From Distributed Machine Learning to Federated Learning: A Survey

TL;DR

This survey reviews federated learning, an approach enabling collaborative machine learning across distributed, privacy-sensitive data sources, highlighting system architectures, techniques, challenges, and future research directions.

Contribution

It provides a comprehensive taxonomy, functional architecture, and analysis of federated learning systems, addressing key challenges and proposing future research avenues.

Findings

Federated learning enables privacy-preserving distributed model training.

A taxonomy of federated learning techniques is proposed.

Limitations include communication efficiency and security concerns.

Abstract

In recent years, data and computing resources are typically distributed in the devices of end users, various regions or organizations. Because of laws or regulations, the distributed data and computing resources cannot be directly shared among different regions or organizations for machine learning tasks. Federated learning emerges as an efficient approach to exploit distributed data and computing resources, so as to collaboratively train machine learning models, while obeying the laws and regulations and ensuring data security and data privacy. In this paper, we provide a comprehensive survey of existing works for federated learning. We propose a functional architecture of federated learning systems and a taxonomy of related techniques. Furthermore, we present the distributed training, data communication, and security of FL systems. Finally, we analyze their limitations and propose future research directions.