Multi-party unsupervised learning joint modeling method based on X86 computing chip

Abstract

The embodiment of the invention provides a multi-party unsupervised learning joint modeling method based on an X86 computing chip, and the method comprises providing a trusted measurement manager based on an X86 architecture chip, a distributed unsupervised machine learning framework and a privacy computing interconnection system, wherein the trusted measurement manager based on the X86 architecture chip comprises a plurality of servers based on the X86 architecture chip, one server serves as a service party, and the other servers serve as calculation parties; realizing multi-party local unsupervised model training based on local data, performing interaction with a parameter server at the same time, realizing multi-party local model safe aggregation, and building a global model. All parties involved in the method carry out data transmission and model parameter aggregation through a privacy computing interconnection system constructed based on MPC, a trusted execution environment or an encrypted connection mode, and the privacy of related data is ensured. And data availability and invisibility, data algorithm credibility measurement and multi-party joint unsupervised learning modeling are ensured.

Description

technical field [0001] The present invention relates to the field of computer technology, more specifically, to a multi-party non-supervised learning joint modeling method based on an X86 computing chip. Background technique [0002] At present, the large amount of data accumulated in various industries requires multi-party data joint for risk control and marketing in order to better mine the value of data. [0003] The way to achieve multi-party data union in the prior art usually includes: 1) summarizing the data of each participant, then performing model training and jointly applying the model; 2) storing the data of each participant in a third party; 3) each Participants realize data sharing through a customized data circulation platform, and each participant agrees on an interface with each other, and calls the corresponding data through the interface. [0004] However, no matter which method is used, there will be a risk of data leakage or the modeling process cannot ...

AI Technical Summary

Problems solved by technology

[0004] However, no matter which method is used, there will be a risk of data leakage or the modeling process cannot be verified and audited, which is very likely to cause major losses to all participants, and may also reduce the trust between the participants.

Moreover, the existing data modeling methods have high customization costs, cumbersome processes, and insufficient transparency

Images