agents are diverse and their intensity changes during the policy learning process. Theoretically, some of these relationships can facilitate cooperative policy learning. However, acquiring these relationships is an intractable problem. To tackle the problem, we propose a diverse effective relationship exploration based multi-agent reinforcement learning (DERE) method. Specifically, a potential fields model is firstly designed to represent relationships between agents. Then to encourage the exploration of effective relationships, we define an information-theoretic objective function. Finally, an intrinsic reward function is designed to optimize the information-theoretic objective, meanwhile, guide agents to learn more effective collaborative policies. Experimental results show that our method outperforms state-of-the-art methods on both super hard StarCraft II micromanagement tasks (SMAC) and Google Research Football (GRF).

Title: An Uncertainty-Aware Imputation Framework for Alleviating the Sparsity Problem in Collaborative Filtering Abstract: ABSTRACT Collaborative Filtering (CF) methods for recommender systems commonly suffer from the data sparsity issue. Data imputation has been widely adopted to deal with this issue. However, existing studies have limitations in the sense that both uncertainty and robustness of imputation have not been taken into account, where there is a high risk that the imputed values are likely to be far from the true values. This paper explores a novel imputation framework, named Uncertainty-Aware Multiple Imputation (UA-MI), which can effectively solve the sparsity issue. Given a (sparse) user-item interaction matrix, our key idea is to quantify uncertainty on each missing entry and then the cells with the lowest uncertainty are selectively imputed. Here, we suggest three strategies for measuring uncertainty in missing user-item interactions, each of which is based on sampling, dropout, and ensemble, respectively. They successfully obtain element-wise mean and variance on the missing entries, where the variance helps determine where in the matrix should be imputed and the corresponding mean values are imputed. Experiments show that our UA-MI framework significantly outperformed the existing imputation strategies

Title: Memory Bank Augmented Long-tail Sequential Recommendation Abstract: ABSTRACTThe goal of sequential recommendation is to predict the next item that a user would like to interact with, by capturing her dynamic historical behaviors. However, most existing sequential recommendation methods do not focus on solving the long-tail item recommendation problem that is caused by the imbalanced distribution of item data. To solve this problem, we propose a novel sequential recommendation framework, named MASR (ie Memory Bank Augmented Long-tail Sequential Recommendation). MASR is an "Open-book'' model that combines novel types of memory banks and a retriever-copy network to alleviate the long-tail problem. During inference, the designed retriever-copy network retrieves related sequences from the training samples and copies the useful information as a cue to improve the recommendation performance on tail items. Two designed memory banks provide reference samples to the retriever-copy network by memorizing the historical samples appearing in the training phase. Extensive experiments have been performed on five real-world datasets to demonstrate the effectiveness of the proposed MASR model. The experimental results indicate that MASR consistently outperforms baseline methods in terms of recommendation performance on tail items.

Title: One Rating to Rule Them All?: Evidence of Multidimensionality in Human Assessment of Topic Labeling Quality Abstract: ABSTRACTTwo general approaches are common for evaluating automatically generated labels in topic modeling: direct human assessment; or performance metrics that can be calculated without, but still correlate with, human assessment. However, both approaches implicitly assume that the quality of a topic label is single-dimensional. In contrast, this paper provides evidence that human assessments about the quality of topic labels consist of multiple latent dimensions. This evidence comes from human assessments of four simple labeling techniques. For each label, study participants responded to several items asking them to assess each label according to a variety of different criteria. Exploratory factor analysis shows that these human assessments of labeling quality have a two-factor latent structure. Subsequent analysis demonstrates that this multi-item, two-factor assessment can reveal nuances that would be missed using either a single-item human assessment of perceived label quality or established performance metrics. The paper concludes by suggesting future directions for the development of human-centered approaches to evaluating NLP and ML systems more broadly.

Title: AutoMARS: Searching to Compress Multi-Modality Recommendation Systems Abstract: ABSTRACTWeb applications utilize Recommendation Systems (RS) to address the problem of consumer over-choices. Recent works have taken advantage of multi-modality or multi-view, input information (such as user interaction, images, texts, rating scores) to boost recommendation system performance compared with using single-modality information. However, the use of multi-modality input demands much higher computational cost and storage capacity. On the other hand, the real-world RS services usually have strict budgets on both time and space for a good customer experience. As a result, the model efficiency of multi-modality recommendation systems has gained increasing importance. While unfortunately, to the best of our knowledge, there is no existing study of a generic compression framework for multi-modality RS. In this paper, we investigate, for the first time, how to compress a multi-modality recommendation system with a fixed budget. Assuming that input information from different modalities are of unequal importance, a good compression algorithm should learn to automatically allocate different resource budgets to each input, based on their importance in maximally preserving recommendation efficacy. To this end, we leverage the tools of neural architecture search (NAS) and distillation and propose Auto Multi-modAlity Recommendation System (AutoMARS), a unified modality-aware model compression framework dedicated to multi-modality recommendation systems. We demonstrate the effectiveness and generality of AutoMARS by testing it on three different Amazon datasets of various sparsity. AutoMARS demonstrates superior multi-modality compression performance than previous state-of-the-art compression methods. For example on the Amazon Beauty dataset, we achieve on average a 20% higher accuracy over previous state-of-the-art methods, while enjoying 65% reduction over baselines. Codes are available at: https://github.com/VITA-Group/AutoMARS.

Title: Multi-Aggregator Time-Warping Heterogeneous Graph Neural Network for Personalized Micro-Video Recommendation Abstract: ABSTRACTMicro-video recommendation is attracting global attention and becoming a popular daily service for people of all ages. Recently, Graph Neural Networks-based micro-video recommendation has displayed performance improvement for many kinds of recommendation tasks. However, the existing works fail to fully consider the characteristics of micro-videos, such as the high timeliness of news nature micro-video recommendation and sequential interactions of frequently changed interests. In this paper, a novel Multi-aggregator Time-warping Heterogeneous Graph Neural Network (MTHGNN) is proposed for personalized news nature micro-video recommendation based on sequential sessions, where characteristics of micro-videos are comprehensively studied, users' preference is mined via multi-aggregator, the temporal and dynamic changes of users' preference are captured, and timeliness is considered. Through the comparison with the state-of-the-arts, the experimental results validate the superiority of our MTHGNN model.

