Layered Model Aggregation based Federated Learning in Mobile Edge Networks
Qiming Cao, Xing Zhang, Yushun Zhang, Yongdong Zhu · 2021
With the continuous improving of performance of the IoT and mobile devices, a new type of machine learning architecture, federated learning came into being. And there is also an increasing need to implement artificial intelligence frameworks on edge nodes. In this paper, we propose a federated learning system deployed in edge computing network, which realizes the server part in distributed form and uses layered model aggregation and dynamic topology to reduce bandwidth usage and time consuming. The experiment shows the effectiveness of federated learning algorithm in our system. And simulation results show that the time cost of our system increases logarithmically with the number of nodes rather than linearly in the traditional system.