A federated learning approach for smart healthcare systems

Abstract

With periodic technology advancements and pandemic-like situations, remote patient health monitoring has increased significantly. The Internet of Things (IoT) devices, including wearables, sensors, and actuators deployed on the human body, detect and regulate physiological data. These systems can establish a trigger mechanism in the event of a possible health incident. Health monitoring using IoT devices generates a large amount of data. Several Machine Learning (ML) strategies have been utilized to analyze the collected data and derive precise predictions. The confidentiality of patient data is one of the essential requirements of these systems. It has been discovered that malicious coordination of ML algorithms might result in a massive attack surface, providing cyber- criminals with an accessible platform. Considering these requirements for using IoT systems for health monitoring, Federated Learning (FL) can solve data privacy challenges by training ML models locally on IoT de- vices without any data transfer to the cloud. FL facilitates information sharing among all IoT devices installed in hospitals through collabora- tive ML model training. This article will examine the significance of em- bedding FL into IoT-enabled smart hospitals and future guidelines for accomplishing this. This article discusses identifying rare diseases and critical care, resolving the problem of insufficient patient health data to train ML models, and preserving patient privacy.

A Comprehensive Survey on Federated Learning in the Healthcare Area: Concept and Applications

“…The federated learning approach trains AI models without sharing raw data, in which different IoT devices contribute their data to train a centralized model. Each IoT device downloads the model, trains it on its data, and encrypts the updates, which are then sent back, decrypted, averaged, and integrated into the central model [71], as shown in Figs. 1 and 11.…”

Image-based crop disease detection with federated learning

“…Recent studies on federated learning, including concepts, challenges, privacy and security, and future research directions have been conducted in [13], [11], [25]. Growing steadily in recent years, federated learning has been applied to solve different types of problems in several domains, including medical [26], [27], [28], distributed networks and systems [24], [29], [30], Internet of Things (IoT) [31], and very recently in the agricultural domain [32], [33], [34].…”

“…They propose the connection of different sources of fragmented health data while preserving confidentiality, and thus improve the quality of care provided. In the article [28], the authors explore the application of federated learning (FL) in intelligent health systems, particularly in the context of remote patient health monitoring. After highlighting the importance of integrating federated learning into IoT-powered smart hospitals, the authors suggest using federated learning to address data privacy challenges by proposing to train machine learning models locally on hospital-installed IoT devices without transferring data to the cloud.…”

Image-based crop disease detection with federated learning

“…Recent studies on federated learning, including concepts, challenges, privacy and security, and future research directions have been conducted in 11 , 13 , 26 . Growing steadily in recent years, federated learning has been applied to solve different types of problems in several domains, including medical 27 – 29 , distributed networks and systems 25 , 30 , 31 , Internet of Things (IoT) 32 , and very recently in the agricultural domain 33 – 35 .…”

“…The authors envision a system in which various sources of health data are interconnected, guaranteeing data confidentiality and subsequently improving the quality of care. The article 29 highlights the integration of federated learning in intelligent healthcare systems, particularly for remote patient monitoring. They highlight the value of merging federated learning and IoT in smart hospitals, suggesting the training of local models on IoT devices to avoid data transfers to the cloud.…”