Secure Event Similarity Detection Using Encoded Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing comparison methods for digital representations of events often reduce resolution and compromise security by requiring strict bit-level matching, and they can lead to information leakage, especially when dealing with personal or sensitive data.
Innovation Solution
A method and system that encode data sets into fixed-length vectors, allowing for secure similarity or dissimilarity comparisons over encrypted data using an approximate equality operator, which minimizes information leakage and maintains security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If strict bit-level matching is used for comparison, then security is improved, but measurement precision deteriorates
Solution Approach 1:
The patent segments the comparison process into two distinct phases: (1) encoding the data sets into fixed-length vectors that preserve similarity relationships, and (2) performing secure comparison operations on the encoded representations. This segmentation allows the system to maintain high measurement precision during encoding while ensuring security during comparison, resolving the contradiction between precision and security.
Solution Approach 2:
The patent introduces encoded vector representations as an intermediary between the original data sets and the comparison operation. These encoded vectors serve as a mediator that preserves the necessary similarity information for accurate comparison while being designed to provide security guarantees during the comparison process, thus resolving the contradiction between measurement precision and security.
2Measurement precision
If data sets are compared in raw form, then measurement precision is improved, but information leakage increases
Solution Approach 1:
The patent uses encoded vector representations as an intermediary that preserves the essential similarity information needed for accurate measurement while preventing direct access to the raw data. This intermediary encoding mechanism allows the system to achieve high measurement precision without exposing the underlying sensitive information, thus resolving the contradiction between precision and information leakage.
Solution Approach 2:
The patent transforms the data from its original form into a different parameter space (encoded vectors) that maintains the similarity relationships necessary for accurate comparison while changing the representation in a way that prevents information leakage. This parameter transformation allows the system to achieve both high measurement precision and reduced information leakage.
3Ease of operation
If resolution is reduced for useful comparison, then ease of operation is improved, but security deteriorates
Solution Approach 1:
The patent segments the comparison system into distinct components: an encoding module that handles the transformation to fixed-length vectors, and a comparison module that operates on the encoded representations. This segmentation allows each component to be optimized independently - the encoding ensures security while the comparison operation maintains usability, resolving the contradiction between ease of operation and security.
Data Source
AI summary
Each of a plurality of clients encodes events as respective vectors and cooperatively choose a joint key. Each client then encrypts its event vector(s) using the joint key to form secret shares of a fixed value and then sends the encoded, encrypted vectors to a service-providing system that selects pairs of the vectors and determines a comparison value from a reconstruction of the secret shares. When the comparison value meets a predetermined criterion, the service-providing system generates a message indicating similarity between the selected pairs of the vectors. The service providing system thus determines a degree of similarity between the events without requiring knowledge of raw data about the events.