PowerAqua: Open Question Answering on the Semantic Web

With the rapid growth of semantic information in the Web, the processes of searching and querying these very large amounts of heterogeneous content have become increasingly challenging. This research tackles the problem of supporting users in querying and exploring information across multiple and heterogeneous Semantic Web (SW) sources. A review of literature on ontology-based Question Answering reveals the limitations of existing technology. Our approach is based on providing a natural language Question Answering interface for the SW, PowerAqua. The realization of PowerAqua represents a considerable advance with respect to other systems, which restrict their scope to an ontology-specific or homogeneous fraction of the publicly available SW content. To our knowledge, PowerAqua is the only system that is able to take advantage of the semantic data available on the Web to interpret and answer user queries posed in natural language. In particular, PowerAqua is uniquely able to answer queries by combining and aggregating information, which can be distributed across heterogeneous semantic resources. Here, we provide a complete overview of our work on PowerAqua, including: the research challenges it addresses; its architecture; the techniques we have realised to map queries to semantic data, to integrate partial answers drawn from different semantic resources and to rank alternative answers; and the evaluation studies we have performed, to assess the performance of PowerAqua. We believe our experiences can be extrapolated to a variety of end-user applications that wish to open up to large scale and heterogeneous structured datasets, to be able to exploit effectively what possibly is the greatest wealth of data in the history of Artificial Intelligence