Automatic Question Pattern Generation for Ontology-based Question Answering

This paper presents an automatic question pattern generation method for ontology-based question answering with the use of textual entailment. In this method, a set of question patterns, called predictive questions, which are predicted to be asked by users in a domain, were generated on the basis of a domain ontology. Their corresponding query templates, which can be used to extract answers to the predictive questions from a knowledge base, were generated as well. The process of producing these question patterns and query templates is described and discussed in the context of the “Movies & Cinemas” domain. An evaluation was carried out to assess the quality of the generated question patterns with the help of a textual entailment engine.