A generic architecture for interactive intelligent tutoring systems

Abstract

This research is focused on developing a generic intelligent architecture for an interactive tutoring system. A review of the literature in the areas of instructional theories, cognitive and social views of learning, intelligent tutoring systems development methodologies, and knowledge representation methods was conducted. As a result, a generic ITS development architecture (GeNisa) has been proposed, which combines the features of knowledge base systems (KBS) with object-oriented methodology. The GeNisa architecture consists of the following components: a tutorial events communication module, which encapsulates the interactive processes and other independent computations between different components; a software design toolkit; and an autonomous knowledge acquisition from a probabilistic knowledge base. A graphical application development environment includes tools to support application development, and learning environments and which use a case scenario as a basis for instruction. The generic architecture is designed to support client-side execution in a Web browser environment, and further testing will show that it can disseminate applications over the World Wide Web. Such an architecture can be adapted to different teaching styles and domains, and reusing instructional materials automatically can reduce the effort of the courseware developer (hence cost and time) in authoring new materials. GeNisa was implemented using Java scripts, and subsequently evaluated at various commercial and academic organisations. Parameters chosen for the evaluation include quality of courseware, relevancy of case scenarios, portability to other platforms, ease of use, content, user-friendliness, screen display, clarity, topic interest, and overall satisfaction with GeNisa. In general, the evaluation focused on the novel characteristics and performances of the GeNisa architecture in comparison with other ITS and the results obtained are discussed and analysed. On the basis of the experience gained during the literature research and GeNisa development and evaluation. a generic methodology for ITS development is proposed as well as the requirements for the further development of ITS tools. Finally, conclusions are drawn and areas for further research are identified.