Machine Unlearning on Tiny-Imagenet Random Forget 20%, γ=0 (test)
95.68FAl1-sparse
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| l1-sparseBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 95.68 | 95.53 | 55.47 | 4.32 | 26.26 | |
| IUBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 95.64 | 95.56 | 55.55 | 4.36 | 26.25 | |
| FTBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 95.62 | 99.98 | 55.27 | 4.38 | 25.07 | |
| GABackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 94.61 | 95.41 | 54.27 | 5.39 | 25.45 | |
| BEBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 94.41 | 94.27 | 49.87 | 5.59 | 24.54 | |
| BSBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 92.26 | 92.45 | 48.25 | 7.74 | 23.64 | |
| SFRonBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 50.23 | 99.98 | 50.11 | 50.77 | 0.83 | |
| RetrainBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 48.87 | 99.98 | 48.51 | 51.13 | 0 | |
| FalWBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 48.6 | 99.98 | 50.67 | 50.4 | 0.79 | |
| SalUnBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 42.74 | 99.13 | 46.59 | 56.26 | 3.51 | |
| RLBackbone=Vgg-16, Random Forget Ratio=20%, γ (Imbalance Factor)=02026.01 | 32.47 | 98.25 | 45.37 | 67.53 | 9.42 |