Title: Bootstrap-based Causal Structure Learning Abstract: ABSTRACTLearning a causal structure from observational data is crucial for data scientists. Recent advances in causal structure learning (CSL) have focused on local-to-global learning, since the local-to-global CSL can be scaled to high-dimensional data. The local-to-global CSL algorithms first learn the local skeletons, then construct the global skeleton, and finally orient edges. In practice, the performance of local-to-global CSL mainly depends on the accuracy of the global skeleton. However, in many real-world settings, owing to inevitable data quality issues (e.g. noise and small sample), existing local-to-global CSL methods often yield many asymmetric edges (e.g., given anasymmetric edge containing variables A and B, the learned skeleton of A contains B, but the learned skeleton of B does not contain A), which make it difficult to construct a high quality global skeleton. To tackle this problem, this paper proposes a Bootstrap sampling based Causal Structure Learning (BCSL) algorithm. The novel contribution of BCSL is that it proposes an integrated global skeleton learning strategy that can construct more accurate global skeletons. Specifically, this strategy first utilizes the Bootstrap method to generate multiple sub-datasets, then learns the local skeleton of variables on each asymmetric edge on those sub-datasets, and finally designs a novel scoring function to estimate the learning results on all sub-datasets for correcting the asymmetric edge. Extensive experiments on both benchmark and real datasets verify the effectiveness of the proposed method.

Title: KiCi: A Knowledge Importance Based Class Incremental Learning Method for Wearable Activity Recognition Abstract: ABSTRACTWearable-based human activity recognition (HAR) is commonly employed in real-world scenarios such as health monitoring, auxiliary diagnosis, etc. As implementing activity recognition is a daunting challenge in an open dynamic environment, incremental learning has become a common method to adapt to variable behavior patterns of users and create dynamic modeling in activity recognition. However, catastrophic forgetting is a significant challenge with incremental learning. This is contrary to our expectations of identifying new activity classes while remembering existing ones. To address this problem, we propose a knowledge importance-based class incremental learning method called KiCi and construct an incremental learning model based on the framework of self-iterative knowledge distillation for dynamic activity recognition. To eliminate the prediction bias of the teacher model on the old knowledge, we utilize the trained weights of previous incremental steps generated by the teacher model as the prior knowledge to obtain knowledge importance. Then use it to make the student model have a reasonable trade-off between old and new knowledge and mitigate catastrophic forgetting by avoiding negative transfer. We conduct extensive experiments on four public HAR datasets and our method consistently outperforms the existing state-of-the-art methods by a large margin.

Title: Learning Hypersphere for Few-shot Anomaly Detection on Attributed Networks Abstract: ABSTRACTThe existence of anomalies is quite common, but they are hidden within the complex structure and high-dimensional node attributes of the attributed networks. As a latent hazard in existing systems, anomalies can be transformed into important instruction information once we detect them, e.g., computer network admins can react to the leakage of sensitive data if network traffic anomalies are identified. Extensive research in anomaly detection on attributed networks has proposed various techniques, which do improve the quality of data in networks, while they rarely cope with the few-shot anomaly detection problem. Few-shot anomaly detection task with only a few dozen labeled anomalies is more practical since anomalies are rare in number for real-world systems. We propose a few-shot anomaly detection approach for detecting the anomaly nodes that significantly deviate from the vast majority. Our approach, based on an extension of model-agnostic meta-learning(MAML), is a Learnable Hypersphere Meta-Learning method running on local subgraphs named LHML. LHML learns on a single subgraph, conducts meta-learning on a set of subgraphs, and maintains the radius of a learnable hypersphere across subgraphs to detect anomalies efficiently. The learnable hypersphere is a changing boundary that can be used to identify anomalies in an unbalanced binary-classification setting and quickly adapt to a new subgraph by a few gradient updating steps of MAML. Furthermore, our model runs across subgraphs, making it possible to identify an anomaly without requiring the whole graph nodes as is usually the way but only a handful of nodes around it, which means LHML can scale to large networks. Experimental results show the effective performance of LHML on benchmark datasets.

Title: Gromov-Wasserstein Multi-modal Alignment and Clustering Abstract: ABSTRACTMulti-modal clustering aims at finding a clustering structure shared by the data of different modalities in an unsupervised way. Currently, solving this problem often relies on two assumptions: i) the multi-modal data own the same latent distribution, and ii) the observed multi-modal data are well-aligned and without any missing modalities. Unfortunately, these two assumptions are often questionable in practice and thus limit the feasibility of many multi-modal clustering methods. In this work, we develop a new multi-modal clustering method based on the Gromovization of optimal transport distance, which relaxes the dependence on the above two assumptions. In particular, given the data of different modalities, whose correspondence is unknown, our method learns the Gromov-Wasserstein (GW) barycenter of their kernel matrices. Driven by the modularity maximization principle, the GW barycenter helps to explore the clustering structure shared by different modalities. Moreover, the GW barycenter is associated with the GW distances between the different modalities to the clusters, and the optimal transport plans corresponding to the GW distances help to achieve the alignment and the clustering of the multi-modal data jointly. Experimental results show that our method outperforms state-of-the-art multi-modal clustering methods, especially when the data are (partially or completely) unaligned. The code is available at https://github.com/rucnyz/GWMAC.

Title: Spatio-temporal Trajectory Learning using Simulation Systems Abstract: ABSTRACTSpatio-temporal trajectories are essential factors for systems used in public transport, social ecology, and many other disciplines where movement is a relevant dynamic process. Each trajectory describes multiple state changes over time, induced by individual decision-making, based on psychological and social factors with physical constraints. Since a crucial factor of such systems is to reason about the potential trajectories in a closed environment, the primary problem is the realistic replication of individual decision making. Mental factors are often uncertain, not available or cannot be observed in reality. Thus, models for data generation must be derived from abstract studies using probabilities. To solve these problems, we present Multi-Agent-Trajectory-Learning (MATL), a state transition model to learn and generate human-like Spatio-temporal trajectory data. MATL combines Generative Adversarial Imitation Learning (GAIL) with a simulation system that uses constraints given by an agent-based model (Aℬℳ). We use GAIL to learn policies in conjunction with the Aℬℳ, resulting in a novel concept of individual decision making. Experiments with standard trajectory predictions show that our approach produces similar results to real-world observations.

Title: Modeling Dynamic Heterogeneous Graph and Node Importance for Future Citation Prediction Abstract: ABSTRACTAccurate citation count prediction of newly published papers could help editors and readers rapidly figure out the influential papers in the future. Though many approaches are proposed to predict a paper's future citation, most ignore the dynamic heterogeneous graph structure or node importance in academic networks. To cope with this problem, we propose a Dynamic heterogeneous Graph and Node Importance network (DGNI) learning framework, which fully leverages the dynamic heterogeneous graph and node importance information to predict future citation trends of newly published papers. First, a dynamic heterogeneous network embedding module is provided to capture the dynamic evolutionary trends of the whole academic network. Then, a node importance embedding module is proposed to capture the global consistency relationship to figure out each paper's node importance. Finally, the dynamic evolutionary trend embeddings and node importance embeddings calculated above are combined to jointly predict the future citation counts of each paper, by a log-normal distribution model according to multi-faced paper node representations. Extensive experiments on two large-scale datasets demonstrate that our model significantly improves all indicators compared to the SOTA models.

