Research map / Graph learning / Graph self-supervised learning
Mitigating the Popularity Bias of Graph Collaborative Filtering: A Dimensional Collapse Perspective
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| replace uniformity loss with LogDet decorrelation penalty (ablation) | GCF LightGCN · Yelp2018 | Recall@20: 0.0713 -> 0.0725 | Helped |
| replace uniformity loss with LogDet decorrelation penalty (ablation) | GCF LightGCN · Amazon-Kindle | Recall@20: 0.2107 -> 0.2151 | Helped |
| replace uniformity loss with LogDet decorrelation penalty (ablation) | GCF LightGCN · Alibaba-iFashion | Recall@20: 0.1136 -> 0.1210 | Helped |
| replace Frobenius soft decorrelation with LogDet penalty (ablation) | GCF LightGCN · Amazon-Kindle | Recall@20: 0.1982 -> 0.2151 | Helped |
| replace Stiefel hard constraint with LogDet penalty (ablation) | GCF LightGCN · Alibaba-iFashion | Recall@20: 0.1071 -> 0.1210 | Helped |
| use LogDet penalty with MLP encoder instead of Frobenius soft penalty (ablation) | MLP encoder · Yelp2018 | Recall@20: 0.0699 -> 0.0721 | Helped |
| add LogDet decorrelation to LightGCN vs plain LightGCN (3 layers) | GCFlogdet · Alibaba-iFashion | Recall@20: 0.1053 -> 0.1210 | Helped |
| use LogDet vs best contrastive baseline SimGCL | GCFlogdet · Yelp2018 | Recall@20: 0.0721 -> 0.0725 | Helped |
| use LogDet penalty for unpopular items | GCFlogdet · Yelp2018 unpopular items | relative Recall@20 improvement over LightGCN: 1.51 | Helped |
Closest papers by meaning
- Hgformer: Hyperbolic Graph Transformer for Collaborative Filtering (ICML 2025 · similarity 0.89)
- Learning-Efficient Yet Generalizable Collaborative Filtering for Item Recommendation (ICML 2024 · similarity 0.87)
- On the Embedding Collapse when Scaling up Recommendation Models (ICML 2024 · similarity 0.86)
- Recommendations with Sparse Comparison Data: Provably Fast Convergence for Nonconvex Matrix Factorization (ICML 2025 · similarity 0.86)
- FACE: A General Framework for Mapping Collaborative Filtering Embeddings into LLM Tokens (NeurIPS 2025 · similarity 0.86)
- How Does Message Passing Improve Collaborative Filtering? (NeurIPS 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.