AI Paper Atlas

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

AcTTA: Rethinking Test-Time Adaptation via Dynamic Activation

Hyeongyu Kim, Geonhui Han, Dosik Hwang

CVPR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Realistic and stable test-time adaptation · 1 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 learnable activation (AcTTA) to TENTTENT · CIFAR10-C WRN-28error rate (%): 18.51 -> 17.03Improved
add AcTTA to SARSAR · ImageNet-C ViT-B/16error rate (%): 53.90 -> 54.17Hurt
add AcTTA to DeYODeYO · ImageNet-C ViT-B/16error rate (%): 47.73 -> 49.64Hurt
add AcTTA to ROIDROID · ImageNet-C ResNet50 (GN)error rate (%): 64.39 -> 55.87Improved
add AcTTA to TENT at batch size 4TENT · ImageNet-C ViT-B/16 (batch 4)error rate (%): 89.92 -> 62.07Improved
freeze BN affine, adapt only activation parameters (ablation)AcTTA · CIFAR10-C WRN-28error rate (%): 18.06 -> 17.03Improved
adapt LN affine jointly with activation parameters (ablation)AcTTA · ImageNet-C ViT-B/16error rate (%): 55.30 -> 52.37Improved
replace activation with PReLU/PGELU (ablation)Tent · CIFAR10-C WRN-28error rate (%): 18.51 -> 29.62Hurt
replace activation with ACON (ablation)Tent · CIFAR10-C WRN-28error rate (%): 18.51 -> 18.46No clear effect
replace activation with PAU (ablation)Tent · ImageNet-C ViT-B/16error rate (%): 53.85 -> 99.96Hurt
make all activations learnable instead of ~50% (ablation)AcTTA · CIFAR100-C WRN-40error rate (%): 34.05 -> 44.97Hurt
pixel-wise instead of channel-wise activation parameters (ablation)AcTTA_TENT · CIFAR10-C WRN-28 (batch 128)error rate (%): 17.03 -> 58.80Hurt

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