Federated Learning as a Mean-Field Game

TL;DR

This paper models federated learning as a mean-field game, providing a new theoretical framework that links machine learning with game theory to analyze large-scale, privacy-preserving distributed algorithms.

Contribution

It introduces a novel perspective by formulating federated learning as a differential game, connecting it with mean-field game theory to analyze equilibrium properties.

Findings

Federated learning is modeled as a differential game.

The properties of the game equilibrium are discussed.

The framework aims to unify research in distributed and privacy-preserving learning.

Abstract

We establish a connection between federated learning, a concept from machine learning, and mean-field games, a concept from game theory and control theory. In this analogy, the local federated learners are considered as the players and the aggregation of the gradients in a central server is the mean-field effect. We present federated learning as a differential game and discuss the properties of the equilibrium of this game. We hope this novel view to federated learning brings together researchers from these two distinct areas to work on fundamental problems of large-scale distributed and privacy-preserving learning algorithms.