Abstract
Healthcare AI has federated learning as a transformational approach in which clinical research can be conducted collaboratively and in patient privacy is preserved. This chapter considers federated learning and how it enables healthcare institutions to learn from sensitive data without centralizing it and without access to that data. It consists of the technological architecture, companding local computational resources, secure communication channels, and model aggregation servers, each with a suitable algorithm for dealing with data heterogeneity between institutions. Its integration with existing healthcare infrastructure (IoMT devices, EHR systems) ensures operational efficiency thus being integrated into the existing system. By balancing innovation with privacy, federated learning is positioned as a cornerstone technology for the development of the personalized medicine — which with technology giants as the players the only type in which patient privacy can be respected while taking advantage of the collective clinical wisdom.

This publication has 33 references indexed in Scilit: