Image Classification on CIFAR-10 (test) (FGSM/PGD Robustness Profile)
95.29Standard Accuracy (SA)ResNet-50
Evaluation results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ResNet-50Backbone=ResNet-502025.12 | 95.29 | 52.68 | 51.54 | 46.01 | 35.04 | 25.46 | 19.26 | 16.25 | 13.87 | 12.14 | 46.2 | 45.68 | 45.53 | 45.44 | 45.45 | |
| DenseNet-169Backbone=DenseNet-1692025.12 | 95.1 | 55.32 | 54.57 | 48.63 | 36.38 | 25.21 | 18.57 | 15 | 12.88 | 11.9 | 47.55 | 46.99 | 46.81 | 46.72 | 46.81 | |
| GoogLeNetBackbone=GoogLeNet2025.12 | 95.01 | 53.83 | 51.53 | 46.96 | 33.59 | 22.7 | 16.61 | 13.83 | 11.69 | 10.91 | 39.6 | 38.59 | 38.53 | 38.48 | 38.49 | |
| Inception-V3Backbone=Inception-V32025.12 | 94.78 | 58 | 56.77 | 52.05 | 41.99 | 29.99 | 21.89 | 16.53 | 13.61 | 11.97 | 47.21 | 45.51 | 45.01 | 44.68 | 44.8 | |
| MobileNet-V2Backbone=MobileNet-V22025.12 | 94.2 | 55.05 | 53.79 | 49.31 | 36.67 | 24.7 | 18.31 | 15.38 | 13.39 | 12.19 | 45.77 | 45.2 | 45.14 | 45.17 | 45.08 | |
| XceptionBackbone=Xception2025.12 | 93.48 | 51.41 | 50.2 | 47.69 | 37.72 | 26.35 | 18.45 | 14.21 | 12.32 | 10.96 | 41.36 | 40.58 | 40.44 | 40.3 | 40.3 | |
| VGG-19Backbone=VGG-192025.12 | 93.18 | 52.08 | 50.92 | 48.52 | 41.53 | 29.22 | 20.22 | 16.06 | 13.7 | 12 | 42.66 | 41.91 | 41.72 | 41.8 | 41.8 | |
| ResNet-18Backbone=ResNet-182025.12 | 92.35 | 52.32 | 48.19 | 40.97 | 30.28 | 21.38 | 15.61 | 12.72 | 11.61 | 11 | 37.72 | 37.28 | 37.1 | 37.02 | 36.95 | |
| DWFBackbone=ResNet-182025.12 | 91.08 | 82.83 | 79.66 | 74.76 | 69.01 | 63.6 | 58.4 | 52.97 | 47.34 | 41.24 | 65.51 | 64.7 | 64.52 | 64.42 | 64.37 | |
| ADVMoEBackbone=ResNet-182025.12 | 88.27 | 79.39 | 62.09 | 40.88 | 23.9 | 16.53 | 15.47 | 15.98 | 16.07 | 16.4 | 47.69 | 44.95 | 44.36 | 44.11 | 44.01 | |
| SoEBackbone=ResNet-182025.12 | 81.69 | 70.23 | 66.22 | 60.27 | 52.65 | 44.62 | 38.26 | 32.96 | 28.74 | 25.74 | 62.27 | 62.34 | 62.31 | 62.31 | 62.31 |