Image Classification on CIFAR-10 AdvL
98.67Accuracydefensive patch generation framework
Evaluation results
| Method | Links | |
|---|---|---|
| defensive patch generation frameworkModel=MNet, Setting=Four models ensemble2022.04 | 98.67 | |
| defensive patch generation frameworkModel=VGG, Setting=Four models ensemble2022.04 | 97.27 | |
| defensive patch generation frameworkModel=SNet, Setting=Four models ensemble2022.04 | 96.07 | |
| defensive patch generation frameworkModel=RNet, Setting=Four models ensemble2022.04 | 96.04 | |
| UnAdvModel=RNet, Setting=Four models ensemble2022.04 | 88.87 | |
| VanillaModel=RNet, Setting=Four models ensemble2022.04 | 85.65 | |
| VanillaModel=MNet, Setting=Four models ensemble2022.04 | 84.13 | |
| UnAdvModel=VGG, Setting=Four models ensemble2022.04 | 83 | |
| TransModel=RNet, Setting=Four models ensemble2022.04 | 81.61 | |
| VanillaModel=VGG, Setting=Four models ensemble2022.04 | 80.4 | |
| VanillaModel=SNet, Setting=Four models ensemble2022.04 | 78.9 | |
| TransModel=MNet, Setting=Four models ensemble2022.04 | 78.65 | |
| TransModel=SNet, Setting=Four models ensemble2022.04 | 73.53 | |
| TransModel=VGG, Setting=Four models ensemble2022.04 | 72.71 | |
| UnAdvModel=SNet, Setting=Four models ensemble2022.04 | 71.79 | |
| UnAdvModel=MNet, Setting=Four models ensemble2022.04 | 62.36 |