Research map / Label-efficient and robust learning / Recognition and long-tail
Robust Learning from Noisily Labeled Long-Tailed Data via Fairness Regularizer
Jiaheng Wei, Zhaowei Zhu, Gang Niu, Tongliang Liu, Sijia Liu, Masashi Sugiyama, Yang 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 fairness regularizer with 2 sub-populations (G2) | CE · CIFAR-10 (long-tailed, Sym noise 0.5, r=100) | accuracy: 36.44 -> 46.33 | Improved |
| add fairness regularizer (G2) | Logit-adj · CIFAR-10 (long-tailed, Imb noise 0.2, r=100) | accuracy: 68.18 -> 72.20 | Improved |
| add fairness regularizer (G2) | Logit-adj · CIFAR-10 (long-tailed, Sym noise 0.5, r=100) | accuracy: 27.32 -> 21.93 | Hurt |
| add fairness regularizer (G2) | LS · CIFAR-10 (long-tailed, Imb noise 0.5, r=50) | accuracy: 36.17 -> 44.11 | Improved |
| add fairness regularizer (G2) | Focal · CIFAR-10 (long-tailed, Imb noise 0.2, r=100) | accuracy: 57.68 -> 56.55 | Hurt |
| add fairness regularizer (G2) | PL · CIFAR-10 (long-tailed, Imb noise 0.5, r=100) | accuracy: 30.13 -> 28.01 | Hurt |
| split sub-populations by KNN clustering into one per class | CE + FR · CIFAR-100 (long-tailed noisy) | accuracy: 46.02 -> 46.18 | No clear effect |
| use 2 sub-populations from a pre-trained model instead of KNN clusters (ablation) | CE + FR · CIFAR-100 (Imb noise 0.5, r=100) | accuracy: 15.19 -> 15.93 | Improved |
| add fairness regularizer (KNN) to CE | CE · CIFAR-10 (long-tailed, Sym noise 0.5, r=100) | accuracy: 36.44 -> 45.63 | Improved |
| vary FR weight theta from 0.0 to 2.0 (ablation) | CE · Clothing1M | accuracy: 72.68 -> 72.99 | No clear effect |
| add FR (theta=1.0) (ablation) | NLS · Clothing1M | accuracy: 74.46 -> 74.49 | No clear effect |
| add FR (theta=1.0) (ablation) | Logit-adj · Clothing1M | accuracy: 72.43 -> 72.67 | Improved |
Closest papers by meaning
- Long-Tailed Learning as Multi-Objective Optimization (AAAI 2024 · similarity 0.88)
- Robust Logit Adjustment for Learning with Long-Tailed Noisy Data (AAAI 2025 · similarity 0.88)
- Enhancing Robust Fairness via Confusional Spectral Regularization (ICLR 2025 · similarity 0.87)
- Breaking Long-Tailed Learning Bottlenecks: A Controllable Paradigm with Hypernetwork-Generated Diverse Experts (NeurIPS 2024 · similarity 0.86)
- Rectifying Soft-Label Entangled Bias in Long-Tailed Dataset Distillation (NeurIPS 2025 · similarity 0.86)
- Learning from Noisy Labels via Conditional Distributionally Robust Optimization (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.
