Research map / Label-efficient and robust learning / Domain adaptation and generalization
Activate and Reject: Towards Safe Domain Generalization under Category Shift
Chaoqi Chen, Luyao Tang, Leitian Tao, Hong-Yu Zhou, Yue Huang, Xiaoguang Han, Yizhou 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 |
|---|---|---|---|
| add unknown activation loss L_UA (ablation) | ART · Four classification benchmarks (average) | H-score (%): 47.9 -> 60.1 | Improved |
| add smoothed cross-entropy L_SCE (ablation) | ART · Four classification benchmarks (average) | H-score (%): 54.1 -> 60.1 | Improved |
| add Test-time Unknown Rejection on top of UGD (ablation) | ART · Four classification benchmarks (average) | H-score (%): 55.2 -> 60.1 | Improved |
| replace TUR with Tent for test-time adaptation (ablation) | UGD · Four classification benchmarks (average) | H-score (%): 55.2 -> 45.3 | Hurt |
| replace TUR with MEMO or TTT (ablation) | UGD · Four classification benchmarks (average) | H-score (%): 55.2 -> 56.9 (MEMO), 55.9 (TTT) | No clear effect |
| add UGD on segmentation (ablation) | ART · GTA5 to Cityscapes | H-score: 50.2 -> 63.1 | Improved |
| add TUR on segmentation (ablation) | ART · GTA5 to Cityscapes | H-score: 62.6 -> 63.1 | Improved |
| add TUR on Pascal VOC to Clipart (ablation) | ART · Pascal VOC to Clipart detection | H-score: 8.6 -> 15.9 | Improved |
| add L_SCE on Pascal VOC to Watercolor (ablation) | ART · Pascal VOC to Watercolor detection | H-score: 20.2 -> 17.6 | Hurt |
| vary order of test data (ablation) | TUR · Four classification benchmarks (average) | H-score (%): 60.0 to 60.3 across orders | No clear effect |
| apply ART instead of previous best OSDG baseline | DGCS classification · PACS | H-score (%): 43.0 -> 52.3 | Improved |
Closest papers by meaning
- Domain Generalization via Model-Agnostic Learning of Semantic Features (NeurIPS 2019 · similarity 0.89)
- Deeper, Broader and Artier Domain Generalization (ICCV 2017 · similarity 0.89)
- Generalize or Detect? Towards Robust Semantic Segmentation Under Multiple Distribution Shifts (NeurIPS 2024 · similarity 0.88)
- Unknown Prompt the only Lacuna: Unveiling CLIP's Potential for Open Domain Generalization (CVPR 2024 · similarity 0.87)
- Object-Aware Domain Generalization for Object Detection (AAAI 2024 · similarity 0.87)
- Category-Aware Active Domain Adaptation (ICML 2024 · 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.
