AI Paper Atlas

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

Understanding and Improving Fast Adversarial Training

Maksym Andriushchenko, Nicolas Flammarion

NeurIPS 2020 · Adversarial robustness and security › Certified and empirical robustness · approach: Adversarial training refinements · 44 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
reduce FGSM step size to 7/255 (no random init) (ablation)FGSM AT · CIFAR-10 eps=8/255PGD-50-10 adversarial accuracy (%): 36.35 -> 45.35Improved
add random initialization step to FGSM (ablation)FGSM AT · CIFAR-10 eps=8/255PGD-50-10 adversarial accuracy (%): 36.35 -> 45.60Improved
add random step to FGSM at larger eps (ablation)FGSM AT · CIFAR-10 eps=10/255adversarial accuracy (%): 0%Hurt
project PGD-10 perturbations onto corners of l_inf ball (ablation)PGD-10 AT · CIFAR-10 eps=8/255adversarial accuracy (%): 50.48 -> 50.64No clear effect
zero out the Laplace-like filter w4 (ablation)Single-layer CNN after catastrophic overfitting · CIFAR-10gradient alignment: 0.08 -> 0.71Improved
add GradAlign regularizer to FGSM ATFGSM AT · CIFAR-10 and SVHN, large epsPGD-50-10 adversarial accuracyImproved
add GradAlign regularizer to PGD-2 ATPGD-2 AT · CIFAR-10 eps=16/255PGD-50-10 adversarial accuracyImproved
add GradAlign regularizer to FGSM ATFGSM AT · CIFAR-10 and SVHN, small epsadversarial accuracyNo clear effect
add GradAlign to FGSM ATFGSM AT · ImageNet eps 2/255 and 4/255adversarial accuracyNo clear effect
add GradAlign to FGSM ATFGSM AT · Training timetraining cost: 3x slowdownHurt

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