Reader-Specific Text Abstraction for Confidential Information Disclosure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems that use machine learning models to output answers based on document data risk leaking confidential information if the output text is viewed by unauthorized readers, while masking confidential information completely hinders effective comprehension.
Innovation Solution
An information processing method that abstracts confidential information in text based on reader attributes using ontology information to define hierarchical relationships and disclosure ranges, replacing sensitive concepts with higher-level concepts accessible to the reader.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If confidential information is completely masked in the output text, then the risk of confidential information leakage is reduced, but the reader cannot effectively comprehend the information in the output text
Solution Approach 1:
The patent applies local quality by differentiating the masking level based on the reader's attribute. Instead of uniformly masking all confidential information, the system selectively abstracts concepts according to the reader's authorization level. For example, a reader with higher clearance may see more specific concepts while readers with lower clearance see more abstracted versions, thus maintaining both security and comprehensibility for each reader type.
Solution Approach 2:
The patent implements dynamics by making the masking level adjustable based on reader attributes. The abstraction level is not fixed but dynamically determined by comparing the reader's attribute with the concept's authorization level. This allows the system to adapt the degree of information disclosure to match the reader's access rights, resolving the contradiction between protection and comprehensibility.
2Loss of information
If all confidential information is disclosed in the output text, then the reader can fully comprehend the information, but there is a risk of confidential information leakage to unauthorized readers
Solution Approach 1:
The system applies local quality by assigning different authorization levels to different concepts in the ontology. Each concept is evaluated individually against the reader's attribute, allowing precise control over what information is disclosed. This granular approach enables full comprehensibility for authorized readers while maintaining protection for unauthorized readers.
Solution Approach 2:
The patent uses feedback by comparing the reader's attribute with the authorization level of each concept. This comparison mechanism provides feedback that determines whether a concept should be disclosed or abstracted. The system continuously evaluates each concept against the reader's clearance level, ensuring appropriate disclosure decisions are made based on real-time attribute matching.
3Device complexity
If a fixed masking level is applied to all readers, then the system is simple to implement, but it cannot appropriately disclose information to different readers with different access levels
Solution Approach 1:
The patent applies universality by creating a standardized ontology structure with authorization levels that can be applied across different readers and concepts. The same masking mechanism serves multiple functions: it protects confidential information, adapts to different reader attributes, and maintains comprehensibility. This universal framework reduces overall system complexity while enabling versatile reader-specific disclosure.
Solution Approach 2:
The system implements dynamics by making the masking level adaptable to each reader's attribute without requiring a completely different system for each reader type. The dynamic comparison between reader attribute and concept authorization level allows the same system to serve multiple reader categories with different access rights, reducing complexity while maintaining adaptability.
Data Source
AI summary
The invention provides a technique capable of effectively reducing leakage of confidential information included in text and appropriately disclosing information to a reader. A confidential-information masking part acquires to-be-processed text that includes an object concept to be concealed. The confidential-information masking part acquires a reader attribute that indicates the attribute of a reader of the to-be-processed text. The confidential-information masking part abstracts, according to the reader attribute, the object concept included in the to-be-processed text by using a conceptual information tree that defines a hierarchical relationship of concepts.