AI Paper Atlas

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

Model-Free Test Time Adaptation for Out-of-Distribution Detection

Yi-Fan Zhang, Xue Wang, Tian Zhou, Kun Yuan, Zhang Zhang, Liang Wang, Rong Jin

TPAMI 2025 · Label-efficient and robust learning › OOD and anomaly detection · approach: Provable and feature-based OOD detection · 10 citations (OpenAlex)

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
add test-time memory augmentation with detected OOD samples to KNNKNN detector · CIFAR-10 (CE-trained)FPR95 avg: 30.77 -> 7.24Improved
add test-time memory augmentation with detected OOD samples to KNNKNN detector · ImageNet-1kFPR95 avg: 53.97 -> 15.97Improved
use supervised contrastive training instead of cross-entropyAdaODD · CIFAR-10FPR95 avg: 7.24 -> 5.51Improved
convert KNN into gradient-based TTA with triplet lossKNN · CIFAR-10 (4 OOD sets)FPR95 avg: 25.74 -> 20.76Improved
convert Mahalanobis into gradient-based TTA with triplet lossMahalanobis · CIFAR-10 (4 OOD sets)FPR95 avg: 39.07 -> 33.84Improved
convert Energy into gradient-based TTA with hinge lossEnergy · CIFAR-10 (4 OOD sets)FPR95 avg: 36.84 -> 38.91Hurt
raise selection margin gamma from 1.0 to 1.2 (fewer test samples added) (ablation)AdaODD · ImageNet-1kFPR95 avg: 15.97 -> 37.81Hurt
use gamma=infinity, adding no test samples (ablation)AdaODD · ImageNet-1kFPR95 avg: 15.97 -> 54.32Hurt
raise gamma from 1.0 to infinity increases ID reevaluation accuracy (ablation)AdaODD · ImageNet-1kID reevaluation accuracy avg: 93.10 -> 95.00Mixed
add feature normalization before kNN search (ablation)AdaODD · ImageNet-1kFPR avg: +66.2Improved
feed OOD datasets sequentially (Stream)AdaODD · CIFAR-10 TextureFPR95: 15.82 -> 3.83Improved
mix several OOD datasets into one test streamAdaODD · CIFAR-10 SVHNFPR95: 2.17 -> 1.13Improved

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.

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

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