国际顶级学术会议WWW2020定在2020年4月20-24日于中国台湾举办。受COVID-19疫情影响(疫情赶紧过去吧)，大会将在线上举行。今天是大会开始的第一天。
本次会议共收到了1129篇论文投稿，录用217篇，录取率仅为19.2%。其中关于推荐系统的论文大约38篇，推荐系统占比17.5%，可见推荐系统的研究受到学术界的广泛关注。另外，值得注意的是，接收的推荐系统论文中大部分都是与工业界合作的产物，因此不管是学术界还是工业界，推荐系统都是研究的热点与重点。
针对这38篇论文，我们进行了梳理分类，如下表所示
分类 | 数量 |
|---|---|
Practical RS | 6 |
Sequential RS | 6 |
Efficient RS | 4 |
Social RS | 3 |
General RS | 3 |
RL for RS | 3 |
POI RS | 2 |
Cold Start in RS | 2 |
Security RS | 2 |
Fairness RS | 2 |
Explianability for RS | 2 |
Cross-domain RS | 1 |
Knowledge Graph RS | 1 |
Conversational RS | 1 |
CTR for RS | 1 |
可见，推荐系统应用的文章以及序列化推荐的文章占比较大；随后是提升推荐效率、社会化推荐、常规推荐以及利用强化学习推荐；其次是兴趣点推荐、冷启动问题研究、推荐系统中的安全性、推荐公平性以及可解释推荐的文章；最后是各有一篇跨域推荐、利用知识图推荐、对话推荐系统以及用于点击率预估的推荐。
2 论文列表
1
Practical RS
- Graph Enhanced Representation Learning for News Recommendation
- Weakly Supervised Attention for Hashtag Recommendation using Graph Data
- Personalized Employee Training Course Recommendation with Career Development Awareness
- Understanding User Behavior For Document Recommendation
- Recommending Themes for Ad Creative Design via Visual-Linguistic Representations
- paper2repo: GitHub Repository Recommendation for Academic Papers
2
Sequential RS
- Adaptive Hierarchical Translation-based Sequential Recommendation
- Attentive Sequential Model of Latent Intent for Next Item Recommendation
- Déjà vu: A Contextualized Temporal Attention Mechanism for Sequential Recommendation
- Intention Modeling from Ordered and Unordered Facets for Sequential Recommendation
- Future Data Helps Training: Modeling Future Contexts for Session-based Recommendation
- Keywords Generation Improves E-Commerce Session-based Recommendation
3
Efficient RS
- Learning to Hash with Graph Neural Networks for Recommender Systems
- LightRec: a Memory and Search-Efficient Recommender System
- A Generalized and Fast-converging Non-negative Latent Factor Model for Predicting User Preferences in Recommender Systems
- Efficient Non-Sampling Factorization Machines for Optimal Context-Aware Recommendation
4
Social RS
- Clustering and Constructing User Coresets to Accelerate Large-scale Top-K Recommender Systems
- The Structure of Social Influence in Recommender Networks
- Few-Shot Learning for New User Recommendation in Location-based Social Networks
5
Explainability for RS
- Directional and Explainable Serendipity Recommendation
- Dual Learning for Explainable Recommendation: Towards Unifying User Preference Prediction and Review Generation
6
POI RS
- Next Point-of-Interest Recommendation on Resource-Constrained Mobile Devices
- A Category-Aware Deep Model for Successive POI Recommendation on Sparse Check-in Data
7
General RS
- Efficient Neural Interaction Function Search for Collaborative Filtering
- Learning the Structure of Auto-Encoding Recommenders
- Deep Global and Local Generative Model for Recommendation
8
Fairness in RS
- Hierarchical Visual-aware Minimax Ranking Based on Co-purchase Data for Personalized Recommendation
- FairRec: Two-Sided Fairness for Personalized Recommendations in Two-Sided Platforms
9
RL for RS
- Off-policy Learning in Two-stage Recommender Systems
- Hierarchical Adaptive Contextual Bandits for Resource Constraint based Recommendation
10
Cross-domain RS
- Exploiting Aesthetic Preference in Deep Cross Networks for Cross-domain Recommendation
11
Knowledge Graph RS
- Reinforced Negative Sampling over Knowledge Graph for Recommendation
12
Conversational RS
- Latent Linear Critiquing for Conversational Recommender Systems
13
CTR for RS
- Adversarial Multimodal Representation Learning for Click-Through Rate Prediction
3 官方Tutorial
最后，WWW2020还进行了两场关于推荐与搜索的Tutorial，分别是利用深度迁移学习的搜索与推荐和可信任的推荐与搜索系统，感兴趣的小伙伴可以学习一下。
