Machine Unlearning on CIFAR-10 Forget 50%
94.18Test AccuracySSD
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SSDBackbone=ResNet-182026.03 | 94.18 | 56.42 | 4.79 | |
| BASEBackbone=ResNet-182026.03 | 94.17 | 56.42 | 4.79 | |
| NEGGRADBackbone=ResNet-182026.03 | 94.05 | 56.59 | 4.8 | |
| RETRAINBackbone=ResNet-182026.03 | 90.31 | 50.24 | 0 | |
| REGUNBackbone=ResNet-182026.03 | 90.11 | 52.1 | 1.48 | |
| AMUNBackbone=ResNet-182026.03 | 89.49 | 51.02 | 1.84 | |
| SALUNBackbone=ResNet-182026.03 | 89 | 52.76 | 2 | |
| ℓ1-SPARSEBackbone=ResNet-182026.03 | 88.82 | 52.81 | 2.09 | |
| NEGGRAD+Backbone=ResNet-182026.03 | 88.62 | 53.19 | 2.62 | |
| FINETUNEBackbone=ResNet-182026.03 | 88.1 | 52.4 | 2.39 |