AI Paper Atlas

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

Boosting Adversarial Transferability via Residual Perturbation Attack

Jinjia Peng, Zeze Tao, Huibing Wang, Meng Wang, Yang Wang

ICCV 2025 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transferable adversarial examples · 5 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
use residual gradient (current minus EMA reference) as perturbation directionflatness-based transfer attack · ImageNet, Inc-v3 surrogate, 8 targetsaverage attack success rate (%): 63.6 -> 64.6 (vs PGN)Improved
combine ResPA with DIMDIM · ImageNet, Res-50 surrogateaverage attack success rate (%): 79.6 -> 91.0Improved
combine ResPA with TIMTIM · ImageNet, Res-50 surrogateaverage attack success rate (%): 70.6 -> 90.9Improved
combine ResPA with SSASSA · ImageNet, Res-50 surrogateaverage attack success rate (%): 82.4 -> 86.6Improved
combine ResPA with AdmixAdmix · ImageNet, Res-50 surrogateaverage attack success rate (%): 70.2 -> 81.8Improved
ResPA in ensemble-surrogate setting vs PGNResPA · ImageNet, ensemble of Res-50/Vgg-19/Den-121average attack success rate (%): 95.8 -> 96.0Improved
ResPA against defense models vs PGNResPA · ImageNet, 7 defense models, Inc-v3 surrogateaverage attack success rate (%): 58.8 -> 59.8Improved
ResPA on Re-ID attack vs TPAResPA · Market-1501 Re-ID, average mAPmAP (%) lower is better: 19.20 -> 18.77Improved
increase sampling number N (ablation)ResPA · ImageNet, Res-50 surrogateattack success rate: stabilizes after N>15 (CNNs), N>25 (ViT/adv-trained)Improved
increase sampling boundary beta beyond 2.5 epsilon (ablation)ResPA · ImageNet, Res-50 surrogateattack success rateHurt
increase penalty coefficient gamma above 0.9 (ablation)ResPA · ImageNet, Res-50 surrogateattack success rateHurt
vary EMA decay theta within [0.2, 1] (ablation)ResPA · ImageNet, Res-50 surrogateattack success rateNo clear effect

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