AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

Adversarial Attacks Already Tell the Answer: Directional Bias-Guided Test-time Defense for Vision-Language Models

Liangsheng Liu, Si Chen, Jiamin Wu, Weiwei Feng, Zhixin Cheng, Xiaotian Yin, Wenfei Yang, Tianzhu Zhang

ICLR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Vision-language model robustness · 0 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 DBD test-time defenseCLIP-ViT-B/32 · 10 fine-grained datasets, PGD-100 eps=4/255robust accuracy: 0.0 -> 91.3 (R-TPT 38.1)Improved
add DBD test-time defenseCLIP-ViT-B/16 · 5 ImageNet-OOD, PGD-100 eps=4/255robust accuracy vs clean: clean 63.7 vs robust 97.7Improved
add DBD test-time defenseCLIP-ViT-B/32 · ImageNet-OOD avgclean accuracy: 52.7 -> 55.0Improved
add linear feature shift along Defense Direction (ablation)DBD with 5 transformation types · 15 datasets, three attack settingsrobust accuracy: 34.8 -> 92.1Improved
add linear feature shift along Defense Direction (ablation)DBD with 5 transformation types · 15 datasets, three attack settingsclean accuracy: 61.5 -> 58.9Hurt
add DB-score threshold two-stream strategy (ablation)DBD · 15 datasets, three attack settingsclean accuracy: 58.9 -> 61.2Improved
add DB-score threshold two-stream strategy (ablation)DBD · 15 datasets, three attack settingsrobust accuracy: 92.1 -> 91.7No clear effect
use multiple transformations instead of a single one (ablation)DBD with feature shift · 15 datasets, three attack settingsclean accuracy: 53.0 -> 58.9 (crop-resize-flip only vs all)Improved
set feature shift lambda from 1.0 to 2.5 (ablation)DBD · 15 datasets, three attack settingsrobust accuracy: 34.8 -> 92.1Improved
set DB-score threshold tau to 0.8 (ablation)DBD · 15 datasets, three attack settingsoverall mean accuracyImproved
use pseudo-labels for attack generationDBD · 6 fine-grained datasets, CLIP-ViT-B/16robust accuracy: 66.8 vs clean 67.5No clear effect

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