© 作者|范欣妍
机构|中国人民大学高瓴人工智能学院
研究方向 | 推荐系统
文章来源 | RUC AI Box
导读
ACM SIGIR 2021是CCF A类会议,人工智能领域智能信息检索( Information Retrieval,IR)方向最权威的国际会议。会议专注于信息的存储、检索和传播等各个方面,包括研究战略、输出方案和系统评估等等。第44届国际计算机学会信息检索大会(The 44rd International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2021)计划于今年7月11日-7月15日以线上会议形式召开。这次会议共收到720篇长文和526篇短文投稿,有151篇长文和145篇短文被录用,录用率约为21%和27%。官方发布的接收论文列表:https://sigir.org/sigir2021/accepted-papers/
本文选取了SIGIR 2021中86篇长文,44篇短文,重点对推荐系统相关论文(87篇)按不同的任务场景和研究话题进行分类整理,也对其他热门研究方向(问答、对话、知识图谱等,43篇)进行了归类,以供参考。
从词云图看今年SIGIR的研究热点:根据长文和短文的标题绘制如下词云图,可以看到今年研究方向主要集中在Recommendation和Retrieval两个方向,也包括Summarization、Conversations等NLP方向;主要任务包括:Ranking、Cross-domain、Multi-Model/Behavior、Few-Shot、User modeling、Personalization等;热门技术包括:Neural Networks、Knowledge Graph、GNN、Attention、Meta Learning等,其中基于Graph的一类方法依旧是今年的研究热点。
本文目录
· 1 按推荐的任务场景划分·
CTR
Collaborative Filtering
Sequential/Session-based Recommendations
Conversational Recommender System
News Recommendations
Cross-domain/Multi-behavior Recommendations
Social Recommendation
基于热门技术
Graph-based
VAE-based
GAN-based
FM-based
Other Tasks
· 2 按推荐的研究话题划分·
Debias in Recommender System
Fairness in Recommender System
Explanation in Recommender System
Long-tail/Cold-start in Recommender System
Attack in Recommender System
Diversity in Recommender System
Ranking in Recommender System
Evaluation
Others
· 3 其他研究方向·
QA
Conversations
Summarization
Generation
Knowledge Graph
Graph Networks
Representation Learning
Multi-Modality
01
按推荐任务场景划分
1.1 CTR
Looking at CTR Prediction Again: Is Attention All You Need?【CTR中Attention是否必要?】
ScaleFreeCTR: Mix Cache-based Distributed Training System for CTR Models with Huge Embedding Table【Embedding Layer通常占很多参数,对于数量巨大的Embedding Table如何训练CTR模型】
A General Method For Automatic Discovery of Powerful Interactions In Click-Through Rate Prediction【自动挖掘有用Interaction的通用方法】
Learning Graph Meta Embeddings for Cold-Start Ads in Click-Through Rate Prediction【冷启动时如何学好图上的元嵌入表示】
GemNN: Gating-enhanced Multi-task Neural Networks with Feature Interaction Learning for CTR Prediction【short paper,CTR中的多任务】
Deep Position-wise Interaction Network For CTR Prediction【short paper】
Deep User Match Network for Click-Through Rate Prediction【short paper,基于用户匹配网络】
Explicit Semantic Cross Feature Learning via Pre-trained Graph Neural Networks for CTR Prediction【short paper,预训练GNN】
RLNF: Reinforcement Learning based Noise Filtering for Click-Through Rate Prediction【short paper,基于强化学习过滤噪声】
1.2 Collaborative Filtering
Bootstrapping User and Item Representations for One-Class Collaborative Filtering
When and Whom to Collaborate with in a Changing Environment: A Collaborative Dynamic Bandit Solution【基于老虎机方法根据环境变化动态调整】
Neural Graph Matching based Collaborative Filtering【基于图匹配网络的CF】
Enhanced Graph Learning for Collaborative Filtering via Mutual Information Maximization【通过MIM来增强图学习】
Set2setRank: Collaborative Set to Set Ranking for Implicit Feedback based Recommendation【集合到集合的排序推荐】
1.3 Sequential/Session-based Recommendations
Category-aware Collaborative Sequential Recommendation【类别感知的SR】
Sequential Recommendation with Graph Convolutional Networks【GCN做SR】
Dual Attention Transfer in Session-based Recommendation with Multi Dimensional Integration【双重Attention迁移】
Unsupervised Proxy Selection for Session-based Recommender Systems【无监督代理选择方法】
StackRec: Efficient Training of Very Deep Sequential Recommender Models by Iterative Stacking【针对非常深的网络的更有效率训练方法】
Counterfactual Data-Augmented Sequential Recommendation【反事实数据增强】
