Private Set Intersection Labeling With Reduced Ciphertext Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing private set intersection (PSI) techniques face challenges in handling large volumes of data due to increased data transmission and computation, leading to inefficiencies and performance degradation, especially in homomorphic encryption-based methods.
Innovation Solution
A method and device that utilize a novel homomorphic encryption approach with a reduced circuit depth by precomputing and encrypting data, employing a linear combination of random numbers and label information to generate final ciphertext, thereby minimizing computation and optimizing the matching function.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If typical PSI techniques (hash-based, public key password-based, circuit-based, oblivious transfer-based) are used to protect data privacy, then privacy protection is improved, but operation speed and scalability deteriorate significantly when comparing large volumes of data
Solution Approach 1:
The patent replaces traditional cryptographic mechanisms (hash functions, public key infrastructure, circuit evaluation, oblivious transfer protocols) with a homomorphic encryption-based polynomial evaluation system. This substitution enables direct computation on encrypted data without the overhead of iterative cryptographic operations, achieving both privacy protection and improved operational efficiency for large-scale PSI operations
Solution Approach 2:
The patent transforms the PSI problem from a set-based comparison operation into a polynomial evaluation problem where parameters are changed from discrete set elements to continuous polynomial coefficients. This parameter transformation allows the use of efficient homomorphic encryption operations and reduces the computational complexity from exponential in traditional methods to polynomial in the new approach
2Quantity of substance
If the number of comparison targets increases to handle larger datasets, then data coverage is improved, but the amount of data transmission and computation increases significantly
Solution Approach 1:
The patent performs preliminary actions by pre-computing polynomial representations of data sets and pre-establishing homomorphic encryption keys before the actual PSI operation. This preliminary preparation enables the main PSI computation to proceed efficiently with minimal real-time computation and transmission overhead, even when handling large volumes of data
Solution Approach 2:
The patent introduces polynomials as an intermediary representation between raw data sets and the PSI computation. This intermediary structure allows efficient homomorphic operations to be performed on encrypted polynomial coefficients rather than directly on encrypted set elements, significantly reducing the computational and transmission overhead for large-scale PSI operations
Data Source
AI summary
A processor-implemented method including generating intersection operation result ciphertext by performing an intersection operation between a first ciphertext, the first ciphertext corresponding to elements of a first data set of a first device, and elements of a second data set of a second device, based on a matching function defined for the elements of the second data set, generating a label ciphertext by encrypting label information obtained from the elements of the second data set, and generating a final ciphertext based on the intersection operation result ciphertext and the label ciphertext.