Research map / Label-efficient and robust learning / Noisy labels
Do We Need to Penalize Variance of Losses for Learning with Label Noise?
Yexiong Lin, Yu Yao, Yuxuan Du, Jun Yu, Bo Han, Mingming Gong, Tongliang Liu
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 VRNL (encourage large loss variance) | Reweight · Fashion-MNIST Sym-90% | accuracy: 45.63 -> 77.74 | Improved |
| add VRNL | Forward · CIFAR-100 Sym-50% | accuracy: 43.45 -> 56.49 | Improved |
| add VRNL | Reweight · CIFAR-10 Sym-80% | accuracy: 51.70 -> 63.24 | Improved |
| add VRNL | BLTM · CIFAR-100 Sym-50% | accuracy: 30.56 -> 31.32 | No clear effect |
| add VRNL | VolMinNet · CIFAR-10 IDN-20% | accuracy: 90.29 -> 90.27 | No clear effect |
| add VRNL | VolMinNet · WebVision val (top-1) | accuracy: 72.96 -> 74.16 | Improved |
| add VRNL | Forward · Clothing1M | accuracy: 71.27 -> 72.43 | Improved |
| add VRNL | Reweight · CIFAR-100 Sym-10% | accuracy: 65.72 -> 67.57 | Improved |
| add VRNL | Forward · CIFAR-10 clean labels (Sym-00%) | accuracy: 92.78 -> 92.71 | No clear effect |
| add linear warmup of alpha (ablation) | VolMinNet-VRNL · CIFAR-10 Sym-80% | accuracy: 59.88 -> 61.23 | Improved |
| add linear warmup of alpha (ablation) | VolMinNet-VRNL · CIFAR-10 Sym-90% | accuracy: 34.79 -> 37.41 | Improved |
| add linear warmup of alpha (ablation) | VolMinNet-VRNL · CIFAR-10 IDN-50% | accuracy: 61.48 -> 64.94 | Improved |
Closest papers by meaning
- Sample Selection with Uncertainty of Losses for Learning with Noisy Labels (ICLR 2022 · similarity 0.89)
- Scalable Penalized Regression for Noise Detection in Learning With Noisy Labels (CVPR 2022 · similarity 0.87)
- Regroup Median Loss for Combating Label Noise (AAAI 2024 · similarity 0.85)
- Mitigating Label Noise through Data Ambiguation (AAAI 2024 · similarity 0.85)
- $\epsilon$-Softmax: Approximating One-Hot Vectors for Mitigating Label Noise (NeurIPS 2024 · similarity 0.84)
- Enhancing Noise-Robust Losses for Large-Scale Noisy Data Learning (AAAI 2025 · 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.
