Document Knowledge Graph for Entity Relationship Exploration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional document retrieval systems face challenges in identifying and presenting relevant information efficiently, often providing false hits or missing relevant documents, and fail to connect entities or explore documents effectively, requiring significant user effort and time.
Innovation Solution
A system and method that generate a user-controllable document knowledge graph, representing entities and their inter-relationships, allowing users to selectively explore specific entities or connections by sending a list of documents to a client device, retrieving relevant data from a database, and providing an interface for interactive exploration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional search systems are used to retrieve documents, then information can be accessed, but the systems provide false hits or miss relevant documents, requiring significant user effort and time to identify relevant information
Solution Approach 1:
The patent introduces an intermediary layer between the user and the document corpus: a knowledge graph that pre-structures information by extracting entities, relationships, and concepts from documents. This intermediary enables semantic search and relationship exploration, allowing users to navigate through structured knowledge rather than manually inspecting unstructured documents, thereby improving retrieval accuracy and reducing time investment
Solution Approach 2:
The system performs preliminary actions by pre-processing documents to extract entities, relationships, and concepts, and organizing them into a knowledge graph structure before user queries are submitted. This pre-structuring of information enables faster and more accurate retrieval during actual use, as the heavy lifting of information extraction and organization has already been completed
2Loss of information
If conventional systems present scattered information from multiple documents, then comprehensive information is available, but users cannot easily identify connections between entities or explore documents together
Solution Approach 1:
The patent merges information from multiple scattered documents into a unified knowledge graph structure that preserves connections between entities. By combining entity extractions, relationship extractions, and concept extractions from various documents into a single interconnected graph, the system maintains comprehensive information while making relationships visually and interactively explorable through the user interface
Solution Approach 2:
The system transforms flat, document-centric information into a multi-dimensional knowledge graph structure where entities, relationships, and concepts exist as interconnected nodes. This dimensional transformation allows users to explore information from multiple perspectives (entities, relationships, concepts) and navigate through connections that would be invisible in traditional document presentations
3Measurement precision
If users manually access and inspect documents to extract relevant information, then thorough analysis is possible, but enormous effort and time are required
Solution Approach 1:
The system performs self-service by automatically extracting entities, relationships, and concepts from documents and organizing them into a structured knowledge graph without requiring user intervention. The automated extraction and organization processes enable thorough information analysis to be performed at machine speed, dramatically improving productivity while maintaining analysis depth through the structured representation
4Loss of information
If conventional systems display documents in traditional formats, then information is presented, but users cannot interactively explore specific entities or relationships
Solution Approach 1:
The system transforms static document presentations into a dynamic knowledge graph interface where users can interactively explore entities, relationships, and concepts. The knowledge graph structure allows for dynamic querying, filtering, and navigation based on user interests, enabling versatile exploration of information while maintaining complete information representation through the underlying graph data structure
Data Source
AI summary
The present invention discloses a system comprising a server and a database storing data pertaining to a plurality of electronic documents, wherein the server is configured to: send a list including one or more electronic documents to a client device for presentation to a user, wherein the list enables the user to select one of the one or more electronic documents; upon selection of a given electronic document from the one or more electronic documents, retrieve data pertaining to the given electronic document from the database, wherein the retrieved data indicates entities and their inter-relationships; generate a document knowledge graph representing the entities and their inter-relationships as specific to the given document, wherein the document knowledge graph is user controllable; and send the document knowledge graph to the client device for presentation to the user, generate and provide an interface that includes the generated document knowledge graph, wherein the document knowledge graph enables the user to selectively look into a specific entity or interrelationship.