A survey on federated learning for security and privacy in healthcare applications

Technological advances in smart devices and applications targeting the Internet of Healthcare Things provide a perfect environment for using Machine Learning-based techniques. However, traditional ML solutions operate on centralized data collection and processing. Federated Learning (FL) is a promising solution to train ML models on multiple decentralized devices without effectively sharing private data. Therefore, FL offers a secure architecture to handle highly sensitive data in the IoHT context. This survey comprehensively reviews emerging data security and privacy applications for FL in IoHT networks. First, we present a background overview of the basic concepts of FL applied in IoHT. In particular, we rigorously investigate and evaluate the main solutions to IoHT data security and privacy issues. In addition, we categorize the most relevant publications related to IoHT data security, whether due to advances in the architecture of FL or data protection. Next, we list several IoHT network datasets for model training. Finally, we highlight the essential lessons from this review, highlighting current challenges and possible directions for future research in data security and privacy in IoHT networks using FL.