AI Paper Atlas

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

Structure Invariant Transformation for better Adversarial Transferability

Xiaosen Wang, Zeliang Zhang, Jianping Zhang

ICCV 2023 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transferable adversarial examples · 118 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
apply different transformations locally to image blocks (SIA) instead of global transformationsMI-FGSM · ImageNet, ResNet-18 to 7 modelsaverage attack success rate (%): 59.7 -> 92.1Improved
add any single block transformation (weakest, DCT) (ablation)SIA · ImageNet, ResNet-18 to 7 modelsaverage attack success rate (%): 59.7 -> 67.3Improved
include each of the ten transformations (adding back any removed one) (ablation)SIA · ImageNet, ResNet-18 to 7 modelsaverage attack success rate (%): 88.3-90.7 -> 92.1Improved
combine two transformations instead of one (ablation)SIA · ImageNet, ResNet-18 to 7 modelsaverage attack success rate (%)Improved
apply transformation per image block rather than to the whole image s x s times (ablation)SIA · ImageNet, ResNet-18 to 7 modelsaverage attack success rate (%)Improved
increase number of blocks s beyond 3 (ablation)SIA · ImageNet, ResNet-18 to 7 modelsattack success rateHurt
increase number of transformed images N from 1 to 20 (ablation)SIA · ImageNet, ResNet-18 to 7 modelsattack success rateImproved
increase N beyond 20 (ablation)SIA · ImageNet, ResNet-18 to 7 modelsattack success rateNo clear effect
integrate SIA into SGM and LinBPinput-transformation attacks · ImageNet, ResNet-18 sourceaverage attack success rate (%): +3.3% (SGM), +3.6% (LinBP) over best baselinesImproved
use SIA instead of best baseline on ensemble modelsensemble attack · ImageNet, 8 modelsattack success rate (%): +6.5% over winner-upImproved

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