Federated Learning for Privacy-Preserving Artificial Intelligence: A Systematic Review

Abstract

Federated Learning (FL) has emerged as a transformative paradigm for enabling privacy-preserving Artificial Intelligence (AI) by allowing machine learning models to be trained collaboratively across decentralized devices and organizations without requiring the exchange of raw data. This systematic review examines the current state of research in federated learning, focusing on its architectures, privacy-preserving mechanisms, applications, challenges, and future research directions.