Research map / Label-efficient and robust learning / OOD and anomaly detection
Activation Matters: Test-time Activated Negative Labels for OOD Detection with Vision-Language Models
Yabin Zhang, Maya Varma, Yunhe Gao, Jean-Benoit Delbrouck, Jiaming Liu, Chong Wang, Curtis Langlotz
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 |
|---|---|---|---|
| replace fixed activation-agnostic negative labels with distribution-adaptive activated labels (ablation) | NegLabel · ImageNet-1k OpenOOD Near-OOD | FPR95: 69.45 -> 61.57 | Improved |
| replace fixed activation-agnostic negative labels with distribution-adaptive activated labels (ablation) | NegLabel · ImageNet-1k OpenOOD Far-OOD | FPR95: 23.73 -> 17.52 | Improved |
| add batch-adaptive activation and activation-aware score on top of distribution-adaptive labels (ablation) | TANL · ImageNet-1k OpenOOD Near-OOD | FPR95: 61.57 -> 60.06 | Improved |
| add batch-adaptive activation and activation-aware score on top of distribution-adaptive labels (ablation) | TANL · ImageNet-1k OpenOOD Far-OOD | FPR95: 17.52 -> 17.21 | No clear effect |
| use activation-aware score Saa instead of standard score Snl with many negative labels (ablation) | TANL · ImageNet-1k OpenOOD | FPR95 vs label number | Improved |
| increase number of negative labels M (ablation) | TANL · ImageNet-1k OpenOOD | FPR95 | Hurt |
| mine negative labels with negative-sample activation instead of positive information alone (ablation) | TANL · ImageNet-1k OpenOOD | FPR95 | Improved |
| lower alpha from 1.0 to 0.95 to mix batch information (ablation) | TANL · ImageNet-1k OpenOOD | FPR95 | Improved |
| lower alpha further below 0.95 (ablation) | TANL · ImageNet-1k OpenOOD | FPR95 | Hurt |
| increase test batch size (ablation) | TANL · ImageNet-1k OpenOOD | FPR95 | Improved |
| use TANL instead of NegLabel | NegLabel · ImageNet-1k, 4 OOD datasets | FPR95: 25.40 -> 9.81 | Improved |
| use TANL instead of NegLabel | NegLabel · Full-spectrum OpenOOD Near-OOD | FPR95: 76.25 -> 68.71 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- AdaNeg — AdaNeg: Adaptive Negative Proxy Guided OOD Detection with Vision-Language Models (NeurIPS 2024) · better on 4 of 4 shared results
- OODD — OODD: Test-time Out-of-Distribution Detection with Dynamic Dictionary (CVPR 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Negative Label Guided OOD Detection with Pretrained Vision-Language Models (ICLR 2024 · similarity 0.90)
- AdaNeg: Adaptive Negative Proxy Guided OOD Detection with Vision-Language Models (NeurIPS 2024 · similarity 0.90)
- ImageNet-OOD: Deciphering Modern Out-of-Distribution Detection Algorithms (ICLR 2024 · similarity 0.89)
- Conjugated Semantic Pool Improves OOD Detection with Pre-trained Vision-Language Models (NeurIPS 2024 · similarity 0.88)
- Rethinking the Evaluation of Out-of-Distribution Detection: A Sorites Paradox (NeurIPS 2024 · similarity 0.88)
- A Noisy Elephant in the Room: Is Your Out-of-Distribution Detector Robust to Label Noise? (CVPR 2024 · similarity 0.88)
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.
