Adversarial Robustness

Defending against adversarial examples.

1
Datasets
3
Results
Robust Accuracy
Canonical metric
Canonical Benchmark

RobustBench CIFAR-10 Linf

RobustBench standardized adversarial robustness benchmark: CIFAR-10 with Linf perturbation eps=8/255, evaluated with AutoAttack.

Primary metric: Robust Accuracy
View full leaderboard

Top 10

Leading models on RobustBench CIFAR-10 Linf.

RankModelRobust AccuracyYearSource
1
Peng et al. 2023 (WRN-70-16)
71.12026paper
2
Wang et al. 2023 (WRN-70-16)
70.72026paper
3
Gowal et al. 2021 (WRN-70-16)
66.12026paper

All datasets

1 dataset tracked for this task.

Related tasks

Other tasks in Adversarial.