Research map / Label-efficient and robust learning / Noisy labels
Beyond Class-Conditional Assumption: A Primary Attempt to Combat Instance-Dependent 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 SEAL instead of the strongest reported baseline at each noise fraction | CNN · MNIST IDN, 10%, 20%, 30%, and 40% | accuracy: 95.77 -> 96.75; 91.07 -> 93.63; 86.20 -> 88.52; 79.30 -> 80.73 | Helped |
| Use SEAL instead of the strongest reported baseline at each noise fraction | Wide ResNet 28×10 · CIFAR-10 IDN, 10%, 20%, 30%, and 40% | accuracy: 91.26 -> 91.32; 87.28 -> 87.79; 84.33 -> 85.30; 78.72 -> 82.98 | Helped |
| Apply SEAL label correction with cross-entropy training | ResNet-50 · Clothing1M | accuracy: 69.07 -> 70.63 | Helped |
| Apply SEAL to DMI training | ResNet-50 with DMI · Clothing1M | accuracy: 72.27 -> 73.40 | Helped |
| Average predictions across training epochs to form instance-specific soft labels | SEAL · Synthetic IDN experiments | distance to true label | Helped |
Closest papers by meaning
- Confidence Scores Make Instance-dependent Label-noise Learning Possible (ICML 2021 · similarity 0.92)
- A Second-Order Approach to Learning With Instance-Dependent Label Noise (CVPR 2021 · similarity 0.92)
- Label Noise: Ignorance Is Bliss (NeurIPS 2024 · similarity 0.91)
- Learning with Instance-Dependent Label Noise: A Sample Sieve Approach (ICLR 2021 · similarity 0.91)
- Safeguarded Dynamic Label Regression for Noisy Supervision (AAAI 2019 · similarity 0.91)
- Hypothesis Testing for Class-Conditional Noise Using Local Maximum Likelihood (AAAI 2024 · 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.