CauseRec: Counterfactual User Sequence Synthesis for Sequential Recommendation【反事实用户序列合成】
Counterfactual Reward Modification for Streaming Recommendation with Delayed Feedback【流推荐中反事实奖励修改】
Package Recommendation with Intra- and Inter-Package Attention Networks【考虑包内和包间关系的package推荐】
The World is Binary: Contrastive Learning for Denoising Next Basket Recommendation【对比学习做SR】
Motif-aware Sequential Recommendation【short paper,基于主题的SR】
Lighter and Better: Low-Rank Decomposed Self-Attention Networks for Next-Item Recommendation【short paper,通过低秩分解自注意力网络,进行模型轻量化】
Augmenting Sequential Recommendation with Pseudo-Prior Items via Reversely Pre-training Transformer【short paper】
ICAI-SR: Item Categorical Attribute Integrated Sequential Recommendation【short paper】
Temporal Augmented Graph Neural Networks for Session-Based Recommendations【short paper,时序增强的GNN做SR】
1.4 Conversational Recommender System
Unified Conversational Recommendation Policy Learning via Graph-based Reinforcement Learning【基于图的强化学习方法】
Learning to Ask Appropriate Questions in Conversational Recommendation【学会提问】
Comparison-based Conversational Recommender System with Relative Bandit Feedback【基于老虎机的反馈】
1.5 News Recommendations
Personalized News Recommendation with Knowledge-aware News Interactions【引入新闻交互知识做个性化新闻推荐】
Joint Knowledge Pruning and Recurrent Graph Convolution for News Recommendation【知识剪枝与递归图卷积】
Empowering News Recommendation with Pre-trained Language Models【short paper,预训练语言模型引入到新闻推荐】
AMM: Attentive Multi-field Matching for News Recommendation【short paper,多领域匹配的新闻推荐】
RMBERT: News Recommendation via Recurrent Reasoning Memory Network over BERT【short paper,递归推理记忆网络】
1.6 Cross-domain/Multi-behavior Recommendations
Federated Collaborative Transfer for Cross-Domain Recommendation【跨领域推荐中的联合协同迁移】
Learning Domain Semantics and Cross-Domain Correlations for Paper Recommendation【通过领域语义和跨领域关联做论文推荐】
Graph Meta Network for Multi-Behavior Recommendation with Interaction Heterogeneity and Diversity【基于图元网络的多行为推荐】
Transfer-Meta Framework for Cross-domain Recommendation to Cold-Start Users【short paper,跨领域推荐中的冷启动】
1.7 Social Recommendation
Social Recommendation with Implicit Social Influence【short paper】
ConsisRec: Enhancing GNN for Social Recommendation via Consistent Neighbor Aggregation【short paper,通过一致用户增强的GNN】
1.8 基于热门技术
1.8.1 Graph-based
Structured Graph Convolutional Networks with Stochastic Masks for Recommender Systems【结构化GCN】
Self-supervised Graph Learning for Recommendation【推荐中的自监督图学习】
DEKR: Description Enhanced Knowledge Graph for Machine Learning Method Recommendation【增强知识图谱】
1.8.2 VAE-based
Variational Autoencoders for Top-K Recommendation with Implicit Feedback【short paper,VAE对Top-K推荐】
Bayesian Critiquing with Keyphrase Activation Vectors for VAE-based Recommender Systems【short paper,基于关键词激活向量的贝叶斯评判】
1.8.3 GAN-based
Info-flow Enhanced GANs for Recommender【short paper】
1.8.4 FM-based
xLightFM: Extremely Memory-Efficient Factorization Machine【高效Memory的因子分解机模型】
1.9 Other Tasks
Clicks can be Cheating: Counterfactual Recommendation for Mitigating Clickbait Issue【反事实推荐中减轻点击诱饵的问题】
UGRec: Modeling Directed and Undirected Relations for Recommendation【在推荐中建立有向和无向关系】
Learning Recommender Systems with Implicit Feedback via Soft Target Enhancement【软目标增强的方法】
PreSizE: Predicting Size in E-Commerce using Transformers【使用Transformer预测size(尺码?)】
Did you buy it already? Detecting Users Purchase-State From Their Product-Related Questions【通过产品相关问题检测用户购买状态,新颖性推荐】
