Research map / Label-efficient and robust learning / Recognition and long-tail
Why Not Hyperparameter-Friendly Optimisation? A Monotonic Adaptive Norm Rescaling Approach For Long-Tailed Recognition
Shuo Zhang, Chenqi Li, Tingting Zhu
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 SAMN to CE classifier retraining | CE · CIFAR100-LT (IF=100) | accuracy: 47.4 -> 54.0 | Improved |
| add SAMN to CE classifier retraining | CE · CIFAR10-LT (IF=100) | accuracy: 80.0 -> 84.1 | Improved |
| add SAMN to SLAS | SLAS · CIFAR100-LT (IF=100) | accuracy: 52.7 -> 54.1 | Improved |
| add SAMN to GLMC | GLMC · CIFAR100-LT (IF=100) | accuracy: 56.8 -> 57.7 | Improved |
| add SAMN to GLMC | GLMC · CIFAR10-LT (IF=10) | accuracy: 95.1 -> 95.2 | No clear effect |
| add SAMN to GLMC | GLMC · ImageNet-LT | accuracy (All): 56.3 -> 57.7 | Improved |
| add SAMN to GLMC | GLMC · iNaturalist2018 | accuracy (All): 72.2 -> 72.7 | Improved |
| add SAMN to GLMC | GLMC · ImageNet-LT | Many-class accuracy: 70.1 -> 69.7 | Hurt |
| add SAMN to GLMC | GLMC · ImageNet-LT | Medium-class accuracy: 52.4 -> 56.1 | Improved |
| add SAMN to weights only vs bias only (ablation) | CE + SAMN · CIFAR100-LT (IF=100) | accuracy: 50.6 -> 53.9 | Improved |
| apply SAMN to both weight and bias instead of weight only (ablation) | CE + SAMN · CIFAR10-LT (IF=100) | accuracy: 83.6 -> 84.1 | Improved |
| use first-stage weight norms instead of class frequency as order metric (ablation) | CE + SAMN · CIFAR100-LT (IF=100) | accuracy: 54.0 -> 53.9 | No clear effect |
Closest papers by meaning
- U-MOF: Uncertainty-Guided Parameter-Efficient Multi-Objective Fine-Tuning for Long-Tailed Recognition (NeurIPS 2026 · similarity 0.89)
- Long-Tailed Recognition via Weight Balancing (CVPR 2022 · similarity 0.89)
- Long-tailed Recognition with Model Rebalancing (NeurIPS 2025 · similarity 0.89)
- Decoupled Optimisation for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.88)
- Exploring Weight Balancing on Long-Tailed Recognition Problem (ICLR 2024 · similarity 0.87)
- Difficulty-aware Balancing Margin Loss for Long-tailed Recognition (AAAI 2025 · similarity 0.86)
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.
