A Knowledge Based Approach to Support Learning Technical Terminology * Vania Dimitrova 1 , Darina Dicheva 2 , Paul Brna 1 , John Self 1 1 Computer Based Learning Unit, University of Leeds, Leeds LS2 9JT UK E-mail: {J.A.Self, P.Brna, V.Dimitrova}@cbl.leeds.ac.uk 2 Faculty of Mathematics and Informatics, Sofia University 5 James Bouchier St.1164 Sofia, Bulgaria E-mail: darinad@fmi.uni-sofia.bg To appear in the Proceedings of the Third Joint Conference on Knowledge Based Software Engineering, Slovakia, September, 1998 Abstract. In this paper we discuss the structure of a domain in a terminological area and some aspects related to using it for computer-based intelligent instruction. The most significant aspect of the suggested terminological model lies in its effective applicability for tutoring. We consider two basic instructional issues: (1) organising the pedagogical process and guiding instruction; (2) building interactive student models. The first issue concerns the work done in the ITELS project whereas the second one presents a research proposal to develop an interactive diagnostic tutor to help students learn technical terminology. We argue that a conceptual graphs terminological knowledge base is useful to fulfil the expert model requirements of a traditional intelligent tutoring system. It provides knowledge allowing the system to take strategic decisions about the instruction including its focus and content. Moreover, conceptual graphs can be used for the design of an interactive diagnostic tutor. They provide a graphical external representation of the student model as well as a medium for system-learner communication and thus appear useful for solving some problems in collaborative diagnosis. 1. Introduction Most of the presently developed terminological knowledge bases support the user’s exploration of the subject area and function as terminological dictionaries and domain encyclopaedias (c.f. [9], [14]). Typically they generate explanations from the terminological knowledge base (KB). However, they do not offer instruction and generally are not concerned with any educational aspects. An intelligent tutor, based on its expert model, should be able to take strategic decisions on the instructional focus and content. It should incorporate appropriate mechanisms for extracting knowledge from the KB necessary for completing the main pedagogical activities, including: answering questions, suggesting feedback, generating learning materials, diagnosing the student’s current knowledge and adapting the instruction to the learner. In this paper we address some aspects related to the use of a terminological KB for computer-based intelligent instruction. We illustrate our ideas on using a Conceptual Graphs (CGs) terminological knowledge base which we have designed and built as part of ITELS - an intelligent tutoring system aimed at helping Bulgarians to learn Computer Science terminology in English [8].