Federated Learning in Healthcare: Enhancing Patient Privacy and Data Security

EasyChair Preprint 14995

Abstract

As the healthcare industry increasingly adopts digital technologies, the importance of data privacy and security has never been more critical. Federated learning (FL) presents a novel approach to training machine learning models across decentralized healthcare data sources while ensuring patient privacy. This paper explores the application of federated learning in healthcare, highlighting its potential to revolutionize data sharing practices without compromising data security. We review the key federated learning algorithms and evaluate their effectiveness in handling the unique challenges of healthcare data, including data heterogeneity, privacy concerns, and regulatory compliance. The study includes a case study on predicting patient outcomes using federated learning across multiple healthcare institutions, demonstrating the balance between privacy preservation and model performance. The findings suggest that federated learning could be a game-changer in healthcare, enabling collaborative research and better patient care without the risks associated with centralized data aggregation.

Keyphrases: Data Security, Federated Learning, Healthcare data, Patient Outcomes, decentralized data, privacy-preserving machine learning

@booklet{EasyChair:14995,
  author    = {Isabella Rossi},
  title     = {Federated Learning in Healthcare: Enhancing Patient Privacy and Data Security},
  howpublished = {EasyChair Preprint 14995},
  year      = {EasyChair, 2024}}