Defense against Data Reconstruction Attack on CIFAR10 (test)
0.006MSEClean
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CleanProtocol=Vertical Federated Learning (VFL)2025.12 | 0.006 | 23.34 | 0.83 | 72.5 | |
| DNPProtocol=Vertical Federated Learning (VFL), δ / λ=0.0552025.12 | 0.013 | 18.91 | 0.622 | 57.7 | |
| InvL-DNPProtocol=Vertical Federated Learning (VFL), δ / λ=0.062025.12 | 0.014 | 18.61 | 0.605 | 58.3 | |
| PruneProtocol=Vertical Federated Learning (VFL), δ / λ=0.92025.12 | 0.021 | 17.76 | 0.664 | 51.8 | |
| DropoutProtocol=Vertical Federated Learning (VFL), δ / λ=0.92025.12 | 0.021 | 17.38 | 0.617 | 30.5 | |
| GNPProtocol=Vertical Federated Learning (VFL), δ / λ=0.72025.12 | 0.021 | 17.16 | 0.57 | 65 | |
| InvL-GNPProtocol=Vertical Federated Learning (VFL), δ / λ=12025.12 | 0.023 | 16.43 | 0.502 | 68.4 |