AI Paper Atlas

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

Defense Against Adversarial Attacks Using Feature Scattering-based Adversarial Training

Haichao Zhang, Jianyu Wang

NeurIPS 2019 · Adversarial robustness and security › Attacks on text and graphs · approach: Defenses against universal perturbations · 118 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
feature-scattering adversarial training vs Madry PGD trainingadversarial training · CIFAR10 PGD20accuracy: 44.9 -> 70.5Improved
feature scattering vs Bilateraladversarial training · CIFAR10 PGD20accuracy: 57.5 -> 70.5Improved
feature scattering vs Madryadversarial training · CIFAR100 PGD20accuracy: 22.6 -> 47.2Improved
feature scattering vs Madryadversarial training · CIFAR10 cleanaccuracy: 85.7 -> 90.0Improved
feature scattering vs Bilateraladversarial training · CIFAR10 cleanaccuracy: 91.2 -> 90.0Hurt
generate perturbations by feature scattering instead of supervised loss (ablation)adversarial training · CIFAR10 PGD20accuracy: 56.0 -> 70.5Improved
feature scattering instead of random perturbation (ablation)adversarial training · CIFAR10 PGD20accuracy: 29.9 -> 70.5Improved
OT matching instead of uniform matching (ablation)feature scattering · CIFAR10 PGD20accuracy: 57.1 -> 70.5Improved
OT matching instead of identity matching (ablation)feature scattering · CIFAR10 PGD20accuracy: 57.5 -> 70.5Improved
IPOT solver instead of Sinkhorn (ablation)feature scattering · CIFAR10 PGD20accuracy: 70.5 -> 69.9No clear effect

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