Skip to content

Latest commit

 

History

History
532 lines (403 loc) · 63.9 KB

2022_WSDM.md

File metadata and controls

532 lines (403 loc) · 63.9 KB

WSDM 2022

Causality

  • Causal Mediation Analysis with Hidden Confounders；Lu Cheng (Arizona State University); Ruocheng Guo (City University of Hong Kong); Huan Liu (Arizona State University)

    • Causal Mediation Analysis; Confounders; Proxy Variable; Latent Variable Model; Fairness

  • Estimating Causal Effects of Multi-Aspect Online Reviews with Multi-Modal Proxies； Lu Cheng (Arizona State University); Ruocheng Guo (City University of Hong Kong); Huan Liu(Arizona State University)

    • Online Reviews; Causal Effect Estimation; Hidden Confounder; Multi-Modality; Multi-Aspect Sentiment
    • 利用多模态代理（multi-modal proxies）如消费者概况、与企业交互等来处理混淆因子（confounder）带来的挑战，从而从细粒度估计某些因子对消费者在线评论的因果效应。

  • Doubly Robust Off-Policy Evaluation for Ranking Policies under the Cascade Behavior Model; Haruka Kiyohara (Tokyo Institute of Technology); Yuta Saito (Cornell University); Tatsuya Matsuhiro (Yahoo Japan Corporation); Yusuke Narita (Yale University); Nobuyuki Shimizu (Yahoo Japan Corporation); Yasuo Yamamoto (Yahoo! Japan)

    • off policy evaluation; slate recommendation; doubly robust; inverse propensity score; cascade model

  • A Counterfactual Modeling Framework for Churn Prediction; Guozhen Zhang (Tsinghua University); Jinwei Zeng (Tsinghua University); Zhengyue Zhao (Tsinghua University); Depeng Jin (Tsinghua University); Yong Li (Tsinghua University)

    • keywords: Churn prediction, social influence, causal information learning
    • 我们开发了一个用于流失预测的反事实建模框架，它可以有效地捕捉社会影响的因果信息，以进行准确且可解释的流失预测。

  • Learning Fair Node Representations with Graph Counterfactual Fairness; Jing Ma(University of Virginia); Ruocheng Guo (City University of Hong Kong); Mengting Wan (Microsoft);Longqi Yang (Microsoft); Aidong Zhang (University of Virginia);Jundong Li (University of Virginia)

    • Counterfactual fairness; graph; fairness; node representation
    • 本文提出了一个基于图的反事实公平性的概念，并设计出了一个基于反事实数据增强的框架，利用每个节点及其邻节点敏感属性扰动进行反事实推理，通过最小化从原始图中学习到的表示和反事实推理之间的差异，进而加强公平性。

  • Uncovering Causal Effects of Online Short Videos on Consumer Behaviors；Ziqi Tan (Zhejiang University); Shengyu Zhang (Zhejiang University); Nuanxin Hong (Zhejiang University); Kun Kuang (Zhejiang University); Yifan Yu (Zhejiang University); Zhou Zhao (Zhejiang University); Jin Yu (Alibaba Group); Hongxia Yang (Alibaba Group)

    • keywords: Short video, Advertising effects, Video subjectivity, Doubly robust
    • 针对短视频场景，对casual effect和subjectivity scores之间的相关性进行分析

Graph mining

  • GAGE: Geometry Preserving Attributed Graph Embeddings; Charilaos Kanatsoulis (University of Pennsylvania); Nicholas D Sidiropoulos (University of Virginia)*

    • keywords: networks, graphs, tensors, representation learning, embedding, multi dimensional scaling
    • 本文提出了一种新的用于属性网络中节点嵌入的张量分解方法，该方法保留了连接和属性的距离。

  • Efficient Graph Convolution for Joint Node Representation Learning and Clustering; Chakib Fettal (Université de Paris); Lazhar Labiod (LIPADE); Mohamed Nadif (Université de Paris)*

    • keywords: Graph Convolutional Networks; Node Embedding; Node Clustering
    • 以往的图卷积神经网络都把聚类任务当成下游任务，本文将聚类目标函数和传统GNN目标函数相结合，提出了更好的图聚类效果。

  • Ada-GNN: Adapting to Local Patterns for Improving Graph Neural Networks; Zihan Luo (Huazhong University of Science and Technology); Jianxun Lian (Microsoft Research Asia); Hong Huang (Huazhong University of Science and Technology); Hai Jin (Huazhong University of Science and Technology); Xing Xie (Microsoft Research Asia)*

    • keywords: Graph Neural Networks; Local Adaption; Meta-learning
    • 将整个图分为若干个不相交的子图，用一个元学习器让基础GNN模型去适应每个subgroup, 然后为每个subgroup生成一个既考虑全局信息，有考虑局部信息的GNN

  • Triangle Graph Interest Network for Click-through Rate Prediction; Wensen Jiang (Alibaba Group); Yizhu Jiao (Fudan University); Qingqin Wang (Fudan University); Chuanming Liang (Alibaba Group); Lijie Guo (Alibaba Group); Yao Zhang (Fudan University); Zhijun Sun (Zhejiang Cainiao Supply Chain Management Co Ltd); Yun Xion*

    • keywords: recommender system, click-through rate prediction, triangle, grap
    • 发现了click-through rate prediction任务中三角形内同质性和三角形间异质性，并基于此提出了一个新颖的图卷积神经网络方法。

  • EvoKG: Jointly Modeling Event Time and Network Structure for Reasoning over Temporal Knowledge Graphs; Namyong Park (Carnegie Mellon University); Fuchen Liu (Microsoft); Purvanshi Mehta (Microsoft); Elena D Cristofor (Microsoft); Christos Faloutsos (CMU); Yuxiao Dong (Microsoft)*

    • keywords: temporal knowledge graph, reasoning over temporal knowledge graphs, graph representation learning
    • 主要同时建模变化的图网络结构，事件时间的估计，递归时间的建模；邻居聚合既在关系层面也在时间层面进行。使用神经网络去估计事件发生时间的概率。

  • Graph Few-shot Class-incremental Learning;Zhen Tan (Arizona State University); Kaize Ding (Arizona State University); Ruocheng Guo (City University of Hong Kong); Huan Liu (Arizona State University)*

    • keywords: Graph Neural Networks, Incremental Learning, Few-shot Learning
    • 使用图神经网络去做少样本学习，类别增量学习

  • Attributed Graph Modeling with Vertex Replacement Grammars; Satyaki Sikdar (University of Notre Dame); Neil Shah (Snap Inc.); Tim Weninger (University of Notre Dame)

    • keywords: graph models, assortativity, graph generation, attributed graphs
    • 主要是做 attributed graph mining 以及 graph generation

  • Adversarial Attack on Graph Neural Networks as An Influence Maximization Problem; Jiaqi Ma (University of Michigan); Junwei Deng (Shanghai Jiao Tong University); Qiaozhu Mei (University of Michigan);

    • keywords: graph neural networks; adversarial attack; influence maximization
    • 本文研究了在 GNN 上的攻击问题。作者将图上节点的 perturbation 操作与 influence maximization 相联系，提供了一种研究图上攻击的新视角，同时也提出了一种有效的攻击手段。

  • Graph Embedding with Hierarchical Attentive Membership; Lu Lin (University of Virginia); Ethan Blaser (University of Virginia); Hongning Wang (University of Virginia);

    • keywords: Representation learning, graph embedding, graph neural network,mixed membership block models
    • (比较推荐做graph的同学看看) 本文研究图上的 Hierarchical Attentive Membership，总的来说就是将图上的节点进行层次化分组，在进行信息聚合时考虑这些 group 的信息，最终在节点分类和链路预测任务上都取得了超过SOTA的效果

  • Linear, or Non-Linear, That is the Question!; Taeyong Kong (Yonsei University); Taeri Kim (Hanyang University); Jinsung Jeon (Yonsei university); Jeongwhan Choi (Yonsei University); Yeon-Chang Lee (Hanyang University); Noseong Park (Yonsei University, Korea); Sang-Wook Kim (Hanyang University, Korea)

    • keywords: Recommender Systems, Collaborative Filtering, Embedding Propagation, Graph Neural Network
    • 这篇paper推荐做图与推荐的同学读一读。本文可以视为 NGCF/LightGCN 的拓展文章，这篇文章的 CF 模型 HMLET 对每个 user/item 使用了一个 gating function决定使用 linear propagation 还是 non-linear propagation，最终在 gowalla/yelp2018/amazon-book 三个数据集上都取得了 SOTA 的效果。