Title: PromptORE - A Novel Approach Towards Fully Unsupervised Relation Extraction Abstract: ABSTRACTUnsupervised Relation Extraction (RE) aims to identify relations between entities in text, without having access to labeled data during training. This setting is particularly relevant for domain specific RE where no annotated dataset is available and for open-domain RE where the types of relations are a priori unknown. Although recent approaches achieve promising results, they heavily depend on hyperparameters whose tuning would most often require labeled data. To mitigate the reliance on hyperparameters, we propose PromptORE, a "Prompt-based Open Relation Extraction" model. We adapt the novel prompt-tuning paradigm to work in an unsupervised setting, and use it to embed sentences expressing a relation. We then cluster these embeddings to discover candidate relations, and we experiment different strategies to automatically estimate an adequate number of clusters. To the best of our knowledge, PromptORE is the first unsupervised RE model that does not need hyperparameter tuning. Results on three general and specific domain datasets show that PromptORE consistently outperforms state-of-the-art models with a relative gain of more than 40% in B3, V-measure and ARI. Qualitative analysis also indicates PromptORE's ability to identify semantically coherent clusters that are very close to true relations.

Title: DP-HORUS: Differentially Private Hierarchical Count Histograms under Untrusted Server Abstract: ABSTRACTHierarchical count histograms is the task of publishing count statistics at different granularity as per hierarchy defined on a dimension table in a data warehouse, which has wide applications in On-line Analytical Processing (OLAP) scenarios. In this paper, we systematically investigate this task subjected to the rigorous privacy-preserving constraint under the untrusted server setting. Our study first reveals that the straightforward baseline approach of the local differential privacy fails to achieve a satisfactory privacy and utility tradeoff. We are thus motivated to propose DP-HORUS, a novel crypto-assisted Differentially Private framework for Hierarchical cOunt histogRams under Untrusted Server. DP-HORUS consists of a series of novel designs, including 1) Encrypted Hierarchical Tree (EHT) structure, which maintains the concept hierarchy in the input data; 2) Random Matrix (RM), which reduces communication and computational cost; 3) To further boosted the utility, we propose DP-HORUS+ encompassing two additional modules of Histograms Structure (HS) and Hierarchical Consistency (HC), which are respectively introduced to reduce the noise caused by data sparsity and to ensure the hierarchy consistency. We provide both theoretical analysis and extensive empirical study on both real-world and synthetic datasets, which demonstrates the superior utility of the proposed methods over the state-of-the-art solutions while ensuring strict privacy guarantee.

Title: MGMAE: Molecular Representation Learning by Reconstructing Heterogeneous Graphs with A High Mask Ratio Abstract: ABSTRACTMasked autoencoder (MAE), as an effective self-supervised learner for computer vision and natural language processing, has been recently applied to molecule representation learning. In this paper, we identify two issues in applying MAE to pre-train Transformer-based models on molecular graphs that existing works have ignored. (1) As only atoms are abstracted as tokens and then reconstructed, the chemical bonds are not decided in the decoded molecule, making molecules with different arrangements of the same atoms indistinguishable. (2) Although a high mask ratio that corresponds to a challenging reconstruction task has been proved beneficial in the vision domain, it cannot be trivially leveraged on molecular graphs as there is less redundancy of information in graph data. To resolve these issues, we propose a novel framework, Molecular Graph Mask AutoEncoder (MGMAE). As the first step in MGMAE, we transform each molecular graph into a heterogeneous atom-bond graph to fully use the bond attributes and design unidirectional position encoding for such graphs. Then we propose a hybrid masking mechanism that exploits the complementary nature between atoms' attributive and spatial features. Meanwhile, we compensate for the mask embedding by a dynamic aggregation representation that exploits the correlations between topologically adjacent tokens. As a result, MGMAE can reconstruct the masked atoms, the masked bonds, and the relative distance among atoms simultaneously, with a high mask ratio. We compare MGMAE with the state-of-the-art methods on various molecular benchmarks and show the competitiveness of MGMAE in both regression and classification tasks.

Title: Aries: Accurate Metric-based Representation Learning for Fast Top-k Trajectory Similarity Query Abstract: ABSTRACTWith the prevalence of location-based services (LBS), trajectories are being generated rapidly. As is widely used in LBS, top-k trajectory similarity query serves as a key operation, deeply empowering applications such as travel route recommendation and carpooling. Given the rise of deep learning, trajectory representation has been well-proven to speed up this operator. However, existing representation-based computing modes remain two major problems understudied: the low quality of trajectory representation and insufficient support for various trajectory similarity metrics, which make them difficult to apply in practice. Therefore, we propose an Accurate metric-based representation learning approach for fast top-k trajectory similarity query, named Aries. Specifically, Aries has two sophisticated modules: (1) An novel trajectory embedding strategy enhanced by the bidirectional LSTM encoder and spatial attention mechanism, which can extract more precise and comprehensive knowledge. (2) A deep metric learning network aggregating multiple measures for better top-k query. Extensive experiments conducted on real trajectory dataset show that Aries achieves both impressive accuracy and lower training time compared with state-of-the-art solutions. In particular, it achieves 5x-10x speedup and 10%-20% accuracy improvement over Euclidean, Hausdorff, DTW, and EDR measures. Besides, our method can maintain stable performance when handling various scenarios, without repeated training in order to adapt to diverse similarity metrics.

Title: Few-Shot Relational Triple Extraction with Perspective Transfer Network Abstract: ABSTRACTFew-shot Relational Triple Extraction (RTE) aims at detecting emerging relation types along with their entity pairs from unstructured text with the support of a few labeled samples. Prior arts use conditional random field or nearest-neighbor matching strategy to extract entities and use prototypical networks for extracting relations from sentences. Nevertheless, they fail to utilize the triple-level information to verify the plausibility of extracted relational triples, and ignore the proper transfer among the perspectives of entity, relation and triple. To fill in these gaps, in this work, we put forward a novel perspective transfer network (PTN) to address few-shot RTE. Specifically, PTN starts from the relation perspective by checking the existence of a given relation. Then, it transfers to the entity perspective to locate entity spans with relation-specific support sets. Next, it transfers to the triple perspective to validate the plausibility of extracted relational triples. Finally, it transfers back to the relation perspective to check the next relation, and repeats the aforementioned procedure. By transferring among the perspectives of relation, entity, and triple, PTN not only validates the extracted elements at both local and global levels, but also effectively handles more realistic and difficult few-shot RTE scenarios such as multiple triple extraction and nonexistence of triples. Extensive experimental results on existing dataset and new datasets demonstrate that our approach can significantly improve performance over the state-of-the-arts.

