Students learning from computers is nothing new, but now Temple University electrical engineering professor Brian P. Butz has devised a way that computers can learn back from the students.
As a student answers a series of questions, Butz's tutorial program--which will be used in several classes at Temple this Fall--determines not only what that student does and does not know, but also how he or she likes to learn. It then creates an individualized program of study. "The intelligent system will sense the direction a student is taking through the interactive material, detect logical flaws the student is making and provide the student with focused tutoring," he explains, adding that the program software is adaptable to a wide variety of subject areas. His project will receive significant support from the National Science Foundation (NSF) over the next two years.
Butz says that computer modules have come a long way from the right-or-wrong "canned remedial responses" of the past. Current programs have more explicit knowledge of the subject matter being taught and can give expanded explanations to the user. His innovation continues this progress by introducing an "expert system" to create an "interactive multimedia intelligent tutoring system." The expert system monitors and records every interaction of the student with the program (selection of an answer, change of selection, computation, and so on). It also "interrupts" the session occasionally to ask the user to clarify a response or explain why that answer was chosen. Once it has gathered enough information about the student's learning patterns, it modifies its questions to focus on the material the student understands least, presented in the way the student learns best.
Butz reports that his expert system could be used for almost any subject, and that "it will be especially helpful for students in introductory and remedial classes. Students who do not quickly master basic skills are sometimes left behind, and become passive observers of material they do not understand." He hopes that "tutorials using the expert system will help teachers face the challenge of meeting the students' academic needs and ensuring that they are active participants in their own education."
Butz will use his tutoring system in electrical engineering courses this coming Fall and Spring. He hopes to find results similar to those of his colleagues Gerardo Mendoza and Dan Reich of the mathematics department. They have written a interactive program which enables math students to work through calculus assignments on a class webpage. Their Calculus on the Web (COW) project, which has also received NSF funding, has resulted in an increase in the retention, enthusiasm, and performance of students in first- and second-semester calculus classes at Temple. They have made their program available to all high school and college students via the internet.
Students supplementing classroom learning with computer tutorials is nothing new, but now Temple electrical engineering professor Brian P. Butz has devised a way that those tutorials can learn back from the students. As each student answers questions, his "expert system" gathers information about his or her learning patterns. It then modifies future questions to focus on the material the student understands least, presented in the way the student learns best.