A Survey on Group Fairness in Federated Learning: Challenges, Taxonomy of Solutions and Directions for Future Research

TL;DR

This survey reviews the challenges, solutions, and future directions for achieving group fairness in federated learning, emphasizing the importance of addressing data heterogeneity and ethical considerations.

Contribution

It provides a comprehensive taxonomy, benchmarks, and analysis of existing approaches to promote fairness in federated learning systems.

Findings

Identifies key challenges in applying group fairness to federated learning.

Proposes a taxonomy based on data partitioning, location, and strategy.

Highlights the need for new methods to handle fairness complexities.

Abstract

Group fairness in machine learning is an important area of research focused on achieving equitable outcomes across different groups defined by sensitive attributes such as race or gender. Federated Learning, a decentralized approach to training machine learning models across multiple clients, amplifies the need for fairness methodologies due to its inherent heterogeneous data distributions that can exacerbate biases. The intersection of Federated Learning and group fairness has attracted significant interest, with 48 research works specifically dedicated to addressing this issue. However, no comprehensive survey has specifically focused on group fairness in Federated Learning. In this work, we analyze the key challenges of this topic, propose practices for its identification and benchmarking, and create a novel taxonomy based on criteria such as data partitioning, location, and strategy. Furthermore, we analyze broader concerns, review how different approaches handle the complexities of various sensitive attributes, examine common datasets and applications, and discuss the ethical, legal, and policy implications of group fairness in FL. We conclude by highlighting key areas for future research, emphasizing the need for more methods to address the complexities of achieving group fairness in federated systems.