Mapping abstract queries to big data web resources for on-the-fly data integration and information retrieval

The emergence of technologies such as XML, web services and cloud computing have helped, the proliferation of databases and their diversity pose serious barriers to meaningful information extraction from these “big databases”. Research in intention recognition has also progressed substantially, yet very little has been done to recognize query intents to search, select, map and extract responses from such enormous pools of candidate databases. Query mapping becomes truly complicated particularly in scientific databases where tools and functions are needed to interpret the database contents, semantics of which are usually hidden inside the functions. In this paper, we present a declarative meta-language, called BioVis, using which biologists potentially are able to express their “intentional queries” with the expectation that a mapping function μ is able to accurately understand the meaning of the queries and map them to the underlying resources appropriately. We show that such a function is technically feasible if we can design a schema mapping function that can tailor itself according to a knowledgebase and recognize entities in schema graphs. We offer this idea as a possible research problem for the community to address.