Tabular Data Re-Anonymization for Multi-Type Leakage Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data anonymization techniques are ineffective in handling multiple types of data columns in tabular data, leading to potential data leakage, especially when using methods like generative adversarial network models.
Innovation Solution
A method involving classification of data types, followed by initial anonymization using a first model and subsequent re-anonymization with a second model when necessary, tailored to specific data types to prevent data leakage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single anonymization model is used to process all data columns, then the processing is simple and fast, but data leakage occurs for certain data types
Solution Approach 1:
The patent divides the anonymization process into multiple stages: an initial anonymization model processes all data columns first, then a second anonymization model processes only the specific data types that require additional anonymization (such as text and unique identifier columns). This segmentation allows the system to achieve thorough anonymization without applying complex processing to all data uniformly, thus resolving the contradiction between reliability and complexity.
Solution Approach 2:
The patent applies different anonymization strategies to different data types based on their specific characteristics. The second anonymization model is selectively applied to data types that are more susceptible to leakage (text columns, unique identifiers) while leaving other data types processed by the first model unchanged. This local quality approach ensures high anonymization effectiveness for vulnerable data types without unnecessarily increasing complexity for all data.
2Reliability
If multiple anonymization models are used for different data types, then data leakage is prevented, but processing time and computational resources increase
Solution Approach 1:
The patent implements a classification step before the second anonymization model to identify which data columns require re-anonymization. This preliminary classification allows the system to apply the computationally intensive second anonymization model only to necessary columns rather than all columns, significantly reducing processing time while maintaining data leakage prevention for vulnerable data types.
Solution Approach 2:
Instead of applying the second anonymization model to all data columns (excessive action), the patent applies it only to specific data types that are more prone to leakage (partial action). This selective approach prevents data leakage for critical data types while avoiding the unnecessary computational overhead of processing all columns with the more rigorous model, thus resolving the time-cost contradiction.
Data Source
AI summary
Embodiments disclosed herein relate to a method, an electronic device, and a computer program product for data anonymization. The method includes: performing classification on data by a classifier to obtain data types of the data. The method further includes: performing anonymization on the data by a first anonymization model to obtain first anonymized data. The method further includes: determining, based on the data types, using an anonymizer whether re-anonymization needs to be performed on the first anonymized data. The method further includes: performing, based on a determination that the re-anonymization needs to be performed, the re-anonymization on the first anonymized data by a second anonymization model to obtain second anonymized data. Accordingly, anonymization processing may be performed on data using different anonymization models for different types of data to obtain the final anonymized data and to ensure that no data leakage occurs.