AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Energy Waveify and Redistribution for Test-Time Adaptation: A Control System Perspective

Zhenbin Wang, Lei Zhang, Lituan Wang, Zhenwei Zhang, Guangwu Qian, Yan Wang, Wei Huang

CVPR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Realistic and stable test-time adaptation

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
adapt with APT (wave-based energy redistribution) instead of Energy TEAWRN-28-10 BatchNorm · CIFAR-10-C severity 5accuracy: 83.34 -> 86.93Improved
adapt with APT instead of Energy TEAWRN-28-10 BatchNorm · CIFAR-100-C severity 1-5accuracy: 71.22 -> 73.57Improved
adapt with APT instead of Energy TEAResNet-50 GroupNorm · TinyImageNet-200-C severity 1-5accuracy: 30.41 -> 35.45Improved
adapt with APT instead of best baseline on average accuracyWRN-28-10 BatchNorm · CIFAR/TinyImageNet-C average over severitiesaccuracy: +2.21% average over best baselineImproved
adapt with APT instead of best baselineResNet-50 GroupNorm · CIFAR/TinyImageNet-C average over severitiesaccuracy: +2.30% average over best baselineImproved
use sampling-free wave-based adaptation instead of top-3 baseline methodsAPT · CIFAR-10-C / CIFAR-100-C, P100 GPUadaptation time: 1/3 to 1/7 of the time of the top three baselinesImproved
adapt with APT instead of Energy TEAWRN-28-10 BatchNorm · CIFAR-10 calibrationMCE: 47.37% -> 42.74%Improved
adapt with entropy-based CRKDWRN-28-10 BatchNorm · CIFAR-10 calibrationMCE: 57.99% -> 58.28%Hurt
adapt with APTPACS single-source · PACS all four source domainsaccuracy: APT averages 29.13, 44.24, 35.01, 27.54Improved
increase adaptation steps (ablation)APT · CIFAR-10-C, CIFAR-100-C, TinyImageNet-200-Cleft/right energy ratio and accuracyImproved

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