  • Interpretable Relation Learning on Heterogeneous Graphs; Qiang Yang (King Abdullah University of Science and Technology); Qiannan Zhang (King Abdullah University of Science and Technology); Chuxu Zhang (Brandeis University); Xiangliang Zhang (University of Notre Dame)

    • keywords: Interpretable relation learning; Heterogeneous graphs; Graph neural networks
    • 本文通过在捕捉到的局部子图上使用 self-supervised GNN，将 node 之间的路径相关性用 meta-path based path encoder 进行建模；从而使得图上节点之间的 relation 得到一个合理的解释

  • AngHNE: Representation Learning for Bipartite Heterogeneous Networks with Angular Loss; Cangqi Zhou (Nanjing University of Science and Technology); Hui Chen (Nanjing University of Science and Technology ); Jing Zhang (Nanjing University of Science and Technology); Qianmu Li (Nanjing University of Science and Technology ); Dianming Hu (Sensedeal.ai)

    • keywords: Representation learning; Bipartite networks; Angular loss
    • 本文研究了二分图上的表征学习问题。现有的异质图表征学习方法（metapath-based, proximity-based and graph neural network-based）使用内积或者向量范数评估在 embedding space 上的相似性，这样要么会破坏三角不等式，要么会对 scaling transformation 相当敏感。本文提出了 angle-based representation 方法来解决以上问题。

  • Deep Graph-level Anomaly Detection by Glocal Knowledge Distillation; Rongrong Ma (University of Technology Sydney); Guansong Pang (University of Adelaide); Ling Chen (” University of Technology, Sydney, Australia”); Anton van den Hengel (University of Adelaide);

    • keywords: Graph-level anomaly detection, Graph neural networks, Knowledge distillation, Deep learning
    • 图异常检测，作者对其考虑到local异常与global异常的特性，通过提出一个图蒸馏框架将两者统一起来。

  • Understanding and Improvement of Adversarial Training for Network Embedding from an Optimization Perspective; Lun Du (Microsoft Research); Xu Chen (Peking University); Fei Gao (Beijing Normal University); Qiang Fu (Microsoft Research Asia); Kunqing Xie (pku); Shi Han (Microsoft Research); Dongmei Zhang (Microsoft Research Asia);

    • keywords: network embedding, adversarial training, optimization method,saturation region problem
    • (推荐做图方向同学学习一下)，本文旨在从优化的角度解释现有Network Embedding中对抗扰动学习（APP）的优势，并且提出一个新的激活方式在多个数据集上均有显著提升。

  • Scalable Graph Topology Learning via Spectral Densification; Yongyu Wang (Michigan Technological University); Zhiqiang Zhao (Stevens Institute of Technology); Zhuo Feng (Stevens Institute of Technology);

    • keywords: graph topology learning; spectral graph theory; spectral clustering; dimensionality reduction
    • 本文解决图表征学习中计算复杂度以及规模化的限制，提出了一个更为高效的模型。(more scalable and higher efficiency)

  • Surrogate Representation Learning with Isometric Mapping for Gray-box Graph Adversarial Attack; Zihan Liu (Westlake University); Yun Luo (Westlake university); Zelin Zang (Westlake University); Stan Z. Li (Westlake University);

    • keywords: graph adversarial attack; gray-box attack; edge perturbation; representation learning; non-euclidean isometric mapping
    • 本文尝试分析图对抗攻击中surrogate representation learning，以保存图拓扑结构信息的重要性。

  • Geometric Inductive Matrix Completion: A Hyperbolic Approach with Unified Message Passing；Chengkun Zhang (the University of Sydney); Hongxu Chen (University of Technology Sydney); Sixiao Zhang (University of Technology Sydney); Guandong Xu (University of Technology Sydney, Australia); Junbin Gao (University of Sydney, Australia)

    • keywords: graph neural networks; inductive matrix completion; hyperbolic space; representation learning.

  • Towards Robust Graph Neural Networks for Noisy Graphs with Sparse Labels；Enyan Dai (The Pennsylvania State University); Wei Jin (Michigan State University); Hui Liu (Michigan State University); Suhang Wang (Pennsylvania State University)

    • Noisy Edges; Robustness; Graph Neural Networks
    • 在带有有限标记节点的有噪图上构造鲁棒GNN的问题

  • Graph Collaborative Reasoning；Hanxiong Chen (Rutgers University); Yunqi Li (Rutgers University); Shaoyun Shi (Tsinghua University); Shuchang Liu (Rutgers University); He Zhu (Rutgers University); Yongfeng Zhang (Rutgers University)

    • Collaborative Reasoning; Relational Reasoning; Neural-Symbolic Learning and Reasoning; GNNs; Recommendation; Link Prediction
    • 从逻辑推理的角度，利用相邻链路信息对图进行关系推理

  • Long Short-Term Temporal Meta-learning in Online Recommendation; Ruobing Xie (Ruobing Xie); Yalong Wang(Ruobing Xie); Rui Wang(Ruobing Xie); Yuanfu Lu(Ruobing Xie); Yuanhang Zou(Ruobing Xie); Feng Xia(Ruobing Xie); Leyu Lin(Ruobing Xie)

    • Recommendation; Temporal meta-learning; Online recommendation
    • 利用图和meta learning捕捉用户实时兴趣

  • HeteroQA: Learning towards Question-and-Answering through Multiple Information Sources via Heterogeneous Graph Modeling; Shen Gao (Peking University); Yuchi Zhang (Alibaba Group); Yongliang Wang (Alibaba group); Yang Dong (antgroup); Xiuying Chen (Peking University); Dongyan Zhao (Peking University); Rui Yan (Peking University)

    • 提出一个问题敏感的异构图transformer将多信息源结合起来，从而对社区用户的问题进行回答

  • MotifClass: Weakly Supervised Text Classification with Higher-order Metadata Information; (RayyanAhmad Khan (Technical University of Munich); Martin Kleinsteuber (Technical University of Munich)*

    • text classification; weak supervision; metadata
    • 利用以后图将文本中的元数据（作者、类型、研究时间等）与文本信息结合，捕捉其中的高阶网络信息从而提升分类性能。

  • Sampling Multiple Nodes in Large Networks: Beyond Random Walks; Omri Ben-Eliezer (MIT); Talya Eden (MIT); Joel Oren (Bosch Center for Artificial Intelligence); Dimitris Fotakis (National Technical University of Athens)

    • Graph and Network Sampling; Node Sampling
    • 提出了一种新的采样方法，通过显式搜索网络中不易访问的组件，绕过了混合时间中的依赖。

  • ComGA:Community-Aware Attributed Graph Anomaly Detection; xuexiong luo (university); Jia Wu (Macquarie University); Amin Beheshti (Macquarie University); Jian Yang (Macquarie University); Xiankun Zhang (Tianjin University of Science and Technology); Yuan Wang (Tianjin University of Science and Technology)

    • Anomaly Detection; Community Structure; Attributed Graphs; Graph Neural Networks
    • 本文研究图异常检测，研究了三种类型的图异常:局部异常、全局异常和结构异常。设计了一个定制深度图卷积网络(tGCN)，用于属性图的异常检测。

  • Finding a Concise, Precise, and Exhaustive Set of Near Bi-Cliques in Dynamic Graphs; Hyeonjeong Shin (KAIST); Taehyung Kwon (KAIST); Neil Shah (Snap Inc.); Kijung Shin (KAIST)

    • Bi-clique; Dynamic Graph; Graph Compression; Pattern Discovery

  • Joint Learning of E-commerce Search and Recommendation with A Unified Graph Neural Network; Jing Dong (The Chinese University of Hong Kong, Shenzhen)*; Ke Li (The Chinese University of Hong Kong, Shenzhen); Shuai Li (Shanghai Jiao Tong University); Baoxiang Wang (The Chinese University of Hong Kong, Shenzhen)

    • keywords: Click-through rate prediction, graph neural network, joint learning, product search and recommendation, e-commerce
    • 本文提出一种综合的图神经网络模型，协同学习推荐和搜索数据，提高训练数据量，提升模型性能

  • Bringing Your Own View: Graph Contrastive Learning without Prefabricated Data Augmentations; Yuning You (Texas A&M University)*; Tianlong Chen (Unversity of Texas at Austin); Zhangyang Wang (University of Texas at Austin); Yang Shen (Texas A&M University)

    • keywords: Graph contrastive learning; graph generative model; information minimization; information bottleneck
    • 利用了图对比学习、信息瓶颈等方法

  • Few-shot Link Prediction in Dynamic Networks; Cheng Yang (Beijing University of Posts and Telecommunications)*; Chunchen Wang (Beijing University of Posts and Telecommunications); Yuanfu Lu (WeChat Search Application Department, Tencent); Xumeng Gong (Beijing University of Posts and Telecommunication)

    • keywords: link prediction, dynamic network, few-shot prediction, meta-learning, graph neural networks
    • 基于meta-learning 的动态图神经网络模型

  • Modeling Scale-free Graphs with Hyperbolic Geometry for Knowledge-aware Recommendation; Yankai Chen (The Chinese University of Hong Kong)*; Menglin Yang (The Chinese University of Hong Kong); Yingxue Zhang (Huawei Technologies Canada); Mengchen Zhao (Huawei Noah’s Ark Lab); Ziqiao Meng (Chinese University of Hong Kong); Jianye Hao (Huawei Noah’s Ark Lab)

    • keywords: Recommender system, Knowledge graph, Hyperbolic geometric
    • 双曲空间知识图谱解决推荐问题

  • Multi-Scale Variational Graph AutoEncoder for Link Prediction; Zhihao Guo (Shanxi University); Feng Wang ()*; Kaixuan Yao (Shanxi University); Jiye Liang (Shanxi University); Zhiqiang Wang (Shanxi University)

    • keywords: Graph AutoEncoder, Self-Supervised Learning, Graph Neural Networks, Link Prediction

  • On Generalizing Static Node Embedding to Dynamic Settings; Di Jin (University of Michigan)*; Sungchul Kim (Adobe); Ryan Rossi (Adobe Research); Danai Koutra (U Michigan)

    • keywords: representation learning; dynamic networks; graph time-series

  • KGNN: Harnessing Kernel-based Networks for Semi-supervised Graph Classification; Wei Ju (Peking University)*; Junwei Yang (Peking University); Meng Qu (Mila); Weiping Song (Peking University); Jianhao Shen (Peking University); Ming Zhang (Peking University)

    • keywords: Graph Classification, Graph Neural Networks, Graph Kernels, Semisupervised Learning

  • A Neighborhood-Attention Fine-grained Entity Typing for Knowledge Graph Completion; Jianhuan Zhuo(University of Chinese Academy of Sciences); Qiannan Zhu(Renmin University of China);Yinliang Yue(University of Chinese Academy of Sciences);Yuhong Zhao(Chinese Academy of Sciences);Weisi Han(Chinese Academy of Sciences)

    • Knowledge Graph Completion, Fine-grained Entity Typing, Knowledge Graph Representation Learning, Entity Type Prediction

  • Friend Story Ranking with Edge-Contextual Local Graph Convolutions; JXianfeng Tang (Amazon); Yozen Liu (Snap Inc); Xinran He (Snap Inc.); Suhang Wang (Pennsylvania State University); Neil Shah (Snap Inc.)*

    • Social networks; graph neural networks; user modeling

  • Predicting Human Mobility via Graph Convolutional Dual-attentive Networks;Weizhen Dang (Tsinghua University); Haibo Wang (Tsinghua University); Shirui Pan (Monash University); Pei ZHANG (Beijing University of Posts and Telecommunications); Chuan Zhou (Chinese Academy of Sciences); Xin Chen (Tsinghua University); Jilong Wang (Tsinghua University) *

    • Mobility Prediction, Graph Convolution, Attention Mechanism,Sequential Modeling

  • Profiling the Design Space for Graph Neural Networks based Collaborative Filtering；Zhenyi Wang (Beijing University of Posts and Telecommunications); Huan Zhao (4Paradigm); Chuan Shi (Beijing University of Posts and Telecommunications)

    • keywords: Collaborative filtering; Graph neural networks; Empirical evaluation
    • 提出了一种泛化能力强的GNN based CF模型，可以高效解决different dimensions对推荐结果的影响

  • Informed Multi-context Entity Alignment；KEXUAN XIN (The University of Queensland); Zequn Sun (Nanjing University); Wen Hua (The University of Queensland); Wei Hu (Nanjing University); Xiaofang Zhou (The Hong Kong University of Science and Technology)

    • keywords: Knowledge Graph, Entity Alignment, Multi-context Transformer, Holistic Reasoning
    • 对多个来源的知识图谱聚合中实体对齐问题进行优化

