Federated Learning paradigm in E-health systems: An overview

Abstract

Every day, professionals generate and use massive healthcare data to save, treat and ameliorate the lives of patients. The healthcare industry has adopted cloud-based solutions to solve several problems in a cost-effective manner. Therefore, privacy and security mechanisms should be deployed to protect valuable medical information from unauthorized access. Much of the work in literature in recent years has focused on using artificial intelligence techniques such as deep learning and federated learning to solve various problems in the health field. Federated learning (FL) is a special technique for machine learning for privacy preservation. This study aims to compare the traditional centralized training approach and FL to show the advantages of using FL in the medical field and prove that FL can be adopted for security and data latency in e-health systems. The results obtained showed the feasibility of FL when compared to traditional methods used in the aspect of securing data and latency in the medical field.

Ähnliche Arbeiten