AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Gray-box Adversarial Training

B. S. Vivek, Konda Reddy Mopuri, R. Venkatesh Babu

ECCV 2018 · Adversarial robustness and security › Black-box and transfer attacks · approach: Robustness benchmarks and defenses · 48 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
train with Gray-box Adversarial Training instead of FGSM adversarial trainingWideResNet-28-10 · CIFAR-100, eps=8/255worst-case accuracy Aw: 44.04 -> 65.43Improved
train with Gray-box Adversarial Training instead of FGSM adversarial trainingWideResNet-28-10 · CIFAR-100, eps=16/255worst-case accuracy Aw: 18.58 -> 67.41Improved
train with Gray-box Adversarial Training instead of FGSM adversarial trainingResNet-18 · CIFAR-10, eps=8/255worst-case accuracy Aw: 56.09 -> 79.28Improved
train with Gray-box Adversarial Training instead of FGSM adversarial trainingResNet-18 · CIFAR-10, eps=16/255worst-case accuracy Aw: 22.87 -> 60.81Improved
train with Gray-box Adversarial Training instead of FGSM adversarial trainingResNet-18 · CIFAR-10, eps=2/255worst-case accuracy Aw: 87.08 -> 89.46Improved
train with Gray-box Adversarial Training instead of ensemble adversarial training (set C)WideResNet-28-10 · CIFAR-100, eps=16/255worst-case accuracy Aw: 15.98 -> 67.41Improved
train with Gray-box Adversarial Training instead of ensemble adversarial training (set B)ResNet-18 · CIFAR-10, eps=16/255worst-case accuracy Aw: 34.84 -> 60.81Improved
adversarial training with FGSM instead of normal trainingWideResNet-28-10 · CIFAR-100, eps=8/255worst-case accuracy Aw: 7.09 -> 44.04Improved
use ensemble adversarial training instead of plain adversarial trainingWideResNet-28-10 · CIFAR-100, eps=8/255worst-case accuracy Aw: 44.04 -> 36.8Hurt
use ensemble adversarial training instead of plain adversarial trainingResNet-18 · CIFAR-10, eps=16/255worst-case accuracy Aw: 22.87 -> 31.01Improved
use FGSM instead of FGSM-LL or FGSM-Rand as the attack (ablation)normally trained WideResNet-28-10 · CIFAR-100accuracy under attackImproved
increase drop-in-loss interval D from 0.2 to 0.4 for saving intermediate models (ablation)GAT ResNet-18 · CIFAR-10robustness plot valleyHurt

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