Research map / Label-efficient and robust learning / OOD and anomaly detection
Out-of-Distribution Identification: Let Detector Tell Which I Am Not Sure
Ruoqi Li, Chongyang Zhang, Hao Zhou, Chao Shi, Yan Luo
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 Attractive-Repulsive loss for feature structuration (ablation) | FOOD-ID-Proto (FCOS) · PASCAL VOC-IO | AUC: 0.705 -> 0.899 | Improved |
| add Attractive-Repulsive loss for feature structuration (ablation) | FOOD-ID-Proto (Faster RCNN) · PASCAL VOC-IO | AUC: 0.687 -> 0.861 | Improved |
| add Attractive-Repulsive loss for feature structuration (ablation) | FOOD-ID-CCNF (FCOS) · PASCAL VOC-IO | AUC: 0.888 -> 0.913 | Improved |
| add Attractive-Repulsive loss for feature structuration (ablation) | FOOD-ID-CCNF (RetinaNet) · PASCAL VOC-IO | AUC: 0.921 -> 0.928 | Improved |
| use class-conditional normalizing flow instead of unconditional flows per class (ablation) | FOOD-ID (FCOS) · PASCAL VOC-IO | AUC: 0.881 -> 0.913 | Improved |
| use class-conditional normalizing flow instead of unconditional flows per class (ablation) | FOOD-ID (Faster RCNN) · PASCAL VOC-IO | TPR@5%FPR: 39.2 -> 57.7 | Improved |
| use CCNF likelihood instead of prototype distance for uncertainty | FOOD-ID (FCOS) · PASCAL VOC-IO | AUC: 0.899 -> 0.913 | Improved |
| use FOOD-ID-CCNF instead of detector confidence score for OOD identification | FCOS · PASCAL VOC-IO | TPR@5%FPR: 43.5 -> 70.1 | Improved |
| use FOOD-ID-CCNF instead of detector confidence score for OOD identification | Faster RCNN · Crack Defect | AUC: 0.872 -> 0.908 | Improved |
| set beta to 0 in CCNF loss (no class-conditioning) (ablation) | FOOD-ID-CCNF (FCOS) · PASCAL VOC-IO | AUC: 0.913 -> 0.500 | Hurt |
| vary beta from 0.5 to 10 in CCNF loss (ablation) | FOOD-ID-CCNF (FCOS) · PASCAL VOC-IO | AUC: 0.911 -> 0.912 | No clear effect |
| exclude 10% of detections identified as OOD with FOOD-ID-CCNF | FCOS · PASCAL VOC-IO | precision of IND set: 0.797 -> 0.836 | Improved |
Closest papers by meaning
- EAT: Towards Long-Tailed Out-of-Distribution Detection (AAAI 2024 · similarity 0.89)
- YolOOD: Utilizing Object Detection Concepts for Multi-Label Out-of-Distribution Detection (CVPR 2024 · similarity 0.89)
- Learning to Shape In-distribution Feature Space for Out-of-distribution Detection (NeurIPS 2024 · similarity 0.88)
- Learning with Mixture of Prototypes for Out-of-Distribution Detection (ICLR 2024 · similarity 0.88)
- Enhancing the Power of OOD Detection via Sample-Aware Model Selection (CVPR 2024 · similarity 0.88)
- A Noisy Elephant in the Room: Is Your Out-of-Distribution Detector Robust to Label Noise? (CVPR 2024 · 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.
