Toward a Natural-Language Interface in Ontologies for Disambiguation

Alejandro Solís Sánchez, Pablo Barraza Cornejo, Juan Carlos Acosta Guadarrama, Juan Ferret
2020 Research on computing science  
With natural-language-interfaces-to-ontology (NLI2O) one can query a specific domain's data, stored into that particular conceptual structure, coded into specialized languages such as Ontology Web Language, OWL. To access such data, one requires complex query languages, such as SPARQL. NLI2Os translate human-language queries into SPARQL ones. In this work, we describe the architecture of an NLI2O. It receives a spoken question from an Android device. The spoken query becomes text and goes to a
more » ... ava Server Page (JSP). The core of our interface runs on a JSP server, which receives the user's text query. It transforms it into a SPARQL query, later used to access, from the ontology to query, the information requested by the user. The interface uses several technologies to achieve that. Those are known as lemmatization, part-of-speech tagging, and tokenization. It also uses knowledge inference from the ontology of interest. It also uses clarification dialogues for disambiguation to answer user questions and revise the query. Besides that, some experiments suggest further studies like the disambiguation of an original text.
dblp:journals/rcs/SanchezCGF20 fatcat:h5bibf2psrak5gfuynda27m6ge