Secure Dataset Analysis via Data Obscuring and Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Companies face a dilemma in sharing data sets for financial gain while maintaining control, as sharing data can lead to loss of control and exposure of proprietary information, and internal data sharing within companies is hindered by fears of data copying and distribution.
Innovation Solution
A system that partially obscures data sets, allowing users to perform statistical operations while limiting exposure by using encryption, de-correlation, and random variation, ensuring that results do not reveal the original data unless necessary, and embedding watermarks for tracking usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is shared with third parties for analysis, then data value and monetization potential increase, but data security and control are compromised
Solution Approach 1:
The data is segmented into two distinct forms: obscured data for sharing with third parties and unobscured data retained by the owner. This segmentation allows the owner to provide access to analytical value while maintaining control over the actual data content, resolving the contradiction between data sharing and data security.
Solution Approach 2:
An obscuring mechanism acts as an intermediary between the data owner and third-party analysts. This intermediary transforms the data into an obscured form that preserves analytical utility while removing sensitive information, enabling monetization without compromising security.
2Reliability
If data is obscured to protect security, then data control is maintained, but analytical utility is reduced
Solution Approach 1:
The obscuring mechanism changes key parameters of the data (such as adding noise, transforming scales, or applying cryptographic transformations) that preserve statistical and analytical properties while removing identifying information. This allows third parties to perform meaningful analysis on obscured data without accessing sensitive details.
3Measurement precision
If complete data access is granted for analysis, then analytical accuracy is improved, but risk of data copying and distribution increases
Solution Approach 1:
The data is pre-obscured before being made available to third parties, creating a protective barrier that prevents copying and distribution of sensitive information. This preliminary protective action maintains analytical utility while preemptively blocking harmful data exfiltration.
Data Source
AI summary
The present application provides a computer system which allows a user to make available a dataset for analysis by others whilst hiding the contents of the dataset.