AI Paper Atlas

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

Test-Time Prompt-Agnostic Decomposition

Junze Wang, Lei Fan, Dezheng Zhang, Donglin Di, Yang Song, Sidong Liu, Cong Cong

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 test-time prompt tuning approach that decomposes prompt updates to be prompt-agnostic, aiming to prevent accumulation of noisy unlabeled objectives across test-time steps.

Problem

Test-time prompt tuning (TPT) adapts large pretrained models to unseen distribution shifts using unlabeled test data; prompt updates are optimized locally at each step, and noisy or biased unlabeled objectives can accumulate across steps and destabilize adaptation.

Approach

Decomposes the prompt update so that it does not depend solely on the current prompt state, reducing error accumulation from reusing updated prompts across test-time steps (based on the abstract only; details unavailable).

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add TPD on top of visual TPT methods (VPT, E2VPT, VFPT, BPT)ViT-based visual prompt tuning · Domain generalization OOD meanaccuracy: OOD mean gains 5.51-8.18%Improved
add TPD on top of CLIP text TPT methods (TPT, C-TPT, R-TPT, O-TPT, CAT, DyPt)CLIP-based text prompt tuning · Domain generalization OOD meanaccuracy: +2.27 to +2.69Improved
add TPD instead of simple stabilization (clipping, EMA, norm control, low-pass filtering) (ablation)VPT · DG meanaccuracy: 44.28 -> 47.84Improved
add TPD instead of low-pass filtering (ablation)TPT · CSD meanaccuracy: 66.23 -> 68.03Improved
use diagonal Koopman operator instead of full operator (ablation)TPD (AKC) · VPT DG meanaccuracy: 43.58 -> 45.92Improved
use Hankel delay embedding instead of plain projection (ablation)TPD (HUR) · VPT DG meanaccuracy: 42.31 -> 45.98Improved
use full routing (persistent 1, oscillatory beta_c, residual beta_n) instead of removing any one routing component (ablation)TPD HUR routing · VPT and TPT DG/CSDaccuracy: 47.06 -> 47.84 (VPT DG, w/o Pres.)Improved
add TPD to prompt-tuning methods for calibrationTPT, C-TPT and others · ImageNet-A ECEECE: 16.42% -> 12.27%Improved
add TPD to TPT on ImageNet-R for calibrationTPT · ImageNet-R ECEECE: 4.36% -> 1.62%Improved
add TPD to VPT across batch sizesVPT · ImageNet-A batch-size sweepaccuracy: 1.14% -> 12.82% at B=1; decrease at B=8Mixed
increase batch size to 128 with TPD (ablation)VPT + TPD · ImageNet-Aaccuracy: 19.17% -> 38.98% at B=128Improved
naive entropy-minimization test-time tuning of visual prompts without stabilizationVPT · Domain generalization OOD meanaccuracy: 46.17 -> 39.66Hurt

Experimental setup

Builds on: test-time prompt tuning

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