Black-box DNN Watermarking on CIFAR-10
93.16CDAClean Model
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Clean ModelBackbone=VGG162026.02 | 93.16 | — | — | |
| EWEBackbone=VGG162026.02 | 93.1 | 100 | 0.3862 | |
| SEW-PreBackbone=VGG16, Enhancement=Pre-specificity2026.02 | 93.03 | 100 | 0.3569 | |
| EWBackbone=VGG162026.02 | 92.92 | 100 | 0.1138 | |
| ContentBackbone=VGG162026.02 | 92.86 | 100 | 0.3505 | |
| AFSBackbone=VGG162026.02 | 92.8 | 100 | 0.2244 | |
| SEW-PostBackbone=VGG16, Enhancement=Post-specificity2026.02 | 92.79 | 100 | 0.0364 | |
| WESBackbone=VGG162026.02 | 92.78 | 100 | 0.0467 | |
| ADIBackbone=VGG162026.02 | 92.51 | 100 | 0.0663 | |
| NoiseBackbone=VGG162026.02 | 92.42 | 100 | 0.3677 | |
| UnrelatedBackbone=VGG162026.02 | 92.42 | 100 | 0.2702 | |
| ISSBABackbone=VGG162026.02 | 92.34 | 100 | 0.0371 | |
| MWBackbone=VGG162026.02 | 87.74 | 100 | 0.1064 |