AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

Principled Out-of-Distribution Detection via Multiple Testing

Akshayaa Magesh, Venugopal V. Veeravalli, Anirban Roy, Susmit Jha

JMLR 2023 · Label-efficient and robust learning › OOD and anomaly detection · approach: Provable and feature-based OOD detection

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 changedWhereMeasuredEffect
Combine Mahalanobis, Gram and Energy scores with BH-inspired test instead of using Mahalanobis alone (ablation)Mahalanobis baseline · CIFAR10 in-dist, SVHN OOD, ResNet34detection power at PF=10% (%): 82.77 -> 97.13Improved
Combine all scores with BH-inspired test instead of Energy alone (ablation)Energy baseline · CIFAR10 in-dist, SVHN OOD, DenseNetdetection power at PF=10% (%): 42.40 -> 94.57Improved
Use BH-inspired combination instead of naive averaging (ablation)Naive averaging · SVHN in-dist, CIFAR10 OOD, DenseNetdetection power at PF=10% (%): 77.22 -> 95.23Improved
Use BH-inspired combination instead of Bonferroni combination (ablation)Bonferroni combination · SVHN in-dist, CIFAR10 OOD, DenseNetdetection power at PF=10% (%): 91.77 -> 95.23Improved
Use BH-inspired combination instead of Bonferroni (ablation)Bonferroni combination · CIFAR10 in-dist, SVHN OOD, ResNet34detection power at PF=10% (%): 96.41 -> 97.13Improved
Add Gram and Energy scores to Mahalanobis in the combination (ablation)Ours - Mahala (5/4) · SVHN in-dist, CIFAR10 OOD, ResNet34detection power at PF=10% (%): 98.31 -> 97.35Hurt
Add Mahalanobis scores to Gram combination (ablation)Ours - Gram (5/4) · SVHN in-dist, CIFAR10 OOD, DenseNetdetection power at PF=10% (%): 72.84 -> 95.23Improved
Add Energy score to Gram combination (ablation)Ours - Gram (5/4) · SVHN in-dist, CIFAR10 OOD, DenseNetdetection power at PF=10% (%): 72.84 -> 78.03Improved
Add Energy score to Mahalanobis combination (ablation)Ours - Mahala (5/4) · CIFAR10 in-dist, ImageNet OOD, DenseNetdetection power at PF=10% (%): 95.19 -> 98.09Improved
Add all three score types (ablation)Mahalanobis-only combination · CIFAR10 in-dist, ImageNet OOD, DenseNetdetection power at PF=10% (%): 95.19 -> 97.20Improved
Combine all scores with BH test across datasetsSingle-score baselines · CIFAR10 in-dist, ResNet34, 4 OOD setsdetection power at PF=10% (%): 82.77 - 90.97 (Mahala) to 97.03 - 98.00Improved
Combine all scores with BH test across datasetsSingle-score baselines · CIFAR10 in-dist, DenseNet, 4 OOD setsdetection power at PF=10% (%): 42.40 - 96.89 (Energy) to 94.57 - 97.78Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy