AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Boosting Generative Adversarial Transferability with Self-supervised Vision Transformer Features

Shangbo Wu, Yu-an Tan, Ruinan Ma, Wencong Ma, Dehua Zhu, Yuanzhang Li

ICCV 2025 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transfer attacks on vision transformers · 4 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
jointly train on DINO (CL) and MAE (MIM) features instead of a single self-supervised ViT (ablation)dSVA · ImageNet black-box transfer (12 target models)fooling rate: +13.70% on average over single-model variantsImproved
use self-supervised ViT features (DINO or MAE) instead of supervised ViT (ablation)dSVA · ImageNet black-box transferfooling rateImproved
use facet-level query/key features instead of whole-layer token output (ablation)dSVA (DINO, MAE) · ImageNet black-box transferfooling rateImproved
add attention saliency regularization to supervised ViT surrogate (ablation)dSVA with supervised ViT · ImageNet black-box transferfooling rateHurt
add attention saliency regularization to DINO and MAE variants (ablation)dSVA (DINO), dSVA (MAE) · ImageNet black-box transferfooling rateImproved
add attention saliency regularization to the joint variant (ablation)dSVA (Joint) · ImageNet black-box transferfooling rateMixed
move feature layer from earlier layers deeper, up to layer 10 (ablation)dSVA · ImageNet black-box transferfooling rateImproved
use final layer 11 instead of penultimate layer 10 (ablation)dSVA · ImageNet black-box transferfooling rateHurt
set lambda to 0.5 rather than 0.9 (more MIM weight) (ablation)dSVA (Joint) · ImageNet black-box transferfooling rateImproved
dSVA (Joint) versus prior attacks on adversarially trained modelsdSVA (Joint) · ImageNet defense modelsfooling rate: +32.98% average margin over compared attacksImproved
dSVA versus BIA and CDA on cross-domain targetsdSVA · coarse and fine-grained cross-domain classificationfooling rate: approximately 6% on averageImproved
use ConvNet surrogates (CDA/BIA) versus dSVA on same-architecture targetsVGG-16, Res-50, Den-121 · ImageNet black-box transferfooling rate: VGG-16: 99.31 vs 96.78Hurt

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