Research map / Label-efficient and robust learning / Recognition and long-tail
AES: Curing Optimizer Blindness in Long-Tailed Recognition via State-Aware Correction
Fanfu Wang, Jiachang Zhan, Zhiheng Gong, Pengkun Wang, Yang Wang
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 SCAF (ablation) | AES · CIFAR-100-LT (IR=100) | top-1 accuracy (All): 56.77 -> 57.85 | Improved |
| add E-PCG (ablation) | AES · CIFAR-100-LT (IR=100) | top-1 accuracy (All): 57.26 -> 57.85 | Improved |
| add ARS (ablation) | AES · CIFAR-100-LT (IR=100) | top-1 accuracy (All): 57.00 -> 57.85 | Improved |
| add ARS (ablation) | AES · CIFAR-100-LT (IR=100) | Tail accuracy: 37.90 -> 34.33 | Hurt |
| add SCAF (ablation) | AES · CIFAR-100-LT (IR=100) | Head accuracy: 78.17 -> 79.23 | Improved |
| replace TS-MOF with AES | TS-MOF · CIFAR-100-LT (IR=50) | top-1 accuracy: 60.2 -> 63.0 | Improved |
| replace TS-MOF with AES | TS-MOF · ImageNet-LT | top-1 accuracy: 56.3 -> 57.4 | Improved |
| replace TS-MOF with AES | TS-MOF · iNaturalist 2018 | top-1 accuracy: 72.3 -> 73.8 | Improved |
| replace TS-MOF with AES | TS-MOF · ImageNet-LT Tail | top-1 accuracy: 42.6 -> 40.4 | Hurt |
| replace TS-MOF with AES | TS-MOF · CIFAR-100-LT (IR=100) Tail | top-1 accuracy: 39.9 -> 33.9 | Hurt |
| replace TS-MOF with AES | TS-MOF · CIFAR-100-LT (IR=100) Medium | top-1 accuracy: 49.2 -> 56.9 | Improved |
| replace TS-MOF with AES | TS-MOF · blindness coefficient | beta: 0.1784 -> 0.1163 | Improved |
Closest papers by meaning
- TailCon: Mitigating Tail Signal Erosion through Memory Consolidation for Long-Tailed Recognition (NeurIPS 2026 · similarity 0.87)
- U-MOF: Uncertainty-Guided Parameter-Efficient Multi-Objective Fine-Tuning for Long-Tailed Recognition (NeurIPS 2026 · similarity 0.86)
- Decoupled Optimisation for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.85)
- Long-tailed Recognition with Model Rebalancing (NeurIPS 2025 · similarity 0.84)
- Multiplicative Logit Adjustment Approximates Neural-Collapse-Aware Decision Boundary Adjustment (ICLR 2025 · similarity 0.84)
- Long-Tailed Learning as Multi-Objective Optimization (AAAI 2024 · similarity 0.84)
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.
