Improving Search on the Semantic Desktop using Associative Retrieval Techniques Peter Scheir (Graz University of Technology, Austria peter.scheir@tugraz.at) Chiara Ghidini (Fondazione Bruno Kessler, Italy ghidini@itc.it) Stefanie N. Lindstaedt (Know-Center Graz, Austria slind@know-center.at) Abstract: While it is agreed that semantic enrichment of resources would lead to better search results, at present the low coverage of resources on the web with semantic information presents a major hurdle in realizing the vision of search on the Semantic Web. To address this problem we investigate how to improve retrieval performance in a setting where resources are sparsely annotated with semantic information. We suggest employing techniques from associative information retrieval to find relevant material, which was not originally annotated with the concepts used in a query. We present an associative retrieval system for the Semantic Desktop and show how the use of associative retrieval increased retrieval performance. Key Words: semantic desktop, associative information retrieval Category: H.3.3, I.2.4, I.2.6, I.2.11 1 Introduction On the one hand it is largely [6] [17] agreed that semantic enrichment of resources on the web or desktop provides for more information to be used during search and that this can lead to higher effectiveness of a retrieval system. One the other hand critics [10] as well as advocates [13] of the Semantic Web agree on the low coverage of resources on the current web with semantic information. The sparse annotation of resources with semantic information presents a major obstacle in realizing search applications for the Semantic Web or the Semantic Desktop, which operate on semantically enriched resources. For this reason we propose the use of techniques from associative information retrieval to find additional relevant material, even if no semantic information is provided for those resources. We describe our approach to information retrieval on the Semantic Desk- top and present a retrieval component developed during the first year of the APOSDLE 1 project. The rest of this paper is organized as follows: in section 2 we introduce the concept of associative information retrieval, in section 3 we present statistical information about the knowledge base used in APOSDLE to 1 http://www.aposdle.org/ Proceedings of I-MEDIA ’07 and I-SEMANTICS ’07 Graz, Austria, September 5-7, 2007