AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

Attack to Fool and Explain Deep Networks

Naveed Akhtar, Mohammad A. A. K. Jalwana, Mohammed Bennamoun, Ajmal Saeed Mian

TPAMI 2021 · Adversarial robustness and security › Attacks on text and graphs · approach: Adversarial perturbations beyond images · 37 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 class-specific attack with leakage suppression instead of enhanced UAN (l_inf)VGG-16 · ImageNet foolingavg fooling ratio: 62.6 -> 80.6Improved
use class-specific attack instead of enhanced UAN (l_inf)MobileNet-V2 · ImageNet foolingavg fooling ratio: 64.6 -> 86.4Improved
use class-specific attack instead of enhanced UAN (l2)MobileNet-V2 · ImageNet foolingavg fooling ratio: 73.4 -> 90.2Improved
use class-specific attack instead of enhanced UAN (l2)Inception-V3 · ImageNet foolingavg fooling ratio: 56.2 -> 70.8Improved
apply original UAN without l_p-ball projection to targeted class foolingUAN · ImageNet foolingfooling ratio: 10-15%Hurt
use explanation-objective attack instead of PGD of Santurkar et al. for image generationrobust ResNet · image generationclassifier confidence: 0.712 -> 0.793Improved
use explanation-objective attack instead of Santurkar et al. for inpaintingrobust ResNet · ImageNet inpaintingPSNR: 20.13 -> 22.31Improved
use explanation-objective attack instead of Santurkar et al. for inpaintingrobust ResNet · ImageNet inpaintingSSIM: 0.79 -> 0.87Improved
use explanation-objective attack instead of Santurkar et al. for inpaintingrobust ResNet · ImageNet inpaintingLPIPS: 0.191 -> 0.188No clear effect
use explanation-objective attack instead of Santurkar et al. for sketch manipulationrobust ResNet · sketch-to-imageclassifier confidence: 0.76 -> 0.79Improved
add perturbation refinement with Otsu-thresholded attention mask (ablation)explanation attack · VGG-16 explanationpattern clarityImproved
increase visual difference between source and target classunbounded attack · MobileNet-V2l2 norm needed: 3586.4 -> 7448.8Hurt

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