On-device Federated Learning with Flower

TL;DR

This paper explores on-device federated learning using the Flower framework, evaluating its implementation on smartphones and embedded devices, and discusses system costs to inform more efficient algorithms.

Contribution

It provides an empirical study of on-device FL deployment on various devices and analyzes system costs to guide future algorithm design.

Findings

Successful implementation of on-device FL on smartphones and embedded devices

Quantification of system costs associated with on-device FL

Insights into designing more efficient FL algorithms based on cost analysis

Abstract

Federated Learning (FL) allows edge devices to collaboratively learn a shared prediction model while keeping their training data on the device, thereby decoupling the ability to do machine learning from the need to store data in the cloud. Despite the algorithmic advancements in FL, the support for on-device training of FL algorithms on edge devices remains poor. In this paper, we present an exploration of on-device FL on various smartphones and embedded devices using the Flower framework. We also evaluate the system costs of on-device FL and discuss how this quantification could be used to design more efficient FL algorithms.