USING GRAMMATICAL INFERENCE TECHNIQUES TO LEARN ONTOLOGIES THAT DESCRIBE THE STRUCTURE OF DOMAIN INSTANCES
See where this sits in the topic map →Summary AI-generated
- TL;DR
- This research introduces a semi-automatic approach to learn ontologies that describe the implicit structure of human-generated information, enabling better computer understanding and search capabilities.
- Problem
- Information created by people usually follows an implicit, agreed-upon structure that is hidden from computer programs, limiting their ability to answer precise search queries.
- Method
- The authors propose a two-step semi-automatic method: first, inferring a grammatical description of the implicit structure from examples using grammar inference techniques; second, transforming that description into an ontology.
- Results
- The approach successfully yields a semi-automatic method for inferring ontologies that encode structural information.
- Contributions
- Not specified in the abstract.
- Limitations
- Not specified in the abstract.
- Takeaways
- Grammatical inference techniques can effectively bridge the gap between human-generated content structure and computer-readable ontologies, as demonstrated in the structuring of technical articles.
- Applications
- Improving search queries on the semantic web by targeting specific sections of documents, such as references in technical articles.
- Topics
- Ontology Learning; Grammatical Inference; Semantic Web; Information Structure
- For industry
- Not specified in the abstract.
- Why it matters
- Not specified in the abstract.
Abstract
Information produced by people usually has an implicit agreed-upon structure. However, this structure is not usually available to computer programs, where it could be used, for example, to aid in answering search queries. For example, when considering technical articles, one could ask for the occurrence of a keyword in a particular part of the article, such as the reference section. This implicit structure could be used, in the form of an ontology, to further the efforts of improving search in the semantic web. We propose a method to build ontologies encoding this structure information by the application of grammar inference techniques. This results in a semi-automatic approach to the inference of such ontologies. Our approach has two main components: (1) the inference of a grammatical description of the implicit structure of the supplied examples, and (2) the transformation of that description into an ontology. We present the application of the method to the inference of an ontology describing the structure of technical articles.