Image Classification on CIFAR-10 (retention set (Dr))
99.6AccuracyBEFORE
Evaluation results
| Method | Links | |
|---|---|---|
| BEFOREk=16, Backbone=ResNet-182026.03 | 99.6 | |
| ORACLEk=16, Backbone=ResNet-182026.03 | 98.74 | |
| STRUCTGUARDk=16, Backbone=ResNet-182026.03 | 91.43 | |
| L2ULk=16, Backbone=ResNet-182026.03 | 85.75 | |
| ADVk=16, Backbone=ResNet-182026.03 | 69.7 | |
| RAWPk=16, Backbone=ResNet-182026.03 | 67.15 | |
| FISHERk=16, Backbone=ResNet-182026.03 | 17.82 | |
| NEGGRADk=16, Backbone=ResNet-182026.03 | 15.79 |