Privacy Preserving Column Binding via Homomorphic Encryption
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing homomorphic encryption methods require data owners to reveal private data when finding intersections between datasets, compromising user privacy.
Innovation Solution
A method that receives encrypted data sets from multiple computing devices, creates a single data set and a common identifier vector indicating shared encrypted identifiers, and transmits this information back to the data owners, allowing for intersection computations without revealing unencrypted data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional column binding algorithms are used to find intersections between datasets, then the intersection can be accurately determined, but private data must be revealed during the process
Solution Approach 1:
The patent introduces encrypted identifiers as an intermediary mechanism. Instead of directly comparing plaintext data to find intersections, the system encrypts identifiers before transmission and uses homomorphic encryption to perform operations on encrypted data. This intermediary layer allows intersection determination without exposing private underlying data, resolving the contradiction between accurate intersection finding and privacy preservation.
Solution Approach 2:
The patent transforms the data representation by changing from plaintext identifiers to encrypted identifiers. By applying encryption as a parameter transformation, the system maintains the ability to perform intersection operations (through homomorphic encryption) while the data exists in an encrypted state throughout the process, preventing private data revelation while preserving intersection accuracy.
2Measurement precision
If all data from multiple data owners is transmitted for intersection computation, then complete intersection analysis can be performed, but data transmission volume and processing operations increase significantly
Solution Approach 1:
The patent extracts only the essential element needed for intersection determination - the identifiers - and transmits them in encrypted form. Instead of transmitting entire datasets containing all attributes and records, the system identifies and transmits only the identifier component that is necessary for matching records across datasets. This extraction principle reduces data transmission volume while maintaining complete intersection analysis capability.
Solution Approach 2:
The patent segments the data transmission process into distinct components: first transmitting encrypted identifiers, then using those identifiers to selectively retrieve and transmit only the specific data records that need to be compared. This segmentation allows the system to perform complete intersection analysis by breaking down the large-scale data transmission into smaller, manageable segments based on identifier matching.
3Productivity
If traditional unencrypted data processing is used, then processing operations are simple and fast, but user privacy is compromised
Solution Approach 1:
The patent changes the state parameter of the data from plaintext to encrypted form throughout the processing pipeline. By maintaining encryption as a persistent state parameter, the system enables privacy-preserving processing. Homomorphic encryption properties allow certain operations to be performed directly on encrypted data without decryption, maintaining productivity while preventing privacy loss that would occur with traditional unencrypted processing.
Solution Approach 2:
The patent substitutes traditional mechanical data processing operations with homomorphic encryption-based operations. Instead of decrypting data and performing operations on plaintext, the system uses mathematical properties of homomorphic encryption to perform computations on ciphertext. This substitution replaces the traditional mechanism with an encrypted mechanism that achieves the same functional result while preserving privacy.
Data Source
AI summary
Methods and systems for improving homomorphic encryption include: receiving, by a computing device, a data set from at least two or more computing devices, each data set comprising: a plurality of encrypted identifiers, and a number indicating the number of the plurality of encrypted identifiers; creating, by the computing device, a single data set including each of the received data sets; creating, by the computing device, a common identifier vector that indicates each encrypted identifier in the single data set that has underlying unencrypted data that is the same in more than one of the received data sets; and transmitting, by the computing device, the common identifier vector to each of the at least two or more computing devices.