Research map / Graph learning / Knowledge and temporal graphs
Knowledge-aware Coupled Graph Neural Network for Social Recommendation
Chao Huang, Huance Xu, Yong Xu, Peng Dai, Lianghao Xia, Mengyin Lu, Liefeng Bo, Hao Xing, Xiaoping Lai, Yanfang Ye
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Jointly encode social relations, item dependencies, multi-typed interactions, and temporal context | KCGN · Epinions, Yelp, and E-commerce interaction prediction | HR@10 and NDCG@10 | Improved |
| remove multi-typed interaction modeling (KCGN-M) (ablation) | KCGN · Recommendation benchmarks in Figure: ablation studies | HR@10 and NDCG@10 | Hurt |
| remove the social relation encoder (KCGN-U) (ablation) | KCGN · Recommendation benchmarks in Figure: ablation studies | HR@10 and NDCG@10 | Hurt |
| remove item dependency knowledge (KCGN-I) (ablation) | KCGN · Recommendation benchmarks in Figure: ablation studies | HR@10 and NDCG@10 | Hurt |
| remove both user- and item-wise relation encoders and coupled mutual-information learning (KCGN-UI) (ablation) | KCGN · Recommendation benchmarks in Figure: ablation studies | HR@10 and NDCG@10 | Hurt |
| remove temporal context encoding (KCGN-T) (ablation) | KCGN · Recommendation benchmarks in Figure: ablation studies | HR@10 and NDCG@10 | Hurt |
| increase propagation depth from 0 or 1 layer to 2 graph layers (ablation) | KCGN · Interaction prediction; datasets shown in hyper-parameter study | accuracy | Improved |
| increase propagation depth from 2 to 3 graph layers (ablation) | KCGN · Interaction prediction; datasets shown in hyper-parameter study | accuracy | Hurt |
| increase embedding dimension initially (ablation) | KCGN · Interaction prediction; datasets shown in hyper-parameter study | accuracy | Improved |
| further increase embedding dimension (ablation) | KCGN · Interaction prediction; datasets shown in hyper-parameter study | accuracy | Hurt |
Closest papers by meaning
- ContextGNN: Beyond Two-Tower Recommendation Systems (ICLR 2025 · similarity 0.88)
- Mind Individual Information! Principal Graph Learning for Multimedia Recommendation (AAAI 2025 · similarity 0.88)
- Bridging the User-side Knowledge Gap in Knowledge-aware Recommendations with Large Language Models (AAAI 2025 · similarity 0.88)
- Semantic Enhanced Heterogeneous Hypergraph Network for Collaborative Filtering (AAAI 2025 · similarity 0.87)
- LS-TGNN: Long and Short-Term Temporal Graph Neural Network for Session-Based Recommendation (AAAI 2025 · similarity 0.87)
- Tail-STEAK: Improve Friend Recommendation for Tail Users via Self-Training Enhanced Knowledge Distillation (AAAI 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
