Mobile user equipment clustering training method for wireless federated learning

Abstract

The invention discloses a wireless federated learning-oriented mobile user equipment clustering training method, which is characterized in that an edge service server and a plurality of mobile user equipment are arranged in a wireless network, and each user equipment belongs to a local data set. Firstly, a DBSCAN algorithm and an LEACH algorithm are used to divide users into a plurality of clusters and select cluster heads. Then, each user uses local data to train respective model parameters and uploads the model parameters to a cluster head of a cluster where the user is located for aggregation, and the cluster head further uploads the aggregated model to an edge server for aggregation; and the edge server distributes the aggregated model parameters to the cluster head, and the cluster head further distributes the aggregated model parameters to the user equipment in the cluster for the next round of training. And repeating the steps until convergence. By using the clustering training method, the communication overhead and the training time delay of wireless federal learning can be reduced.

Description

technical field [0001] The present invention is oriented to the technical field of federated learning under the Internet of Things environment, and is especially aimed at improving the situation that the centralized aggregation server of federated learning may stop working due to physical damage or security attacks by malicious users, thereby interrupting the training process. Background technique [0002] At present, artificial intelligence (AI) has entered all aspects of life. As we all know, the core supporting AI training is data, especially high-quality data that is accurate and representative of the distribution. In real life, except for a few giant companies that can meet the requirements, most enterprises have the problem of small amount of data and poor data quality, which is not enough to support the realization of artificial intelligence technology; at the same time, domestic and foreign regulatory environments are gradually strengthening data protection, and succe...

AI Technical Summary

Problems solved by technology

In addition, the traditional centralized FL aggregation server may stop working due to security attacks or physical damage, and with the client training delay and client upload delay, it will cause slow convergence of machine learning. The distributed architecture The server can improve the system's high concurrency, high availability, and scalable performance

Images