Automated Document Redaction with Contextual Conflict Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional software applications for document redaction lack contextual capabilities, requiring manual review to ensure accuracy, as they cannot determine when words should be redacted or not based on phrases or other contextual structures, and do not allow users to specify content to be kept in the document.
Innovation Solution
An automated redaction process using pre-configured lists of content to be redacted and content to be kept, with conflict resolution processing for words or phrases appearing in both lists, allowing users to specify proximity data for matching and selecting content for redaction, thereby reducing the need for manual updates and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If conventional redaction tools are used to redact sensitive information, then the redaction process can be automated to some extent, but manual review is still necessary to ensure accuracy
Solution Approach 1:
The system presents a list of identified redaction candidates to the user and incorporates user feedback by allowing selection or rejection of candidates. This feedback loop enables the system to learn from user decisions and improve future automated redaction accuracy, resolving the contradiction between automation extent and reliability.
Solution Approach 2:
The system performs preliminary identification of potential redaction candidates using multiple data sources (document text, metadata, embedded objects) before presenting them to the user for confirmation. This preliminary filtering reduces the manual review workload while maintaining accuracy, addressing both automation and reliability concerns.
2Reliability
If manual selection of content for redaction is required, then redaction accuracy can be maintained, but the time and effort required for document processing increases significantly
Solution Approach 1:
The system performs partial automation by identifying and presenting only the most likely redaction candidates to the user, rather than requiring manual review of every possible candidate. This approach maintains accuracy for critical items while reducing overall processing time through selective automation.
Solution Approach 2:
The system universally searches multiple data sources (text, metadata, embedded objects) simultaneously to identify redaction candidates, consolidating what would otherwise require multiple separate manual processes into a single automated operation, thereby reducing time loss while maintaining comprehensive accuracy.
3Stability of the object's composition
If conventional redaction methods are used, then the original document structure is preserved, but the user cannot specify content that should be kept in the document
Solution Approach 1:
The system segments the redaction decision-making process into distinct categories (must-redact, must-keep, optional) and presents them separately to the user. This segmentation allows users to specify content retention requirements while maintaining document structure, resolving the contradiction between stability and adaptability.
Solution Approach 2:
The system dynamically adjusts the redaction process based on user specifications, allowing users to define content that must be kept while preserving the original document structure. The system adapts its candidate selection and presentation based on these user-defined constraints, achieving both structural stability and content flexibility.
Data Source
AI summary
A redaction process obtains redaction data indicating content to be redacted in a document. In addition, the redaction process obtains non-redaction data indicating content not to be redacted in the document. Furthermore, the redaction process obtains proximity data indicating proximate expressions to be matched against the document. In this manner, the redaction process processes the redaction data, non-redaction data and proximity data against the document to produce a redact list and a potential list. Upon receiving user selections from both the redact and the potential lists, the redaction process applies a redaction function to the document to produce a redacted version of the document.