Research map / Label-efficient and robust learning / Noisy labels
Can Less be More? When Increasing-to-Balancing Label Noise Rates Considered Beneficial
Yang Liu, Jialu Wang
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 |
|---|---|---|---|
| Balance class-dependent label noise before training with cross-entropy | Adult one-layer perceptron · Adult, e_-=0.0, e_+=0.1 | accuracy: 72.63±0.29 -> 73.62±0.37 | Improved |
| Balance class-dependent label noise before training with peer loss | Compas one-layer perceptron · Compas, e_-=0.0, e_+=0.1 | accuracy: 68.06±0.70 -> 68.22±0.68 | No clear effect |
| Balance class-dependent label noise before training with peer loss | Fairface one-layer perceptron · Fairface, e_-=0.0, e_+=0.1 | accuracy: 87.58±0.15 -> 88.49±0.12 | Improved |
| Balance class noise rates before training with peer loss | MNIST MLP · MNIST, noise gap 0.1 | accuracy: 88.12 -> 89.19 | Improved |
| Balance class noise rates before training with cross-entropy | MNIST MLP · MNIST, noise gap 0.3 | accuracy: 85.24 -> 81.89 | Hurt |
| Balance class noise rates before training with peer loss | CIFAR-10 classifier · CIFAR-10, noise gap 0.1 | accuracy: 89.66 -> 89.90 | No clear effect |
| Balance class noise rates before training with cross-entropy | CIFAR-10 classifier · CIFAR-10, noise gap 0.3 | accuracy: 89.78 -> 87.98 | Hurt |
| Balance noise rates before training with cross-entropy | Clothing1M ResNet-50 · Clothing1M | test accuracy: 68.94% -> 70.37% | Improved |
| Balance group-dependent noise before fairness-constrained training | Adult logistic regression · Adult, e_a=0.2, e_b=0.4 | fairness violation: 6.48 -> 3.16 | Improved |
| Balance group-dependent noise before fairness-constrained training | Fairface logistic regression · Fairface, e_a=0.2, e_b=0.4 | fairness violation: 5.87 -> 1.38 | 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.
- Co-Teaching — Co-teaching: Robust training of deep neural networks with extremely noisy labels (NeurIPS 2018) · better on 1 of 1 shared result
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
- Foster Adaptivity and Balance in Learning with Noisy Labels (ECCV 2024 · similarity 0.90)
- Mitigating Label Bias in Machine Learning: Fairness through Confident Learning (AAAI 2024 · similarity 0.89)
- Toward Robustness in Multi-Label Classification: A Data Augmentation Strategy against Imbalance and Noise (AAAI 2024 · similarity 0.88)
- NoiseGPT: Label Noise Detection and Rectification through Probability Curvature (NeurIPS 2024 · similarity 0.87)
- Label Noise: Ignorance Is Bliss (NeurIPS 2024 · similarity 0.87)
- Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization (NeurIPS 2025 · similarity 0.87)
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.
