AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Stylized Adversarial Defense

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

TPAMI 2022 · Adversarial robustness and security › Certified and empirical robustness · approach: Adversarial training refinements · 21 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 10-step stylized attack instead of single step (ablation)SAT · CIFAR10 AutoAttackrobust accuracy: 13.4 -> 53.9Improved
replace Trades with SAT (n=10)WideResNet · CIFAR10 AutoAttackrobust accuracy: 53.1 -> 53.9Improved
replace Trades with SAT (n=10) clean accuracyWideResNet · CIFAR10 testclean accuracy: 84.9 -> 91.3Improved
replace feature scattering with SAT (n=10)WideResNet · CIFAR10 AutoAttackrobust accuracy: 36.6 -> 53.9Improved
replace PCL with SAT (n=1)ResNet · CIFAR10 PGDrobust accuracy: 46.7 -> 83.5Improved
add stylized adversarial training instead of Trades/FS for corruptionsResNet/WRN · CIFAR10-Cmean accuracy: 76.6 -> 89.0Improved
add CutMix augmentation to SAT (n=10) vs Gowal et al.WRN · CIFAR10 AutoAttackrobust accuracy: 57.50 -> 58.61Improved
train SAT for 200 epochs instead of 100 (ablation)SAT (n=1) · CIFAR10PGD accuracy: 77.5 -> 81.1Improved
train SAT for 250 epochs instead of 200 (ablation)SAT (n=1) · CIFAR10PGD accuracy: 81.1 -> 81.1No clear effect
use equal loss weights alpha=gamma=beta=1 (ablation)SAT (n=1) · CIFAR10 val PGD-10robust accuracy: 79.1 -> 82.5Improved
use untargeted attack in SAT instead of stylized target samples (ablation)SAT (n=1) · CIFAR10 PGD-10robust accuracy: 81.1 -> 66.4Hurt
use targeted attack in PCL (ablation)PCL · CIFAR10 PGD-10robust accuracy: 46.7 -> 38.7Hurt

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