Question Answering System Supplementing Answers via Knowledge Graph
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Question Answering (QA) systems fail to provide links between related subjects, returning accurate answers without information on connections between them, unless a specific question about the link is asked.
Innovation Solution
A QA system that indexes and stores previous question-answer sets in a database with a knowledge graph, identifies related concepts, and supplements candidate answers with content from previous question-answer sets, restructuring them to include the question as a statement for natural language context.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a QA system returns accurate answers from structured sources, then answer precision is improved, but information completeness deteriorates because links between related subjects are not provided
Solution Approach 1:
The patent introduces an intermediary component that analyzes the QA system's answer and identifies related unstructured sources. This intermediary acts as a bridge between the precise but isolated answer and the broader context that would provide information completeness, allowing the system to maintain precision while reducing information loss by selectively incorporating relevant external information.
Solution Approach 2:
The patent segments the information retrieval process into distinct components: the core QA system that provides precise answers, and an additional module that separately identifies and retrieves related unstructured sources. This segmentation allows each component to optimize for its specific function while collectively achieving both answer precision and information completeness.
2Loss of information
If a QA system searches through large sets of sources to analyze content, then information coverage is improved, but processing time deteriorates
Solution Approach 1:
The patent applies partial action by having the QA system first process structured sources to obtain a precise answer, then selectively applying additional processing only to identify and retrieve relevant unstructured sources related to that answer. This partial approach to searching unstructured content reduces overall processing time while still achieving improved information coverage through targeted retrieval.
Solution Approach 2:
The system performs preliminary action by first obtaining the core answer from structured sources before initiating the search for related unstructured content. This preliminary step allows the system to have a focal point for the subsequent search, reducing the time required to process large sets of unstructured sources by limiting the search to content specifically related to the already-obtained answer.
3Ease of operation
If a QA system provides only direct answers without contextual links, then response simplicity is improved, but user understanding deteriorates
Solution Approach 1:
The patent applies local quality by maintaining the simplicity of direct answers for the core response while selectively adding contextual links and related unstructured information in specific locations or formats that enhance user understanding. This allows the system to preserve response simplicity where needed while providing additional context where it aids comprehension, thereby reducing the loss of information without sacrificing overall simplicity.
Data Source
AI summary
Candidate answers are generated by a question-answering system in response to a question from a user. One or more generated candidate answers are compared to previous question-answer sets. The previous question-answer sets are indexed and stored in a database which includes a knowledge graph on the previous questions-answer sets. A previous question-answer set is identified as correlating with a first generated candidate answers. The previous question-answer set is identified using relationships of the database. The first generated candidate answer is restructured using the question as a statement which is embedded with the first generated candidate answer. The restructured generated candidate answer which correlates with the previous question-answer set is supplemented with content from the previous question-answer set.