Research map / Label-efficient and robust learning / Noisy labels
Rethinking Label Refurbishment: Model Robustness under Label Noise
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 |
|---|---|---|---|
| Use label smoothing with α close to 1.0 (extreme label smoothing) (ablation) | 4-layer CNN and ResNet34 · MNIST and CIFAR-10 with symmetric label noise | test accuracy | Helped |
| Use bootstrapping loss with a suitably tuned α | 4-layer CNN and ResNet34 · MNIST and CIFAR-10 with label noise | test accuracy | Helped |
| Use backward correction with a suitably tuned α | 4-layer CNN and ResNet34 · MNIST and CIFAR-10 with label noise | test accuracy | Helped |
| Use reversed label smoothing with negative α (ablation) | 4-layer CNN and ResNet34 · MNIST and CIFAR-10 with symmetric and asymmetric label noise | test accuracy | Helped |
| Use reversed energy regularized loss with negative α (ablation) | 4-layer CNN and ResNet34 · MNIST and CIFAR-10 with symmetric and asymmetric label noise | test accuracy | Helped |
| Increase BL α beyond 0.7 (ablation) | Bootstrapping loss · MNIST and CIFAR-10 with symmetric label noise | test accuracy | Hurt |
| Constrain network parameters close to initialization using ELS (ablation) | Neural network trained with noisy labels · CIFAR-10, symmetric noise η = 0.4 | test accuracy | Helped |
| Constrain network parameters close to initialization using DTI (ablation) | Neural network trained with noisy labels · CIFAR-10, symmetric noise η = 0.4 | test accuracy | Helped |
Closest papers by meaning
- Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization (NeurIPS 2025 · similarity 0.89)
- Combating Semantic Contamination in Learning with Label Noise (AAAI 2025 · similarity 0.89)
- Soften to Defend: Towards Adversarial Robustness via Self-Guided Label Refinement (CVPR 2024 · similarity 0.89)
- Mitigating Label Noise through Data Ambiguation (AAAI 2024 · similarity 0.89)
- Learning with Structural Labels for Learning with Noisy Labels (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.