Title: MonitorLight: Reinforcement Learning-based Traffic Signal Control Using Mixed Pressure Monitoring Abstract: ABSTRACTAlthough Reinforcement Learning (RL) has achieved significant success in the Traffic Signal Control (TSC), most of them focus on the design of RL elements while the impact of the phase duration is neglected. Due to the lack of exploring dynamic phase duration, the overall performance and convergence rate of RL-based TSC approaches cannot be guaranteed, which may result in poor adaptability of RL methods to different traffic conditions. To address these issues, in this paper, we formulate a novel phase-duration-aware TSC (PDA-TSC) problem and propose an effective RL-based TSC approach, named MonitorLight. Our approach adopts a new traffic indicator, mixed pressure, which enables RL agents to simultaneously analyze the impacts of stationary and moving vehicles on intersections. Based on the observed mixed pressure of intersections, RL agents can autonomously determine whether or not to change the current signals in real-time. In addition, MonitorLight can adjust the control method for scenarios with different real-time requirements and achieve excellent results in different situations. Extensive experiments on both real-world and synthetic datasets demonstrate that MonitorLight outperforms the current state-of-the-art IPDALight by up to 2.84% and 5.71% in average vehicle travel time, respectively. Moreover, our method significantly speeds up the convergence, leading IPDALight by 36.87% and 34.58% in the start to converge episode and jumpstart performance, respectively.

Title: Smart Contract Scams Detection with Topological Data Analysis on Account Interaction Abstract: ABSTRACTThe skyrocketing market value of cryptocurrencies has prompted more investors to pour funds into cryptocurrencies to seek asset hedging. However, the anonymity of blockchain makes cryptocurrency naturally a tool of choice for criminals to commit smart contract scams. Consequently, smart contract scam detection is particularly critical for investors to avoid economic loss. Previous methods mainly leverage specific code logic of smart contracts and/or design rules based on abnormal transaction behaviors for scam detection. Although these methods gain success at detecting particular scams, they perform worse when applied to scams with highly similar codes. Besides, well-designed decision rules rely on expert knowledge and tedious data collection steps, which causes poor flexibility. To combat these challenges, we consider the problem of smart contract scam detection via mining topological features of account interaction information that dynamically evolves. We adopt interactive features extracted from dynamic interaction information of accounts and propose a framework named TTG-SCSD to utilize the features and Topological Data Analysis for smart contract scams detection. The TTG-SCSD constructs discrete dynamic interaction graphs for each contract and designs interactive features that characterize account behaviors. The features are modeled combined with a topology quantification mechanism to capture contract intentions in transactions. Experimental results on real-world transaction datasets from Ethereum show that TTG-SCSD obtains better generalizability and improves the performance of the bare versions of the comparison methods.

Title: Detecting Significant Differences Between Information Retrieval Systems via Generalized Linear Models Abstract: ABSTRACTBeing able to compare Information Retrieval(IR) systems correctly is pivotal to improving their quality. Among the most popular tools for statistical significance testing, we list t-test and ANOVA that belong to the linear models family. Therefore, given the relevance of linear models for IR evaluation, a great effort has been devoted to studying how to improve them to better compare IR systems. Linear models rely on assumptions that IR experimental observations rarely meet, e.g. about the normality of the data or the linearity itself. Even though linear models are, in general, resilient to violations of their assumptions, departing from them might reduce the effectiveness of the tests. Hence, we investigate the use of the Generalized Linear Models (GLM) framework, a generalization of the traditional linear modelling that relaxes assumptions about the distribution and the shape of the models. To the best of our knowledge, there has been little or no investigation on the use of GLMs for comparing IR system performance. We discuss how GLM work and how they can be applied in the context of IR evaluation. In particular, we focus on the link function used to build GLMs, which allows for the model to have non-linear shapes. We conduct thorough experimentation using two TREC collections and several evaluation measures. Overall, we show how the log and logit links are able to identify more and more consistent significant differences (up to 25% more with 50 topics) than the identity link used today and with a comparable, or slightly better, risk of publication bias.

Title: Federated K-Private Set Intersection Abstract: ABSTRACTPrivate set intersection (PSI) is a popular protocol that allows multiple parties to evaluate the intersection of their sets without revealing them to each other. PSI has numerous practical applications, including privacy preserving data mining and location-based services. In this work, we develop a new approach for the PSI problem within the federated analytics framework. In particular, we consider a setting where a server wants to determine (query) which among its local set of data identifiers appears coupled with the same value in at least K of the N parties. Applications for this framework include but are not limited to: double-filing insurance verification, credit scoring and password checkup on an institutional level. To address the proposed setting, we propose a new protocol Fed-K-PSI that allows the server to answer this query while being oblivious to the data of identifiers that do not satisfy the distributed query at the parties. In addition, Fed-K-PSI also maintains the anonymity of the parties by hiding which K parties satisfied the query, or which value associated with the identifier which caused the query to be successful. Our proposed setting does not lend itself directly to state-of-the-art approaches in PSI based on Oblivious Transfer, since the server does not have a complete representation of a datapoint (only the identifier, but no value). Our proposed approach tackles this problem by constructing a distributed function at the parties, which encodes the datapoints and returns a deterministic known property if and only if the value for a given identifier is the same in at least K of the N parties. We show that Fed-K-PSI achieves a strong information-theoretic privacy guarantee and is resilient to collusion scenarios among honest-but-curious parties. We also evaluate Fed-K-PSI via extensive experiments to study the effect of the different system parameters.

Title: GDOD: Effective Gradient Descent using Orthogonal Decomposition for Multi-Task Learning Abstract: ABSTRACTMulti-task learning (MTL) aims at solving multiple related tasks simultaneously and has experienced rapid growth in recent years. However, MTL models often suffer from performance degeneration with negative transfer due to learning several tasks simultaneously. Some related work attributed the source of the problem is the conflicting gradients. In this case, it is needed to select useful gradient updates for all tasks carefully. To this end, we propose a novel optimization approach for MTL, named GDOD, which manipulates gradients of each task using an orthogonal basis decomposed from the span of all task gradients. GDOD decomposes gradients into task-shared and task-conflict components explicitly and adopts a general update rule for avoiding interference across all task gradients. This allows guiding the update directions depending on the task-shared components. Moreover, we prove the convergence of GDOD theoretically under both convex and non-convex assumptions. Experiment results on several multi-task datasets not only demonstrate the significant improvement of GDOD performed to existing MTL models but also prove that our algorithm outperforms state-of-the-art optimization methods in terms of AUC and Logloss metrics.

Title: Weakly-Supervised Online Hashing with Refined Pseudo Tags Abstract: ABSTRACTWith the rapid development of social media, various types of tags uploaded by social users are attached to the images. Compared to clean labels marked by experts, although user-provided tags are imperfect, e.g., wrong tags, reduplicative tags, or missing tags, they are more diverse, fine-grained, and informative. Currently, there exist several weakly-supervised hashing methods attempting to learn hash codes using tags as supervision. Although they could benefiting from the rich information contained in tags, most of them may defy the nature of social media data. In real scenarios, social media data appears in streaming fashion, but most weakly-supervised hashing methods are just batch-based which cannot effectively handle streaming data. To this end, only one weakly-supervised online hashing method has been proposed, but it is still far from enough to alleviate the negative effects of tags. In this paper, to address the above problems, we propose a new method, termed Weakly-Supervised Online Hashing with Refined Pseudo Tags (RPT-WOH). To improve the quality of weakly-supervised tags, we design the real-valued pseudo tag matrix and learn it by exploiting the correlation between the previous and new tags. Furthermore, we propose a memory-based similarity learning which could effectively maintain the semantic correlation between old and new data. In addition, we propose an effective and efficient discrete online optimization algorithm making RPT-WOH easily scalable to large-scale data. Extensive experiments conducted on two benchmark datasets demonstrate that RPT-WOH offers satisfactory performance.

Title: Efficient Trajectory Similarity Computation with Contrastive Learning Abstract: ABSTRACTThe ubiquity of mobile devices and the accompanying deployment of sensing technologies have resulted in a massive amount of trajectory data. One important fundamental task is trajectory similarity computation, which is to determine how similar two trajectories are. To enable effective and efficient trajectory similarity computation, we propose a novel robust model, namely Contrastive Learning based Trajectory Similarity Computation (CL-TSim). Specifically, we employ a contrastive learning mechanism to learn the latent representations of trajectories and then calculate the dissimilarity between trajectories based on these representations. Compared with sequential auto-encoders that are the mainstream deep learning architectures for trajectory similarity computation, CL-TSim does not require a decoder and step-by-step reconstruction, thus improving the training efficiency significantly. Moreover, considering the non-uniform sampling rate and noisy points in trajectories, we adopt two type of augmentations, i.e., point dowm-sampling and point distorting, to enhance the robustness of the proposed model. Extensive experiments are conducted on two widely-used real-world datasets, i.e., Porto and ChengDu, which demonstrate the superior effectiveness and efficiency of the proposed model.

Title: When Should We Use Linear Explanations? Abstract: ABSTRACTThe increasing interest in transparent and fair AI systems has propelled the research in explainable AI (XAI). One of the main research lines in XAI is post-hoc explainability, the task of explaining the logic of an already deployed black-box model. This is usually achieved by learning an interpretable surrogate function that approximates the black box. Among the existing explanation paradigms, local linear explanations are one of the most popular due to their simplicity and fidelity. Despite their advantages, linear surrogates may not always be the most adapted method to produce reliable, i.e., unambiguous and faithful explanations. Hence, this paper introduces Adapted Post-hoc Explanations (APE), a novel method that characterizes the decision boundary of a black-box classifier and identifies when a linear model constitutes a reliable explanation. Besides, characterizing the black-box frontier allows us to provide complementary counterfactual explanations. Our experimental evaluation shows that APE identifies accurately the situations where linear surrogates are suitable while also providing meaningful counterfactual explanations.

Title: Scaling Up Maximal k-plex Enumeration Abstract: ABSTRACTFinding all maximal k-plexes on networks is a fundamental research problem in graph analysis due to many important applications, such as community detection, biological graph analysis, and so on. A k-plex is a subgraph in which every vertex is adjacent to all but at most k vertices within the subgraph. In this paper, we study the problem of enumerating all large maximal k-plexes of a graph and develop several new and efficient techniques to solve the problem. Specifically, we first propose several novel upper-bounding techniques to prune unnecessary computations during the enumeration procedure. We show that the proposed upper bounds can be computed in linear time. Then, we develop a new branch-and-bound algorithm with a carefully-designed pivot re-selection strategy to enumerate all k-plexes, which outputs all k-plexes in O(n2?k n) time theoretically, where n is the number of vertices of the graph and ? k is strictly smaller than 2. In addition, a parallel version of the proposed algorithm is further developed to scale up to process large real-world graphs. Finally, extensive experimental results show that the proposed sequential algorithm can achieve up to 2× to 100× speedup over the state-of-the-art sequential algorithms on most benchmark graphs. The results also demonstrate the high scalability of the proposed parallel algorithm. For example, on a large real-world graph with more than 200 million edges, our parallel algorithm can finish the computation within two minutes, while the state-of-the-art parallel algorithm cannot terminate within 24 hours.

Title: Meta-Path-based Fake News Detection Leveraging Multi-level Social Context Information Abstract: ABSTRACTFake news, false or misleading information presented as news, has a significant impact on many aspects of society, such as in politics or healthcare domains. Due to the deceiving nature of fake news, applying Natural Language Processing (NLP) techniques to the news content alone is insufficient. Therefore, more information is required to improve fake news detection, such as the multi-level social context (news publishers and engaged users in social media) information and the temporal information of user engagement. The proper usage of this information, however, introduces three chronic difficulties: 1) multi-level social context information is hard to be used without information loss, 2) temporal information of user engagement is hard to be used along with multi-level social context information, and 3) news representation with multi-level social context and temporal information is hard to be learned in an end-to-end manner. To overcome all three difficulties, we propose a novel fake news detection framework, Hetero-SCAN. We use Meta-Path, a composite relation connecting two node types, to extract meaningful multi-level social context information without loss. We then propose Meta-Path instance encoding and aggregation methods to capture the temporal information of user engagement and learn news representation end-to-end. According to our experiment, Hetero-SCAN yields significant performance improvement over state-of-the-art fake news detection methods.

Title: An Empirical Study on How People Perceive AI-generated Music Abstract: ABSTRACTMusic creation is difficult because one must express one's creativity while following strict rules. The advancement of deep learning technologies has diversified the methods to automate complex processes and express creativity in music composition. However, prior research has not paid much attention to exploring the audiences' subjective satisfaction to improve music generation models. In this paper, we evaluate human satisfaction with the state-of-the-art automatic symbolic music generation models using deep learning. In doing so, we define a taxonomy for music generation models and suggest nine subjective evaluation metrics. Through an evaluation study, we obtained more than 700 evaluations from 100 participants, using the suggested metrics. Our evaluation study reveals that the token representation method and models' characteristics affect subjective satisfaction. Through our qualitative analysis, we deepen our understanding of AI-generated music and suggested evaluation metrics. Lastly, we present lessons learned and discuss future research directions of deep learning models for music creation.

