Key management system and method for federated learning security audit, and storage medium

Abstract

The invention relates to a key management system and method for federated learning security audit, and a storage medium. The key management system comprises: a user interface module used for receiving a request of a user; a key service processing module which is configured to generate a visible key space specific to the user, wherein a key in the visible key space specific to the user can be used for responding to a request of the user; a key generation module; and a key storage module. The state of each secret key stored by the secret key storage module is determined according to whether the federated learning task corresponding to the secret key is finished or not and whether the security audit requirement corresponding to the federated learning task is met or not. The method not only can meet the requirements of various federated learning frameworks and software products for normalization and security, but also considers the requirements of subsequent security auditing and the differences of different individuals of the same participant or different participants in the aspects of permission, privacy, isolation requirements and the like.

Description

technical field [0001] This application relates to the technical field of data security and privacy protection, and in particular to a key management system, method and storage medium for federated learning security audit. Background technique [0002] With the development of application fields such as artificial intelligence and big data mining analysis, the demand for data volume is increasing. For example, training artificial intelligence application models requires a large amount of training data with appropriate data labels or feature values. High-quality data often come from the application data generated and accumulated in business activities. However, application data is often scattered in the hands of different organizations and individuals. For example, transaction data is scattered in various financial institutions, and medical diagnosis data is scattered in various medical institutions. Cross-industry and cross-field application data are often scattered. For ex...

AI Technical Summary

Problems solved by technology

[0003] With the development of federated learning, more and more participants began to collaboratively optimize the federated learning model through the federated learning framework, but each participant often generates its own keys, including public keys for plaintext encryption and private keys for decryption , and after completing the encryption and decryption tasks of the gradient model, the private key necessary for decryption may be lost, which will cause obstacles to subsequent security audits and also pose a threat to the normative and security of the entire federated learning

Moreover, the same participant may have different individuals who create or delete keys on behalf of the participant at different times, and these individuals may have different permissions or isolation requirements or lack effective key management records

Images