Data Sample Anonymization for Secure Classification Review
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Discovering and classifying target data across multiple data assets in computing systems is computationally resource-intensive and exposes data to significant risks of data breaches and loss, particularly when mapping and transferring data between systems.
Innovation Solution
A computer-implemented process that intermingles real data samples with supplemental anonymizing data samples, using machine-learning classification to label and anonymize data elements, minimizing the risk of data exposure by generating fictitious occurrences similar to real data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data is transferred between computing systems for discovery and classification, then data mapping accuracy is improved, but data security and risk of data breaches worsen
Solution Approach 1:
The patent introduces an intermediary system that sits between the source computing system and the target computing system. This intermediary receives data from the source system, performs classification and mapping operations, and then transmits only the classified results to the target system. This mediator architecture allows accurate data mapping to be achieved while preventing direct exposure of sensitive data between systems, thereby reducing data breach risks.
Solution Approach 2:
The patent creates copies of data for classification purposes rather than transferring the original sensitive data. The system generates classified representations or metadata about the data that can be shared between systems without exposing the actual sensitive information. This copying approach maintains mapping accuracy while protecting data security.
2Loss of information
If data is transferred between computing systems for classification, then classification completeness is improved, but data loss risk worsens
Solution Approach 1:
The patent performs classification operations in advance before data is transferred or exposed to multiple systems. By pre-classifying data in the source system and transmitting only the classification results or metadata, the system achieves complete classification coverage while minimizing the risk of data loss during transfer. The preliminary classification action ensures all necessary information is captured before any potential data exposure occurs.
3Adaptability or versatility
If multiple computing systems are used to store and process data, then data discovery capability is improved, but computing resource requirements worsen
Solution Approach 1:
The patent implements a universal classification framework and standardized data schemas that can be applied across multiple computing systems. By establishing common classification standards, metadata formats, and processing protocols, the system enables data discovery capabilities across distributed systems without requiring each system to independently perform redundant classification operations. This multi-functional approach reduces overall computing resource requirements while maintaining broad data discovery capability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In general, various aspects of the present invention provide methods, apparatuses, systems, computing devices, computing entities, and/or the like for mapping the existence of target data within computing systems in a manner that does not expose the target data to potential data-related incidents. In accordance with various aspects, a method is provided that comprises: receiving a source dataset that comprises a label assigned to a data element used by a data source in handling target data that identifies a type of target data and data samples gathered for the data element; determining, based on the label, that the data samples are to be anonymized; generating supplemental anonymizing data samples associated with the label that comprise fictitious occurrences of the type of the target data; generating a review dataset comprising the supplemental anonymizing data samples intermingled with the data samples; and sending the review dataset to a review computing system.