Database Data Anonymization with Sensitivity-Based Noise Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing differential privacy algorithms apply homogeneous noise to all sensitive data, leading to the destruction of information in data with lower amplitudes and resulting in anonymized databases that deviate significantly from the original, thus losing valuable information.
Innovation Solution
A differential privacy algorithm that applies noise levels based on the inherent sensitivity of the data, partitioning sensitive data into subsets and adjusting noise amplitude accordingly, ensuring higher noise for highly sensitive data and lower noise for less sensitive data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If homogeneous noise is applied to all sensitive data, then the anonymity level is maintained, but the information quality in the anonymized database deteriorates
Solution Approach 1:
The patent applies different noise levels to different subsets of sensitive data based on their sensitivity characteristics. Highly sensitive data receives higher noise levels to maintain anonymity, while less sensitive data receives lower noise levels to preserve information quality. This local differentiation resolves the contradiction by allowing anonymity and information quality to be optimized for each data subset independently.
Solution Approach 2:
The patent segments the set of sensitive data into multiple subsets based on sensitivity levels. By dividing the data into groups with different anonymity requirements, the system can apply tailored noise levels to each subset, thereby maintaining overall anonymity while preserving valuable information in less sensitive data regions.
2Reliability
If high noise levels are applied to ensure anonymity, then the anonymity level improves, but the deviation from original data increases
Solution Approach 1:
The patent implements local quality by assigning different noise levels to different data subsets based on their sensitivity. This allows high anonymity to be achieved for sensitive data while maintaining data accuracy for less sensitive data, resolving the contradiction between anonymity and precision.
Solution Approach 2:
The patent changes the noise level parameter dynamically based on data sensitivity characteristics. By adjusting the noise parameter locally rather than applying a uniform value, the system achieves both high anonymity where needed and high accuracy where possible.
3Ease of manufacture
If uniform noise amplitude is used across all data, then the implementation simplicity is maintained, but the data utility is reduced
Solution Approach 1:
The patent applies local quality by using different noise amplitudes for different data subsets. This improves data utility by preserving information in less sensitive areas while still maintaining anonymity in sensitive areas, overcoming the limitation of uniform noise approaches.
Solution Approach 2:
By segmenting data into subsets with different sensitivity levels, the patent enables differentiated noise application that improves overall data utility compared to uniform noise, while maintaining reasonable implementation complexity through systematic segmentation.
Data Source
Figure 1~3
Figure 4~5
AI summary
Method and device for anonymizing data stored in a database. The method according to the invention makes it possible to anonymize so-called initial data stored in a database of a computing system and resulting from an aggregation of personal data relating to a plurality of individuals. It comprises: ¾ a step (E20) of identification among the initial data of a set of sensitive data liable to be affected by the addition or the removal in the database of personal data relating to an individual; ¾ a step (E30) of partitioning of the set of sensitive data into a plurality of subsets as a function of a sensitivity level of the sensitive data; ¾ a step (E40) of determination of a sensitivity level for each subset; and ¾ a step (E50, E80) of anonymization of the initial data comprising, for each subset, a noising of the sensitive data of this subset according to a noise level dependent on the sensitivity level determined for the subset.