SWSE: Answers Before Links! Andreas Harth, Aidan Hogan, Renaud Delbru, J¨ urgen Umbrich, Sean O’Riain, and Stefan Decker National University of Ireland, Galway Digital Enterprise Research Institute firstname.lastname@deri.org Abstract. We present a system that improves on current document- centric Web search engine technology; adopting an entity-centric per- spective, we are able to integrate data from both static and live sources into a coherent, interlinked information space. Users can then search and navigate the integrated information space through relationships, both existing and newly materialised, for improved knowledge discovery and understanding. 1 Introduction Todays best in class search engines such as Google, Yahoo and MSN offer search over web documents. Paramount to their success is their data gather- ing prowess and their ability to allow consumers quickly and efficiently find documents matching a set of keywords. Results in such sites are simply links to documents; users are required to manually traverse these documents in order to achieve the answers to their information needs. Ranking tries to present relevant documents first, but little effort is made to enhance, extract or integrate data to provide precise answers, such as “who are the friends of Rudi Studer”, “what organism pertains to the Protein SHNF1”, or “where is SK Telekom located”? In addition, most current search engines offer little support for disambigua- tion or refinement of results: adroit keyword query construction and reformula- tion is required to avoid trawling through a quagmire of irrelevant results. More recent developments in the search space have seen Vivisimio 1 offering search and clustering capabilities, where search is initially driven by keyword but also fea- tures filtering using generated clusters which provides the ’context’ of the results. Quintura 2 also offers dynamic cluster creation based on keyword queries, with analysis of the contextual relationships between keywords. The result is a visual semantic keyword cloud that the user may call upon to further filter the result set. However, all of these tools are still based upon the traditional document- centric view of knowledge rather than real-world entities and their relationships. On the other hand, search engines operating on structured data sources, such as 1 http://vivisimo.com/ 2 http://quintura.com/