A Survey on Federated Learning for Privacy-Preserving AI in Edge Computing Environments

Federated Learning (FL) has emerged as a decentralized machine learning paradigm that enables the training of AI models across distributed devices or servers while keeping raw data localized.This approach is particularly relevant in edge computing environments, where data are often generated at the edge of the network (e.g., mobile devices and IoT nodes) and privacy concerns prevent central aggregation.In this survey, we examine the latest work on federated learning, especially on how to keep data private, how to use it on edge devices, and how to improve the performance of the models.We discuss modern frameworks, important privacy methods such as differential privacy and secure multiparty computation, and ways to make training faster with less communication.We also grouped recent studies based on their application, such as in healthcare, smart cities, self-driving systems, and industrial IoT.Along with the progress made, we also point out the problems that still need to be solved for using federated learning on a large scale in edge computing.