Image Classification on CIFAR100 (test) (Adversarial Robustness Evaluation)
10.099L^e Start Bound (x10^-7)Normal
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| NormalModel Architecture=ConvNet2024.03 | 10.099 | 6.14 | 48.39 | 0.05 | |
| NormalModel Architecture=ResNet182024.03 | 12.044 | 4.716 | 41.24 | 0 | |
| SEPposModel Architecture=ConvNet2024.03 | 13.65 | 9.932 | 37.41 | 17.98 | |
| SEPposModel Architecture=ResNet182024.03 | 19.217 | 12.744 | 34.83 | 15.16 | |
| MATModel Architecture=ConvNet2024.03 | 20.642 | 13.65 | 36.4 | 17.35 | |
| SEPnegModel Architecture=ConvNet2024.03 | 22.506 | 14.97 | 36.17 | 17.43 | |
| MATModel Architecture=ResNet182024.03 | 33.456 | 22.623 | 36.14 | 15.7 | |
| SEPnegModel Architecture=ResNet182024.03 | 35.525 | 24.289 | 34.8 | 15.87 |