|
Natural Language Systems, Department of Informatics, University of Hamburg
|
NavigationMainCoursesSE3: LogikprogrammierungProseminar: Verarbeitung gesprochener SpracheProjekt Dokumentenmanagement IMoreResearchConstruction GrammarINCOMConstraint Dependency GrammarLT4eLPROLIVITeachManagelexG.E.R.L.Recent EventsEAMT08LSP 2007Workshop RANLP 2007ELSNET 2006Ontology Workshop 06MoreInformation for exchange studentsDeutschEnglishEspañolOn-line DictionaryNatsWikiSandboxSystem |
Ill Defined DomainsError diagnosis plays an important role in Intelligent Tutoring Systems (ITS) because diagnostic information is the foundation for modelling the state of student’s knowledge and for initiating an appropriate instructional action. At present, several ITS apply simple diagnostic approach which presents a problem to the students and provide the possibility to submit a solution in the form of multi-choice or templates. Diagnostic approaches supported by this type of solution submission might be used in tutoring systems aiming at helping students to be familiar with bisic concepts of the domain. However, most of current systems provide students with domains of problems which are referred to as well-defined. What about ill-defined domains? Is it possible to build ITS for ill-defined domains?In my talk, I will give you an overwiew about chracteristics of well-defined domains and present two case studies: logic programming and UML class diagrams. After that, I will demonstrate how I apply the constraint-based approach to model the two domains mentioned and make them well-defined for diagnostic purposes. Nguyen-Thinh Le -- LeNguyenThinh -- 30 May 2006
|