AI Paper Atlas

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

Dual Distribution Estimation for Zero-shot Noisy Test-Time Adaptation with VLMs

Wenjie Zhu, Yabin Zhang, Liang Xu, Xin Jin, Wenjun Zeng, Lei Zhang

ECCV 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Pseudo-label test-time adaptation · 3 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
replace inclusion-only PFDE with PFDE using inclusion and exclusion (ablation)DDE · ImageNet noisy streamAcc_H: 72.77 -> 73.52Improved
add NLDE (ablation)DDE (no PFDE) · ImageNet noisy streamAcc_N: 83.96 -> 95.99Improved
add NLDE to PFDE (in+ex) (ablation)DDE · ImageNet noisy streamAcc_H: 73.52 -> 76.82Improved
add adaptive lambda (ablation)DDE · ImageNet noisy streamAcc_H: 76.82 -> 78.00Improved
increase noise rate from 0.1 to 0.7 (ablation)DDE · ImageNet noisy streamAcc_H: 78.00 -> 76.87No clear effect
scale backbone from RN50 to ViT-L (ablation)DDE · ImageNet noisy streamAcc_H: 81.20 (ViT-L), +10.12% over RN50Improved
increase number of negative labels M from 500 to 1000 (ablation)DDE · ImageNet noisy streamAcc_H: 78.00 -> 71.28Hurt
reduce scaling factor rho from 0.005 to 0.0005 (ablation)DDE · ImageNet noisy streamAcc_H: 78.00 -> 77.68No clear effect
increase exclusion GDA weight beta up to 0.5 (ablation)DDE · ImageNet noisy streamAcc_H: 77.42 -> 78.00Improved
increase cache size Q to 5000 (ablation)DDE · ImageNet noisy streamAcc_HImproved
use DDE instead of AdaNDAdaND · ImageNet noisy TTAAcc_H: 73.09 -> 76.79Improved
use DDE instead of AdaND for OOD detectionAdaND · ImageNet OOD detectionFPR95: 16.00 -> 9.80Improved

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