Federated-Unlearning-under-Plausible-Deniability

July 13, 2025 ยท View on GitHub

This is a repository for the paper 'Efficient Federated Unlearning under Plausible Deniability' to be presented in ACML 2024 - journal category (Springer Machine Learning).

Flowchart of the perturbed k-IPfedAvg:

Screenshot 2024-08-14 at 15 15 52

Experimental details:

Screenshot 2024-08-14 at 14 40 11

Comparison of the test accuracy fedEraser, fedAvg_retrained, and our approach (perturbed k-IPfedAvg):

Screenshot 2024-08-14 at 14 40 55

Disk Space comparison between fedEraser and our model:

Screenshot 2024-08-14 at 14 41 27

Running wall-clock time comparison:

Screenshot 2024-08-14 at 14 41 51