Model training method under federated learning network and related equipment thereof

Abstract

The embodiment of the invention belongs to the field of artificial intelligence. The method is applied to the field of smart communities, and relates to a model training method under a federated learning network and related equipment thereof, the federated learning network comprising a central client and a plurality of nodes is established, a control node receives an initialization model as a local model, and the control node uses local data to train the local model to obtain gradient information; controlling the central client to generate global information according to the gradient information; the control node obtains gradient information of other nodes according to the global information, uses the gradient information to test a local model of the current node to obtain accuracy, adjusts the global information according to the accuracy, and updates the local model of the current node; obtaining a result model until the model converges; and inputting the user data received by the node into a result model corresponding to the node to obtain recommendation information output by the result model. The gradient information of each node may be stored in a blockchain node. According tothe invention, personalized training of local models of different nodes is realized.

Description

technical field [0001] This application relates to the field of artificial intelligence technology, in particular to a model training method and related equipment under a federated learning network. Background technique [0002] Federated machine learning refers to a machine learning framework that can effectively help multiple nodes perform data usage and machine learning modeling while meeting the requirements of data privacy protection and data security. [0003] At present, the optimization methods of federated learning include FedSGD, FedAvg, FedProx, FedMA, SCAFFOLD, etc. However, these methods all update the model on the central client, and the final training models of each participant are basically the same, which makes it impossible to achieve personalized training; there is a certain loss in the Non-IID (independent and identical distribution) data distribution, and the accuracy rate is not high enough , and when some nodes use meaningless data to maliciously part...

AI Technical Summary

Problems solved by technology

However, these methods all update the model on the central client, and the final training models of each participant are basically the same, which makes it impossible to achieve personalized training; there is a certain loss in the Non-IID (independent and identi...

Images