AI Paper Atlas

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

Training-Free Active Test-Time Adaptation for Vision-Language Models

Jihwan Bang, Sumyeong Ahn, Hwanjun Song, Jae-Gil Lee

NeurIPS 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Pseudo-label test-time adaptation · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a training-free active test-time adaptation method for vision-language models that selectively queries external labels for hard shifted samples instead of relying solely on self-generated pseudo-labels.

Problem

Test-time adaptation for VLMs is typically fully unsupervised, relying on the model's own pseudo-labels, confidence, or entropy, which is unreliable on hard shifted samples; supervised online adaptation assumes labels for all samples, which is impractical.

Approach

Introduces active, training-free test-time adaptation in which a small budget of external label queries is spent on the most uncertain shifted samples, sitting between fully unsupervised TTA and fully supervised online adaptation. (Details inferred from the abstract framing only.)

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add ACTOR retrieval-based correction with queried labels to TDATDA · Cross-domain, ViT-B/16average accuracy: 68.14 -> 72.66Improved
add ACTOR to TDATDA · Cross-domain, ResNet-50average accuracy: 60.78 -> 66.16Improved
add ACTOR to TDATDA · OOD benchmark (5 datasets)average accuracy: 65.22 -> 66.40 (ViT-B/16); 49.64 -> 50.80 (RN50)Improved
add ACTOR to TDA on ImageNet-V2TDA RN50 · ImageNet-V2top-1 accuracy: 55.10 -> 54.99No clear effect
use training-free ACTOR instead of supervised Online-CoOp / Replay-CoOp with same queried labelsCLIP ViT-B/16 · 15 datasetsaverage accuracy: 65.15 / 65.20 -> 67.65Improved
add Verifier consistency check, full consistency-and-reliability rule vs 5-NN majority vote (ablation)ACTOR retrieval correction · 5-dataset subset, TDA ViT-B/16accuracy: 67.22 -> 69.50Improved
add reliability threshold on top of consistency (ablation)ACTOR retrieval correction · 5-dataset subsetharmful-correction rate: 11.99 -> 5.28Improved
add consistency check on top of reliability-only reuse (ablation)ACTOR retrieval correction · 5-dataset subsetaccuracy: 67.30 -> 69.50Improved
increase number of Verifiers from 0 to 1 (ablation)ACTOR · retrieval correction accuracycorrection accuracyImproved
increase number of Verifiers from 1 to 2 (ablation)ACTOR · overall accuracyaccuracyHurt
increase label latency from 1 to 1000 samples (ablation)ACTOR · DTDaccuracy: 46 -> 42Hurt
increase label latency from 1 to 1000 samples (ablation)TDA+ACTOR · ImageNet-Saccuracy: around 43No clear effect

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