Homomorphic Encryption for Private Correlation Coefficient Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing correlation analysis methods for massive data, particularly in federal learning models, lead to leakage of private information due to data sharing, compromising data security and accuracy.
Innovation Solution
A method involving homomorphic encryption, where data is encrypted using an associated key agreed upon by participation nodes, allowing for correlation coefficient calculations without decrypting the data, ensuring privacy and security while maintaining accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data sharing is performed for correlation analysis in federal learning models, then correlation coefficients can be calculated, but private information leakage occurs compromising data security
Solution Approach 1:
An auxiliary node is introduced as an intermediary to perform homomorphic operations on encrypted data. The auxiliary node receives encrypted data from participation nodes, performs correlation coefficient calculations without decrypting the data, and returns encrypted results. This mediator approach allows accurate correlation analysis while preventing private information leakage, as the auxiliary node never accesses plaintext data.
Solution Approach 2:
The patent replaces traditional mechanical data sharing and processing mechanisms with homomorphic encryption technology. Instead of sharing plaintext data for correlation analysis, the system uses homomorphically encrypted data that can undergo mathematical operations while remaining encrypted. This substitution maintains calculation accuracy while fundamentally securing data privacy throughout the computation process.
2Reliability
If homomorphic encryption is applied to protect data privacy, then data security is improved, but calculation complexity increases
Solution Approach 1:
The system segments the homomorphic encryption process into distinct functional components: data encryption at participation nodes, homomorphic operations at the auxiliary node, and result decryption. This segmentation allows each component to be optimized independently and reduces the computational burden on any single node, making the overall system more manageable despite the inherent complexity of homomorphic encryption.
3Reliability
If data is encrypted before transmission and calculation, then transmission safety is improved, but processing speed decreases
Solution Approach 1:
The homomorphic encryption process is designed to maintain data encryption throughout the entire computation pipeline without requiring decryption intermediaries. The auxiliary node performs all correlation calculations on encrypted data continuously, eliminating the need for decryption/encryption cycles during processing. This continuous encrypted processing maintains transmission safety while minimizing the time overhead associated with repeated encryption/decryption operations.
Data Source
AI summary
Provided are a correlation coefficient acquisition method, an electronic device, and a non-transitory computer readable storage medium. The implementation scheme is as follows: first original data is acquired, the first original data is homomorphically encrypted by using an associated key to determine first transmission data, where the associated key is jointly agreed by the first participation node and a second participation node; the first transmission data is sent to an auxiliary node so that the auxiliary node receives the first transmission data and performs a homomorphic operation on the first transmission data and second transmission data to obtain correlation coefficients between the first original data and second original data, where the second transmission data is determined by the second participation node homomorphically encrypting the second original data by using the associated key; and the correlation coefficients fed back by the auxiliary node is received.