In the era of decentralized data analysis, Federated Learning has emerged as a powerful approach that enables collaborative machine learning while preserving data privacy. Despite the rapid growth of Federated Learning techniques, it is still challenging for researchers and academics to obtain a thorough grasp of current approaches, difficulties, and future prospects due to the fragmented nature of the existing research on privacy preservation. This study aims to provide a comprehensive systematic review of the latest advancements in Federated Learning, its challenges, and future directions, along with some common pre-processing techniques and datasets, focusing on privacy preservation. It systematically examines the applications, techniques, and models that characterize the state of privacy preservation in Federated Learning. A total of 392 papers were initially found through searches, from where 153 publications, of which more than 90% were Scopus-indexed Q1 papers, met the eligibility requirements and were included in the final analysis. The paper ends with a thorough examination of particular challenges in privacy preservation in Federated Learning and future directions for further research. With an emphasis on current advancements in the field, this paper attempts to present a fair assessment of privacy preservation models in Federated Learning. The findings highlight the importance of integrating advanced privacy measures to ensure safe and ethical deployment of Federated Learning in real-world applications.
Sum et al. (Wed,) studied this question.