Clustered Federated Learning with Non-IID Data: Mitigating Accuracy Overestimates Through Hold-Out Model Selection and Evaluation.

Davoud Gholamiangonabadi, Katarina Grolinger

Browse the full ICMLA paper archive.