Debiasing、fairness、long-tail、ODD、domain adaption、鲁棒性、降噪 （包括推荐和非推荐的都算）

  • Towards Fair Classifiers Without Sensitive Attributes: Exploring Biases in Related Features; Tianxiang Zhao (Penn State University); Enyan Dai (The Pennsylvania State University); Kai Shu (Illinois Institute of Technology); Suhang Wang (Pennsylvania State University)
    • keywords: Fairness; Social mining; Data learning
    • 本文解决的问题是：当训练数据中直接影响 fairness 的 sensitive attributes 不可得时如何利用一些 non-sensitive attributes 来使得模型学的比较 fair。
  • Efficient two-stage label noise reduction for retrieval-based tasks; Mengmeng Kuang (Tencent Holdings Ltd.); Weiyan Wang (HKUST); Zhenhong Chen (Tencent Holdings Ltd. ); Lie Kang (Tencent Holdings Ltd. ); Qiang Yan (Tencent);
    • keywords: Data cleaning, noisy labels, text dataset, retrieval-based tasks
    • 本文针对检索任务中噪声标签的影响，提出一个两步走的策略进行噪声降低。
  • k-Clustering with Fair Outliers; Matteo Almanza (Sapienza University); Alessandro Epasto (Google); Alessandro Panconesi (Sapienza, University of Rome); Giuseppe Re (Sapienza University of Rome);
    • keywords: Clustering; outliers; fairness; 𝑘−means; disparate impact; coreset.
    • 本文针对聚类问题中边界case的公平性问题进行分析，提出对边界坏样本进行移除可有效改善聚类的下游任务。
  • Introducing the Expohedron for efficient Pareto-optimal fairness-utility amortizations in repeated rankings; Till Kletti (NaverLabs Europe); Jean-Michel Renders (NAVER LABS Europe); Patrick Loiseau (Inria);
    • keywords: ranking, fairness, amortization, pareto-optimal, muli-objective optimization, Carathéodory, expohedron, GLS, balanced words
    • 本文针对大规模推荐的效率性问题，从几何的角度尝试完成用户检索效率最大化以及商家曝光均匀性的目标。
  • Unsupervised Cross-Domain Adaptation for Response Selection Using Self-Supervised and Adversarial Training；Jia Li (Peking University); Chongyang Tao (Microsoft); Huang Hu (Microsoft); Can Xu (microsoft); Yining Chen (Microsoft); Daxin Jiang (Microsoft, Beijing, China)*
    • Multi-turn response selection; retrieval-based chatbot; unsupervised cross-domain adaptation; deep neural network; matching
  • It Is Different When Items Are Older: Debiasing Recommendations When Selection Bias and User Preferences are Dynamic; Jin Huang (University of Amsterdam); Harrie Oosterhuis (Radboud University); Maarten de Rijke (University of Amsterdam & Ahold Delhaize)*
    • Selection bias; Dynamic user preferences
    • 拓展IPS解决动态选择偏差并推断用户兴趣；理论证明了在动态场景下，选择偏差和用户兴趣是动态的
  • Multi-Sparse-Domain Collaborative Recommendation via Enhanced Comprehensive Aspect Preference Learning; Xiaoyun Zhao (Sichuan University); Ning Yang (Sichuan University); Philip S Yu (UIC)*
    • Cross-Domain Recommendation, Dual-Target CDR, Multi-Target CDR, Transfer Learning
    • 利用多个相关的domain结合对抗学习提升交叉域推荐性能
  • An Ensemble Model for Combating Label Noise; Yangdi Lu (mcmaster university); Yang Bo (McMaster University); Wenbo He (McMaster University)
    • noisy labels; image classification; ensemble learning; weakly supervised learning
  • **AdaptKT: A Domain Adaptable Method for Knowledge Tracing；*Song Cheng (University of Science and Technology of China); Qi Liu (” University of Science and Technology of China, China”); Enhong Chen (University of Science and Technology of China); Kai Zhang (University of Science and Technology of China); Zhenya Huang (University of Science and Technology of China); Yu Su (iFLYTEK Research)
    • domain adaptation; knowledge tracing; deep learning
    • 通过MMD跨域进行知识抽取
  • RecGURU: Adversarial Learning of Generalized User Representations for Cross-Domain Recommendation; Chenglin Li (University of Alberta); Mingjun Zhao (University of Alberta); huanming zhang (South China University of Technology); Chenyun YU (City University of Hong Kong); Lei Cheng (Tencent); Guoqiang Shu (Tencent); Jianming Yang (Tencent ); Di Niu (University of Alberta)
    • Cross-Domain Recommendation; Sequential Recommendation; Learning Representation; Autoencoder; Adversarial Learning
    • 提出了RecGURU算法框架来生成一个包含跨领域用户信息的广义用户表示即使在两个领域中存在最小或没有公共用户。利用了self-attentive autoencoder，adversial learning等技术。
  • MtCut: A Multi-Task Framework for Ranked List Truncation; Jianxin Li (Beihang University); Wang Dong (Beihang University); Tianchen Zhu (Beihang University); Qishan Zhu (Beihang University); Yuxin Wen (Beihang University); Piao Hongming (Beihang University)
    • keywords: Information Retrieval, Multi-Task Learning, Ranked List Truncation, Bias Correction
    • 我们研究了更改截断时检索偏差的特征，并提出了一个多任务截断模型 MtCut。 它采用了两个辅助任务来使互补偏差互补.
  • External Evaluation of Ranking Models under Extreme Position-Bias; Yaron Fairstein (Technion); Elad Haramaty (Amazon); Arnon Lazerson (Amazon)*; Liane Lewin-Eytan (Amazon)
    • keywords: voice search, position bias, off-policy evaluation
    • 一个推荐系统评估模型，用于解决position-bias问题下，推荐系统的离线评估问题
  • Fighting Mainstream Bias in Recommender Systems via Local Fine Tuning; Ziwei Zhu (Texas A&M University)*; James Caverlee (Texas A&M University)
    • keywords: recommender systems; mainstream bias; local models
  • Personalized Transfer of User Preferences for Cross-domain Recommendation; Yongchun Zhu (Institute of Computing Technology); Zhenwei Tang (Institute of Computing Technology);Yudan Liu (WeChat Search Application Department);Fuzhen Zhuang (Institute of Artificial Intelligence);, Ruobing Xie (WeChat Search Application Department) ;Xu Zhang (WeChat Search Application Department);Leyu Lin (WeChat Search Application Department) ; Qing He (Institute of Computing Technology);
    • Cross-domain Recommendation; Cold-start Problem; Meta Net�work; Personalized Transfer
  • Improving the Applicability of Knowledge-Enhanced Dialogue Generation Systems by Using Heterogeneous Knowledge from Multiple Sources；Sixing Wu (Peking University);Minghui Wang (Peking University);Ying Li (Peking University); Dawei Zhang (Peking University);Zhonghai Wu(Peking University)
    • keywords: knowledge-enhanced dialogue generation, multi-source knowledge
    • 本文利用异构的外部知识来辅助对话生成系统的学习。
  • The Datasets Dilemma: How Much Do We Really Know About Recommendation Datasets?；Jin Yao Chin (Nanyang Technological University); Yile Chen (Nanyang Technological University); Gao Cong (Nanyang Technological Univesity)
    • keywords: Item Recommendation; Evaluation; Datasets; Data Characteristics
    • 对推荐系统训练使用的数据集进行分析，提出要重视数据集选择对结果鲁棒性的影响
  • Towards Unbiased and Robust Causal Ranking for Recommender Systems；Teng Xiao (Pennsylvania State University); Suhang Wang (Pennsylvania State University)
    • keywords: Causal Inference; Recommender Systems; Counterfactual Learning
    • 提出一种general and theoretically rigorous的因果推荐框架
  • Diversified Subgraph Query Generation with Group Fairness；Sheng Guan (Case Western Reserve University); Hanchao Ma (Case Western Reserve University); Yinghui Wu (Case Western Reserve University)
    • keywords: Attributed graph, Query suggestion, Fairness
    • 本文研究对于subgraph query generation问题，能够同时满足diversity和fairness的解决方案并提出算法

交互式推荐、Bandit、强化学习

  • Show Me the Whole World: Towards Entire Item Space Exploration for Interactive Personalized Recommendations; Yu Song (Huazhong University of Science and Technology); Shuai Sun (Huazhong University of Science and Technology); Jianxun Lian (Microsoft Research Asia); Hong Huang (Huazhong University of Science and Technology); Yu Li (University of Electronic Science and Technolgy of China)*
    • keywords: Recommender System, Contextual Bandit, Interest Exploration
    • 引入了两个简单但有效的分层 CB 算法，使经典的 CB 模型能够探索用户对整个项目空间的兴趣，而不受限于小项目集。 我们提出了一种分层CB（HCB）算法来探索用户对分层树的兴趣。
  • Supervised Advantage Actor-Critic for Recommender Systems; Xin Xin (Shandong University); Alexandros Karatzoglou (Google Research); Ioannis Arapakis (Telefonica Research); Joemon M Jose (University of Glasgow)
    • keywords: Recommendation; Reinforcement Learning; Actor-Critic; Q-learning; Advantage Actor-Critic; Negative Sampling
    • 这篇文章主要解决在 session/sequential 推荐里面用 RL 存在 action space 较大，reward 较稀疏等问题。作者结合 supervised sequential learning 提出了一种新的负采样策略解决这些问题。
  • Heterogeneous Global Graph Neural Networks for Personalized Session-based Recommendation; Yitong Pang (Tongji University); Lingfei Wu (JD.COM Silicon Valley Research Center); Qi Shen (Tongji University); Yiming Zhang (Tongji University); Zhihua Wei (Tongji University); Fangli Xu (College of William and Mary); Ethan Chang (Middlesex School); Bo Long (JD.com); Jian Pei (Simon Fraser University)
    • keywords: Recommendation system; Session-based recommendation; Graph neural network
    • 本文针对 Session-based 推荐提出了一种异质图神经网络，使得模型可以在所有 session 内挖掘 item transitions，从而利用当前以及历史 session 更好的建模用户兴趣。
  • Choosing the Best of All Worlds: Accurate, Diverse, and Novel Recommendations through Multi-Objective Reinforcement Learning; Du_an Stamenkovi_ (Department of Mathematics and Informatics, University of Novi Sad); Alexandros Karatzoglou (Google Research); Ioannis Arapakis (Telefonica Research); Xin Xin (Shandong University); Kleomenis Katevas (Telefonica Research);
    • keywords: Recommendation; Reinforcement Learning; Multi-Objective Reinforcement Learning; Diversity; Novelty
    • 多目标强化学习，作者提出了一个新的框架可满足多个原则: accuracy, diversity, and novelty of recommendations.
  • Toward Pareto Efficient Fairness-Utility Trade-off in Recommendation through Reinforcement Learning; Yingqiang Ge (Rutgers University); Xiaoting Zhao (Etsy); Lucia Yu (Etsy); Saurabh Paul (Etsy); Diane Hu (Etsy); Yongfeng Zhang (Rutgers University);
    • keywords: Recommender System; Multi-Objective Reinforcement Learning; Pareto Efficient Fairness; Unbiased Recommendation
    • 本文通过强化学的以保证推荐系统中fairness以及CTR的trade-off
  • A Cooperative-Competitive Multi-Agent Framework for Auto-bidding in Online Advertising; Chao Wen (Alibaba Group, China); Miao Xu (Alibaba Group, China); Zhilin Zhang (Alibaba Group); Zhenzhe Zheng (Shanghai Jiao Tong University); Yuhui Wang (MIIT Key Laboratory of Pattern Analysis and Machine Intelligence); Xiangyu Liu(Alibaba Group); Yu Rong(Alibaba Group); Dong Xie(Alibaba Group); Xiaoyang Tan (MIIT Key Laboratory of Pattern Analysis and Machine Intelligence); Chuan Yu Jian Xu (Alibaba Group); (Alibaba Group); Fan Wu (Shanghai Jiao Tong University); Guihai Chen (Shanghai Jiao Tong University); Xiaoqiang Zhu(Alibaba Group);
    • Auto-bidding; Bid Optimization; Multi-Agent Reinforcement Learning; Online Advertising
  • Combinatorial Bandits under Strategic Manipulations; Jing Dong (The Chinese University of Hong Kong, Shenzhen)*; Ke Li (The Chinese University of Hong Kong, Shenzhen); Shuai Li (Shanghai Jiao Tong University); Baoxiang Wang (The Chinese University of Hong Kong, Shenzhen)
    • keywords: multi-armed bandits, strategic manipulations, crowdsourcing, online information maximization, recommendation systems
  • Non-stationary Continuum-armed Bandits for Online Hyperparameter Optimization；Shiyin Lu (Nanjing University); Yu-Hang Zhou (Alibaba Group); Jing-Cheng Shi (Nanjing University; Alibaba Group); wenya zhu (Alibaba); Qingtao Yu (Alibaba); Qing-Guo Chen (Alibaba); Qing Da (Alibaba Group); Lijun Zhang (Nanjing University)*
    • continuum-armed bandits, non-stationary environments, hyperparameter optimization
    • 针对当前推荐系统处于一种动态的（在线）的场景，提出一种基于Bandits思想的超参数优化方法。
  • C2-CRS: Coarse-to-Fine Contrastive Learning for Conversational Recommender System；Yuanhang Zhou (Renmin University of China); Kun Zhou (Renmin University of China); Wayne Xin Zhao (Renmin University of China); Cheng Wang (Kuaishou Inc); Peng Jiang (Kuaishou Inc.); He Hu (Renmin University of China)
    • keywords: Conversational Recommender System; Contrastive Learning
    • 用多粒度对比学习框架，使对话推荐系统可以适用于multi-type external data