Title: Finding Heterophilic Neighbors via Confidence-based Subgraph Matching for Semi-supervised Node Classification Abstract: ABSTRACTGraph Neural Networks (GNNs) have proven to be powerful in many graph-based applications. However, they fail to generalize well under heterophilic setups, where neighbor nodes have different labels. To address this challenge, we employ a confidence ratio as a hyper-parameter, assuming that some of the edges are disassortative (heterophilic). Here, we propose a two-phased algorithm. Firstly, we determine edge coefficients through subgraph matching using a supplementary module. Then, we apply GNNs with a modified label propagation mechanism to utilize the edge coefficients effectively. Specifically, our supplementary module identifies a certain proportion of task-irrelevant edges based on a given confidence ratio. Using the remaining edges, we employ the widely used optimal transport to measure the similarity between two nodes with their subgraphs. Finally, using the coefficients as supplementary information on GNNs, we improve the label propagation mechanism which can prevent two nodes with smaller weights from being closer. The experiments on benchmark datasets show that our model alleviates over-smoothing and improves performance.

Title: Explainable Link Prediction in Knowledge Hypergraphs Abstract: ABSTRACTLink prediction in knowledge hypergraphs has been recognized as a critical issue in various downstream tasks for knowledge-enabled applications, from question answering to recommender systems. However, most existing approaches are primarily performed in a black-box fashion, which learn low-dimensional embeddings for inference, thus cannot provide human-understandable interpretation. In this paper, we present HyperMLN, an n-ary, mixed, and explainable framework that interprets the path-reasoning process with first-order logic, which provides a knowledge-enhanced interpretable prediction framework, in which domain knowledge in the logic rules improves the performance of embedding models, while semantic information in the embedding space can optimize the weight of the logic rules in turn. To provide benchmark rule sets for explainable link prediction methods, three types of meta-logic rules in each popular dataset are mined for interpreting results. While achieving explainability, our framework also realizes an average improvement of 3.2% on [email protected] compared to the state-of-the-art knowledge hypergraph embedding method. Our code is available at https://github.com/zirui-chen/HyperMLN.

Title: ReLAX: Reinforcement Learning Agent Explainer for Arbitrary Predictive Models Abstract: ABSTRACTCounterfactual examples (CFs) are one of the most popular methods for attaching post-hoc explanations to machine learning (ML) models. However, existing CF generation methods either exploit the internals of specific models or depend on each sample's neighborhood, thus they are hard to generalize for complex models and inefficient for large datasets. This work aims to overcome these limitations and introduces ReLAX, a model-agnostic algorithm to generate optimal counterfactual explanations. Specifically, we formulate the problem of crafting CFs as a sequential decision-making task and then find the optimal CFs via deep reinforcement learning (DRL) with discrete-continuous hybrid action space. Extensive experiments conducted on several tabular datasets have shown that ReLAX outperforms existing CF generation baselines, as it produces sparser counterfactuals, is more scalable to complex target models to explain, and generalizes to both classification and regression tasks. Finally, to demonstrate the usefulness of our method in a real-world use case, we leverage CFs generated by ReLAX to suggest actions that a country should take to reduce the risk of mortality due to COVID-19. Interestingly enough, the actions recommended by our method correspond to the strategies that many countries have actually implemented to counter the COVID-19 pandemic.

Title: Task Publication Time Recommendation in Spatial Crowdsourcing Abstract: ABSTRACTThe increasing proliferation of networked and geo-positioned mobile devices brings about increased opportunities for Spatial Crowdsourcing (SC), which aims to enable effective location-based task assignment. We propose and study a novel SC framework, namely Task Assignment with Task Publication Time Recommendation. The framework consists of two phases, task publication time recommendation and task assignment. More specifically, the task publication time recommendation phase hybrids different learning models to recommend the suitable publication time for each task to ensure the timely task assignment and completion while reducing the waiting time of the task requester at the SC platform. We use a cross-graph neural network to learn the representations of task requesters by integrating the obtained representations from two semantic spaces and utilize the self-attention mechanism to learn the representations of task-publishing sequences from multiple perspectives. Then a fully connected layer is used to predict suitable task publication time based on the obtained representations. In the task assignment phase, we propose a greedy and a minimum cost maximum flow algorithm to achieve the efficient and the optimal task assignment, respectively. An extensive empirical study demonstrates the effectiveness and efficiency of our framework.

Title: GCF-RD: A Graph-based Contrastive Framework for Semi-Supervised Learning on Relational Databases Abstract: ABSTRACTRelational databases are the main storage model of structured data in most businesses, which usually involves multiple tables with key-foreign-key relationships. In practice, data analysts often want to pose predictive classification queries over relational databases. To answer such queries, many existing approaches perform supervised learning to train classification models, which heavily rely on the availability of sufficient labeled data. In this paper, we propose a novel graph-based contrastive framework for semi-supervised learning on relational databases, achieving promising predictive classification performance with only a handful of labeled data. Our framework utilizes contrastive learning to exploit additional supervision signals from massive unlabeled data. Specifically, we develop two contrastive graph views that are 1) advantageous for modeling complex relationships and correlations among structured data in a relational database, and 2) complementary to each other for learning robust representations of structured data to be classified. We also leverage label information in contrastive learning to mitigate its negative effect in knowledge transfer on the supervised counterpart. We conduct extensive experiments on three real-world relational databases and the results demonstrate that our framework is able to achieve the state-of-the-art predictive performance in limited labeled data settings, compared with various supervised and semi-supervised learning approaches.

Title: Towards Self-supervised Learning on Graphs with Heterophily Abstract: ABSTRACTRecently emerged heterophilous graph neural networks have significantly reduced the reliance on the assumption of graph homophily where linked nodes have similar features and labels. These methods focus on a supervised setting that relies on labeling information heavily and presents the limitations on general graph downstream tasks. In this work, we propose a self-supervised representation learning paradigm on graphs with heterophily (namely HGRL) for improving the generalizability of node representations, where node representations are optimized without any label guidance. Inspired by the designs of existing heterophilous graph neural networks, HGRL learns the node representations by preserving the node original features and capturing informative distant neighbors. Such two properties are obtained through carefully designed pretext tasks that are optimized based on estimated high-order mutual information. Theoretical analysis interprets the connections between HGRL and existing advanced graph neural network designs. Extensive experiments on different downstream tasks demonstrate the effectiveness of the proposed framework.

Title: Learning to Generalize in Heterogeneous Federated Networks Abstract: ABSTRACTWith the rapid development of the Internet of Things (IoT), the need to expand the amount of data through data-sharing to improve the model performance of edge devices has become increasingly compelling. To effectively protect data privacy while leveraging data across silos, federated learning has emerged. However, in the real world applications, federated learning inevitably faeces both data and model heterogeneity challenges. To address the heterogeneity issues in federated networks, in this work, we seek to jointly learn a global feature representation that is robust across clients and potentially also generalizable to new clients. More specifically, we propose a personalized Federated optimization framework with Meta Critic (FedMC) that efficiently captures robust and generalizable domain-invariant knowledge across clients. Extensive experiments on four public datasets show that the proposed FedMC outperforms the competing state-of-the-art methods in heterogeneous federated learning settings. We have also performed detailed ablation analysis on the importance of different components of the proposed model.

