Ontology-Based Interpretation of Keywords for Semantic Search Thanh Tran, Philipp Cimiano, Sebastian Rudolph, and Rudi Studer Institute AIFB, Universität Karlsruhe, Germany Abstract. Current information retrieval (IR) approaches do not formally capture the explicit meaning of a keyword query but provide a comfortable way for the user to specify information needs on the basis of keywords. Ontology-based ap- proaches allow for sophisticated semantic search but impose a query syntax more di cult to handle. In this paper, we present an approach for translating keyword queries to DL conjunctive queries using background knowledge available in on- tologies. We present an implementation which shows that this interpretation of keywords can then be used for both exploration of asserted knowledge and for a semantics-based declarative query answering process. We also present an evalua- tion of our system and a discussion of the limitations of the approach with respect to our underlying assumptions which directly points to issues for future work. 1 Introduction Part of the Semantic Web vision is to provide web-scale access to semantically de- scribed content. In particular, this implies understanding users’ information needs ac- curately enough to allow for retrieving a precise answer using semantic technologies. Currently, most web search engines are however based on purely statistical techniques. While they are not able to figure out the meaning of a query, they can provide answers by returning the statistically most appropriate answer to a user’s query—based on some measures for computing similarity in vector space (cf. [1]). Information Retrieval (IR) techniques applied to the Web have gained a reasonable degree of maturity which is clearly corroborated by the success of search engines such as Google, Yahoo and the like. These search engines are in fact providing a baseline quite di cult to outperform. Due to the nature and the maturity of the underlying statistical techniques, they are more robust and scale to the size of the Web, as opposed to semantic technologies. For restricted domains which can be formalized using ontologies, there is neverthe- less hope that semantic technologies can be put into work to allow for more semantics- based search. One of the crucial steps within such an endeavor is to precisely capture the user’s information need (see also [2]). But how does the user express his information need? If we look at the wide-spread usage of web search engines, we can conclude that users are definitely used to express their information need via simple queries based on keywords. However, while there is substantial recent work on interpreting full natural language questions semantically w.r.t. an ontology (cf. [3], [4]) or database schema [5], not as much work has been carried out with respect to the formal interpretation of key- word queries. A notable exception is the approach described in [6], which we discuss further in the related work section. K. Aberer et al. (Eds.): ISWC ASWC 2007, LNCS 4825, pp. 523–536, 2007. c Springer-Verlag Berlin Heidelberg 2007