Path-based Deep Network for Candidate Item Matching in Recommenders【基于路径的深度网络做候选item匹配】
How Powerful are Interest Diffusion on Purchasing Prediction: A Case Study of Taocode【基于Taocode的购买预测分析】
A Study of Defensive Methods to Protect Visual Recommendation Against Adversarial Manipulation of Images【可视化推荐】
02
按推荐研究话题划分
2.1 Debias in Recommender System
AutoDebias: Learning to Debias for Recommendation【自动debias】
Mitigating Sentiment Bias for Recommender Systems【消除情感偏差】
Causal Intervention for Leveraging Popularity Bias in Recommendation【流行度偏差】
Enhanced Doubly Robust Learning for Debiasing Post-Click Conversion Rate Estimation
Dual Unbiased Recommender Learning for Implicit Feedback【short paper,双重无偏的推荐学习】
2.2 Fairness in Recommender System
Personalized Counterfactual Fairness in Recommendation【个性化反事实公平性】
TFROM: A Two-sided Fairness-Aware Recommendation Model for Both Customers and Providers【对消费者和商家都公平的推荐模型】
The Winner Takes it All: Geographic Imbalance and Provider (Un)fairness in Educational Recommender Systems【short paper,教育推荐系统中的地域不平衡】
2.3 Explanation in Recommender System
On Interpretation and Measurement of Soft Attributes for Recommendation【对推荐中软属性的解释和测量】
ReXPlug: Explainable Recommendation using Plug-and-Play Language Model【使用即插即用语言模型做可解释性推荐】
User-Centric Path Reasoning towards Explainable Recommendation
Counterfactual Explanations for Neural Recommenders【short paper,反事实解释】
2.4 Long-tail/Cold-start in Recommender System
Learning Graph Meta Embeddings for Cold-Start Ads in Click-Through Rate Prediction【学习图元嵌入表示】
FORM: Follow the Online Regularized Meta-Leader for Cold-Start Recommendation【在线正则化的元学习器】
Privileged Graph Distillation for Cold-start Recommendation【优先图蒸馏】
Learning to Warm Up Cold Item Embeddings for Cold-start Recommendation with Meta Scaling and Shifting Networks【如何warm up】
Fairness among New Items in Cold Start Recommender Systems【冷启动中新物品的公平性比较】
Long-Tail Hashing【长尾Hash】
Transfer-Meta Framework for Cross-domain Recommendation to Cold-Start Users【short paper】
Cluster-Based Bandits: Fast Cold-Start for Recommender System New Users【short paper,基于聚类的老虎机模型来快速冷启动】
Sequential Recommendation for Cold-start Users with Meta Transitional Learning【short paper,元过渡学习】
Decoupling Representation and Regressor for Long-Tailed Information Cascade Prediction【short paper,解耦表示和回归方法】
2.5 Attack in Recommender System
Fight Fire with Fire: Towards Robust Recommender Systems via Adversarial Poisoning Training【以牙还牙,用有毒的对抗网络增强模型鲁棒性】
2.6 Diversity in Recommender System
Enhancing Domain-Level and User-Level Adaptivity in Diversified Recommendation【从领域层面和用户层面增强推荐多样性】
2.7 Ranking in Recommender System
New Insights into Metric Optimization for Ranking-based Recommendation【排序类推荐模型中的指标优化】
2.8 Evaluation
Standing in Your Shoes: External Assessments for Personalized Recommender Systems【外部评估】
Evaluation measures based on preference graphs【基于偏好图的评估】
2.9 Others
A Guided Learning Approach for Item Recommendation via Surrogate Loss Learning【替代损失学习】
Neural Representations in Hybrid Recommender Systems: Prediction versus Regularization【short paper】
Underestimation Refinement: A General Enhancement Strategy for Exploration in Recommendation Systems【short paper,推荐中的探索问题】
Cross-Batch Negative Sampling for Training Two-Tower Recommenders【short paper,双塔模型中的负采样问题】
03
其他研究方向
3.1 QA
Reinforcement Learning from Reformulations in Conversational Question Answering over Knowledge Graphs【强化学习+KG】
