AI Paper Atlas

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

Boosting Transferability in Vision-Language Attacks via Diversification along the Intersection Region of Adversarial Trajectory

Sensen Gao, Xiaojun Jia, Xuhong Ren, Ivor Tsang, Qing Guo

ECCV 2024 · Adversarial robustness and security › Attacks on text and graphs · approach: Vision-language model robustness · 37 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 image augmentation to SGA (ablation)SGA · Flickr30K ITR, ALBEF to TCLTR R@1 ASR: 70.07 -> 87.88Improved
sample from intersection region of adversarial trajectory with text-guided selection (vs SGA)SGA · Flickr30K ITR, ALBEF to CLIP_ViTTR R@1 ASR: 36.69 -> 46.26Improved
use proposed method instead of SGA on same-architecture transferSGA · Flickr30K ITR, ALBEF to TCLTR R@1 ASR: 87.88 -> 91.57Improved
use proposed method instead of SGA in white-box settingSGA · Flickr30K ITR, ALBEF white-boxTR R@1 ASR: 99.9 -> 99.9No clear effect
use proposed method instead of SGA for cross-architecture transfer to CLIP_CNNSGA · Flickr30K ITR, CLIP_ViT to CLIP_CNNTR R@1 ASR: 53.26 -> 64.88Improved
increase clean-image weight lambda in adversarial text generation (ablation)adversarial text loss · Flickr30K ITR, ALBEF to CLIP_ViTTR R@1 ASR: 42.21 (0.2) -> 46.26 (0.6) -> 46.13 (0.8)Mixed
increase number of intersection-region samples N beyond 5 (ablation)proposed attack · Flickr30K ITR, ALBEF to CLIP_ViTTR R@1 ASR: 46.26 (N=5) -> 45.4 (N=7)No clear effect
use proposed attack instead of SGA for cross-task transfer to captioningSGA · ITR to MSCOCO image captioningCIDEr (lower is stronger attack): 91.4 -> 88.3Improved
add intersection-region diversification and text deviationattack time cost · Flickr30K ITRtime per adversarial example: 1.58-2.28 times longer than SGAHurt
remove diversification along intersection region vs remove text deviation (ablation)proposed attack · Flickr30K ITR, ALBEF to other modelsASRMixed

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