Federated Learning for Digital Healthcare Systems

Modern healthcare systems facilitate the collection of critical medical data for statistical evaluation and inference using machine learning, however, the application of ML in healthcare data analytics has not been fully exploited due to the proliferation of security and privacy concerns. The potential of machine learning is also limited by insufficient data, posing a significant impediment to the transition from research to clinical practice. Over the past five years, Federated Learning has been introduced to strengthen the performance of machine learning. In federated learning, artificial intelligence models are trained with data from multiple sources. In this case, data anonymity, security, privacy and integrity are maintained, thus removing potential barriers to data sharing. Additionally, models trained by federated learning have shown favorable progress in the agreement with models obtained from centrally hosted data sets. A successfully implemented federated learning model can produce unbiased decisions which facilitate better-informed decision making in precision medicine. Federated Learning for Digital Healthcare Systems critically examines the key factors that contribute to the problem of applying machine learning in healthcare systems and investigates how federated learning can be employed to address the problem. The book discusses, examines, and compares the applications of federated learning solutions in emerging digital healthcare systems, providing a critical look in terms of the required resources, computational complexity, and system performance. In the first section, chapters examine how to address critical security and privacy concerns and how to revamp existing machine learning models. In subsequent chapters, authors review recent advances to tackle emerging efficient and lightweight algorithms and protocols to reduce computational overheads and communication costs in wireless healthcare systems. Consideration is also given to government and economic regulations as well as legal considerations when federated learning is applied to digital healthcare systems.