Method and device for processing model parameters in federated learning process and related equipment

Abstract

The invention discloses a method for processing model parameters in a federated learning process, which is applied to the technical field of artificial intelligence and is used for solving the technical problem that in the federated learning process, user privacy data is easy to leak when mutual information representation between the gradient of the model parameters and sample data is too large. The method provided by the invention comprises the following steps: sampling a sample data set for training a local model to obtain a plurality of samples; obtaining a gradient of a parameter obtained by training the local model through the plurality of samples; inputting the sampling sample and the gradient of the parameter into a pre-trained statistical model to obtain a mutual information value between the sample data set and the gradient; when the mutual information value is greater than or equal to a preset value, sending a risk prompt that privacy disclosure exists in the gradient for uploading the parameter; and when the mutual information value is smaller than the preset value, uploading the gradient of the parameter to a server.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a method, device, computer equipment and storage medium for processing model parameters in the process of federated learning. Background technique [0002] At present, the federated learning system is mainly divided into horizontal federated learning and vertical federated learning. In the horizontal federated learning system, each participant uses local data to train the local model, each participant uploads the gradient of the parameters of the trained local model to the central server, and the central server aggregates the gradient of the parameters of each participant to the model Update, and finally return the updated model to each participant. This method can avoid the leakage of user's private data due to the theft of user sample data in the process of transmitting user sample data by directly transmitting the gradient of model parameters. [0003]...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a method, device, computer equipment, and storage medium for processing model parameters in the process of federated learning, so as to solve the problem that the mutual information between the gradient of the model parameters and the sample data is too large in the process of federated learning Technical issues that easily leak user privacy data

Images