Analyzing the Impact of Personalization on Fairness in Federated Learning for Healthcare

Abstract

Machine learning (ML) models have the potential to generate biased outcomes, which may exacerbate existing health disparities. With privacy regulations leading to data silos in health data, federated learning (FL) has emerged as a promising solution for this issue by enabling collaborative ML without patient data sharing. Personalization within FL aims to handle performance degradation that arises due to heterogeneous data distributions across organizations. However, the relationship between personalized FL and fairness remains unclear. This paper aims to investigate and analyze the potential impact of personalized FL on fairness for healthcare. Our analysis is expected to have significant implications for fairness in federated learning and healthcare.

Ähnliche Arbeiten