Research map / Label-efficient and robust learning / OOD and anomaly detection
VRA: Variational Rectified Activation for Out-of-distribution Detection
Mingyu Xu, Zheng Lian, Bin Liu, Jianhua Tao
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 plain activations with VRA rectification | DenseNet-101 · CIFAR-10 OOD average | FPR95: 20.84 -> 17.74 | Improved |
| add intermediate-activation amplification (VRA+) (ablation) | VRA · ImageNet OOD average | FPR95: 25.49 -> 23.31 | Improved |
| add intermediate-activation amplification (VRA+) (ablation) | VRA · CIFAR-100 OOD average | FPR95: 47.12 -> 43.34 | Improved |
| combine rectified features with logit energy (VRA++) (ablation) | VRA+ · CIFAR-100 OOD average | FPR95: 43.31 -> 35.20 | Improved |
| use fixed thresholds instead of adaptive quantile thresholds (ablation) | VRA · CIFAR-100 OOD average | FPR95: 47.12 -> 56.35 | Hurt |
| use fixed thresholds instead of adaptive quantile thresholds (ablation) | VRA · CIFAR-100 OOD average | AUROC: 90.21 -> 86.09 | Hurt |
| estimate the optimal activation using 1% sampled ID and OOD data rather than practical VRA (ablation) | VRA · CIFAR-10 OOD average | AUROC: 96.47 -> 97.75 | Improved |
| combine VRA feature scoring with logit energy | VRA++ · ImageNet OOD average on BiT | FPR95: 41.47 -> 30.67 | 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 3 of 3 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
- Leveraging Perturbation Robustness to Enhance Out-of-Distribution Detection (CVPR 2025 · similarity 0.86)
- Scaling for Training Time and Post-hoc Out-of-distribution Detection Enhancement (ICLR 2024 · similarity 0.86)
- Detecting Out-of-Distribution Through the Lens of Neural Collapse (CVPR 2025 · similarity 0.85)
- Kernel PCA for Out-of-Distribution Detection (NeurIPS 2024 · similarity 0.85)
- Discriminability-Driven Channel Selection for Out-of-Distribution Detection (CVPR 2024 · similarity 0.85)
- The Best of Both Worlds: On the Dilemma of Out-of-distribution Detection (NeurIPS 2024 · similarity 0.85)
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.
