Machine Unlearning on CIFAR-10 Forget 10%
94.29Test AccuracyBASE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BASEBackbone=ResNet-182026.03 | 94.29 | 59.5 | 3.88 | |
| SSDBackbone=ResNet-182026.03 | 94.29 | 59.5 | 3.88 | |
| NEGGRADBackbone=ResNet-182026.03 | 93.89 | 59.47 | 3.82 | |
| RETRAINBackbone=ResNet-182026.03 | 93.81 | 50.19 | 0 | |
| NEGGRAD+Backbone=ResNet-182026.03 | 93.02 | 59.1 | 3.71 | |
| AMUNBackbone=ResNet-182026.03 | 91.97 | 52.63 | 1.46 | |
| SALUNBackbone=ResNet-182026.03 | 91.59 | 53.45 | 2.48 | |
| ℓ1-SPARSEBackbone=ResNet-182026.03 | 90.63 | 53.01 | 2.49 | |
| REGUNBackbone=ResNet-182026.03 | 90.6 | 51.01 | 2 | |
| FINETUNEBackbone=ResNet-182026.03 | 90.23 | 53.92 | 2.79 |