Comparison of Cognitive Model Uses in Intelligent Training Systems

Abstract

Cognitive models are human performance models that represent human knowledge and internal information manipulation processes. Many uses of cognitive models in training systems have been proposed in the literature, but actual applications have lagged behind, and comparative assessments are rare. To fill this gap, experiences in applying cognitive models in several different intelligent tutoring/training systems are reviewed and compared. All the applications were developed using the same cognitive modeling framework (COGNET) and software (iGEN™). The models fall into two main categories:

expert performance models, used for tracing knowledge and actions, and

instructional agents, used to predict, observe, and diagnose trainee use of specific knowledge sets and skills.

Comparisons focus on the model development process and the efficacy of the resulting system. A set of preliminary conclusions on the selection and development of cognitive models in training systems is offered.