Answer Complex Questions: Path Ranker Is All You Need【路径排序】
IDRQA: Iterative Document Reranking for Open-domain Multi-hop Question Answering【开放域多跳问答】
LPF: A Language-Prior Feedback Objective Function for De-biased Visual Question Answering【short paper】
3.2 Conversations
Initiative-Aware Self-Supervised learning for Knowledge-Grounded Conversations【自监督学习对知识型对话】
Conversations Powered by Cross-Lingual Knowledge【跨语言知识】
Partner Matters! An Empirical Study on Fusing Personas for Personalized Response Selection in Retrieval-Based Chatbots【个性化回复的选择很重要】
One Chatbot Per Person: Creating Personalized Chatbots based on Implicit User Profiles【针对用户画像建立个性化对话机器人】
3.3 Summarization
Multi-Modal Supplementary-Complementary Summarization using Multi-Objective Optimization【多模型下的补充摘要】
DeepQAMVS: Query-Aware Hierarchical Pointer Networks for Multi-Video Summarization【指针网络,多视频摘要】
Transformer Reasoning Network for Personalized Review Summarization【基于Transformer推理网络的评论摘要】
Leveraging Lead Bias for Zero-shot Abstractive News Summarization【zero-shot的新闻摘要】
DSGPT: Domain-Specific Generative Pre-Training of Transformers for Text Generation in E-commerce Title and Review Summarization【short paper,在电商领域标题和评论摘要中的文本生成】
Unsupervised Extractive Text Summarization with Distance-Augmented Sentence Graphs【short paper,无监督抽取文本摘要】
3.4 Generation
Multi-type Textual Reasoning for Product-aware Answer Generation【回复生成】
Knowledge-based Review Generation by Coherence Enhanced Text Planning【评论生成】
3.5 Knowledge Graph
Relational Learning with Gated and Attentive Neighbor Aggregator for Few-Shot Knowledge Graph Completion【few-shot的知识图谱补全】
TIE: A Framework for Embedding-based Incremental Temporal Knowledge Graph Completion【时序知识图谱补全】
MetaP: Meta Pattern Learning for One-Shot Knowledge Graph Completion【short paper,one-shot的知识图谱补全】
Knowledge Graph Embedding via Metagraph Learning【short paper,知识图谱嵌入表示】
DOZEN: Cross-Domain Zero Shot Named Entity Recognition with Knowledge Graph【short paper,跨领域zero-shot命名实体识别】
3.6 Graph Networks
Graph Similarity Computation via Differentiable Optimal Assignment【图相似度计算】
Should Graph Convolution Trust Neighbors? A Simple Causal Inference Method【图卷积中邻居的可靠性】
A Graph-Convolutional Ranking Approach to Leverage the Relational Aspects of User-Generated Content【图卷积排序方法】
Retrieving Complex Tables with Multi-Granular Graph Representation Learning【多粒度图表示学习】
Meta-Inductive Node Classification across Graphs【节点分类】
WGCN: Graph Convolutional Networks with Weighted Structural Features【带权重结构特征的GCN】
3.7 Representation Learning
Temporal Knowledge Graph Reasoning Based on Evolutional Representation Learning【进化表征学习】
Legal Judgment Prediction with Multi-Stage Case Representation Learning in the Real Court Setting【法院判决里的表征学习】
One Person, One Model, One World: Learning Lifelong User Representation without Forgetting【一个user表示用一生】
Inductive Representation Learning in Temporal Networks via Mining Neighborhood and Community Influences【short paper,时序网络中的表征学习】
Enhanced Representation Learning for Examination Papers with Hierarchical Document Structure【short paper,分层文档结构中的表示学习】
3.8 Multi-Modality
GilBERT: Generative Vision-Language Pre-Training for Modality-Incomplete Visual-Linguistic Tasks【视觉到语言的跨模态】
Hierarchical Cross-Modal Graph Consistency Learning for Video-Text Retrieval【视频-文字检索】
????
现在,在「知乎」也能找到我们了
进入知乎首页搜索「PaperWeekly」
点击「关注」订阅我们的专栏吧
关于PaperWeekly
PaperWeekly 是一个推荐、解读、讨论、报道人工智能前沿论文成果的学术平台。如果你研究或从事 AI 领域,欢迎在公众号后台点击「交流群」,小助手将把你带入 PaperWeekly 的交流群里。