Research map / Label-efficient and robust learning / Recognition and long-tail
Open Long-Tailed Recognition in a Dynamic World
Ziwei Liu, Zhongqi Miao, Xiaohang Zhan, Jiayun Wang, Boqing Gong, Stella X. Yu
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 |
|---|---|---|---|
| use OLTR++ instead of plain model | ResNet-10 · ImageNet-LT (closed-set) | overall top-1 accuracy: 20.9 -> 39.6 | Improved |
| use OLTR++ instead of plain model | ResNet-10 · ImageNet-LT (open-set) | F-measure: 0.295 -> 0.446 | Improved |
| add OpenMax on top of Range Loss | Range Loss ResNet-10 · ImageNet-LT (open-set) | F-measure: 0.373 -> 0.368 | No clear effect |
| use two-headed ResNet-50 backbone with OLTR++ | OLTR++ · CIFAR-100-LT (imbalance ratio 100) | overall accuracy: 41.2 -> 47.4 | Improved |
| use two-headed ResNet-50 backbone with OLTR++ | OLTR++ · iNaturalist-18 | overall accuracy: 65.3 -> 71.9 | Improved |
| use OLTR++ instead of plain model | ResNet-50 · MegaFace (one-shot) | identification rate: 84.60 -> 87.60 | Improved |
| use OLTR++ instead of plain model | ResNet-50 · MegaFace (zero-shot) | identification rate: 77.72 -> 79.50 | Improved |
| use OLTR++ over MetaModelNet | ResNet-152 · SUN-LT | top-1 accuracy: 57.3 -> 58.7 | Improved |
| add reachability calibration for open-set detection (ablation) | OLTR++ · CIFAR100 + TinyImageNet open-set detection | FPR at 95% TPR: 82.2 -> 51.5 | Improved |
| add memory feature and concept selector (ablation) | OLTR++ · ImageNet-LT | accuracy on many/medium/few-shot | Improved |
| use modulated attention instead of spatial attention on feature maps (ablation) | OLTR++ · ImageNet-LT | accuracy | Improved |
| make the dataset more imbalanced (lower alpha) (ablation) | OLTR++ · ImageNet-LT | accuracy | No clear effect |
Closest papers by meaning
- Feature Fusion from Head to Tail for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.88)
- Robust Visual Recognition with Class-Imbalanced Open-World Noisy Data (AAAI 2024 · similarity 0.88)
- Out-of-Distribution Detection in Long-Tailed Recognition with Calibrated Outlier Class Learning (AAAI 2024 · similarity 0.87)
- EAT: Towards Long-Tailed Out-of-Distribution Detection (AAAI 2024 · similarity 0.87)
- Decoupled Optimisation for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.87)
- OW-OVD: Unified Open World and Open Vocabulary Object Detection (CVPR 2025 · similarity 0.87)
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.
