Research map / Label-efficient and robust learning / Noisy labels
A Second-Order Approach to Learning With Instance-Dependent Label Noise
Check 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 |
|---|---|---|---|
| Add covariance correction to CORES² | CORES² · CIFAR10, η=0.2 | test accuracy (%): 91.14±0.46 -> 92.01±0.75 | Helped |
| Add covariance correction to CORES² | CORES² · CIFAR10, η=0.4 | test accuracy (%): 83.67±1.29 -> 84.96±1.25 | Helped |
| Add covariance correction to CORES² | CORES² · CIFAR10, η=0.6 | test accuracy (%): 77.68±2.24 -> 79.82±2.56 | Helped |
| Add covariance correction to CORES² | CORES² · CIFAR100, η=0.2 | test accuracy (%): 66.47±0.45 -> 69.11±0.46 | Helped |
| Add covariance correction to CORES² | CORES² · CIFAR100, η=0.4 | test accuracy (%): 58.99±1.49 -> 63.17±1.40 | Helped |
| Add covariance correction to CORES² | CORES² · CIFAR100, η=0.6 | test accuracy (%): 38.55±3.25 -> 43.58±3.30 | Helped |
| Add covariance correction to CORES² | CORES² · Clothing1M | accuracy: 73.24 -> 74.17 | Helped |
| Add covariance term while keeping the peer term (ablation) | Peer Only · CIFAR10, η=0.2 | test accuracy (%): 91.85 -> 92.69 | Helped |
| Add covariance term while keeping the peer term (ablation) | Peer Only · CIFAR10, η=0.4 | test accuracy (%): 84.41 -> 85.55 | Helped |
| Add covariance term while keeping the peer term (ablation) | Peer Only · CIFAR10, η=0.6 | test accuracy (%): 78.74 -> 81.54 | Helped |
| Use covariance term without the peer term (ablation) | No Covariance, No Peer · CIFAR10, η=0.4 | test accuracy (%): 82.56 -> 78.49 | Hurt |
Closest papers by meaning
- Learning with Instance-Dependent Label Noise: A Sample Sieve Approach (ICLR 2021 · similarity 0.94)
- On the Role of Label Noise in the Feature Learning Process (ICML 2025 · similarity 0.89)
- Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization (NeurIPS 2025 · similarity 0.89)
- Forming Auxiliary High-confident Instance-level Loss to Promote Learning from Label Proportions (CVPR 2025 · similarity 0.88)
- Self-cognitive Denoising in the Presence of Multiple Noisy Label Sources (ICML 2024 · similarity 0.88)
- Label Noise: Ignorance Is Bliss (NeurIPS 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.