ABSTRACT
This chapter discusses the connection between the ACT-R theory and intelligent tutoring. It also discusses the model-tracing approach to maximizing learning and its basis in the ACT-R theory. The chapter provides an overview of the field of intelligent tutoring, focusing on contrasts between the modal assumptions of the field and our current approach. It reviews specific criticisms of the model-tracing approach. The authors select new problems that will offer maximal opportunity for the student to learn, and stop presenting new problems when they believe the student has reached a mastery level of performance on all productions in a particular section. The chapter reflects on the artificial intelligence (AI) roots of the field. Contrasting authors' approach with the traditional AI approach serves to highlight what is unique about their approach. The chapter describes the goals involved approaches to bug diagnosis, human emulation, and exploratory systems that stand in contrast with model-tracing tutors.
