Federated learning communication traffic optimization method and system based on core data set

Abstract

The invention relates to the field of federated machine learning, and discloses a federated learning communication traffic optimization method and system based on a core data set. The method comprisesthe following steps that: firstly, each terminal user screens out core data from local training data in parallel, a cloud center constructs a sparse global model according to a set sparse proportion,and each terminal user performs local model training according to the screened local core data to obtain a local model update; then, in order to enable the global model to adapt to local core data better, the cloud center adjusts the network structure of the global model according to global model update obtained by an aggregating local model update, including the two steps of removing unimportantconnection and adding important connection; and finally, the cloud center distributes the adjusted global model to each terminal user, and iterates the above steps until the global model converges. By screening the core data from the terminal user and deploying the adaptive sparse network model, the uploading of the model parameters of the terminal user and the cloud center is reduced, and the problem of high communication cost caused by frequent transmission of high-dimensional update parameters between the terminal user and the cloud center in the federated learning technology is essentially solved.

Description

technical field [0001] The present invention relates to the field of federated machine learning, and more specifically, to a method and system for optimizing federated learning traffic based on core data sets, which are used to solve the problem of frequently transmitting high-dimensional update parameters between end users / equipment and cloud centers in federated learning technology The problem of high communication cost caused. Background technique [0002] As an important branch of artificial intelligence, machine learning has been successfully and widely used in various fields such as pattern recognition, data mining and computer vision. Due to the limited computing resources of terminal devices, the current training of machine learning models usually adopts a cloud-based method. In this method, the data collected by terminal devices, such as pictures, videos, or personal location information, must be uploaded to the cloud. Center, and the cloud center completes the tra...

AI Technical Summary

Problems solved by technology

[0008] In order to more effectively solve the high communication cost problem of federated learning, the...

Images