넓고 얕게 GNN 기반 추천 훑어보기
Keword: Why GNN?, Survey, GCN, CF, Inductive Learning, Web-Scale, Graph Attention, Knowledge Graph
Table of Contents
- Schedule
- 1. OT + Understanding Convolutions on Graphs
- 2. Why GNN for RS? & Survey
- 3. Neural Graph Collaborative Filtering
- 4. LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation
- 5. Inductive Representation Learning on Large Graphs
- 6. Graph Convolutional Neural Networks for Web-Scale Recommender Systems
- 7. Graph Attention Networks
- 8. KGAT: Knowledge Graph Attention Network for Recommendation
- 9. Cross Domain & Session based Recommendation
- 10. Pytorch Geometric & Bundle Recommendation
진행방식
- Resource를 각자 보고 옵니다.
- 사전에 질문이나 이야기해볼 거리를 Issue에 남깁니다.
- 진행 시간은 1시간을 기본으로 합니다.
-
Learning and Reasoning on Graph for Recommendation
- Part1 까지
-
Graph Neural Networks for Recommender Systems: Challenges, Methods, and Directions
- Paper, 인용수 313이상, SIGIR'20
- Resource, Slide
- Code: torch
- 추가자료
- UltraGCN: Ultra Simplification of Graph Convolutional Networks for Recommendation
- GNN으로 Cross-Domain 맛보기
- GNN in Session-based Recommendation