How knowledge drives understanding - matching medical ontologies with the needs of medical language processing

In this article, we introduce a knowledge-based approach to medical text un derstanding. From an in-depth consideration of deep sentence and text under standing we distill basic requirements for an adequate knowledge representa tion framework. These requirements are then matched with currently availabl e medical ontologies (thesauri, terminologies, etc.). A fundamental trade-o ff is recognized between large-scale conceptual coverage on the one hand, a nd formal mechanisms for integrity preservation and conceptual expressivene ss on the ether hand. We discuss various shortcomings of the most wide-spre ad ontologies to capture medical knowledge in-the-large. As a result, we ar gue for the need of a formally sound and expressive model along the lines o f KL-ONE-style terminological representation systems in the format of descr iption logics. These provide an adequate methodology for designing more sop histicated, flexible medical ontologies serving the needs of 'deep' knowled ge applications which are by no means restricted to medical language proces sing. (C) 1999 Elsevier Science B.V. All rights reserved.