Adversarial Defense On Cifar 10
평가 지표
Accuracy
Robust Accuracy
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | Accuracy | Robust Accuracy | Paper Title | Repository |
|---|---|---|---|---|
| Diffusion Classifier | 89.85 | 75.67 | Robust Classification via a Single Diffusion Model | |
| Ours (Stochastic-LWTA/PGD/WideResNet-34-1) | 81.87 | - | Stochastic Local Winner-Takes-All Networks Enable Profound Adversarial Robustness | |
| Stochastic-LWTA/PGD/WideResNet-34-10 | 84.3 | - | Stochastic Local Winner-Takes-All Networks Enable Profound Adversarial Robustness | |
| Ours (Stochastic-LWTA/PGD/WideResNet-34-5) | 83.4 | - | Stochastic Local Winner-Takes-All Networks Enable Profound Adversarial Robustness | |
| ResNet18 (TRADES-ANCRA/PGD-40) | 81.70 | 82.96 | Enhancing Robust Representation in Adversarial Training: Alignment and Exclusion Criteria | |
| PCL (against PGD, white box) | 46.7 | - | Adversarial Defense by Restricting the Hidden Space of Deep Neural Networks | |
| WRN-28-10 | 90.03 | 71.68 | Language Guided Adversarial Purification | - |
| Stochastic-LWTA/PGD/WideResNet-34-5 | - | - | Stochastic Local Winner-Takes-All Networks Enable Profound Adversarial Robustness |
0 of 8 row(s) selected.