Privacy-Preserving Table Generation via Natural-Language Intermediates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating shared table data using latent embedding are non-transparent, leading to uncertainty, data distortion, and low consistency and accuracy, especially in protecting privacy and maintaining data integrity.
Innovation Solution
Perform semantic segmentation to block cells with predetermined attributes, convert the blocked table to a readable text, and generate a synthetic table based on natural language processing to ensure privacy protection and improve data accuracy and consistency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If latent embedding methods are used to generate shared table data, then data sharing is enabled, but transparency is lost and data distortion occurs
Solution Approach 1:
The patent introduces an intermediary representation layer between the original table data and the generated synthetic data. This intermediary layer processes the table data through a series of transformations including cell blocking, text conversion, and natural language processing, creating a bridge that enables data sharing while maintaining transparency and reducing distortion.
Solution Approach 2:
The patent segments the table data processing into distinct stages: identifying and blocking sensitive cells, converting blocked tables to text format, processing through natural language models, and generating final synthetic tables. This segmentation allows each stage to be optimized independently and maintains transparency throughout the data transformation pipeline.
2Object-affected harmful factors
If cells with predetermined attributes are blocked to protect privacy, then privacy protection is improved, but data completeness is reduced
Solution Approach 1:
The patent creates synthetic copies of the original table data that mimic the statistical properties and patterns without containing actual sensitive information. The blocking mechanism identifies and masks sensitive cells, then the system generates synthetic tables that replicate the overall data distribution and relationships while excluding the blocked sensitive attributes, thus protecting privacy while maintaining data utility.
Solution Approach 2:
The patent transforms the table data through parameter changes at multiple levels: blocking specific cell values, converting table structure to text representation, transforming through natural language processing parameters, and regenerating with modified parameters. This multi-stage parameter transformation enables privacy protection while preserving essential data characteristics for analysis.
3Object-affected harmful factors
If semantic segmentation is performed to block cells, then privacy protection is enhanced, but processing complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-defining the blocking strategy and cell identification criteria before processing begins. The system pre-processes the table data to identify and block sensitive cells, then converts the blocked table to text format. These preliminary actions simplify subsequent processing steps and make the complex privacy protection process more manageable and automated.
Data Source
AI summary
A method in an illustrative embodiment includes blocking cells with predetermined attributes in a first table by performing semantic segmentation on the first table, to obtain a blocked first table. The method further includes converting the blocked first table to a first text. The method further includes generating a second text in a form of natural language based on an understanding of the first text. The method further includes generating a second table corresponding to the first table based on the second text. In this way, the cells containing sensitive information can be identified and blocked, thereby retaining other useful information of the table while protecting privacy. Moreover, using the text in a form of natural language as an intermediate processing result can increase the transparency of table generation, lower the risk of data distortion, and improve the accuracy and consistency of synthetic data while protecting data privacy.