Type-Specific Generative Models for Sensitive Data Desensitization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Privacy Preserving Data Publishing (PPDP) algorithms are ineffective in processing unstructured and semi-structured data, leaving sensitive information vulnerable to leakage.
Innovation Solution
A method utilizing distinct generative models for tabular, image, and text types to detect and desensitize sensitive data, including generative adversarial networks for tabular and image data, and neural networks for text data, to convert and generate synthetic data without sensitive information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single PPDP algorithm is used for data processing, then the processing method is simple, but it is ineffective for unstructured and semi-structured data, leaving sensitive information vulnerable
Solution Approach 1:
The patent segments the data processing system into multiple specialized components: a data type identification module that categorizes input data, and multiple specialized generative models (GAN for tabular data, VAE for semi-structured data, transformer for text data) that process different data types. This segmentation allows each component to be optimized for its specific function, improving overall reliability while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The patent creates a universal data processing framework that can handle multiple data types (tabular, semi-structured, text) through a single system. The framework uses a unified architecture with data type identification and conditional model selection, allowing one system to perform multiple desensitization functions across different data formats, thereby improving reliability without requiring separate systems for each data type.
2Measurement precision
If multiple specialized generative models are used for different data types, then desensitization accuracy improves, but the system complexity increases
Solution Approach 1:
The patent introduces a data type identification module as an intermediary between the input data and the multiple generative models. This intermediary automatically categorizes the input data type and routes it to the appropriate specialized model (GAN, VAE, or transformer), thereby achieving high desensitization accuracy through precise model matching while keeping the system complexity manageable through automated routing logic.
Solution Approach 2:
The patent performs preliminary data type identification and classification before selecting and applying the appropriate generative model. This preliminary action ensures that the correct specialized model is chosen based on the data type, improving desensitization accuracy while organizing the complexity into a structured, predictable sequence of operations that simplifies system management.
3Productivity
If direct disclosure of data is performed, then data accessibility is high, but sensitive information leakage occurs
Solution Approach 1:
The patent extracts and removes sensitive information from the data through specialized generative models that generate synthetic data preserving statistical properties while eliminating sensitive content. The models identify and extract sensitive features, replacing them with synthetic alternatives, thereby maintaining data utility and processing efficiency while preventing sensitive information leakage through mathematical transformation rather than simple deletion.
Solution Approach 2:
The patent creates synthetic copies of the original data that preserve its statistical properties, structure, and utility while containing no actual sensitive information. The generative models generate realistic synthetic data that can be used for analysis and processing at full speed, eliminating the need to handle or expose sensitive original data, thus maintaining productivity while preventing information leakage.
Data Source
AI summary
Embodiments of the present disclosure involve a method, an electronic device, and a computer program product for processing data. The method includes detecting a type of the data, wherein the data includes sensitive data. The method further includes selecting, according to the detected type, a generative model corresponding to the detected type from a plurality of generative models. The method further includes processing the data by using the selected generative model to desensitize the sensitive data from the data. The method for processing data according to embodiments of the present disclosure processes the data by combining various types of generative models, and the application potentials of various generative models can be fully exploited, so that desensitization processing of a plurality of types of data can be realized more quickly and accurately, and leakage of sensitive information can be avoided, thereby ensuring the security of various types of data.