The Evolution of intelligent tutoring systems : Dimensions of Design (1)

The Evolution of intelligent tutoring systems : Dimensions  of Design

Hugh Burns
Universuty of Texas at Austin

James W. Parlett
Air Force Human resources Laboratory

Acts of instruction – training, tutoring, teaching – must be integral acts, indivisible, whole. Yet to a designer, they are composed of many instructions/personal, highly collaborative interactions, filled with moments of increasing expertise, of understanding, what was not understood, of being collage to know important things in personally useful and publicly usable ways.

In the 21st century, professional credibility will depend in part on how well educators have kept up with technology in general and the development and use intelligent tutoring systems (ITS) in particular. Plainly stated, modern educators cannot afford artificial intelligence illiteracy in tomorrow’s electronic schoolhouses. They cannot afford it for their professional lives; they certainly cannot afford it for their students’ futures. An “intelligent” computer is clearly on the practical horizon. Educators and trainers need to be able to exploit it, and – better yet – need to influence the design and evaluation of ITSs. In fact, the credibility of educators as master teachers or tutors depends on moving their future classrooms to singularly, scholarly, social conversations with one human at a time. This in itself, should not be surprising.

What should be surprising to us today is how far some of the research and developments have come. The literature and scholarship of ITSs is maturing on an international level. To complicate the matter, instructors trained before the widespread use of the microcomputer are often particularly wary of computer-based instruction of any sort. They perceive computer-based instruction, first, as a technology whose time peaked in the 1960s and 1970s and, second, as a symbol of all that mechanizes individual human performance. This bias is unfortunate, but real in all too many instances. This apprehension about technology stands in opposition to another roadblock facing ITS designers, researchers, and implementers, all too often, educators, administrators, and managers of technical training respond to the push of technology overzealously and none to wisely. Although it is disheartening to encounter a schoolhouse of any sort where technology is shunned, feared, or ignored, it is almost equally disheartening to enter a schoolhouse where administrators have responded to the demands of technology by hurling fistfuls of money at the problem with no real conceptual understanding of it. The result, predictably, is a schoolhouse full of fancy hardware platforms, but no design/development staff (or funding for such a beast) in sight.

Designing, developing, and evaluating ITSs are not so much strategic matters – what things to do – as they are an integrated technical enterprise – how to do it. Figure 1.1 portrays the overall evolving architecture of a practical intelligent tutoring system.

It is difficult to separate the dancer from the dance; nevertheless. In Foundations of Intelligent Tutoring Systems (Polson & Richardson, 1988), the foundational anatomy was use to discuss research issues within each of the separate components. These components were an expert module, a student diagnostic module, an instructional module, an intelligent interface, and a user. This classification has allowed the research community to focus attention on issues; for example, representing expert knowledge, designing student bug libraries, developing rules for teacher intervention, presenting intuitive computer work spaces and, for some preliminary evaluation of ITSs since 1987. Because designing intelligent tutors is such an interdisciplinary activity, attending to the anatomy piece by piece often ignored the synergy that a wholly integrated system could achieve. Now we are wiser and more ambitious. This companion piece to Foundations of Intelligent Tutoring Systems broadens the issues to emphasize the interactivity of the major components in the design and further explores the dimensions of communication, instruction, and expertise.

Intelligent tutoring system architecture