Backdoor Detection on CIFAR-100 (test)
100Top-1 ASRLoRAcle
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LoRAcleModel=ResNet18, Attack Type=Blend, Poison Rate=0.22026.01 | 100 | 100 | |
| LoRAcleModel=VGG19, Attack Type=BadNets, Poison Rate=0.22026.01 | 100 | 100 | |
| LoRAcleModel=VGG19, Attack Type=Blend, Poison Rate=0.22026.01 | 100 | 100 | |
| LoRAcleModel=DenseNet, Attack Type=Blend, Poison Rate=0.22026.01 | 80 | 80 | |
| LoRAcleModel=ResNet18, Attack Type=WaNet, Poison Rate=0.22026.01 | 80 | 80 | |
| LoRAcleModel=DenseNet, Attack Type=BadNets, Poison Rate=0.22026.01 | 75 | 75 | |
| LoRAcleModel=DenseNet, Attack Type=WaNet, Poison Rate=0.22026.01 | 66.7 | 66.7 | |
| LoRAcleModel=ResNet18, Attack Type=BadNets, Poison Rate=0.22026.01 | 60 | 100 | |
| LoRAcleModel=VGG19, Attack Type=WaNet, Poison Rate=0.22026.01 | 40 | 100 |