AI Paper Atlas

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

CONGA:Confidence-and-Gradient-Aware Learning Rate Schedule for Test Time Adaptation

Lv Shaoran, Xinyao Li, Jingjing Li

ICML 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Pseudo-label 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
add CONGA scheduleTent (ViT-Base) · ImageNet-C standard TTAaccuracy (%): 61.1 -> 64.3Improved
add CONGA scheduleTent (ResNet-50) · ImageNet-C standard TTAaccuracy (%): 43.2 -> 45.9Improved
add CONGA scheduleDeYO (ViT-Base) · ImageNet-C standard TTAaccuracy (%): 66.2 -> 68.5Improved
add CONGA scheduleEATA (ResNet-50) · ImageNet-C CTTAaccuracy (%): 41.8 -> 42.2Improved
add CONGA scheduleTent (ViT-Base) · ImageNet-C CTTAaccuracy (%): 62.4 -> 62.6No clear effect
add CONGA scheduleReservoir (ViT-Base) · ImageNet-C CTTAaccuracy (%): 66.0 -> 66.6Improved
add CONGA scheduleDeYO (ViT-Base) · ImageNet-Raccuracy (%): 66.1 -> 68.3Improved
use cosine decay instead of constant LR (ablation)Tent (ViT-Base) · ImageNet-Sketchaccuracy (%): 49.1 -> 50.1Improved
use exponential decay instead of constant LR (ablation)Tent (ViT-Base) · ImageNet-Sketchaccuracy (%): 49.1 -> 48.4Hurt
use LR decay instead of LR warmup (ablation)FOA with CONGA · ImageNet-Sketchaccuracy (%)Improved
add Learning Gate on top of Cycle Regulator (ablation)Tent + CONGA (ViT) · ImageNet-Sketchaccuracy (%): 49.2 -> 49.5Improved
add Safety Gate on top of Cycle Regulator and Learning Gate (ablation)Tent + CONGA (ViT) · ImageNet-Sketchaccuracy (%): 49.5 -> 50.7Improved

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