Title: User Recommendation in Social Metaverse with VR Abstract: ABSTRACTSocial metaverse with VR has been viewed as a paradigm shift for social media. However, most traditional VR social platforms ignore emerging characteristics in a metaverse, thereby failing to boost user satisfaction. In this paper, we explore a scenario of socializing in metaverse with VR, which brings major advantages over conventional social media: 1) leverage flexible display of users' 360-degree viewports to satisfy individual user interests, 2) ensure the user feelings of co-existence, 3) prevent view obstruction to help users find friends in crowds, and 4) support socializing with digital twins. Therefore, we formulate the Co-presence, and Occlusion-aware Metaverse User Recommendation (COMUR) problem to recommend a set of rendered players for users in social metaverse with VR. We prove COMUR is an NP-hard optimization problem and design a dual-module deep graph learning framework (COMURNet) to recommend appropriate users for viewport display. Experimental results on real social metaverse datasets and a user study with Occulus Quest 2 manifest that the proposed model outperforms baseline approaches by at least 36.7% of solution quality.

Title: Contrastive Cross-Domain Sequential Recommendation Abstract: ABSTRACTCross-Domain Sequential Recommendation (CDSR) aims to predict future interactions based on user's historical sequential interactions from multiple domains. Generally, a key challenge of CDSR is how to mine precise cross-domain user preference based on the intra-sequence and inter-sequence item interactions. Existing works first learn single-domain user preference only with intra-sequence item interactions, and then build a transferring module to obtain cross-domain user preference. However, such a pipeline and implicit solution can be severely limited by the bottleneck of the designed transferring module, and ignores to consider inter-sequence item relationships. In this paper, we propose C2DSR to tackle the above problems to capture precise user preferences. The main idea is to simultaneously leverage the intra- and inter- sequence item relationships, and jointly learn the single- and cross- domain user preferences. Specifically, we first utilize a graph neural network to mine inter-sequence item collaborative relationship, and then exploit sequential attentive encoder to capture intra-sequence item sequential relationship. Based on them, we devise two different sequential training objectives to obtain user single-domain and cross-domain representations. Furthermore, we present a novel contrastive cross-domain infomax objective to enhance the correlation between single- and cross- domain user representations by maximizing their mutual information. Additionally, we point out a serious information leak issue in prior datasets. We correct this issue and release the corrected datasets. Extensive experiments demonstrate the effectiveness of our approach C2DSR.

Title: Imitation Learning to Outperform Demonstrators by Directly Extrapolating Demonstrations Abstract: ABSTRACTWe consider the problem of imitation learning from suboptimal demonstrations that aims to learn a better policy than demonstrators. Previous methods usually learn a reward function to encode the underlying intention of the demonstrators and use standard reinforcement learning to learn a policy based on this reward function. Such methods can fail to control the distribution shift between demonstrations and the learned policy since the learned reward function may not generalize well on out-of-distribution samples and can mislead the agent to highly uncertain states, resulting in degenerated performance. To address this limitation, we propose a novel algorithm called Outperforming demonstrators by Directly Extrapolating Demonstrations(ODED). Instead of learning a reward function, ODED trains an ensemble of extrapolation networks that generate extrapolated demonstrations, i.e., demonstrations that may be induced by a good agent, based on provided demonstrations. With these extrapolated demonstrations, we can use an off-the-shelf imitation learning algorithm to learn a good policy. Guided by extrapolated demonstrations, the learned policy avoids visiting highly uncertain states and therefore controls the distribution shift. Empirically, we show that ODED outperforms suboptimal demonstrators and achieves better performance than state-of-the-art imitation learning algorithms on the MuJoCo and DeepMind Control Suite tasks.

Title: Memory Graph with Message Rehearsal for Multi-Turn Dialogue Generation Abstract: ABSTRACTMulti-turn dialogue system has attracted increasing attention in both academic and industry community. Multi-turn dialogue generation task is a challenging work as the relations among words, utterances and external knowledge are extremely complex. However, the existing methods only focus on constructing the relations between current utterance and historical utterances, and they even oversimplify the relation mining process. Moreover, with the accumulation of dialogue information, the deep semantic information is difficult to understand so that it needs a mechanism with the ability of reasoning and digesting information repeatedly, which is ignored by previous methods. In order to solve the above problems, we propose a Memory Graph with Message Rehearsal (MGMR) for dialogue generation based on the cognitive process of human memory. MGMR contains three main modules: sensory memory, short-term memory and long-term memory. Sensory memory converts the current utterance into embeddings from both word-level and sentence-level. We design a message rehearsal module in short-term memory to extract valuable information of current utterance deeply and repeatedly combined with the relative historical dialogue information and external knowledge stored in long-term memory. Furthermore, we innovatively design a memory graph in long-term memory to construct the relations among words, utterances and knowledge. The memory graph achieves three goals: extracting accurate relations between current utterance and historical utterances, updating the historical dialogue information, and achieving knowledge precipitation by expanding memory graph with the key words and relevant external knowledge of current utterance. We evaluate our model on real-world datasets and achieve better performance compared with the existing state-of-the-art methods.

Title: DocSemMap 2.0: Semantic Labeling based on Textual Data Documentations Using Seq2Seq Context Learner Abstract: ABSTRACTMethods for automated semantic labeling of data are an indispensable basis for increasing the usability of data. On the one hand, they contribute to the homogenization of the annotations and thus to the increase in quality; on the other hand, they reduce the modeling effort, provided that the quality of the used methodology is sufficient. In the past, research has focused primarily on data- and label-based methods. Another approach that has received recent attention is the incorporation of textual data documentations to support the automatic mapping of datasets to a knowledge graph. However, upon deeper analysis, our recent approach called DocSemMap gives away potential in a number of places. In this paper, we extend the current state of the art approach by uncovering existing shortcomings and presenting our own improvements. Using a sequence-to-sequence model (Seq2Seq), we exploit the context of datasets. An additional introduced classifier provides the linkage of documentation and labels for prediction. Our extended approach achieves a sustainable improvement in comparison to the reference approach.

Title: Samba: Identifying Inappropriate Videos for Young Children on YouTube Abstract: ABSTRACTYouTube videos are one of the most effective platforms for disseminating creative material and ideas, and they appeal to a diverse audience. Along with adults and older children, young children are avid consumers of YouTube materials. Children often lack means to evaluate if a given content is appropriate for their age, and parents have very limited options to enforce content restrictions on YouTube. Young children can thus become exposed to inappropriate content, such as violent, scary or disturbing videos on YouTube. Previous studies demonstrated that YouTube videos can be classified into appropriate or inappropriate for young viewers using video metadata, such as video thumbnails, title, comments, etc. Metadata-based approaches achieve high accuracy, but still have significant misclassifications, due to the reliability of input features. In this paper, we propose a fusion model, called Samba, which uses both metadata and video subtitles for content classification. Using subtitles in the model helps better infer the true nature of a video improving classification accuracy. On a large-scale, comprehensive dataset of 70K videos, we show that Samba achieves 95% accuracy, outperforming other state-of-the-art classifiers by at least 7%. We also publicly release our dataset.

