Model processing method, system and device for shared learning and electronic equipment

Abstract

The invention provides a model processing method for shared learning. The method comprises the following steps: obtaining a plurality of candidate model update values for an initial neural network model; performing quality evaluation on the plurality of candidate model update values to obtain weights of the plurality of candidate model update values; adopting a median mechanism with weight for the plurality of candidate model update values, and determining a model update value for updating the initial neural network model; and updating the initial neural network model based on the determined update value. By firstly obtaining the plurality of candidate model update values, then obtaining weights of the plurality of candidate model update values through quality evaluation, and then determining the model update value for updating the initial neural network model based on the weighted median mechanism, based on the weighted median mechanism, it can be avoided that a false trained neural network model is adopted to update the initial neural network model, and the problem that an updated neural network model does not conform to a real situation by adopting an existing model updating method can be solved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a model processing method, system, device and electronic equipment for shared learning. Background technique [0002] With the continuous development of neural network technology, the application of neural network technology in various fields is becoming more and more extensive. Especially in shared learning, the combination of shared learning technology and neural network technology can facilitate the protection of user privacy data. The combination of shared learning technology and neural network technology is mainly based on the following idea: the cloud sends the initial neural network model to be trained to multiple users, and the user can realize the training of the initial neural network model through the data set at the local end, and then the user will train The neural network model is sent to the cloud. Since this process prevents the user from sending the...

AI Technical Summary

Problems solved by technology

[0004] This application provides a model processing method for shared learning to solve the problem that the existing model update method cannot avoid the malicious behavior of some users in order to increase their own weight, and finally make the updated neural network model not conform to the real situation.

Images