Model joint training method, device and equipment based on asymmetric federated learning

Abstract

The embodiment of the present application discloses a model joint training method, device and equipment based on asymmetric federated learning, belonging to the field of computer technology. The method includes: acquiring a plurality of locally stored first sample data and corresponding first sample identifiers and indication identifiers, classifying the plurality of first sample data, determining first classification information, and sending the first classification information to the first The device sends the obtained multiple first sample identifiers and corresponding encrypted indication identifiers, receives the second classification information sent by the first device, and trains the classification based on the first classification information and the second classification information Model. The embodiment of the present application provides a way of jointly training the classification model. In the process of training the classification model, the classification information shared by other providers to the current provider only includes the encryption indicator corresponding to the sample data, so that the current provider The sample data cannot be inferred based on the classification information, which avoids information leakage and improves security.

Description

technical field [0001] The embodiments of the present application relate to the field of computer technology, and in particular to a model joint training method, device and equipment based on asymmetric federated learning. Background technique [0002] In the current information age, users can generate user data in various scenarios, such as user credit data, user medical data, user consumption data, etc., and these data are stored by their respective providers. Currently, a method for jointly training a classification model is proposed, which can jointly train a classification model based on sample data owned by multiple providers, thereby classifying user data based on the classification model. [0003] In the process of multi-party joint training of the classification model, each provider will classify according to the sample data it owns, and share the classification information with other providers. The classification information can indicate which sample data belongs t...

AI Technical Summary

Problems solved by technology

[0003] In the process of multi-party joint training of the classification model, each provider will classify according to the sample data it owns, and share the classification information with other providers. The classification information can indicate which sample data belongs to which category, resulting in information leakage, resulting in poor security

Images