Research map / Label-efficient and robust learning / Noisy labels
Robust Loss Functions under Label Noise for Deep Neural Networks
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 MAE instead of CCE | neural network classifier · MNIST, symmetric noise η=60% | accuracy: 0.5888 -> 0.9799 | Helped |
| Use MAE instead of CCE | neural network classifier · RCV1, symmetric noise η=60% | accuracy: 0.7905 -> 0.8637 (± 0.0201) | Helped |
| Use MAE instead of CCE | neural network classifier · CIFAR-10, symmetric noise η=60% | accuracy: 0.3083 -> 0.5328(± 0.0251) | Helped |
| Use MAE instead of CCE | neural network classifier · IMDb, symmetric noise η=30% | accuracy: 0.6466 -> 0.7352(± 0.0145) | Helped |
| Use MAE instead of CCE | neural network classifier · Newswire, symmetric noise η=60% | accuracy: 0.4670 -> 0.6357(± 0.0106) | Helped |
| Use MAE instead of CCE | neural network classifier · 20 Newsgroups, symmetric noise η=60% | accuracy: 0.6435 -> 0.7319 | Helped |
| Use MAE instead of CCE | neural network classifier · class-conditional noise, six datasets | accuracy | Helped |
| Use bounded MSE instead of unbounded CCE | neural network classifier · label noise, across datasets | accuracy | Mixed |
| Use MAE instead of CCE | neural network classifier · noise-free training data, across datasets | accuracy | Mixed |
| Increase dropout rate | CCE-trained neural network · label-noisy datasets | accuracy | Helped |
Closest papers by meaning
- Asymmetric Loss Functions for Learning with Noisy Labels (ICML 2021 · similarity 0.94)
- Normalized Loss Functions for Deep Learning with Noisy Labels (ICML 2020 · similarity 0.93)
- L_DMI: A Novel Information-theoretic Loss Function for Training Deep Nets Robust to Label Noise (NeurIPS 2019 · similarity 0.91)
- Learning from Noisy Labels with Complementary Loss Functions (AAAI 2021 · similarity 0.91)
- Enhancing Noise-Robust Losses for Large-Scale Noisy Data Learning (AAAI 2025 · similarity 0.90)
- $\epsilon$-Softmax: Approximating One-Hot Vectors for Mitigating Label Noise (NeurIPS 2024 · similarity 0.90)
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.