CACS FACULTY AND THEIR RESEARCH INTERESTS

Dr. MIHAI CIOCOIU

mihaic@cacs.louisiana.edu

Assistant Professor of Computer Science. B.S., Computer Science, Polytechnic University, Bucharest, 1995; M.S., Computer Science, Polytechnic University, Bucharest, 1996; Ph.D., Computer Science, University of Maryland, College Park, 2001.

Dr. Ciocoiu's research interests are in the area of Artificial Intelligence. His main research focus is on knowledge representation and knowledge translation, in particular on the problem of automatically generating efficient translators between pairs of declarative languages, based on their ontologies. His recent results include a new kind of semantics for declarative languages, called Ontology-Based Semantics, that can be used for reasoning about knowledge translation in a multi-language environment, and a methodology for automatically generating concept/relation level translators based on the notion of Translation as Approximation.

If we only have one (first order) language, we

can formalize what we mean for a set of sentences to be a translation of another, by requiring that the two sets share the same models. However, in order to formalize translation for the case where the two sets of sentences are of different languages, we need a different notion of semantics, capable of overcoming the language barrier. Ontology-Based Semantics was introduced with this goal in mind. Ontologies can be used to make implicit assumptions explicit, and they are integrated into Ontology-Based Semantics in order to restrict the set of models a set of sentences has. Ontology-Based Models can be used to define formally knowledge translation for the different language case in a similar way ordinary models are used to define translation for the one language situation.

Translation as Approximation is an approach to translation in which the ontologies of the languages we're translating among are described by terminological axioms in a Description Logic, and translation is done as approximation at the concept/relation level.

In this setting, at translator generation time concepts of one language are approximated by concept expressions of the second, and the procedure takes advantage of the relatively cheap subsumption operation available for various Description Logics. At translation time, sentences of the first language are translated to sentences of the second based on the computed concept approximations by applying a fixed set of rules, which leads to very efficient translators.

SELECTED PUBLICATIONS

Ciocoiu, M., Gruninger, M., and Nau, D. Ontologies for Integrating Engineering Applications. Journal of Computing and Information Science in Engineering, Vol. 1, No. 1, pages 12-22, ASME International, March 2001.
Ciocoiu, M. and Nau, D. Ontology-Based Semantics. In Cohn, A., Giunchiglia, F., and Selman, B., editors, Principles of Knowledge Representation and Reasoning. Proceedings of the Seventh International Conference, pages 539-546, Breckenridge, Colorado, April 11-15 2000. Morgan Kaufmann.
Schlenoff, C., Gruninger, M., Ciocoiu, M., and Lee, J. The Essence of the Process Specification Language. In Marefat, M. and Olson, J., editors, Transactions of the Society for Computer Simulation International, Special Issue on Modeling and Simulation in Manufacturing Systems. Vol. 16, No. 4, pages 204-216, December 1999.
Schlenoff, C., Ciocoiu, M., Libes, D., and Gruninger, M. Process Specification Language: Results of the First Pilot Implementation. In Proceedings of the 1999 ASME International Mechanical Engineering Congress and Exposition, Nashville, Tennessee, November 14-19 1999.
Ciocoiu, M. and Bruda, S. GULiveR: Generalized Unification Based LR Parser for Natural Languages. In Tufis, D. and Andersen, P., editors, Recent Advances in Romanian Language Technology, pages 94-104, Romanian Academy Publishing House, 1997.
Bruda, S. and Ciocoiu, M. Mediu de tip LR Generalizat pentru Gramatici de Unificare. In Tufis, D., editor, Limbaj si Tehnologie, pages 189-195, Romanian Academy Publishing House, 1996.