其他方面的推荐

  • S-Walk: Accurate and Scalable Session-based Recommendation with Random Walks; Minjin Choi (Sungkyunkwan University); Jinhong Kim (Sungkyunkwan University ); Joonseok Lee (Google Research & Seoul National University); Hyunjung Shim (Yonsei University); Jongwuk Lee (Sungkyunkwan University)

    • keywords：Collaborative filtering; Session-based recommendation; Random walks; Closed-form solution
    • 为了兼顾准确性和可扩展性，我们提出了一种新颖的基于会话的随机游走推荐，通过使用带重启的随机游走处理项目之间的高阶关系来捕获会话内和会话间的相关性。

  • ANTHEM: Attentive Hyperbolic Entity Model for Product Search;Nurendra Choudhary (Virginia Tech); Nikhil Rao (Amazon); Sumeet Katariya (Amazon); Karthik Subbian (Amazon); Chandan K Reddy (Virginia Tech)

    • keywords: E-commerce, productsearch, query representation, hyperbolic space
    • 提出基于注意力机制的双曲建模方法，能够更好的匹配查询和商品

  • Leveraging World Events to Predict E-Commerce Consumer Demand under Anomaly;Dan Kalifa (Technion – Israel Institute of Technology); Uriel Singer (Technion, Israel Institute of Technology); Ido Guy (eBay); Guy D. Rosin (Technion – Israel Institute of Technology); Kira Radinsky (Ebay)*

    • keywords: world events; anomalies; e-commerce; forecasting
    • 主要通过收集大量的世界信息和他们对应的文本，来预测市场的用户购买需求（宏观层面的需求分析）

  • Sequential Modeling with Multiple Attributes for Watchlist Recommendation in E-Commerce;Uriel Singer (Technion, Israel Institute of Technology); Haggai Roitman (IBM Research Haifa); yotam eshel (eBay); Alexander Nus (eBay); Ido Guy (eBay); Or Levi (eBay); Idan Hasson (eBay ); Eliyahu Kiperwasser (eBay)*

    • keywords: Watchlist, Sequential-Model, Transformers, E-Commerce
    • 本文主要研究用户的候选列表，从候选列表中选择用户最需要关注的商品，采用序列建模的方法来研究 -POLE: Polarized Embedding for Signed Networks;Zexi Huang (University of California, Santa Barbara); Arlei Silva (Rice University); Ambuj K Singh (UCSB)*
    • keywords:Representation learning; Signed embedding; Social polarization
    • 通过随机游走，有符号的嵌入来解决社交媒体的两极分化问题

  • Hierarchical Imitation Learning via Subgoal Representation Learning for Dynamic Treatment Recommendation; Lu Wang (East China Normal University); Ruiming Tang (Huawei Noah’s Ark Lab); Xiaofeng He (ECNU); Xiuqiang He (Huawei Noah’s Ark Lab)

    • keywords: Dynamic Treatment Recommendation, Hierarchical Imitation Learning, Subgoal Representation Learning
    • 本文通过 Hierarchical Imitation Learning 做 Treatment Recommendation。

  • A Peep into the Future: Adversarial Future Encoding in Recommendation; Ruobing Xie (WeChat Search Application Department, Tencent); Shaoliang Zhang (Tencent); Rui Wang (Tencent); Feng Xia (WeChat Search Application Department, Tencent); Leyu Lin (WeChat Search Application Department, Tencent);

    • keywords: recommendation, future information, GAN
    • 本文在推荐中将未来动作序列特征进行考虑，利用对抗的方式进行学习。

  • Variational User Modeling with Slow and Fast Features; Ghazal Fazelnia (Spotify); Eric Simon (Spotify); Ian Anderson (Spotify); Benjamin Carterette (Spotify); Mounia Lalmas (Spotify);

    • keywords:User Modeling, Latent Representation, Music Streaming
    • 本文在推荐中考虑fast and slow-moving features以更符合特定推荐场景（Music Streaming）。

  • Leaving No One Behind: A Multi-Scenario Multi-Task Meta Learning Approach for Advertiser Modeling; qianqian zhang (Alibaba); Xinru Liao (Alibaba Group); Quan Liu (Alibaba Group); Jian Xu (Alibaba Group); Bo Zheng (Alibaba Group);

    • keywords: Advertiser Modeling, Multi-Task Learning, Meta Learning, MultiBehavior Learning
    • 本文认为广告推荐存在多场景的复杂性，提出多场景多任务元学习方式以满足多个实际需求。

  • Learning-To-Ensemble by Contextual Rank Aggregation in E-Commerce;Xuesi Wang(Alibaba Group), Guangda Huzhang(Alibaba Group), Qianying Lin(Alibaba Group), Qing Da(Alibaba Group);

    • keywords: Learning-To-Ensemble, Rank Aggregation, Contextual Black-Box Optimization, Evaluator and Generator
    • 在电子商务场景下，本文在给定给定的RA（Rank Aggregator）模型的情况下，为子模型找到最优权值，解决了在线上下文黑盒优化(contextualBlack-Box Optimization)任务。

  • Scope-aware Re-ranking with Gated Attention in Feed;Hao Qian (Ant Services Group); Qintong Wu (Ant Group ); Kai Zhang (University of Science and Technology of China); Zhiqiang Zhang (Ant Group); Lihong Gu (Ant Group); Xiaodong Zeng (Ant Services Group ); Jun Zhou (Ant Financial); Jinjie Gu (Ant Group)*

    • Recommender System; Learing To Rank; Re-ranking
    • 按照用户的视窗范围让re-ranking模型更注意相邻物品

  • Improving Knowledge Tracing with Collaborative Information; Ting Long (Shanghai Jiao Tong University); Jiarui Qin (Shanghai Jiao Tong University); Jian Shen (Shanghai Jiao Tong University); Weinan Zhang (Huawei Noah’s Ark Lab); Wei Xia (Huawei Noah’s Ark Lab); Ruiming Tang (Huawei Noah’s Ark Lab); Xiuqiang He; Yong Yu (Huawei Noah’s Ark Lab);

    • knowledge tracing; sequence retrieval; correctness prediction

  • PipAttack: Poisoning Federated Recommender Systems for Manipulating Item Promotion; shijie zhang (The University of Queensland); Hongzhi Yin (The University of Queensland); Tong Chen (The University of Queensland); Zi Helen Huang (University of Queensland); Quoc Viet Hung Nguyen (Griffith University); Lizhen Cui (ShanDong University)*

    • Federated Recommender System; Poisoning Attacks; Deep Learning
    • 通过联邦学习就行去中心化，减轻poisoning attack的影响

  • PLdFe-RR: Personalized Long-distance Fuel-efficient Route Recom mendation Based On Historical Trajectory；Zhan Wang (Shandong University); Zhaohui Peng (Shandong University); Senzhang Wang (Central South University); Qiao Song (Shandong University)

