Research map / Label-efficient and robust learning / Noisy labels
Robust Training under Label Noise by Over-parameterization
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 |
|---|---|---|---|
| Model label corruption with a sparse over-parameterized term and train it with implicit regularization | SOP versus CE · CIFAR-10, 80% symmetric noise | test accuracy: 59.28 ± 0.97 -> 68.32 ± 0.77 | Hurt |
| Model label corruption with a sparse over-parameterized term and train it with implicit regularization | SOP versus CE · CIFAR-10, 80% symmetric noise | test accuracy: 73.86 ± 0.61 -> 68.32 ± 0.77 | Hurt |
| Use sparse over-parameterization for noisy labels | SOP versus CE · CIFAR-10, 20% symmetric noise | test accuracy: 86.32 ± 0.18 -> 93.18 ± 0.57 | Helped |
| Use sparse over-parameterization for noisy labels | SOP versus CE · CIFAR-100, 80% symmetric noise | test accuracy: 20.23 ± 0.82 -> 30.20 ± 0.63 | Helped |
| Add consistency and class-balance regularization | SOP to form SOP+ · CIFAR-10, 80% symmetric noise | test accuracy: 93.2 -> 94.0 | Helped |
| Use sparse over-parameterization with regularizers | SOP+ versus CE · CIFAR-10, 40% asymmetric noise | test accuracy: 82.2 -> 93.8 | Helped |
| Use sparse over-parameterization with regularizers | SOP+ versus ELR+ · CIFAR-10N Aggregate | test accuracy: 94.83 ± 0.10 -> 95.61 ± 0.13 | Helped |
| Use sparse over-parameterization with regularizers | SOP+ versus ELR+ · CIFAR-100N Noisy | test accuracy: 66.72 ± 0.07 -> 67.81 ± 0.23 | Helped |
| Use sparse over-parameterization for noisy labels | SOP versus ELR · Clothing-1M | test accuracy: 72.9 -> 73.5 | Helped |
| Use sparse over-parameterization for noisy labels | SOP versus ELR · WebVision | test accuracy: 76.2 -> 76.6 | Helped |
| Optimize the noise variables v with MSE rather than cross-entropy (ablation) | SOP · Image classification with label noise | noise-variable learning | Helped |
Closest papers by meaning
- Coresets for Robust Training of Deep Neural Networks against Noisy Labels (NeurIPS 2020 · similarity 0.91)
- Training Deep Neural Networks with Virtual Smoothing Classes (AAAI 2025 · similarity 0.88)
- $\epsilon$-Softmax: Approximating One-Hot Vectors for Mitigating Label Noise (NeurIPS 2024 · similarity 0.88)
- Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization (NeurIPS 2025 · similarity 0.88)
- Soften to Defend: Towards Adversarial Robustness via Self-Guided Label Refinement (CVPR 2024 · similarity 0.88)
- On the Role of Label Noise in the Feature Learning Process (ICML 2025 · 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.