Extensions of intelligent tutoring paradigms to support collaborative learning

Abstract

Intelligent training systems with rich underlying knowledge in systematic form can readily be extended to incorporate collaborative learning possibilities. We have realized this as we have built two generations of intelligent coached practice environments for learning to diagnose failures of complex electronic systems. Using standard artificial intelligence and object-oriented programming approaches, we built a work domain simulation in which students could diagnose electronic system failures. Intelligent, context-sensitive advice is available, and it is tailored to a model of the student’s developing knowledge. Reflection opportunities after solving a problem include a variety of tools for examining one’s performance and comparing it to that of an expert. With this base, it is straightforward to extend self-critique to peer critique, to support teams of students working together, and to allow students to pose problems to each other. We discuss these possibilities and show what is needed to extend intelligent training systems to include them.