Federated Learning Meets Fairness and Differential Privacy

TL;DR

This paper introduces an ethical federated learning model that simultaneously integrates fairness and differential privacy, analyzing their combined impact on model performance across multiple datasets.

Contribution

It presents the first model combining fairness, privacy, and federated learning, and explores their empirical interplay in real-world datasets.

Findings

Trade-offs between accuracy, fairness, and privacy are demonstrated.

The model achieves balanced performance across multiple ethical metrics.

Empirical results highlight challenges and benefits of integrating these measures.

Abstract

Deep learning's unprecedented success raises several ethical concerns ranging from biased predictions to data privacy. Researchers tackle these issues by introducing fairness metrics, or federated learning, or differential privacy. A first, this work presents an ethical federated learning model, incorporating all three measures simultaneously. Experiments on the Adult, Bank and Dutch datasets highlight the resulting ``empirical interplay" between accuracy, fairness, and privacy.