Federated Learning Techniques for 5G Mobile Networks
Bohdan Shubyn, Dariusz Mrozek, Ludvig Fabry, Taras Maksymyuk, El Mehdi Amhoud, Juraj Gazda · 2022 IEEE 16th International Conference on Advanced Trends in Radioelectronics, Telecommunications and Computer Engineering (TCSET) · 2022
The growing adoption of AI/ML in 5G mobile networks results in the increased amount of the sensitive data sharing by the end users, which may cause certain privacy and security concerns. Therefore, the general efforts in the direction of privacy protection by AI/ML community so far are focused on the federated learning techniques. In this paper, we consider the main advantages of the decentralized federated learning approach in the case of 5G mobile networks, in terms of the data privacy for subscribers. We implement the federated learning model to the task of network load prediction. Simulation results show that proposed federated learning model provided the same accuracy as conventional learning model, while reducing the redundancy and ensuring the privacy of user's data.