Targeted Adversarial Attack on ImageNet-1K (val)
9.18ResNet-18 ResultGaussian Noise
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Gaussian NoiseSurrogate=None, Training set=None, l_inf-norm bound=10/2552022.03 | 9.18 | — | — | — | — | — | 10.2 | |
| GD-UAPSurrogate=ResNet-20, Training set=CIFAR-100, l_inf-norm bound=10/2552022.03 | 12.72 | — | — | — | — | — | 10.22 | |
| GD-UAPSurrogate=VGG-11, Training set=CIFAR-100, l_inf-norm bound=10/2552022.03 | 12.93 | — | — | — | — | — | 10.39 | |
| GD-UAPSurrogate=ResNet-20, Training set=CIFAR-10, l_inf-norm bound=10/2552022.03 | 14.09 | — | — | — | — | — | 11.44 | |
| GD-UAPSurrogate=VGG-11, Training set=CIFAR-10, l_inf-norm bound=10/2552022.03 | 14.63 | — | — | — | — | — | 12.38 | |
| SSAHSurrogate=VGG-11, Training set=CIFAR-10, l_inf-norm bound=10/2552022.03 | 16.92 | — | — | — | — | — | 17.52 | |
| SSAHSurrogate=ResNet-20, Training set=CIFAR-10, l_inf-norm bound=10/2552022.03 | 17.66 | — | — | — | — | — | 16.87 | |
| SSAHSurrogate=VGG-11, Training set=CIFAR-100, l_inf-norm bound=10/2552022.03 | 17.92 | — | — | — | — | — | 19.14 | |
| SSAHSurrogate=ResNet-20, Training set=CIFAR-100, l_inf-norm bound=10/2552022.03 | 18.31 | — | — | — | — | — | 18.68 | |
| AdvDropIteration=150, RunTime (s)=469682022.03 | — | 99.76 | 14.95 | 0.06 | 11.28 | 5.67 | — | |
| BIMIteration=10, RunTime (s)=21662022.03 | — | 98.31 | 25.18 | 0.03 | 39.61 | 10.44 | — | |
| C&W l2Iteration=1000, RunTime (s)=> 1000002022.03 | — | 97.83 | 1.85 | 0.04 | 12.93 | 0.84 | — | |
| MIMIteration=10, RunTime (s)=23582022.03 | — | 99.98 | 92.86 | 0.03 | 81.62 | 39.93 | — | |
| PGDIteration=10, RunTime (s)=29732022.03 | — | 98.65 | 53.84 | 0.03 | 37.21 | 16.87 | — | |
| SSAIteration=200, RunTime (s)=317422022.03 | — | 98.64 | 4.31 | 0.02 | 8.64 | 1.94 | — | |
| SSAHIteration=200, RunTime (s)=300502022.03 | — | 98.06 | 3.38 | 0.02 | 6.42 | 0.47 | — |