    • keywords: Route recommendation, Spatiotemporal data, Trajectory data mining, Genetic algorithm
    • 路径推荐算法，首次加入用户偏好特征，适用于长距离路径推荐问题

  • Reinforcement Learning over Sentiment-Augmented Knowledge Graphs towards Accurate and Explainable Recommendation；Sung-Jun Park (Hanyang University); Dong-Kyu Chae (Hanyang University); Hong-Kyun Bae (Hanyang University); Sumin Park (Hanyang University); Sang-Wook Kim (Hanyang University, Korea)

    • keywords: Explainable recommendation, knowledge graph, sentiment analysis
    • 用基于强化学习的sentiment analysis，增加可解释性推荐的准确性和解释质量

  • Learning Multi-granularity Consecutive User Intent Unit for Session-based Recommendation；Jiayan Guo (Peking University); Yaming Yang (MSRA); Xiangchen Song (Carnegie Mellon University); Yuan Zhang (Peking University); Yujing Wang (MSRA); Jing Bai (Microsoft); Yan Zhang (Peking University)

    • keywords: Recommender System, Session-based Recommendation, Graph Neural Networks
    • 一种Session-based Recommendation模型，捕捉多粒度user intent，提高推荐效果

  • Contrastive Meta Learning with Behavior Multiplicity for Recommendation；Wei Wei (South China University of Technology); Chao Huang (University of Hong Kong); Lianghao Xia (South China University of Technology); Yong Xu (South China University of Technology); Jiashu Zhao (Wilfrid Laurier University); Dawei Yin (Baidu)

    • keywords: Collaborative filtering, Self-Supervised Learning, Multi-Behavior Recommendation, Meta Learning, Graph Neural Network
    • 研究利用推荐系统中user和item之间的multi-behavior的算法

搜索

  • Deep-QPP: A Pairwise Interaction-based Deep Learning Model for Supervised Query Performance Prediction; Suchana Datta (University College Dublin); Debasis Ganguly (University of Glasgow); Derek Greene (University College Dublin); Mandar Mitra (Indian Statistical Institute, Kolkata)
    • keywords: Supervised Query Performance Prediction, Interaction-based Models, Convolutional Neural Networks
    • 提出了一个fully data-driven的query performance prediction (QPP)模型，用以估计用户输入query的质量。
  • Lightweight Composite Re-Ranking for Efficient Keyword Search with BERT; Yingrui Yang (University of California, Santa Barbara); Yifan Qiao (University of California, Santa Barbara); Jinjin Shao (UCSB); Xifeng Yan (University of California, Santa Barbara); Tao Yang (UC Santa Barbara)
    • keywords: CPU-friendly inference; embedding composition; integration of feature-based and neural models
    • 考虑 BERT-based reranking 模型的效率问题。
  • Differential Query Semantic Analysis: Discovery of Explicit Interpretable Knowledge from E-Com Search Logs; Sahiti Labhishetty (University of Illinois Urbana-Champaign); ChengXiang Zhai (University of Illinois at Urbana-Champaign); Min Xie (Instacart); Lin Gong (WalmartLabs); Rahul Sharnagat (WalmartLabs); Satya Chembolu (WalmartLabs)
    • keywords: Query word lexicon; E-Com Search Logs; Query difference analysis
    • 本文从搜索日志中发掘 query 中的可解释信息，并将其对应到 product 中的 attribute 上。
  • A GNN-based Multi-task Learning Framework for Personalized Video Search; Li Zhang (University of Sheffield); Lei Shi ( Baidu); Jiashu Zhao (Wilfrid Laurier University); Juan Yang (Baidu); Tianshu Lyv (Baidu); Dawei Yin (Baidu); Haiping Lu (University of Sheffield)
    • keywords: Personalized Video Search, Graph Neural Networks, Multi-Task Learning, User-query Graph, Query-document Click Graph.
    • 现有的 personalized search methods (PSMs) 通过用户的 feedback 数据进行训练，作者发现这样会导致模型只能找到吸引人的 video 而找不到与 query 匹配的video，此外 user click 数据通常也很 sparse。因此作者提出了一个 multi-task 的 GNN 模型，既能对 user 点击行为建模又能对 query-video 之间的 relevance 建模。
  • Modeling Users’ Contextualized Page-wise Feedback for Click-Through Rate Prediction in E-commerce Search; Zhifang Fan (Alibaba Group); Dan Ou (Alibaba Group); Yulong Gu (Alibaba Group); Bairan Fu (Nanjing University); Xiang Li (Alibaba Group); WenTian Bao (alibaba); Xin-yu Dai (Nanjing University); Xiaoyi Zeng (Alibaba Group); Tao Zhuang (Alibaba Group);
    • keywords:user sequence modeling, neural networks, personalized search
    • 本文将搜索行为视为页面行为序列，以捕捉用户兴趣。
  • Beyond NED: Fast and Effective Search Space Reduction for Complex Question Answering over Knowledge Bases; Philipp Christmann (MPI for Informatics); Rishiraj Saha Roy (Max Planck Institute for Informatics); Gerhard Weikum (Max-Planck-Institut fur Informatik);
    • keywords: Question Answering, Knowledge Bases, Entity Linking
    • 本文针对KBQA中检索的效率问题以及多义性问题，提出一种剪枝方案进行优化。
  • Using Conjunctions for Faster Disjunctive Top-k Queries; Michal Siedlaczek (New York University); Antonio Mallia (New York University); Torsten Suel (“New York Univ., USA”);
    • keywords: query processing; top-k retrieval; candidate generation
    • 本文针对检索中效率问题进行优化
  • Leveraging Multi-view Inter-passage Interactions for Neural Document Ranking; Chengzhen Fu (huawei)*; EnRui Hu (ZheJiang University); letian feng (huawei.com); Zhicheng Dou (Remin University of China); Yantao Jia (Huawei Technologies Co., Ltd); Lei Chen (Hong Kong University of Science and Technology); Fan Yu (Huawei Technologies Co., Ltd)
    • keywords: Document ranking; Inter-passage attention; Intra-passage attention
    • 本文提出一种新的深度学习网络模型，用于弥补transformer等模型无法处理长文档的弱点，建模文档之间的关联关系，优化文档排序结果
  • Improving Session Search by Modeling Multi-Granularity Historical Query Change; Xiaochen Zuo (Renmin University of China)*; Zhicheng Dou (Remin University of China)
    • keywords: Session Search, Query Change, Document Ranking
    • 利用用户搜索会话中搜索词的变化来更好地学习搜索排序
  • `It’s on the tip of my tongue’: A new dataset for Known-Item Retrieval; Samarth Bhargav (University of Amsterdam)*; Georgios Sidiropoulos (University of Amsterdam); Evangelos Kanoulas (University of Amsterdam)
    • keywords: Known Item Retrieval; Tip of the tongue known item retrieval;
    • 一个新的搜索数据集
  • GraSP: Optimizing Graph-based Nearest Neighbor Search with Subgraph Sampling and Pruning; Minjia Zhang (Microsoft AI and Research)*; Wenhan Wang (Microsoft); Yuxiong He (Microsoft)
    • keywords: Vector management and search, search efficiency, graph sampling
  • Multi-Resolution Attention for Personalized Item Search; Furkan Kocayusufoglu (UC, Santa Barbara)*; Tao Wu (Google Research); Anima Singh (Google); Georgios Roumpos (Google Research); Heng-Tze Cheng (Google Research); Sagar Jain (Google); Ed H. Chi (Google); Ambuj K Singh (UCSB)
    • keywords: item search, personalization, temporal attention, multi-resolution attention, recommender systems
  • A Cooperative Neural Information Retrieval Pipeline with Knowledge Enhanced Automatic Query Reformulation；Xiangsheng Li (Tsinghua University); Jiaxin Mao (Renmin University of China); Weizhi Ma (Tsinghua University); Zhijing Wu (Tsinghua University); Yiqun LIU (Tsinghua University); Min Zhang (Tsinghua University); Shaoping Ma (Tsinghua University); Zhaowei
    • keywords: Neural IR; Query reformulation; Knowledge graph
    • 将knowledge information融合进query reformulation，提高检索性能
  • Learning Relevant Questions for Conversational Product Search using Reinforcement Learning；Ali Montazeralghaem (University of Massachusetts Amherst); James Allan (University of Massachusetts Amherst)
    • keywords: Conversational Product Search, Reinforcement Learning, Relevant Questions, Intelligent Assistants
    • 利用强化学习，生成对话产品搜索中所用的对话question

