AQUEOS: A System for Question Answering over Semantic Data

This paper presents a methodology to automatically answer natural language questions by querying an underlying domain ontology. This methodology is made up of a three-phase process, where the ontological data is firstly read and indexed, the input question is then processed by means of lexical analysis and associated with a specific question type, and finally the corresponding SPARQL queries are generated and executed in order to return the answer to the original question. This process focuses on single-verb phrases in order to guarantee a highest level of precision in providing its answer, and deals with critical lexical aspects like comparatives and superlatives by relying upon language-specific lexicons; nonetheless, it is also able to take into account more complex questions with multiple verbs, provided they meet certain specific criteria. Such a methodology has been implemented in a research prototype and is being currently experimented upon by asking questions either in the English or the Italian language, and could be applied on a number of ontology-driven applications, including advanced help desk support systems, biomedical knowledge bases and intelligent e-learning solutions.