Document Type : Original Article
Authors
1 Senior student of Sports Physiology and Health, Islamic Azad University, Science and Research Branch, Tehran
2 Associate Professor, Department of Exercise Physiology, Islamic Azad University, Science and Research Branch
Abstract
Smart wearable technologies hold immense potential for continuous collection of physiological data and the enhancement of personal health. However, challenges related to data privacy, security, and the integration of data from fragmented sources (data silos) hinder the full realization of this potential. Federated Learning (FL), as an emerging machine learning paradigm, offers a promising solution by enabling decentralized data analysis without the need to transfer raw data to a central server.
This systematic review aims to examine the applications, opportunities, and challenges of using federated learning for analyzing physiological data collected by smart wearables. Through a comprehensive search of reputable scientific databases, relevant studies are identified and analyzed.
Key opportunities include preserving user privacy, enhancing data security, overcoming data fragmentation, and achieving scalability across a large number of devices. On the other hand, the main challenges involve data heterogeneity across different devices, communication overhead, model aggregation complexities, and vulnerability to specific attacks unique to federated learning.
The findings of this review suggest that federated learning represents a promising pathway for advancing artificial intelligence in the domain of wearable health technologies. However, further research and innovation are required to address practical limitations and to ensure the efficiency and security of such systems at scale.
Keywords
Main Subjects