TRADES-ANCRA/ResNet18

Reported on 3 benchmarks across 2 tasks · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Adversarial3 results

Adversarial AttackonCIFAR-10
Attack: AutoAttack· 2023-10-05
59.7
best: 78.13 (3-ensemble of multi-resolution self-ensembles)
SOTA
Enhancing Robust Representation in Adversarial Training: Alignment and Exclusion Criteria arXiv:2310.03358
Adversarial RobustnessonCIFAR-10
Accuracy· 2023-10-05
81.7
best: 95.23 (Mixed classifier)
Enhancing Robust Representation in Adversarial Training: Alignment and Exclusion Criteria arXiv:2310.03358
Adversarial RobustnessonCIFAR-10
Attack: AutoAttack· 2023-10-05
59.7
best: 82.6 (Stochastic-LWTA/PGD/WideResNet-34-10)
Enhancing Robust Representation in Adversarial Training: Alignment and Exclusion Criteria arXiv:2310.03358