Obfuscation Table N-gram Mapping for Privacy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data protection methods fail to effectively obfuscate sensitive information while maintaining the statistical properties necessary for research and analysis, particularly in handling large volumes of personal and private data across various domains like healthcare and finance.
Innovation Solution
A data obfuscation system using a computer system with processors and memories, which retrieves a data stream of n-grams, maps them to tokens using an obfuscation table, and generates a second data stream for analysis, allowing for secure research and analysis without exposing confidential data, employing methods like one-to-one mapping and encryption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional data protection methods are used to obfuscate sensitive information, then privacy protection is improved, but statistical properties necessary for research and analysis are lost
Solution Approach 1:
The patent creates a synthetic copy of the original data that preserves statistical properties while eliminating identifiable information. The system generates artificial data records that replicate the statistical characteristics (means, variances, distributions) of the original dataset without containing actual sensitive values, thereby enabling research analysis while protecting privacy.
Solution Approach 2:
The patent transforms data by changing parameters from identifiable formats to synthetic representations. It applies statistical transformations that modify the parameter structure of the data - converting specific sensitive values into synthetic equivalents that maintain statistical relationships and distributions while removing direct identifiability.
2Reliability
If data is obfuscated to protect privacy, then confidentiality is improved, but the ability to perform research and analysis is reduced
Solution Approach 1:
The system generates a synthetic copy of the dataset that can be freely used for research and analysis without compromising confidentiality. This synthetic replica maintains the statistical structure, relationships, and patterns needed for meaningful analysis while containing no actual sensitive information that could be de-anonymized.
Solution Approach 2:
The patent introduces synthetic data as an intermediary between the original sensitive data and research analysts. This intermediary layer allows researchers to perform statistical analysis and pattern recognition on data that statistically mirrors the original while being completely decoupled from identifiable information, thus enabling productivity without sacrificing confidentiality.
3Reliability
If traditional encryption methods are applied to data, then security is improved, but the data cannot be used for statistical analysis
Solution Approach 1:
Instead of applying cryptographic encryption that preserves exact values, the patent applies statistical parameter transformations that change the data representation fundamentally. It transforms sensitive values into synthetic equivalents that preserve statistical parameters (distributions, correlations, moments) while making reverse-engineering to original values computationally infeasible, thus maintaining security while preserving analytical value.
Solution Approach 2:
The system creates a statistical copy of the encrypted data that retains analytical properties. Rather than encrypting the actual data, it generates synthetic data that copies the statistical structure and relationships, enabling analysis on the copy while the original sensitive data remains protected without being directly accessed or decrypted.
Data Source
AI summary
At least some aspects of the present disclosure feature systems and methods for obfuscating data. The method includes the steps of receiving or retrieve an input data stream including a sequence of n-grams, mapping at least some of the sequence of n-grams to corresponding tokens using an obfuscation table, and disposing the corresponding tokens to an output data stream.