AI Document Relation Extraction for Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document relation determination methods fail to accurately identify relationships between document elements, even when their contents are not similar, leading to inefficiencies in determining the necessity of updates or changes in documents.
Innovation Solution
An information processing apparatus using AI machine learning to extract relation descriptions from document elements and generate relation information, enabling the identification of relationships between document elements based on extracted descriptions, regardless of content similarity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If document relation determination is based on content similarity between document elements, then the determination process is simple and straightforward, but it fails to accurately identify relationships when contents are not similar
Solution Approach 1:
The patent introduces an intermediary description (e.g., 'refers to', 'based on') that mediates the relationship between document elements. Instead of directly comparing content similarity, the system extracts these descriptive phrases that explicitly indicate relationships, allowing accurate relation determination even when contents differ significantly. This intermediary approach resolves the contradiction by providing a new pathway for relation detection that doesn't rely on content similarity.
2Reliability
If traditional relation extraction methods are used, then the system operates with simple processing, but it cannot identify relationships between document elements with low content similarity
Solution Approach 1:
The patent extracts specific relation-describing phrases from document elements rather than performing comprehensive content comparison. By taking out only the critical relational information (e.g., 'refers to', 'based on') from the full content, the system achieves reliable relation identification without the computational burden of comparing entire document contents, thus maintaining processing efficiency while improving reliability.
3Measurement precision
If content similarity comparison is performed for all document elements, then comprehensive analysis is achieved, but the processing time and computational resources increase significantly
Solution Approach 1:
The patent extracts only the essential relational phrases from document elements instead of performing comprehensive content similarity comparison. This extraction approach detects relationships quickly by focusing on key indicator phrases like 'refers to' or 'based on', significantly reducing processing time while maintaining accurate relation detection capability.
Data Source
AI summary
An information processing apparatus includes a processor configured to extract a description including a phrase indicating a relation with a second document element from a first document element, and generate relation information corresponding to information on the description extracted from the first document element, by an AI which has learned, in advance, by machine learning to generate the relation information indicating a relation between the first document element and the second document element from the information on the description.