Research map / Label-efficient and robust learning / Self-supervised and contrastive
Debiased Contrastive Learning
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 |
|---|---|---|---|
| Debias negative sampling with τ+ = 0.01 | CMC · ImageNet-100 | Top-1: 73.58 -> 74.6 | Helped |
| Debias negative sampling with τ+ = 0.01 | CMC · ImageNet-100 | Top-5: 92.06 -> 92.08 | Helped |
| Debias negative sampling | QT · CR | accuracy: 81.3 -> 82.9 | Helped |
| Debias negative sampling | QT · SUBJ | accuracy: 86.6 -> 86.9 | Helped |
| Debias negative sampling | QT · MPQA | accuracy: 93.4 -> 93.7 | Helped |
| Debias negative sampling | QT · MSRP | accuracy: 73.6 -> 74.7 | Helped |
| Debias negative sampling | QT · MSRP | F1: 81.8 -> 82.7 | Helped |
| Debias negative sampling | QT · MR | accuracy: 76.8 -> 76.5 | Hurt |
| Debias negative sampling | QT · TREC | accuracy: 89.8 -> 89.1 | Hurt |
| Debias negative sampling | CURL · Finger Spin | score: 310±33 -> 381±24 | Helped |
| Debias negative sampling | CURL · Cartpole Swingup | score: 850±20 -> 866±7 | Helped |
| Debias negative sampling | CURL · Reacher Easy | score: 918±96 -> 927±99 | Helped |
Closest papers by meaning
- Difficulty-Based Sampling for Debiased Contrastive Representation Learning (CVPR 2023 · similarity 0.92)
- Statistical Consistency and Generalization of Contrastive Representation Learning (ICML 2026 · similarity 0.90)
- Self-Supervised Contrastive Learning is Approximately Supervised Contrastive Learning (NeurIPS 2025 · similarity 0.90)
- $\mathbb{X}$-Sample Contrastive Loss: Improving Contrastive Learning with Sample Similarity Graphs (ICLR 2025 · similarity 0.89)
- Weakly-Supervised Contrastive Learning for Imprecise Class Labels (ICML 2025 · similarity 0.88)
- Your contrastive learning problem is secretly a distribution alignment problem (NeurIPS 2024 · similarity 0.88)
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.