Encrypted Vector Matching for Secure Fuzzy Data Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current secure intersection algorithms in privacy computing, such as those based on RSA encryption, can only match data successfully when both parties' data are exactly the same, limiting their application scenarios and business scope.
Innovation Solution
A data matching method using semi-homomorphic encryption and pre-trained vector conversion models to generate encrypted vectors, allowing for fuzzy matching by calculating an encrypted distance and determining a target distance using a private key, enabling matching even when data types and character counts differ.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional secure intersection algorithms (e.g., RSA encryption) are used, then data security is ensured, but data matching is limited to exact matches only, reducing adaptability
Solution Approach 1:
The patent changes the parameter of encryption from traditional RSA to semi-homomorphic encryption, which allows mathematical operations on encrypted data. This enables distance calculation on encrypted vectors, facilitating fuzzy matching while maintaining security. The semi-homomorphic encryption parameter allows the system to compute encrypted distances without decrypting the data, resolving the contradiction between security and matching flexibility.
Solution Approach 2:
The patent introduces an intermediary mechanism using semi-homomorphic encryption to enable distance calculation on encrypted data. The encryption algorithm acts as a mediator that allows the second device to compute encrypted distances based on encrypted vectors without accessing the actual data, thus maintaining security while enabling flexible matching. This intermediary approach allows fuzzy matching capabilities to emerge from the encryption process itself.
2Measurement precision
If data must be exactly the same for matching, then matching precision is high, but the business scope and application scenarios are limited
Solution Approach 1:
The patent introduces dynamic matching capability through semi-homomorphic encryption, allowing the system to adapt between exact matching and fuzzy matching based on needs. The encrypted distance calculation enables flexible matching thresholds, transforming the static exact-match requirement into a dynamic adjustable parameter. This dynamic approach maintains high precision when needed while expanding application scenarios where approximate matching is acceptable.
Solution Approach 2:
The patent performs preliminary vector conversion on data before encryption, transforming data into vector representations that can be processed mathematically. This preliminary action enables subsequent distance calculations and fuzzy matching operations. By preparing data in vector form beforehand, the system can perform flexible matching operations without compromising the original data integrity or security.
3Adaptability or versatility
If semi-homomorphic encryption is used for fuzzy matching, then adaptability is improved, but computational complexity increases
Solution Approach 1:
The patent segments the data matching process into distinct phases: vector conversion, encryption, distance calculation, and decryption. By dividing the complex task into manageable segments, each handled by a specific component or device, the system reduces overall computational complexity. The first device handles vector conversion and encryption, while the second device performs distance calculation on encrypted data, distributing the computational burden and reducing complexity at any single point.
Data Source
AI summary
Disclosed in the present disclosure are a data matching method. In the present disclosure, a first vector corresponding to first data and a second vector corresponding to second data are respectively obtained; a first encrypted vector obtained by means of encrypting the first vector and a second encrypted vector obtained by means of encrypting the second vector are acquired; a first encrypted distance is calculated on the basis of the first encrypted vector and the second encrypted vector; a target distance between the first vector and the second vector is determined on the basis of the first encrypted distance and a first target private key; and on the basis of the target distance and a first preset distance threshold value, it is determined whether the first data matches the second data.