Research map / Label-efficient and robust learning / Noisy labels
RML++: Regroup Median Loss for Combating Label Noise
Fengpeng Li, Kemou Li, Qizhou Wang, Bo Han, Jinyu Tian, Jiantao Zhou
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 |
|---|---|---|---|
| add loss processing for sample selection (ablation) | RML · CIFAR-10 instance-0.4 | accuracy (%): 88.47 -> 90.73 | Improved |
| add loss processing for sample selection (ablation) | RML · CIFAR-100 instance-0.4 | accuracy (%): 65.43 -> 71.12 | Improved |
| add loss processing for sample selection (ablation) | RML · CIFAR-100 pairflip-0.45 | accuracy (%): 56.36 -> 64.67 | Improved |
| use median operation instead of mean loss (ablation) | RML · CIFAR-10 instance-0.4 | accuracy (%): 85.32 -> 90.73 | Improved |
| increase median window size n from 2 to 6 (ablation) | RML · CIFAR-10 instance-0.4 | accuracy (%): 87.41 -> 90.73 | Improved |
| increase window size n from 6 to 10 (ablation) | RML · CIFAR-10 instance-0.4 | accuracy (%): 90.73 -> 89.71 | Hurt |
| use RML instead of PES | ResNet · CIFAR-100 sym-0.5 | accuracy (%): 58.90 -> 67.05 | Improved |
| use RML instead of PES | ResNet · CIFAR-10 sym-0.2 | accuracy (%): 92.38 -> 92.89 | Improved |
| add semi-supervised strategy (MixMatch) to RML | RML · WebVision | accuracy (%): 73.32 -> 81.34 | Improved |
| use RML-Semi instead of PES | ResNet-50 · Clothing1M | accuracy (%): 74.99 -> 75.14 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DivideMix — DivideMix: Learning with Noisy Labels as Semi-supervised Learning (ICLR 2020) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Regroup Median Loss for Combating Label Noise (AAAI 2024 · similarity 0.99)
- Sample-wise Label Confidence Incorporation for Learning with Noisy Labels (ICCV 2023 · similarity 0.89)
- Robust Classification via Regression for Learning with Noisy Labels (ICLR 2024 · similarity 0.88)
- Enhanced Sample Selection with Confidence Tracking: Identifying Correctly Labeled Yet Hard-to-Learn Samples in Noisy Data (AAAI 2025 · similarity 0.87)
- Mitigating Label Noise through Data Ambiguation (AAAI 2024 · similarity 0.86)
- NoiseGPT: Label Noise Detection and Rectification through Probability Curvature (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.
