AI Paper Atlas

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

On Generating Transferable Targeted Perturbations

Muzammal Naseer, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Fatih Porikli

ICCV 2021 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transferable adversarial examples · 92 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 augmentations during generator training (ablation)TTP · Per-target average, VGG16 (SIN) target, ResNet50 sourcetarget top-1 accuracy (%): 17.8 -> 30.3Improved
add augmentations during generator training (ablation)TTP · Per-target average, VGG16 (natural) target, ResNet50 sourcetarget top-1 accuracy (%): 63.5 -> 78.8Improved
use a ResNet ensemble as source instead of ResNet50 (ablation)TTP · Per-target average, VGG16 (SIN) targettarget top-1 accuracy (%): 30.3 -> 53.6Improved
use a ResNet ensemble as source instead of ResNet50 (ablation)TTP · Per-target average, VGG16 (natural) targettarget top-1 accuracy (%): 78.8 -> 89.1Improved
train with distribution matching loss only instead of cross-entropy (GAP) or relativistic cross-entropy (CDA) (ablation)generator attack · ResNet50 to VGG16 (SIN)target transferability: more than 100%Improved
add Gaussian smoothing operator W (3x3) (ablation)TTP · ResNet50 to VGG16 (SIN)target transferabilityImproved
add augmentations with both distribution matching losses (ablation)TTP · ResNet50 to VGG16 (SIN)target transferabilityImproved
add neighbor similarity loss (ablation)TTP · ResNet50 to VGG16 (SIN)target transferabilityImproved
use TTP instead of CDAgenerative attack · 10-targets all-source, ResNet50 to VGG19_BNtarget top-1 accuracy (%): 64.58 -> 78.15Improved
use TTP instead of CDA against stylized-trained ResNet50 (100 targets)generative attack · 100-targets, SIN targettarget top-1 accuracy (%): 7.6 -> 31.0Improved
raise perturbation budget from 16 to 32TTP · 10-targets, SIN targettarget top-1 accuracy (%): 30.17 -> 41.52Improved
apply median blur defense on clean images (VGG19_BN)VGG19_BN · ImageNet val clean accuracytop-1 accuracy (%): 74.24 -> 53.86Hurt

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