Neural Network Verification on CIFAR 1000 images
51.6Verified Accuracyβ-CROWN FSB
Evaluation results
| Method | Links | ||
|---|---|---|---|
| β-CROWN FSBModel=ConvBig2021.03 | 51.6 | 15.3 | |
| PRIMAModel=ConvBig2021.03 | 48.3 | 176 | |
| β-CROWN FSBModel=ConvSmall2021.03 | 46.3 | 6.8 | |
| kPolyModel=ConvBig2021.03 | 45.9 | 346 | |
| PRIMAModel=ConvSmall2021.03 | 44.6 | 13 | |
| CROWN/DeepPolyModel=ConvBig2021.03 | 42.1 | 43 | |
| kPolyModel=ConvSmall2021.03 | 39.9 | 86 | |
| OptC2VModel=ConvSmall2021.03 | 39.8 | 105 | |
| CROWN/DeepPolyModel=ConvSmall2021.03 | 35.9 | 4 | |
| PRIMAModel=ResNet2021.03 | 24.8 | 1.7 | |
| β-CROWN FSBModel=ResNet2021.03 | 24.8 | 1.6 | |
| kPolyModel=ResNet2021.03 | 24.5 | 91 | |
| CROWN/DeepPolyModel=ResNet2021.03 | 24.1 | 1 |