Research map / Label-efficient and robust learning / OOD and anomaly detection
Revisit PCA-based Technique for Out-of-Distribution Detection
Xiaoyuan Guan, Zhouwu Liu, Wei-Shi Zheng, Yuren Zhou, Ruixuan 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 |
|---|---|---|---|
| fuse regularized PCA reconstruction error into Energy score (ablation) | Energy, ResNet50 · ImageNet-1K OOD average | FPR95: 58.41 -> 49.50 | Improved |
| fuse regularized reconstruction error into ReAct (ablation) | ReAct, ResNet50 · ImageNet-1K OOD average | FPR95: 31.43 -> 18.66 | Improved |
| fuse regularized reconstruction error into MSP (ablation) | MSP, ResNet50 · ImageNet-1K OOD average | FPR95: 66.95 -> 62.71 | Improved |
| fuse regularized reconstruction error into BATS (ablation) | BATS, MobileNet-v2 · ImageNet-1K SUN/Places | FPR95 and AUROC: iNaturalist FPR95 49.57 -> 50.51; Places 64.48 -> 66.43; average 52.91 -> 48.90 | Mixed |
| fuse regularized reconstruction error into Energy (ablation) | Energy, ResNet34 · CIFAR-10 OOD average | FPR95: 24.71 -> 25.19 | No clear effect |
| fuse regularized reconstruction error into ReAct (ablation) | ReAct, ResNet34 · CIFAR-10 OOD average | FPR95: 25.90 -> 27.49 | Hurt |
| fuse regularized reconstruction error into ReAct (ablation) | ReAct, ResNet34 · CIFAR-100 OOD average | FPR95: 54.11 -> 43.64 | Improved |
| combine reconstruction error with LogitNorm on CIFAR-10 | LogitNorm, ResNet34 · CIFAR-10 OOD average | FPR95: 15.03 -> 13.06 | Improved |
| combine reconstruction error with LogitNorm on CIFAR-100 WideResNet | LogitNorm, WideResNet28-10 · CIFAR-100 OOD average | FPR95: 51.23 -> 51.79 | No clear effect |
| reduce subspace dimension k so that less feature signal is preserved (ablation) | regularized reconstruction error · CIFAR-100 and ImageNet-1K | AUROC | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ReAct — ReAct: Out-of-distribution Detection With Rectified Activations (NeurIPS 2021) · better on 5 of 5 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
- Kernel PCA for Out-of-Distribution Detection (NeurIPS 2024 · similarity 0.90)
- Test-Time Linear Out-of-Distribution Detection (CVPR 2024 · similarity 0.88)
- Exploiting Discrepancy in Feature Statistic for Out-of-Distribution Detection (AAAI 2024 · similarity 0.87)
- What If the Input is Expanded in OOD Detection? (NeurIPS 2024 · similarity 0.87)
- Enhancing the Power of OOD Detection via Sample-Aware Model Selection (CVPR 2024 · similarity 0.87)
- Mining In-distribution Attributes in Outliers for Out-of-distribution Detection (AAAI 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.
