Abstract

Storage and transfer of information as well as interfaces for accessing this information have undergone a remarkable evolution. Nevertheless, information systems are still not `intelligent' in the sense that they `understand' the information they store, manipulate, and present to their users. A case in point is the world wide web and search engines allowing to access the vast amount of information available there. Web-pages are mostly written for human consumption and the mark-up provides only rendering information for textual and graphical information. Search engines are usually based on keyword search and often provide a huge number of answers, many of which are completely irrelevant, whereas some of the more interesting answers are not found. In contrast, the vision of a `semantic web' aims for machine-understandable web resources, whose content can then be comprehended and processed both by automated tools, such as search engines, and by human users.
The content-based representation of information requires representation formalisms with a well-defined formal semantics since otherwise there cannot be a common understanding of the represented information. This semantics can elegantly be provided by a translation into an appropriate logic or the use of a logic-based formalism in the first place. This logical approach has the additional advantage that logical inferences can then be used to reason about the represented information, thus detecting inconsistencies and computing implicit information. However, in this setting there is a fundamental tradeoff between the expressivity of the representation formalism on the one hand, and the efficiency of reasoning with this formalism on the other hand.
This motivates the `engineering of logics', i.e., the design of logical formalisms that are tailored to specific representation tasks. This also encompasses the formal investigation of the relevant inference problems, the development of appropriate inferences procedures, and their implementation, optimization, and empirical evaluation. Another important topic in this context is the combination of logics and their inference procedures since a given application my require the use of more than one specialized logic. The talk will illustrate this approach with the example of so-called Description Logics and their application as ontology languages for the semantic web.

About Franz Baader

Franz Baader is professor for Theoretical Computer Science at the RWTH Aachen, Germany. His research interests include knowledge representation (in particular, description logics, nonmonotonic logics, and modal logics) and automated deduction (in particular, unification theory, term rewriting systems, and combination of constraint solving methods). He has participated in ESPRIT Basic Research Projects and Working Groups, and he currently has several project grants from the German national science foundation (DFG). He has been in the program committee of various national and international conferences in the areas automated deduction, artificial intelligence, knowledge representation, and logics, has organized many international workshops, and is in the editorial board of three electronic AI journals. He has published about 80 refereed articles in major journals and conferences.