Machine Unlearning on Tiny-Imagenet Random Forget 30%, γ=1/3 (test)
95.85FAIU
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| IUBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 95.85 | 95.46 | 55.67 | 4.15 | 29.29 | |
| l1-sparseBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 95.83 | 95.46 | 55.17 | 4.17 | 29.15 | |
| BEBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 94.89 | 94.94 | 51.43 | 5.11 | 27.88 | |
| BSBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 94.33 | 94.59 | 50.81 | 5.67 | 27.53 | |
| FTBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 90.44 | 99.74 | 51.19 | 9.56 | 24.39 | |
| GABackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 77.98 | 93.73 | 48.67 | 22.02 | 19.03 | |
| SFRonBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 51.2 | 99.79 | 45.49 | 48.8 | 3.39 | |
| FalWBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 46.94 | 99.98 | 46.11 | 53.06 | 1.31 | |
| RetrainBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 44.57 | 99.98 | 45.61 | 55.43 | 0 | |
| SalUnBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 39.88 | 99.95 | 43.47 | 60.12 | 2.89 | |
| RLBackbone=Vgg-16, Random Forget Ratio=30%, γ (Imbalance Factor)=1/32026.01 | 31.76 | 97.13 | 41.25 | 68.24 | 8.21 |