Qualitative models for interactive articulated learning environments

Abstract

behaviour, in particular how manipulations of some aspect of the system will effect its behaviour. In order to teach these behavioural characteristics, quantitative simulations of these systems are often used in computer-based learning environments. However, some behavioural fea~ tures are hard to communicate by a computer program that is based on such a quantitative simulation. Among others, generating causal explanations of the systems' behaviour, reasoning from structure (e.g., deriving the behaviour of a given structural description), and qualitativeness in general (e.g. a vocabulary for reasoning about behaviour in qualitative terms) cannot be dealt with adequately. A large part of the research on qualitative reasoning originated from efforts trying to cope with the limitations that followed from using quantitative simulators for teaching purposes. This talk will first discuss the idea of interactive learning environments. In particular, it will focus on the idea of learning by interacting with a simulation model. Next, the talk will present ideas on how qualitative models can be used for this purpose. During the past 10 years a large number of promising results have been achieved by the qualitative reasoning community, whereas at the same time the limitations of the current techniques are well understood. The talk will particularly focus on the use of causal models for curriculum planning, cognitive diagnosis and explanation. Finally, the use of digital multimedia technology provides opportunities to broaden the scope of communication between learner and computer. As such multimedia related issues are important for the construction of interactive articulated learning environments. Interesting is the notion of knowledge-based multimedia: automated generation and adaptation of a certain medium and its contents by a computer program in order to optimise the interaction with a learner or a trainee. The idea of automatic visualisation of qualitative models will be discussed.