Research map / Label-efficient and robust learning / Recognition and long-tail
Influence-Balanced Loss for Imbalanced Visual Classification
Seulki Park, Jongin Lim, Younghan Jeon, Jin Young Choi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 influence-balanced fine-tuning | CE baseline ResNet-32 · CIFAR-10 long-tailed rho=100 | accuracy: 70.87 -> 78.26 | Improved |
| add influence-balanced fine-tuning | CE baseline ResNet-32 · CIFAR-100 long-tailed rho=100 | accuracy: 38.05 -> 42.14 | Improved |
| add influence-balanced fine-tuning | CE baseline ResNet-50 · iNaturalist 2018 | top-1 accuracy: 57.30 -> 65.39 | Improved |
| add influence-balanced fine-tuning | CE baseline ResNet-18 · Tiny ImageNet long-tailed rho=100 | accuracy: 38.52 -> 42.65 | Improved |
| use L1 norm instead of L2 for influence magnitude (ablation) | IB loss · CIFAR-10 long-tailed rho=100 | accuracy: 75.67 -> 78.41 | Improved |
| use L1 norm instead of L-infinity for influence magnitude (ablation) | IB loss · CIFAR-10 long-tailed rho=100 | accuracy: 77.23 -> 78.41 | Improved |
| add the IB weighting factor to the epsilon constant (ablation) | IB loss · CIFAR-10 long-tailed rho=100 | accuracy: 64.91 -> 78.17 | Improved |
| use epsilon 1e-3 instead of 1e-8 (ablation) | IB loss · CIFAR-10 long-tailed rho=100 | accuracy: 76.03 -> 78.17 | Improved |
| use epsilon 1e-3 instead of 1e-2 (ablation) | IB loss · CIFAR-10 long-tailed rho=100 | accuracy: 77.55 -> 78.17 | No clear effect |
| combine IB with focal loss in fine-tuning | IB loss · CIFAR-10 long-tailed rho=100 | accuracy: 78.26 -> 79.76 | Improved |
| combine IB with CB effective number | IB loss · CIFAR-10 long-tailed rho=100 | accuracy: 78.26 -> 78.04 | No clear effect |
| combine IB with CB effective number | IB loss · CIFAR-100 long-tailed rho=100 | accuracy: 42.14 -> 41.31 | Hurt |
Closest papers by meaning
- Balancing the Scales: A Theoretical and Algorithmic Framework for Learning from Imbalanced Data (ICML 2025 · similarity 0.88)
- Improved Balanced Classification with Theoretically Grounded Loss Functions (NeurIPS 2025 · similarity 0.87)
- A Unified Loss for Handling Inter-Class and Intra-Class Imbalance in Medical Image Segmentation (AAAI 2025 · similarity 0.86)
- Learning Imbalanced Data with Beneficial Label Noise (ICML 2025 · similarity 0.86)
- Inverse Weight-Balancing for Deep Long-Tailed Learning (AAAI 2024 · similarity 0.86)
- Rebalancing Multi-Label Class-Incremental Learning (AAAI 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.
