AI Paper Atlas

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

Bilateral Adversarial Training: Towards Fast Training of More Robust Models Against Adversarial Attacks

Jianyu Wang, Haichao Zhang

ICCV 2019 · Adversarial robustness and security › Certified and empirical robustness · approach: Adversarial training refinements · 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
add random start to FGSM-based training (ablation)one-step adversarial training · CIFAR10 CE7 attackaccuracy (%): 0.0 -> 48.0Improved
add random start to LL targeted training (ablation)one-step adversarial training · CIFAR10 CE7 attackaccuracy (%): 0.0 -> 36.0Improved
use most-confusing target instead of least-likely target in training (ablation)R-LL · CIFAR10 CE7 attackaccuracy (%): 36.0 -> 48.4Improved
add label smoothing (ablation)R-FGSM · CIFAR10 CE20accuracy (%): 42.9 -> 49.0Improved
add label smoothing (ablation)R-MC · CIFAR10 CE20accuracy (%): 43.4 -> 53.3Improved
replace label smoothing with adversarial label (ablation)R-MC-LS · CIFAR10 CE20accuracy (%): 53.3 -> 55.3Improved
add label adversary to multi-step training (ablation)Madry* · CIFAR10 CE20accuracy (%): 47.6 -> 53.2Improved
longer training (ablation)R-MC-LA · CIFAR10 CE20accuracy (%): 55.3 -> 57.8Improved
train with perturbation budget 4 instead of 8 (ablation)R-MC-LA · CIFAR10 CE100accuracy (%): 52.9 -> 60.1Improved
use weaker PGD2-8 attack in training instead of PGD7-2 (ablation)Madry adversarial training · CIFAR10 PGD100-2accuracy (%): 47.2 -> 46.2No clear effect
increase training images per class from 0.5K to 5K (ablation)R-MC-LA · CIFAR10 CE20accuracy (%): 25.1 -> 55.3Improved
use main-paper label solution instead of proportional-gradient solution (ablation)R-MC-LA · CIFAR10 CE100accuracy (%): 51.1 -> 52.9Improved

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