Privacy-Preserving Computation Data Integrity Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing privacy-preserving computation techniques fail to verify the accuracy and completeness of data when two or more participants are involved, as they prioritize data privacy over accountability, leading to a lack of transparency and verifiability in computational results.
Innovation Solution
Implementing a framework that enables computations over sets of data from multiple participants while preserving privacy, using techniques such as oblivious transfers, random shares, and Beaver triple arithmetic, which allows for data integrity checks and verification of computation results without compromising data protection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If privacy-preserving computation techniques are used to protect data during computations, then data privacy is improved, but data verifiability and accountability deteriorate
Solution Approach 1:
The patent segments the computation process into distinct phases: data submission phase where participants submit data with proofs of correctness, computation phase where the result is calculated, and verification phase where the result and data accuracy are verified. This segmentation allows privacy to be maintained during computation while verifiability is ensured through separate proof mechanisms.
Solution Approach 2:
The patent introduces an intermediary verification mechanism that acts as a mediator between data privacy and verifiability. This intermediary system validates data accuracy through proofs of correctness without revealing the actual data content, thus preserving privacy while enabling verification.
2Reliability
If data is obfuscated using differential privacy techniques for large populations, then data privacy is improved, but applicability to small participant groups deteriorates
Solution Approach 1:
The patent creates a universal verification framework that functions across different participant group sizes. The proof of correctness mechanism and verification protocol are designed to be adaptable whether there are two participants or many participants, making the system universally applicable rather than limited to specific scenarios.
Solution Approach 2:
The patent changes the verification parameters based on the number of participants. For large populations, the system can use aggregated verification methods, while for small groups like two participants, it employs individual proof validation. This parameter adjustment allows the same framework to serve different scales effectively.
3Measurement precision
If participants are required to provide complete and accurate data for accountability, then data accuracy is improved, but data protection and privacy deteriorate
Solution Approach 1:
The patent uses copying in the form of cryptographic proofs of correctness that replicate the accuracy verification function without copying the actual data. These proofs serve as substitutes that verify data accuracy while the original private data remains protected and undisclosed.
Solution Approach 2:
The patent replaces the mechanical system of data inspection (where accuracy would be verified by examining actual data) with a cryptographic verification system. Instead of mechanically checking data content for accuracy, the system uses mathematical proofs that guarantee accuracy without requiring data exposure.
Data Source
AI summary
Methods, systems and computer program products for data analytics. An information ecosystem comprises a plurality of participants and a plurality of data sets associated with the participants. An event initiates performance of a computation over different obfuscated data sets to determine an obfuscated computational result. An integrity value pertaining to constituent data of the different obfuscated data sets and, correspondingly, an integrity value pertaining to the computational result itself, is quantified by checking if the earlier offered data set or any constituents thereof are consistent with one or more aspects of later retrieved data. Certain variations of methods, systems and computer program products are used for verifying data accuracy in privacy-preserving computations that are performed in a health ecosystem where the data sets pertain to health information associated with the participants. When the integrity value is below a threshold, the data is deemed to include falsified or inaccurate data.