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:
Experimental details:
Comparison of the test accuracy fedEraser, fedAvg_retrained, and our approach (perturbed k-IPfedAvg):
Disk Space comparison between fedEraser and our model:
Running wall-clock time comparison:
