Question Answering Query Translation for Relational Questions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional question-answering systems struggle to effectively answer relational questions due to the difficulty in combining information from unstructured documents and the limitations of Boolean operators in handling complex queries.
Innovation Solution
Integrating a database management system within the question-answering system, transforming natural-language queries into structured queries executable by a database, and leveraging its capabilities to handle relational questions, while also utilizing a knowledge base for complementary answers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a text-based knowledge base is used to answer questions, then simple lookup questions can be answered easily, but relational questions requiring information combination become difficult to answer
Solution Approach 1:
The patent combines a text-based knowledge base with a database management system to create a hybrid architecture. The knowledge base handles unstructured text retrieval while the database component manages structured data and relational operations, allowing the system to answer both simple lookup questions and complex relational questions that require information combination across multiple sources.
2Ease of operation
If Boolean operators are used to control search, then the answer space can be limited, but complex operations like obtaining averages or identifying largest numbers become difficult
Solution Approach 1:
The patent introduces a database management system as an intermediary between the user's natural language query and the information retrieval process. This intermediary translates queries into database operations that can perform complex computational functions such as calculating averages, finding maximum values, and other mathematical operations that Boolean operators cannot handle, while still maintaining the ability to limit and control the answer space.
3Adaptability or versatility
If information is stored in unstructured documents, then flexibility in document format is maintained, but ease of locating and combining information decreases
Solution Approach 1:
The patent segments information storage into two distinct components: unstructured text documents stored in the knowledge base and structured data stored in the database management system. This segmentation allows each component to optimize for its specific strength - the knowledge base maintains document format flexibility while the database provides efficient structured data retrieval and combination, enabling the system to leverage both advantages simultaneously.
Data Source
AI summary
A question-answering system that receive a natural-language question includes a database to provide a basis for that answer and a structured-query generator that constructs a structured query from the question and uses it to obtain an answer to the question from the database.