Research map / Label-efficient and robust learning / Recognition and long-tail
Rethinking Collaborative Metric Learning: Toward an Efficient Alternative Without Negative Sampling
Shilong Bao, Qianqian Xu, Zhiyong Yang, Xiaochun Cao, Qingming Huang
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 |
|---|---|---|---|
| replace uniform negative sampling with sampling-free full-sample training | CML · MovieLens-100k | NDCG@5: 17.40 -> 24.65 | Improved |
| replace uniform negative sampling with sampling-free full-sample training | CML · CiteULike | MAP: 4.67 -> 6.70 | Improved |
| replace uniform negative sampling with sampling-free full-sample training | CML · CiteULike | AUC: 86.62 -> 83.44 | Hurt |
| replace popularity negative sampling with sampling-free training | CML · CiteULike | NDCG@5: 9.72 -> 10.29 | Improved |
| replace popularity negative sampling with sampling-free training | CML · MovieLens-100k | NDCG@5: 13.93 -> 24.65 | Improved |
| replace hard negative sampling with sampling-free training | CML · MovieLens-100k | MAP: 15.94 -> 18.00 | Improved |
| use SFCML instead of sampling-free neural EHCF | SFCML vs EHCF · CiteULike | NDCG@5: 6.29 -> 10.29 | Improved |
| use fast graph-based SFCML instead of naive full-sample CML | SFCML · CiteULike | all metrics | Improved |
| use graph-accelerated SFCML instead of naive full-sample CML | SFCML · MovieLens-100k | MAP: 17.62 -> 18.00 | No clear effect |
| replace sampling-based translation CML (CRML) with SFCML | CRML · MovieLens-100k | MAP: 16.33 -> 18.00 | Improved |
Closest papers by meaning
- Improved Diversity-Promoting Collaborative Metric Learning for Recommendation (TPAMI 2024 · similarity 0.93)
- The Minority Matters: A Diversity-Promoting Collaborative Metric Learning Algorithm (NeurIPS 2022 · similarity 0.89)
- Adaptive Hardness Negative Sampling for Collaborative Filtering (AAAI 2024 · similarity 0.86)
- Learning-Efficient Yet Generalizable Collaborative Filtering for Item Recommendation (ICML 2024 · similarity 0.85)
- From Pairwise to Ranking: Climbing the Ladder to Ideal Collaborative Filtering with Pseudo-Ranking (AAAI 2025 · similarity 0.84)
- PSL: Rethinking and Improving Softmax Loss from Pairwise Perspective for Recommendation (NeurIPS 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.
