ALBERT CORBETT
Associate Research Professor, Human Computer Interaction
www

My interests are in the application of artificial intelligence to education, in particular, the use of cognitive models in computer-based tutoring environments. Our approach is to build tutoring programs around rule-based models of how people perform skills, such as programming. Such a model allows the tutoring program to solve practice exercises along with the student, providing help as needed. My research has focused recently on estimating students' mastery of the knowledge represented in the cognitive model. By integrating learning and performance assumptions with the cognitive rules, we attempt to predict students' test performance, shape practice appropriately and improve the model. I am also interested in related issues in the design of computer-based practice environments, for example the content and timing of help messages and their impact on students' performance.