Machine Unlearning on SVHN (remaining set Dr)
99.84AccuracyFine-tune
Evaluation results
| Method | Links | |
|---|---|---|
| Fine-tuneUnlearning Approach=Fine-tune, Proportion of remaining data Dr used (@)=0.32026.06 | 99.84 | |
| Certified HessianUnlearning Approach=Cert.-Hess., Proportion of remaining data Dr used (@)=0.352026.06 | 99.78 | |
| Relabel with Fine-tuneUnlearning Approach=Relab.+FT, Proportion of remaining data Dr used (@)=0.352026.06 | 99.75 | |
| Adversarial RetrainUnlearning Approach=Adv. Retr., Proportion of remaining data Dr used (@)=0.32026.06 | 99.63 | |
| Pre-unlearnUnlearning Approach=Pre-unlearn2026.06 | 99.29 | |
| HessianUnlearning Approach=Hessian, Proportion of remaining data Dr used (@)=0.32026.06 | 99.08 | |
| RetrainUnlearning Approach=Retrain, Proportion of remaining data Dr used (@)=0.32026.06 | 98.79 | |
| Gradient Ascent with Fine-tuneUnlearning Approach=GA+FT, Proportion of remaining data Dr used (@)=0.32026.06 | 98.06 | |
| FisherUnlearning Approach=Fisher, Proportion of remaining data Dr used (@)=0.352026.06 | 96.44 | |
| Gradient AscentUnlearning Approach=GA, Proportion of remaining data Dr used (@)=0.052026.06 | 43.92 | |
| RelabelUnlearning Approach=Relabel, Proportion of remaining data Dr used (@)=0.052026.06 | 42.75 |