Research map / Label-efficient and robust learning / OOD and anomaly detection
Low-Dimensional Gradient Helps Out-of-Distribution Detection
Yingwen Wu, Tao Li, Xinwen Cheng, Jie Yang, Xiaolin Huang
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 feature input with low-dimensional gradient | MSP · ImageNet, ResNet50 | FPR95: 64.76 -> 32.38 | Improved |
| replace feature input with low-dimensional gradient | Energy · ImageNet, ResNet50 | FPR95: 57.47 -> 27.90 | Improved |
| replace feature input with low-dimensional gradient | ReAct · ImageNet, ResNet50 | FPR95: 30.70 -> 23.03 | Improved |
| replace feature input with low-dimensional gradient | KNN · ImageNet, ResNet50 | FPR95: 56.30 -> 43.43 | Improved |
| replace feature input with low-dimensional gradient | Mahalanobis · ImageNet, ResNet50 | FPR95: 81.51 -> 87.30 | Hurt |
| replace feature input with low-dimensional gradient | ReAct · ImageNet, ResNet50, SUN OOD | FPR95: 24.07 -> 45.34 | Hurt |
| replace feature input with low-dimensional gradient | KNN · ImageNet, MobileNet-v2 | FPR95: 86.50 -> 65.32 | Improved |
| replace feature input with low-dimensional gradient | Energy · ImageNet, DenseNet121 | FPR95: 50.41 -> 50.56 | No clear effect |
| add forward-backward score ensemble (ablation) | gradient-based ReAct · ImageNet, ResNet50 | FPR95: 23.03 -> 19.55 | Improved |
| add forward-backward score ensemble (ablation) | gradient-based KNN · CIFAR10, ResNet18 | FPR95: 21.55 -> 20.24 | Improved |
| use 200-dim PCA subspace instead of 10-dim PCA (ablation) | gradient + KNN · CIFAR10, ResNet18 | FPR95: 36.12 -> 21.55 | Improved |
| use 10-dim PCA subspace instead of 10-dim average-gradient subspace (ablation) | gradient + KNN · CIFAR10, ResNet18 | FPR95: 39.14 -> 36.12 | Improved |
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 1 of 1 shared result
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
- Gradient-based Out-of-Distribution Detection (ECCV 2024 · similarity 0.93)
- Splitting & Integrating: Out-of-Distribution Detection via Adversarial Gradient Attribution (ICML 2025 · similarity 0.92)
- Leveraging Perturbation Robustness to Enhance Out-of-Distribution Detection (CVPR 2025 · similarity 0.89)
- Double Descent Meets Out-of-Distribution Detection: Theoretical Insights and Empirical Analysis on the Role of Model Complexity (NeurIPS 2025 · similarity 0.89)
- Mysteries of the Deep: Role of Intermediate Representations in Out of Distribution Detection (NeurIPS 2025 · similarity 0.89)
- Detecting Out-of-Distribution Through the Lens of Neural Collapse (CVPR 2025 · similarity 0.89)
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.
