Research map / Label-efficient and robust learning / Recognition and long-tail
Trading Personalization for Accuracy: Data Debugging in Collaborative Filtering
Long Chen, Yuan Yao, Feng Xu, Miao Xu, Hanghang Tong
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 |
|---|---|---|---|
| modify top 10% maverick ratings with CFDEBUG | matrix factorization CF · MovieLens | RMSE: 0.9134 -> 0.8785 | Improved |
| modify top 10% maverick ratings with CFDEBUG | matrix factorization CF · Douban | RMSE: 0.8751 -> 0.8247 | Improved |
| use incremental update instead of full retraining (ablation) | CFDEBUG · MovieLens and Douban | RMSE | No clear effect |
| replace CFDEBUG detection with eMF at 10% modify | CF · MovieLens | RMSE: 0.9134 -> 0.9246 | Hurt |
| replace CFDEBUG detection with NrMF at 10% modify | CF · MovieLens | RMSE: 0.9134 -> 0.9495 | Hurt |
| add NoiseCorrection at 10% delete | CF · Douban | RMSE: 0.8751 -> 0.8758 | No clear effect |
| apply CFDEBUG modify with 5% max-injected noisy ratings | CF · MovieLens injected noise | RMSE: 0.9495 -> 0.8946 | Improved |
| apply CFDEBUG modify with 2% random-injected noisy ratings | CF · MovieLens injected noise | RMSE: 0.9191 -> 0.8772 | Improved |
| apply CFDEBUG delete with 5% min-injected noisy ratings | CF · MovieLens injected noise | RMSE: 0.9469 -> 0.9169 | Improved |
| order candidates randomly (Random + CFDEBUG) instead of CFDEBUG ordering (ablation) | CFDEBUG · MovieLens | RMSE | Hurt |
| order candidates by distance to item average instead of CFDEBUG ordering (ablation) | CFDEBUG · MovieLens | RMSE | Hurt |
| replace modified ratings with item average instead of CFDEBUG values (ablation) | CFDEBUG · MovieLens | RMSE | Hurt |
Closest papers by meaning
- Echoes in Filter Bubble: Diagnosing and Curing Popularity Bias in Generative Recommender Systems (NeurIPS 2026 · similarity 0.85)
- Preference Diffusion for Recommendation (ICLR 2025 · similarity 0.85)
- Addressing Hidden Confounding with Heterogeneous Observational Datasets for Recommendation (NeurIPS 2024 · similarity 0.85)
- Relaxing the Accurate Imputation Assumption in Doubly Robust Learning for Debiased Collaborative Filtering (ICML 2024 · similarity 0.85)
- Building Higher-Order Abstractions from the Components of Recommender Systems (AAAI 2024 · similarity 0.84)
- Debiased Collaborative Filtering with Kernel-Based Causal Balancing (ICLR 2024 · similarity 0.84)
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.
