Federated Learning: Advancing Healthcare through Collaborative Artificial Intelligence

Abstract

Abstract More and more healthcare data are becoming easily accessible from clinical institutions, patients, insurance companies and the pharmaceutical industry, amongst others, due to the quick development of computer software and hardware technologies. With this access, data science technologies have a never-before-seen chance to generate data-driven insights and raise the standard of healthcare delivery. However, healthcare data are frequently fragmented and private, making it challenging to produce reliable results across populations. The electronic health records of various patient populations, for instance, are owned by multiple hospitals, and because of their sensitive nature, it is challenging for hospitals to share these records. This poses a substantial obstacle to creating generalisable, effective analytical methods that require various ‘big data’. Federated learning offers an excellent opportunity to integrate disparate healthcare data sources while protecting privacy. Federated learning uses a central server to train a standard global model while retaining all the sensitive data in local institutions where it belongs.

Ähnliche Arbeiten