Secure joint calculation method and system based on multi-party data

Abstract

The invention relates to a secure joint calculation method and system based on multi-party data. The method comprises the steps that feature processing is carried out on data of a local party, a partner and other parties, and heterogeneous information network diagrams are respectively generated; a cooperative communication module inputs data of the heterogeneous information network diagrams of the data of the local party, the partner and other parties after the feature processing into a main control end for joint calculation; and evaluation and optimization are performed after the joint calculation is completed, and optimized parameters are input into feature processing modules of all the parties again. According to the invention, the problem of the joint calculation among a plurality of non-trusted participants on the premise of protecting respective data privacy security is solved, and cross-domain joint calculation of multi-party data and evaluation of calculation are realized on the premise of ensuring independence of multi-party input and privacy of the data; and meanwhile, the accuracy of the joint calculation is continuously optimized and iterated through a calculation evaluation module. The invention has broad applicability.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a multi-party data-based safe joint computing method and system. Background technique [0002] Joint computing is a computing framework for realizing distributed computing proposed to solve the problem of multi-party data security in computing. It can be used for joint computing of multi-party device-side data, and each computing device performs a meaningful calculation. , and gather the calculation results with others through the network. [0003] The more popular in joint computing is secure multi-party computing, which is a multi-party computing method "for securely performing multi-party collaborative computing without a trusted third party. That is, in a distributed network, multiple The participants participating in the multi-party calculation each input the data participating in the multi-party calculation. Each participant hopes to jointly complete the calculation of a...

AI Technical Summary

Problems solved by technology

Existing technical measures can only deal with structured citizen data, and cannot handle unstructured or semi-structured citizen data, but more and more "unstructured data" citizen data have begun to be applied on a large scale, including social media, email , text, audio, video, photos, web logs, etc., and the existing data governance cannot realize the trusted sharing and joint computing functions of these unstructured citizen data or semi-structured citizen data

[0008] Secondly, the current technology cannot provide security guarantees for citizen data in the process of cross-domain joint computing. Traditional data governance only includes information resource directory management, metadata management, and data quality management, but lacks security for citizen data in the computing process. Guarantee, especially in the process of cross-department data joint calculation, due to the insecurity of the sharing mechanism, which leads to data leakage in the calculation process, so the existing data sharing mechanism cannot provide security for the joint calculation of citizen data

[0009] Thirdly, even if citizen data is shared through the alliance chain, the lack of a suitable security mechanism on the alliance chain will also cause a large-scale data leakage of all citizen data when it is transmitted on the alliance chain

[0010] To sum up, the existing citizen data sharing mechanism based on data governance cannot realize the security issues in the cross-domain joint computing of citizen data

Images