Question Answering over Knowledge Bases

The fast growth of the Semantic Web has unleashed its potentialities, leading to the development of many tools and services that can exploit the huge amount of information it contains. As more semantic information is available online, mainly in the form of ontology based Knowledge Bases the process of searching and querying this content has become more and more challenging. Question Answering, which defines the task of retrieving an answer to a question formulated using natural language, can make the Semantic Web easily accessible by anyone, even by those users that do not know how to use a specific data query language or are unaware of the structure of the KB they want to query. Moreover, in the same way the Semantic Web can benefit in its spread from Question Answering, also question Answering systems can improve their outcome since Knowledge Bases can be exploited to retrieve a concise answer for complex questions. Although several approaches have been proposed by the research community in the field, the experimental results show that the performance are still far from optimal. Following the future directions presented in the latest works about this field, we outline an approach for Question Answering over structured data applicable to Knowledge Bases whose aim is to overcome the main issues that affect the research in this area.