Distributed Collaborative Learning in Wireless Mobile Communication

Communication, especially wireless mobile communication is a rapidly evolving field, with new technologies being developed all the time. One of the key challenges facing this field is the need to collect and analyze large amounts of data in order to improve the performance of wireless networks. The traditional approach of collecting and storing data in centralized servers is limited by the risk of data breaches, the slow and costly nature of transferring large amounts of data over the network, and the potential for privacy concerns. Collaborative and distributed Machine Learning (ML) techniques offer a solution to these challenges by allowing multiple entities to train ML models together, without the need to transfer the data to a central server. Two of the most promising such ML techniques for wireless mobile communication are Federated Learning (FL) and Split Learning (SL), both of which follow a model-to-data approach whereby clients train and test ML models without sharing raw data. While FL has gotten some attention in the field of wireless communication, SL, which has better performance than FL in certain instances, is relatively unexplored. This work investigates the application and comparison of SL, FL, and their combination, SplitFed Learning (SFL), in wireless communication. It presents a comprehensive exploration of the most suitable methods for addressing the main tasks in this area. Additionally, the study provides an in-depth comparison of these three approaches applied in wireless network traffic analysis. The experimental results highlight the strengths and limitations of each approach, aiding in the selection of the most suitable technique for network traffic analysis. This research shows the importance of using distributed ML approaches in wireless communication and contributes to the advancement of this field.