Abstract
This paper considers the problem of decentralized, personalized federated learning. For centralized personalized federated learning, a penalty that measures the deviation from the local model and its average, is often added to the objective function. However, in a decentralized setting this penalty is expensive in terms of communication costs, so here, a different penalty — one that is built to respect the structure of the underlying computational network — is used instead. We present lower bounds on the communication and local computation costs for this problem formulation and we also present provably optimal methods for decentralized personalized federated learning. Numerical experiments are presented to demonstrate the practical performance of our methods.
•We consider the problem of decentralized, personalized federated learning.•In this setting, we penalise the difference between local models that are neighbours in a decentralized network.•We present lower bounds on the communication and local computation costs for this problem formulation.•We present provably optimal methods which meet lower bounds.•Numerical experiments demonstrate the practical performance of our methods.