其他

  • Identifying Cost-effective Debunkers for Multi-stage Fake News Mitigation Campaigns;Xiaofei Xu (RMIT University); KE DENG (RMIT University); Xiuzhen Zhang (RMIT University)*

    • keywords: Fake News Mitigation, Social Network, Reinforcement Learning, Multivariate Hawkes Process
    • 通过强化学习来寻找人群中的defunker去提供正确信息，最大化提升缓解虚假信息的作用。

  • Wikipedia Reader Navigation: When Synthetic Data is Enough; Akhil Arora (EPFL); Martin Gerlach (Wikimedia Foundation); Tiziano Piccardi (EPFL); Alberto Garcia-Duran (EPFL); Robert West (EPFL)

    • keywords: Wikipedia clickstream, Wikipedia server logs, User navigation

  • Differentially Private Ensemble Classifiers for Data Streams; Lovedeep Singh Gondara (Simon Fraser University); Ke Wang (SFU); Ricardo Silva Carvalho (Simon Fraser University)

    • keywords: Differential privacy; data streams; ensembles; concept drift
    • 研究连续 data stream 上学习的差分隐私问题

  • An Adaptive Unified Allocation Framework for Guaranteed Display Advertising; Xiao Cheng (Alibaba); Chuanren Liu (University of Tennessee); Liang Dai (Alibaba); Peng Zhang (Alibaba); Zhen Fang (Alibaba); Zhonglin Zu (alibaba)

    • keywords: E-commerce Marketing; Guaranteed Display Advertising; Optimal Allocation; Computational Advertising.
    • 针对一类广告投放场景（Guaranteed Display Advertising）提出的投放算法

  • Collaborative Curating for Discovery and Expansion of Visual Clusters;Dung Duy Le (College of Engineering and Computer Science, VinUniversity); Hady Lauw (Singapore Management University)

    • keywords: Collaborative Curating; Visual Curation; Visual Discovery

  • Fast Learning of MNL Model From General Partial Rankings with Application to Network Formation Modeling; Jiaqi Ma (University of Michigan); Xingjian o Zhang (University of Michigan); Qiaozhu Mei (University of Michigan)

    • keywords: Learning to rank; multinomial logit model; Plackett-Luce model; partial ranking; network formation modeling

  • Speaker and Time-aware Joint Contextual Learning for Dialogue-act Classification in Counselling Conversations; Ganeshan Malhotra (Lcs2); Abdul Waheed (Maharaja Agrasen Institute of Technology, New Delhi); ASEEM SRIVASTAVVA (IIIT Delhi); Md Shad Akhtar (IIIT Delhi); Tanmoy Chakraborty (Indraprastha Institute of Information Technology Delhi (IIIT-D), India )

    • keywords: Dialogue-Act Classification; Mental-health Counselling
    • 本文跟对话系统比较相关。提出了一个新的心理咨询对话行为分类数据集 HOPE，并在此之上提出了一个 transformer-based 模型。

  • Towards Understanding and Answering Comparative Questions; Alexander Bondarenko (Martin-Luther-Universität Halle-Wittenberg); Yamen Ajjour (Martin-Luther-Universität Halle-Wittenberg); Valentin Dittmar (Martin-Luther-Universität Halle-Wittenberg); Niklas Homann (Martin-Luther-Universität Halle-Wittenberg);

    • keywords: Comparative questions; Question intent understanding; Comparison objects and aspects; Answer stance detection
    • 有关对比性问题回答研究，作者提出一个分类器进行对比性问题检测，以及重点语义识别。

  • MAF: A General Matching and Alignment Framework for Multimodal Named Entity Recognition; Bo Xu (Donghua University); Shizhou Huang (Donghua University); Chaofeng Sha (Fudan University); Hongya Wang (Donghua University);

    • keywords: multimodal named entity recognition; contrastive learning
    • 本文针对现有多模态实体识别模型，利用对比学习解决不同模态间的语义差异以及连续性问题。

  • Outside In: Market-aware Heterogeneous Graph Neural Network for Employee Turnover Prediction; Jinquan Hang (Peking University); Zheng Dong (Baidu); Hengshu Zhu (Baidu); Hongke Zhao (Tianjin University); Xin Song (Baidu Inc.); Peng Wang (Baidu Inc.);

    • keywords: Employee Turnover Prediction; Skill Graph; Graph Embedding Neural Network
    • 本文针对员工流动预测问题，提出从 internal graph and external graph进行考虑，从而完善现有方案的视角。

  • Crowdsourcing-based Multi-Device Communication Cooperation for Mobile High-Quality Video Enhancement; Xiaotong Wu (Nanjing Normal University); Lianyong Qi (Qufu Normal University); Xiaolong Xv (Nanjing University of Information Science and Technology); Shui Yu (University of Technology Sydney (UTS)); Wanchun Dou (Nanjing University); Xuyun Zhang (Macquarie University);

    • keywords: Mobile Videos; D2D Communication; Movement Control; Utility Optimization
    • 本文提出通过众包以促进移动控制决策的优化。

  • DAME: Domain Adaptation for Matching Entities; Mohamed Trabelsi (Lehigh University); Jeff Heflin (Lehigh University); Jin Cao (Nokia Bell Labs);

    • keywords: entity matching, transfer learning, domain adaptation
    • 本文针对实体对齐中不同domain均匀性的不合理，提出一个泛化能力更好的，迁移性能更佳的框架。

  • Knowledge Enhanced Sports Game Summarization；Jiaan Wang (Soochow University); Zhixu Li (Fudan University); Tingyi Zhang (Soochow University); Duo Zheng (Beijing University of Posts and Telecommunications); JIanfeng Qu (Soochow University); An Liu (Soochow University); Lei Zhao (Soochow University)

    • datasets, sports game summarization, text summarization
    • 一个新的用于体育比赛摘要自动生成的数据集

  • How Do You Test a Test?: A Multifaceted Examination of Significance Tests；Nicola Ferro (University of Padova); Mark Sanderson (RMIT University)

    • statistical significance testing; comparing tests; ANOVA; prediction
    • 对三种统计显著性检验方法的有效性比较

  • A Simple but Effective Bidirectional Framework for Relational Triple Extraction；Feiliang Ren (Northeastern University); Longhui Zhang (Northeastern University); Shujuan Yin (Northeastern University); Xiaofeng Zhao (Northeastern University); Shilei Liu (Northeastern University); Bochao Li (Northeastern University)

    • relational triple extraction, joint extraction of entities and relations, overlapping triple issue, bidirectional extraction framework, convergence rate inconsistency issue, share-aware learning mechanism
    • 提出了一种双向的RTE模型。RTE：relational triple extraction，旨在从非结构化文本（通常是句子）中提取三元组

  • MAVE: A Product Dataset for Multi-source Attribute Value Extraction; Li Yang (Google Research); Qifan Wang (Google Research); Zac Yu (Google); Anand Kulkarni (Google); Sumit Sanghai (Google LLC); Bin Shu (Google); Jonathan Elsas (Google); Bhargav Kanagal (Google)*

    • attribute value extraction; open tag extraction; zero-shot learning
    • 提出了一个新的用于属性抽取的数据集

  • Reconfiguration Problems on Submodular Functions; Naoto Ohsaka (NEC Corporation); Tatsuya Matsuoka (NEC Corporation)*

    • reconfiguration; submodular functions; approximation algorithms; influence maximization; determinantal point processes

  • Translating Human Mobility Forecasting through Natural Language Generation; Hao Xue (RMIT University); Flora D. Salim (RMIT University); Yongli Ren (RMIT University); Charles Clarke (University of Waterloo)*

    • temporal forecasting; natural language; human mobility prediction
    • 从语言翻译角度整合各种上下文信息(如每个兴趣地点的语义类别信息)来解决将人类迁移预测问题

  • ESC-GAN: Extending Spatial Coverage of Physical Sensors; Xiyuan Zhang (University of California, San Diego); Ranak Roy Chowdhury (University of California, San Diego); Jingbo Shang (University of California, San Diego); Rajesh Gupta (UC San Diego); Dezhi Hong (UC San Diego)*

    • keywords: Spatio-temporal data; imputation; super resolution; self-attention; generative adversarial network
    • 使用GAN做时空插补任务

  • Information Retrieval; Web Searching; Linear Bandits; Private Recommender Systems; Ryoma Sato (Kyoto University)

    • keywords：Information Retrieval; Web Searching; Linear Bandits; Private Recommender Systems
    • 文章提出一个新的任务setting：我们考虑了对图像数据库访问受限的用户如何使用他们自己的黑盒函数检索图像。

  • ConsistSum: Unsupervised Opinion Summarization with the Consistency of Aspect, Sentiment and Semantic; Wenjun Ke (Institute Of Computing Technology Chinese Academy Of Sciences); Jinhua Gao (Institute of Computing Technology, Chinese Academy of Sciences); Huawei Shen (Institute of Computing Technology, Chinese Academy of Sciences); Xueqi Cheng (Institute of Computing Technology, Chinese Academy of Sciences)*

    • keywords: opinion summarization, unsupervised method, consistency enhancement
    • 文章一种无监督的意见摘要方法，致力于捕捉评论和摘要之间的方面和情感的一致性

  • An Unsupervised Detection Framework for Chinese Jargons in Darknet; Liang Ke (Sichuan University); Xinyu Chen (Sichuan University); Haizhou Wang (Sichuan University)*

    • keywords:Jargon detection; underground economy; NLP; pretrained models
    • 我们提出了一个基于无监督学习的中文行话检测框架。 主要思想是将相似度与来自不同语料库的高维词嵌入特征进行比较以找到行话。使用了预训练模型。

  • Directed Network Embedding with Virtual Negative Edges； Hyunsik Yoo (Hanyang University); Yeon-Chang Lee (Hanyang University); Kijung Shin (KAIST); Sang-Wook Kim (Hanyang University, Korea)

    • keywords：network embedding; directed networks; virtual negative edges
    • directed network embedding希望找到一种embedding保持节点之间的不对称关系。本文提出虚拟负边的概念去表示节点之间潜在的负面关系。

  • Structure Meets Sequences: Predicting Network of Co-evolving Sequences; Yaojing Wang (NJU)*; Yuan Yao (Nanjing University); Feng Xu (Nanjing University); Yada Zhu (IBM Research); Hanghang Tong (University of Illinois at Urbana-Champaign)

    • keywords: Co-evolving sequences, sequence prediction, network structure
    • 研究相关序列的预测问题 network of co-evolving sequences (NoCES)

  • MonLAD: Money Laundering Agents Detection in Transaction Streams; Xiaobing Sun (Institute of Computing Technology, CAS, China)*; Wenjie Feng (National University of Singapore ); Shenghua Liu (Institute of Computing Technology, CAS, China); Yuyang Xie (Tsinghua University); Siddharth Bhatia (National University of Singapore)

    • keywords: Anomaly detection, Money laundering, Stream algorithm
    • 洗钱行为检测

  • Sentiment Analysis of Fashion Related Posts in Social Media; Yifei Yuan (The Chinese University of Hong Kong)*; Wai Lam (The Chinese University of Hong Kong)

    • keywords: Multimodal Sentiment Analysis; Social Media Mining; Fashion Sentiment Analysis
    • 结合图片和文本，检测社交网络贴文是否与流行物品相关

  • Learning Transferable Node Representations for Attribute Extraction from Web Documents; Yichao Zhou (UCLA)*; Ying Sheng (Google); Nguyen Vo (Google); Nick Edmonds (Google); Sandeep Tata (“Google, USA”)

    • keywords: structured data extraction, web information extraction
    • 从网页中提取关键信息表示，用于搜索、推荐等下游任务

  • A Sequence-to-Sequence Model for Large-scale Chinese Abbreviation Database Construction; Chao Wang (Fudan University)*; Jingping Liu (Fudan University); Tianyi Zhuang (Fudan university); Jiahang Li (Fudan University); Juntao Liu (Fudan University); Yanghua Xiao (Fudan University); Wei Wang (” Fudan University, China”); Rui Xie (Meituan)

    • keywords: Chinese abbreviation, Sequence-to-sequence model

  • *CMT-Net: A Mutual Transition Aware Framework for Taxicab Pick-ups and Drop-offs Co-Prediction*; Yudong Zhang (University of Science and Technology of China);Binwu Wang(University of Science and Technology of China) ; Ziyang Shan(University of Science and Technology of China); Zhengyang Zhou (University of Science and Technology of China);Yang Wang (University of Science and Technology of China);

    • Taxicab demand prediction, Spatiotemporal data mining, Cluster-based partition, Mutual transition aware
    • 设计了一种新的基于相互过渡感知的协同预测框架，来预测复杂情况下乘客上下车的时空信息，可用于提高人们对于城市人类流动性的理解，也可以用于帮助出租车的调度等场景。

  • Assessing Algorithmic Biases for Musical Version Identification; Furkan Yesiler(Pompeu Fabra University);Marius Miron(Pompeu Fabra University);Joan Serrà(Dolby Laboratories);Emilia Gómez(Joint Research Centre, European Commission)

    • information retrieval, version identification, algorithmic bias

  • Harvesting More Answer Spans from Paragraph beyond Annotation; Qiaoben Bao (Fudan University); Jiangjie Chen (Fudan University); Linfang Liu (Fudan University); Jingping Liu (East China University of Science and Technology); Jiaqing Liang (Fudan University); Yanghua Xiao (Fudan University)

    • keywords: Information extraction; Positive-unlabeled learning

  • Semi-supervised Stance Detection of Tweets Via Distant Network Supervision；Subhabrata Dutta (Jadavpur University, India); Samiya Caur (IIIT-Delhii, India); Soumen Chakrabarti (IIT Bombay, India); Tanmoy Chakraborty (IIIT-Delhi, India)

    • keywords: tweet-level stance detection, semi-supervised, distant supervision signal, homophily assumption, minority stance labels and noisy text
    • 用于推特级别数据集的半监督立场检测模型

  • Dy-HIEN: Dynamic Evolution based Deep Hierarchical Intention Network for Membership Prediction；Zhenyun Hao (Shandong University); Jianing Hao (Shandong University); Zhaohui Peng (Shandong University); Senzhang Wang (Central South University); Philip S Yu (UIC); Xue Wang (Shandong University); Jian Wang (Shandong University)

    • keywords: Membership prediction, Dynamic embedding learning, Hierarchical intention evolution

  • Keyword Assisted Embedded Topic Model；Bahareh Harandizadeh (University of Southern California); J. Hunter Priniski (University of California, Los Angeles); Fred Morstatter (University of Southern California)

    • keywords: Topic models, Guided Topic Modeling, Embedded Topic Modeling, prior knowledge, Clustering, Human-in-the-Loop and Collaborative
    • 一种新的topic model

  • Community Trend Prediction on Heterogeneous Graph in E-commerce；Jiahao Yuan (East China Normal University); Zhao Li (Alibaba Group); Pengcheng Zou (Alibaba Group); Xuan Gao (Alibaba Group); Jinwei Pan (East China Normal University); Wendi Ji (East China Normal University); Xiaoling Wang (East China Normal University)

    • keywords: community trend, heterogeneous graph, e-commerce, dynamic evolution
    • 一种community trend预测模型，主要针对attribute tags进行预测

  • A Personalized Cross-Platform Post Style Transfer Method Based on Transformer and Bi-Attention Mechanism；Zhuo Chen (Fudan University); Baoxi Liu (Fudan University); Peng Zhang (Fudan University); Tun Lu (Fudan University); Ning Gu (Fudan University)

    • keywords: Cross-Platform Content Sharing; Text Style Transfer
    • 研究分享内容在不同平台之间的文本风格迁移

  • Time Masking for Temporal Language Models；Guy D. Rosin (Technion); Ido Guy (eBay); Kira Radinsky (Ebay)

    • keywords: temporal semantics; semantic change detection; language models
    • 提出TempoBERT模型，利用time-masking将时序信息加入contextual language model

  • MTLVS: A Multi-Task Framework to Verify and Summarize Crisis-Related Microblogs；Rajdeep Mukherjee (IIT Kharagpur); Uppada Vishnu (IIT Kharagpur); Chandana Peruri (IIT Kharagpur); Sourangshu Bhattacharya (IIT KHARAGPUR); niloy ganguly (IIT Kharagpur); Pawan Goyal ( Indian Institute of Technology Kharagpur); Koustav Rudra (Leibniz University Hannover)

    • keywords: Trustworthy Summarization, Rumour Detection, Disaster, Twitter
    • 研究对于Twitter级disaster-specific summarizers的谣言检测工作