From Novice to Expert Learner Models: Authoring adaptive user-interface behaviour to convey training needs

1 Abstract No past research has addressed the connection between learner models and interaction models. Nor has any previous work accounted for the dynamic modelling of interface contents based on the underlying pedagogic directives. This paper describes how cognitive and computational concepts can be applied to building multimodal interface and learner models a nd for maintaining model consistency throughout long -term tutorial interactions. We show an important influence of an experiment and our cognitive approach to the design of intelligent tutoring systems. This approach includes not only the notion of multipl e dimensions of domain expertise but also the appropriate dialogue factors that can promote the significant aspects of fluency in human -centred, multi modal dialogues. A substantial part of the theoretical findings has been implemented in the RUI prototype system.