Mixed classifier

Reported on 3 benchmarks across 1 task · 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 RobustnessonCIFAR-10
Accuracy· uses extra data· 2023-01-29
95.23
SOTA
Improving the Accuracy-Robustness Trade-Off of Classifiers via Adaptive Smoothing arXiv:2301.12554
Adversarial RobustnessonCIFAR-10
Attack: AutoAttack· uses extra data· 2023-01-29
68.06
best: 82.6 (Stochastic-LWTA/PGD/WideResNet-34-10)
Improving the Accuracy-Robustness Trade-Off of Classifiers via Adaptive Smoothing arXiv:2301.12554
Adversarial RobustnessonCIFAR-10
Robust Accuracy· uses extra data· 2023-01-29
68.06
best: 84.3 (Stochastic-LWTA/PGD/WideResNet-34-10)
Improving the Accuracy-Robustness Trade-Off of Classifiers via Adaptive Smoothing arXiv:2301.12554