2004 IRISH SCIENTIST YEAR BOOK

Home Page

Table of contents

Index by Author

Index by topics

Search


University of Ulster
David Patterson, Mykola Galushka, Niall Rooney & Vladimir Dobrynin
Trouble finding information on-line? Then discover it with SOPHIA!

The Northern Ireland Knowledge Engineering Laboratory (NIKEL) is a research and technology transfer organisation founded in 1992 at the University of Ulster. NIKEL's objectives are to produce high quality research in Artificial Intelligence and to help companies gain competitive advantage through the application of advanced software technology. NIKEL's core expertise centres around data mining technologies which enable the transformation of corporate operational data into useful knowledge that the organisation can then use to improve its decision making processes.

Recently, NIKEL has enhanced its expertise further in the related area of information retrieval. In conjunction with a group at the State University of St Petersburg, NIKEL has been developing a tool it believes to be the next generation in search engine technology. Its name is SOPHIA, and it is initially aimed at the Corporate Intranet market. Its design has been motivated by the fact that many users of conventional search engine technology, such as Google and Alta Vista, often become frustrated with the long lists of links returned as the result of a query. Frequently, the most useful links are buried deep in these lists and the user has to invest time to locate them. Additionally, whenever a useful link is found, there is no relationship between its relevancy and the relevancy of the links above and below it, which leaves the user little option but to begin the search again or to continue to scan the list.


Example document tree for search query

What SOPHIA offers is a new way to search for information that is visual, intuitive and interactive. The result of a query returns a versatile tree-like representation of the relationships between documents deemed relevant by the system, whereby documents linked directly are very similar in nature (contain similar content) while those at a distance are less similar. Documents are automatically labelled succinctly to enable users ascertain initial relevancy to their needs. The user can then browse through the tree by clicking on nodes and previewing the document content. Key concepts are highlighted within the document to draw the user's attention to them in order to help effectively determine document relevancy. Due to the structuring of the documents, once one document is identified as useful, the user knows that those documents within its vicinity are also useful (red nodes in diagram). This is a major benefit over conventional search systems. Additionally, NIKEL believes that known user expertise within a domain, combined with knowledge of previous search behaviour, should be used to enhance the search experience further to meet the specific needs and requirements of users.

Contact: NIKEL at http://nikel.ulster.ac.uk ;
E-mail: [email protected]