A system architecture for flexible, knowledge-based, multimedia CBT-applications

One objective of the system architecture described in this paper is to increase the reusability of multimedia resources. For modern CBT applications, designing and creating multimedia resources such as text, graphics, pictures, animations, audio, and video sequences requires a considerable amount of time and effort. The project covers the development of methods and tools for authors. These tools enable authors to reconfigure existing modules for different educational settings, e.g. teaching basics, brush-up, training, and user classes, e.g. subject experts and novices. Besides reusability, increasing flexibility to adapt to each user's preferences, motivation, and experiences is another objective. The architecture is based on a model combining elements of hypermedia with those of semantic networks. Hypermedia concepts use and link small, multimedia teaching modules, whereas semantic networks provide an interpretable description of these modules. For each user, a specific course is dynamically created at run-time, using a run-time controller developed within the project. This controller contains a fuzzy logic component for representing pedagogical knowledge by fuzzy rules and neural networks. The latter one is included to represent experiences and decisions of former users (other learners or even authors) for their successors. In order to adapt learning paths and styles to individual users' needs (not only for user classes), a user model with static and dynamic data is integrated.<>