The Watson Logo

FABilT - Finding Answers in a Billion Triples

This page provides a brief description of our submission to the Billion Triple Challenge 2008. This submission does not present one system, but two, the first one being the underlying infrastructure component enabling the second one. Watson is a Gateway to the Semantic Web: it crawls and indexes semantic data online to provide a variety of access mechanisms for human users and applications. We show here how we indexed most of the datasets provided for the challenge, obtaining an infrastructure (web services, API, web interface, etc.) for supporting the exploration of these datasets. PowerAqua is an open domain question answering system which allows users to pose natural language queries to the Semantic Web. PowerAqua is one of the many applications that are made possible by Watson, as it relies on its ability to crawl and index Semantic Web sources. We call the system composed of Watson and PowerAqua using the billion triples FABilT.

Watson on a Billion Triples

Interface: The standard Watson web interface has been reproduced to create a billion triple specific version, available at the following address: http://kmi-web06.open.ac.uk:8081/WatsonWUI/

API: As for the interface, the same Web services used for the standard version of Watson have been deployed for the billion triple version. The jar file containing the Watson API using the billion triple version of the services can be downloaded from (http://kmi-web06.open.ac.uk:8081/watson-ws/watson-client-api-btc.jar).

PowerAqua using the Billion Triple Watson

A prototype version of PowerAqua relying on the billion triple version of Watson (i.e. FABilT) will be made available from the address: http://kmi.open.ac.uk/technologies/poweraqua/demo.html.