Edge-based federated learning model cleaning and equipment clustering method, system and equipment and readable storage medium

Abstract

The invention provides an edge-based federated learning model cleaning and equipment clustering method, system and equipment and a readable storage medium. The method comprises the steps of carrying out the clustering of equipment according to the addresses of local area networks where the equipment is located, and deploying a mobile edge node server in each local area network, enabling the terminal equipment participating in training to receive the global model sent by the cloud and performing training on the local data to obtain a local updating model, calculating cosine similarity between the locally updated model parameter and the global model parameter of the terminal equipment, judging whether the cosine similarity is greater than a set threshold value or not, and if the cosine similarity is greater than the set threshold value, transmitting the locally updated model to a mobile edge node server to participate in edge aggregation to obtain a cluster model, and sending the clustermodel of the local area network to the cloud to participate in global aggregation to obtain a global aggregation model. According to the method provided by the invention, the federated learning communication efficiency can be improved under the conditions of reducing unnecessary communication overhead and avoiding transmission delay caused by high concurrent access of the server.

Description

Technical field[0001]The invention relates to the field of edge intelligence federated learning, in particular to an edge-based federated learning model cleaning and device clustering method, system, device and readable storage medium.Background technique[0002]As a result of industry competition and data privacy protection, in most industries, data often exists in the form of islands. Even in the same company, data integration between different departments is facing huge resistance, not to mention the integration of data from various agencies, which is almost impossible in reality. In addition, with the further development of big data, the emphasis on data privacy and security has become a global trend. Therefore, the traditional machine learning method of sending terminal data to the cloud for deep learning is facing great challenges. As the core technology of Artificial Intelligence (AI), Federated Learning (FL) is a promising method to solve this challenge. In the learning proces...

AI Technical Summary

Problems solved by technology

In contrast, there are usually two problems in the network between the terminal device and the cloud server: 1) the bandwidth of the network is limited, and the server for high-bandwidth services is expensive; 2) the network connection between the local and the cloud has asymmetric characteristics : The uplink speed of the network is usually much slower than the downlink speed

In other words, the existing methods to improve the efficiency of federated learning are too one-sided, and do not take into account the multi-dimensional differences of model parameters and the delay caused by high server concurrency

Images