AI Paper Atlas

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

Improving the Transferability of Adversarial Examples with Resized-Diverse-Inputs, Diversity-Ensemble and Region Fitting

Junhua Zou, Zhisong Pan, Junyang Qiu, Xin Liu, Ting Rui, Wei Li

ECCV 2020 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transfer attacks on face recognition · 84 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
replace DIM with RDIM (no probability, larger diversity size, resize back)TI-DIM, Inc-v3 source · Inc-v3ens3 black-box, single modelattack success rate (%): 46.6 -> 59.1Improved
replace DIM with RDIMNI-TI-DIM, Inc-v3 source · Inc-v3ens3 black-box, single modelattack success rate (%): 50.0 -> 53.4Improved
replace DIM with RDIMTI-DIM, Res-v2-152 source · HGD black-box, single modelattack success rate (%): 52.6 -> 53.4Improved
replace SIM with DEM multi-scale diversity ensembleSI-TIM, Inc-v3 source · Inc-v3ens3 black-box, single modelattack success rate (%): 48.4 -> 70.1Improved
replace SIM with DEMSI-NI-TIM, Inc-v3 source · Inc-v3ens3 black-box, single modelattack success rate (%): 52.1 -> 66.4Improved
replace SIM with DEMSI-TIM, IncRes-v2 source · HGD black-box, single modelattack success rate (%): 60.6 -> 80.0Improved
replace DIM with RDIM in ensemble attackTI-DIM, ensemble of four models · IncRes-v2ens black-box, ensembleattack success rate (%): 78.5 -> 79.1Improved
replace DIM with RDIM in ensemble attackTI-DIM, ensemble of four models · HGD black-box, ensembleattack success rate (%): 83.0 -> 82.1Hurt
add region fitting (step size epsilon instead of alpha) (ablation)TI-RDIM, ensemble of four models · Inc-v3ens3 black-box, ensembleattack success rate (%): 85.0 -> 91.3Improved
add region fitting (ablation)DE-TIM, ensemble of four models · Inc-v3ens3 black-box, ensembleattack success rate (%): 91.2 -> 94.7Improved
add region fitting (ablation)DE-TIM, ensemble of four models · IncRes-v2ens black-box, ensembleattack success rate (%): 88.2 -> 89.1Improved
add Nesterov momentum (NI) to DE-TIMDE-TIM, ensemble of four models · IncRes-v2ens black-box, ensembleattack success rate (%): 88.2 -> 69.8Hurt

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