Machine Unlearning on CIFAR-100 Random Forget 30% (forget set)
97.02FA (%)IU
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| IUBackbone=ResNet-18, Gamma=1/42026.01 | 97.02 | 96.89 | 70.41 | 2.98 | 23.5 | |
| FTBackbone=ResNet-18, Gamma=1/42026.01 | 96.32 | 99.95 | 69.8 | 3.68 | 22.23 | |
| l1-sparseBackbone=ResNet-18, Gamma=1/42026.01 | 95.64 | 95.38 | 66.65 | 4.36 | 22.25 | |
| GABackbone=ResNet-18, Gamma=1/42026.01 | 85.51 | 89.24 | 61.4 | 14.49 | 17.41 | |
| BEBackbone=ResNet-18, Gamma=1/42026.01 | 75.6 | 79.53 | 48.09 | 24.4 | 17.13 | |
| FalWBackbone=ResNet-18, Gamma=1/42026.01 | 57.73 | 99.98 | 60.67 | 41.27 | 0.91 | |
| RetrainBackbone=ResNet-18, Gamma=1/42026.01 | 57.13 | 99.96 | 59.24 | 42.87 | 0 | |
| BSBackbone=ResNet-18, Gamma=1/42026.01 | 56.79 | 58.94 | 38.46 | 43.21 | 15.62 | |
| SalUnBackbone=ResNet-18, Gamma=1/42026.01 | 55.19 | 99.96 | 59.98 | 39.41 | 1.54 | |
| SFRonBackbone=ResNet-18, Gamma=1/42026.01 | 54.46 | 99.97 | 60.72 | 45.54 | 1.71 | |
| RLBackbone=ResNet-18, Gamma=1/42026.01 | 44.24 | 96.34 | 58.08 | 65.76 | 10.14 |