Federated Unlearning on MNIST
93.49Pretraining Accuracyf-FUM (KL-KL)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| f-FUM (KL-KL)N (Number of clients)=10, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 93.49 | 97.75 | |
| f-FUM (KL-JS)N (Number of clients)=10, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 93.49 | 98.81 | |
| f-FUM (KL-χ²)N (Number of clients)=10, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 93.49 | 97.33 | |
| NoTN (Number of clients)=10, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 93.49 | 86.58 | |
| HalimiN (Number of clients)=10, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 93.49 | 97.8 | |
| f-FUM (KL-KL)N (Number of clients)=5, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 84.06 | 95.69 | |
| f-FUM (KL-JS)N (Number of clients)=5, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 84.06 | 98.41 | |
| f-FUM (KL-χ²)N (Number of clients)=5, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 84.06 | 94.05 | |
| NoTN (Number of clients)=5, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 84.06 | 81.64 | |
| HalimiN (Number of clients)=5, Removed (%)=2%, Backbone=LeNet5, Unlearning Scenario=Confuse2026.02 | 84.06 | 73.52 |