Automating Text Definitions from Ontologies
Author Information
Author(s): Robert Stevens, James Malone, Sandra Williams, Richard Power, Allan Third
Primary Institution: The University of Manchester
Hypothesis
Can natural language generation (NLG) be used to automatically create text definitions from logical descriptions in ontologies?
Conclusion
The generated definitions from the ontology were found acceptable and have been incorporated into the Experimental Factor Ontology (EFO).
Supporting Evidence
- The generated definitions were found to convey the intended meaning of the axiomatisation satisfactorily.
- Two rounds of surveys indicated improvements in the generated text definitions.
- The developers of EFO incorporated the generated definitions into their ontology.
Takeaway
This study created a tool that can turn complex scientific definitions into simple English, making it easier for people to understand.
Methodology
The study used a natural language generation tool to create definitions from the Experimental Factor Ontology and evaluated them through surveys.
Limitations
The tool's coverage is limited to a subset of OWL 2, and the generated text may not always follow domain conventions.
Participant Demographics
Participants were from the OWL ontology community and included a range of expertise.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website