Cross-device ECG federal privacy classification framework method and system and storage medium

Abstract

The invention discloses a cross-device ECG federal privacy classification framework method and system and a storage medium, and relates to the technical field of federal learning. The method comprises the following specific steps: a server sends an initial model parameter to a client; the client uses the training data set to train the initial model to obtain basic type identification model parameters, and the basic type identification model parameters are sent to the server in a differential mode; the server receives the basic type identification model parameters for clustering to obtain a clustering result; the method also includes performing weighted average on the clustering result to obtain updating parameters of a final server model, and finishing training of the initial model; and identifying the ECG data through the trained initial model, and outputting respective ECG data type identification results on each client. According to the invention, the data identification precision can be improved while the user data privacy is protected in the ECG health monitoring process.

Description

technical field [0001] The present invention relates to the technical field of federated learning, and more specifically relates to a cross-device ECG federated privacy classification framework method, system and storage medium. Background technique [0002] With the increase in the incidence of cardiovascular diseases and the development of IoT software and hardware devices, the demand for long-term home ECG health monitoring equipment has increased rapidly. At the same time, user privacy and security issues exposed in the ECG health monitoring process have become more and more prominent. Although the centralized ECG diagnosis model can avoid user data leakage, it cannot incorporate more rare types of data into the model in time. In order to balance the identification and diagnosis performance and privacy properties of the model, the existing best ECG privacy-preserving classification scheme is a collaborative learning strategy, including federated learning (Federated Lear...

AI Technical Summary

Problems solved by technology

[0003] However, when the federated learning framework is used in the ECG type recognition task, the cross-device user data will have extreme Non-IID performance, resulting in a sharp decline in the predictive ability of the federated model for rare types of clients.

Existing work cannot adapt to the scenario of extreme Non-IID user data, because the parameters from different clients to the server are treated equally, and it is impossible to achieve an effective knowledge transfer mechanism between different clients and a single client's global awareness

Images