Seeded Data Perturbation for Repeatable Privacy-Preserving Services
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data services face challenges in providing sensitive information while preserving privacy and ensuring repeatability, as small data sets can lead to inference of personal data and rule-based systems risk exposing sensitive information, especially over small groups.
Innovation Solution
A data service system that generates a seed for each request using a hash function, which seeds a pseudorandom number generator to produce a consistent perturbation, ensuring that repeated requests yield the same perturbed result without storing state information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If data is anonymized before being provided, then privacy is protected, but personally identifying data can still be inferred from small changing data sets
Solution Approach 1:
The patent applies preliminary action by adding perturbation to data results before they are returned to users. The system modifies the raw query results through controlled perturbation operations, ensuring privacy protection is built into the data delivery process itself rather than relying solely on anonymization. This preliminary modification prevents inference attacks while maintaining data utility.
Solution Approach 2:
The patent introduces perturbation as an intermediary between the raw data and the user. Instead of directly providing anonymized data that remains vulnerable to inference, the system uses perturbation as a mediating layer that transforms the data into a form that protects individual identities while preserving aggregate insights. This intermediary transformation makes it computationally infeasible to reverse-engineer individual records.
2Object-affected harmful factors
If a perturbation is applied to data results, then privacy is protected, but repeated requests may yield different results
Solution Approach 1:
The patent applies dynamics by making the perturbation behavior adaptive rather than static. The system adjusts the perturbation magnitude and characteristics based on factors such as query type, data sensitivity, and request patterns. This dynamic approach allows the system to maintain consistency for legitimate repeated queries while providing sufficient variability to protect against inference attacks.
Solution Approach 2:
The patent uses parameter changes by modifying perturbation parameters such as noise magnitude, perturbation type, and application frequency based on system state and query characteristics. By changing these parameters dynamically, the system achieves a balance between result consistency for legitimate users and sufficient variability to prevent inference, resolving the contradiction between stability and privacy protection.
3Stability of the object's composition
If state information is stored to ensure repeatability, then consistent results are achieved, but storage needs and computing resources increase
Solution Approach 1:
The patent applies self-service by enabling the system to generate consistent perturbation results without relying on external state storage. The deterministic nature of the perturbation algorithm allows it to reproduce the same results for identical inputs without needing to remember previous states. This self-sufficient approach maintains repeatability while eliminating the need for additional storage infrastructure.
Solution Approach 2:
The patent uses cheap short-living objects by replacing persistent state storage with transient computational state. Instead of storing perturbation parameters in databases or memory, the system computes them on-demand using lightweight algorithms that require minimal computational resources. This disposable approach to state management reduces both storage requirements and computing overhead while maintaining result consistency.
Data Source
AI summary
A processor may receive a request to perform an operation. The processor may generate a seed derived from data required to perform the operation. The processor may generate a perturbation based on inputting the seed into a pseudorandom number generator. The processor may generate the actual result based on performing the operation. The processor may generate a perturbed result, wherein generating the perturbed result may comprise performing a second operation based on the actual result and the perturbation. The processor may return the perturbed result in response to the request.