Title: KRAF: A Flexible Advertising Framework using Knowledge Graph-Enriched Multi-Agent Reinforcement Learning Abstract: ABSTRACTBidding optimization is one of the most important problems in online advertising. Auto-bidding tools are designed to address this problem and are offered by most advertising platforms for advertisers to allocate their budgets. In this work, we present a Knowledge Graph-enriched Multi-Agent Reinforcement Learning Advertising Framework (KRAF). It combines Knowledge Graph (KG) techniques with a Multi-Agent Reinforcement Learning (MARL) algorithm for bidding optimization with the goal of maximizing advertisers' return on ad spend (ROAS) and user-ad interactions, which correlates to the ad platform revenue. In addition, this proposal is flexible enough to support different levels of user privacy and the advent of new advertising markets with more heterogeneous data. In contrast to most of the current advertising platforms that are based on click-through rate models using a fixed input format and rely on user tracking, KRAF integrates the heterogeneous available data (e.g., contextual features, interest-based attributes, information about ads) as graph nodes to generate their dense representation (embeddings). Then, our MARL algorithm leverages the embeddings of the entities to learn efficient budget allocation strategies. To that end, we propose a novel coordination strategy based on a mean-field style to coordinate the learning agents and avoid the curse of dimensionality when the number of agents grows. Our proposal is evaluated on three real-world datasets to assess its performance and the contribution of each of its components, outperforming several baseline methods in terms of ROAS and number of ad clicks.

Title: Generative Adversarial Zero-Shot Learning for Cold-Start News Recommendation Abstract: ABSTRACTNews recommendation models extremely rely on the interactive information between users and news articles to personalize the recommendation. Therefore, one of their most serious challenges is the cold-start problem (CSP). Their performance is dropped intensely for new users or new news. Zero-shot learning helps in synthesizing a virtual representation of the missing data in a variety of application tasks. Therefore, it can be a promising solution for CSP to generate virtual interaction behaviors for new users or new news articles. In this paper, we utilize the generative adversarial zero-shot learning in building a framework, namely, GAZRec, which is able to address the CSP caused by purely new users or new news. GAZRec can be flexibly applied to any neural news recommendation model. According to the experimental evaluations, applying the proposed framework to various news recommendation baselines attains a significant AUC improvement of 1% - 21% in different cold start scenarios and 1.2% - 6.6% in the regular situation when both users and news have a few interactions.

Title: On Smoothed Explanations: Quality and Robustness Abstract: ABSTRACTExplanation methods highlight the importance of the input features in taking a predictive decision, and represent a solution to increase the transparency and trustworthiness in machine learning and deep neural networks (DNNs). However, explanation methods can be easily manipulated generating misleading explanations particularly under visually imperceptible adversarial perturbations. Recent work has identified the decision surface geometry of DNNs as the main cause of this phenomenon. To make explanation methods more robust against adversarially crafted perturbations, recent research has promoted several smoothing approaches. These approaches smooth either the explanation map or the decision surface. In this work, we initiate a very thorough evaluation of the quality and robustness of the explanations offered by smoothing approaches. Different properties are evaluated. We present settings in which the smoothed explanations are both better, and worse, than the explanations derived by the commonly-used (non-smoothed) Gradient explanation method. By making the connection with the literature on adversarial attacks, we demonstrate that such smoothed explanations are robust primarily against additive attacks. However, a combination of additive and non-additive attacks can still manipulate these explanations, revealing important shortcomings in their robustness properties.

Title: AutoForecast: Automatic Time-Series Forecasting Model Selection Abstract: ABSTRACTIn this work, we develop techniques for fast automatic selection of the best forecasting model for a new unseen time-series dataset, without having to first train (or evaluate) all the models on the new time-series data to select the best one. In particular, we develop a forecasting meta-learning approach called AutoForecast that allows for the quick inference of the best time-series forecasting model for an unseen dataset. Our approach learns both forecasting models performances over time horizon of same dataset and task similarity across different datasets. The experiments demonstrate the effectiveness of the approach over state-of-the-art (SOTA) single and ensemble methods and several SOTA meta-learners (adapted to our problem) in terms of selecting better forecasting models (i.e., 2X gain) for unseen tasks for univariate and multivariate testbeds.

Title: Spherical Graph Embedding for Item Retrieval in Recommendation System Abstract: ABSTRACTOne of the challenging problems in large-scale recommendation systems is to retrieve relevant candidates accurately and efficiently. Graph-based retrievals have been widely deployed in industrial recommendation systems. Previous graph-based methods depend on integrated graph infrastructures because of inherent data dependency in graph learning. However, it could be expensive to develop a graph infrastructure. In this paper, we present a simple and effective graph-based retrieval method, which does not need any graph infrastructures. We conduct extensive offline evaluations and online tests in a real-world recommendation system. The results show that the proposed method outperforms the existing methods. The source code of our algorithm is available online.

Title: Modeling Price Elasticity for Occupancy Prediction in Hotel Dynamic Pricing Abstract: ABSTRACTIn this paper, we propose a novel elastic demand function that captures the price elasticity of demand in hotel occupancy prediction. We develop a price elasticity prediction model (PEM) with a competitive representation module and a multi-sequence fusion model to learn the dynamic price elasticity from a complex set of affecting factors. Moreover, a multi-task framework consisting of room- and hotel-level occupancy prediction tasks is introduced to PEM to alleviate the data sparsity issue. Extensive experiments on real-world datasets show that PEM outperforms other state-of-the-art methods for both occupancy prediction and dynamic pricing. PEM model has been successfully deployed at Fliggy and shown good performance in online hotel booking services.

Title: Dialogue State Tracking Based on Hierarchical Slot Attention and Contrastive Learning Abstract: ABSTRACTDialogue state is a key information in traditional task-oriented dialogue systems, which represents the user's dialogue intention at each moment through a set of (slot, value). The recent methods model the slot and the dialogue context to keep track of the state, but there is a lack of refinement of context information. They do not consider the influence of dialogue context in different scenarios. Our proposed approach utilizes a fine-grained representation of each slot at multiple levels and incorporates an interaction mechanism to obtain a weight of past memory, present utterance and relevance of the slots. Besides, to address the problem that the dialogue utterance is semantically distant from the corresponding slot value, we introduce the contrastive learning to make the utterance embedding mapped under each slot name more suitable with the ground truth value and away from other slot values. This improves the accuracy of mapping between feature space and semantic space. In the predefined ontology-based approaches, our model achieves leading results with both MultiWOZ2.0 and MultiWOZ2.1 datasets.