Adversarial Robustness Accuracy on SVHN (FGSM, PGD, BIM, APGD, Jitter)
71.06Robustness Acc (FGSM, ε=8/255)Zubov-Net
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Zubov-Net2025.09 | 71.06 | 69.49 | 72.83 | 72.58 | 92.06 | 89.36 | 72.26 | 71.75 | 67.38 | 63.95 | 74.27 | |
| FxTS-Net2025.09 | 66.1 | 61.35 | 72.23 | 72.07 | 90.22 | 82.48 | 72.15 | 71.64 | 63.53 | 58.09 | 71.01 | |
| LyaNet2025.09 | 65.26 | 60.84 | 72.76 | 72.23 | 84.45 | 70.39 | 71.65 | 71.14 | 61.19 | 53.48 | 68.34 | |
| Proj-NODE2025.09 | 65.04 | 60.21 | 72.17 | 71.87 | 82.88 | 68.09 | 71.47 | 71.17 | 59.57 | 51.59 | 67.4 | |
| SODEF2025.09 | 64.09 | 59.25 | 71.8 | 71.47 | 83.33 | 69.74 | 70.58 | 70.17 | 60.86 | 53.11 | 67.44 | |
| Neural ODE2025.09 | 63.86 | 57.93 | 70.48 | 70.05 | 82.84 | 67.55 | 69.38 | 69.08 | 59.17 | 51.02 | 66.14 | |
| ResNet182025.09 | 62.96 | 57.23 | 69.87 | 69.56 | 82.76 | 67.26 | 69.01 | 68.56 | 58.41 | 50.48 | 65.61 |