Jul 2026
Encryption-Compatible Clustered Federated Learning via Distributed Expectation-Maximization over Metadata
FLAMECHE is proposed, which reformulates metadata-based CFL as a distributed Expectation-Maximization (EM) procedure, restricting server updates to additive operations while preserving efficiency and enabling compatibility with practical secure FL schemes.
Michael Ben Ali, I. Megdiche, A. Péninou et al.
· arXiv.org · 0 citations