Research map / Label-efficient and robust learning / OOD and anomaly detection
Exploiting Local Flatness for Efficient Out-of-Distribution Detection
Seonghwan Park, Hyunji Jung, Dongyeop Lee, Namhoon Lee
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 partial feature normalization (0<alpha<1) to the baselines and Fold (ablation) | MSP, EBO, ReAct, ASH, Fold variants · CIFAR-10 | AUROC: 87.13 -> 91.13 | Improved |
| add partial feature normalization (0<alpha<1) to the baselines and Fold (ablation) | MSP, EBO, ReAct, ASH, Fold variants · ImageNet-200 | AUROC: 87.64 -> 89.91 | Improved |
| use full normalization (alpha=1) instead of partial on ImageNet-200 (ablation) | MSP, EBO, ReAct, ASH, Fold variants · ImageNet-200 | AUROC: 89.91 -> 79.02 | Hurt |
| use full normalization (alpha=1) instead of none on CIFAR-10 (ablation) | Fold and baselines · CIFAR-10 | AUROC: 87.13 -> 91.34 | Improved |
| compute curvature in feature space instead of logit space (ablation) | Fold · Four benchmarks average | AUROC: 86.02 -> 86.81 | Improved |
| use feature Hessian instead of logit Hessian on CIFAR-10 (ablation) | Fold · CIFAR-10 | AUROC: 92.52 -> 92.46 | No clear effect |
| use feature Hessian instead of logit Hessian on ImageNet-200 (ablation) | Fold · ImageNet-200 | AUROC: 86.35 -> 88.65 | Improved |
| add partial normalization, raising Cohen's d on ImageNet-200 (ablation) | Fold · ImageNet-200 | Cohen's d: 1.605 -> 1.689 | Improved |
| full normalization collapses score separability (ablation) | Fold · ImageNet-200 | Cohen's d: 1.605 -> 0.520 | Hurt |
| combine Fold with ReAct (Fold-r) | Fold · Four benchmarks average | AUROC: 87.20 -> 87.91 | Improved |
| combine Fold with ASH (Fold-a) on CIFAR-10 | Fold · CIFAR-10 | AUROC: 92.46 -> 89.43 | Hurt |
| replace grid-searched alpha (OOD validation) with AutoFold ID logit-masking tuning | Fold · Four benchmarks average | AUROC: 87.20 -> 86.24 | Mixed |
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 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
- Detecting Out-of-Distribution Through the Lens of Neural Collapse (CVPR 2025 · similarity 0.89)
- Is Complex Training Necessary for Long-Tailed OOD Detection? A Re-think from Feature Geometry (NeurIPS 2026 · similarity 0.88)
- Image Background Serves as Good Proxy for Out-of-distribution Data (ICLR 2024 · similarity 0.88)
- A Systematic Analysis of Out-of-Distribution Detection Under Representation and Training Paradigm Shifts (NeurIPS 2026 · similarity 0.88)
- Double Descent Meets Out-of-Distribution Detection: Theoretical Insights and Empirical Analysis on the Role of Model Complexity (NeurIPS 2025 · similarity 0.87)
- Improving Out-of-Distribution Detection via Dynamic Covariance Calibration (ICML 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.
