Research map / Label-efficient and robust learning / Recognition and long-tail
Class-Balanced Loss Based on Effective Number of Samples
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 effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-10, imbalance 200 | classification error rate: 36.00 -> 31.11 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-10, imbalance 100 | classification error rate: 30.41 -> 25.43 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-10, imbalance 50 | classification error rate: 23.29 -> 20.73 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-10, imbalance 20 | classification error rate: 16.77 -> 15.64 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-10, imbalance 10 | classification error rate: 12.97 -> 12.51 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-100, imbalance 200 | classification error rate: 64.38 -> 63.77 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-100, imbalance 100 | classification error rate: 61.22 -> 60.40 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-100, imbalance 50 | classification error rate: 55.68 -> 54.68 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-100, imbalance 20 | classification error rate: 48.05 -> 47.41 | Helped |
| Add effective-number class-balanced weighting to the selected loss | ResNet-32 · Long-Tailed CIFAR-100, imbalance 10 | classification error rate: 44.22 -> 42.01 | Helped |
| Replace softmax cross-entropy with class-balanced focal loss | ResNet-50 · iNaturalist 2017 validation | top-1 error rate: 45.38 -> 41.92 | Helped |
| Replace softmax cross-entropy with class-balanced focal loss | ResNet-50 · iNaturalist 2018 validation | top-1 error rate: 42.86 -> 38.88 | Helped |
Closest papers by meaning
- Inverse Weight-Balancing for Deep Long-Tailed Learning (AAAI 2024 · similarity 0.87)
- Exploring Weight Balancing on Long-Tailed Recognition Problem (ICLR 2024 · similarity 0.87)
- Improved Balanced Classification with Theoretically Grounded Loss Functions (NeurIPS 2025 · similarity 0.86)
- Long-Tailed Learning as Multi-Objective Optimization (AAAI 2024 · similarity 0.86)
- Difficulty-aware Balancing Margin Loss for Long-tailed Recognition (AAAI 2025 · similarity 0.85)
- Balancing the Scales: A Theoretical and Algorithmic Framework for Learning from Imbalanced Data (ICML 2025 · similarity 0.85)
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.