Ontology Graph Query Mapping for Accurate Natural Language Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Natural language queries cannot be directly parsed by computers, and converting them to structured query language statements often results in incorrect or irrelevant database responses due to the loss of implicit ontological relationships.
Innovation Solution
A multi-layer graph data structure is used to parse natural language queries, linking terms to nodes and edges with associated tags defining ontological relationships, and converting these into structured query language statements through a set of paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If natural language queries are directly parsed by computer systems, then ease of operation is improved, but measurement precision deteriorates because the queries cannot be properly understood
Solution Approach 1:
The patent introduces an intermediary system that includes a parser, entity resolution model, and query generator that acts as a mediator between natural language queries and database queries. The parser extracts entities and relationships from natural language, the entity resolution model resolves ambiguities by matching entities to database schemas, and the query generator converts resolved entities into proper database query syntax, thereby maintaining both ease of operation and measurement precision
2Measurement precision
If structured query language statements are manually formed, then measurement precision is improved, but ease of operation deteriorates because users are not familiar with query language syntax
Solution Approach 1:
The patent inverts the traditional approach by instead of requiring users to learn structured query language syntax, the system automatically generates proper database queries from natural language. The inversion transforms the problem from 'user learns syntax to formulate queries' to 'system learns syntax to translate natural language', thereby improving ease of operation while maintaining measurement precision through automated query generation
3Device complexity
If natural language queries are converted to structured queries without ontological knowledge, then device complexity is reduced, but reliability deteriorates because implicit relationships are lost
Solution Approach 1:
The patent applies preliminary action by pre-processing the natural language query to extract entities, relationships, and semantic information before converting to database query syntax. The parser identifies entities and their relationships in advance, the entity resolution model pre-resolves ambiguities by matching to database schemas, and only then is the final query generated, ensuring reliability while managing complexity through staged processing
Data Source
AI summary
A method including parsing a natural language query to generate terms. The method also includes linking the terms to entities of a graph data structure including a first layer of nodes connected by edges. The entities are selected from among the nodes and the edges. The graph data structure further includes a meta layer which has tags associated with the edges and the nodes. The tags define an ontology for the entities. A term in the terms is linked to an entity in the entities when the term matches the entity. The method also includes generating a set of paths between selected tags in the meta layer. Each of the selected tags is associated with a corresponding edge in the graph data structure that matches a corresponding term extracted from the natural language query. The method also includes converting the set of paths into a structured query language statement.