Consumer Record Matching via Encrypted Attribute Hashing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for matching consumer information across different entities often face challenges due to privacy concerns and inefficiencies in data sharing, leading to missed matches despite the existence of corresponding information.
Innovation Solution
A method involving the receipt and processing of encrypted identification and attribute strings from multiple compute devices, where records with matching hashed attribute strings are combined, and subsequent filtering to create a matched record set, ensuring privacy compliance and accurate matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If consumer information is shared across entities for matching, then matching accuracy improves, but privacy compliance deteriorates
Solution Approach 1:
The patent introduces a trusted third-party computing environment (secure enclave, homomorphic encryption, or differential privacy mechanism) as an intermediary that enables matching operations without direct exposure of raw consumer data. This intermediary layer allows entities to contribute data for matching while maintaining privacy compliance, resolving the contradiction between matching accuracy and privacy protection.
Solution Approach 2:
The patent creates cryptographic copies or transformed representations of consumer data (encrypted identifiers, hashed attributes, synthetic data) that can be shared and processed for matching purposes without exposing the original sensitive information. These copies enable accurate matching while preserving privacy, as the original data remains protected in its source systems.
2Object-affected harmful factors
If limited information is used for matching, then privacy compliance improves, but matching accuracy deteriorates
Solution Approach 1:
The patent segments consumer information into different components with varying privacy sensitivities (identifiers, demographics, behavior data) and processes them through different privacy-preserving mechanisms. This segmentation allows maximization of usable information for matching while applying appropriate privacy protection to each segment, overcoming the limitation of having to use only limited information.
Solution Approach 2:
The patent transforms data parameters through cryptographic operations (encryption, hashing, differential privacy noise addition) that change the form of information to enable matching. These parameter changes allow the system to use far more information than limited approaches would permit, while the transformed parameters maintain privacy compliance through mathematical guarantees.
3Reliability
If multiple data sources are integrated for matching, then matching completeness improves, but system complexity deteriorates
Solution Approach 1:
The patent implements a universal matching framework that can integrate multiple data sources through a common set of privacy-preserving operations and protocols. This multi-functional system handles different data types, encryption methods, and computing environments (secure enclaves, homomorphic encryption, differential privacy) through unified interfaces, reducing the complexity that would otherwise arise from managing multiple separate integration systems.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In some embodiments, a method includes receiving a first record set from a first compute device and a second record set from a second compute device. Each record from the first record set has a first attribute string; and each record from the second record set has a first attribute string. The method includes defining a third record set to include each record from the first record set that has a first attribute string equal to a first attribute string of a record from the second record set. The method also includes repeating the above steps for a fourth record set from the first compute device and a fifth record set from the second compute device to further define the third record set. Each record from the fourth record set has a second attribute string and each record from the fifth record set has a second attribute string.