INTELLIGENT TUTORING AND DYNAMIC ASSESSMENT FOR EDUCATION AND TRAINING

Laveen N. Kanal,Ph.D
Professor Emeritus,
Univ. of Maryland, College Park, MD/
LNK Corporation, Inc., Riverdale, MD 207837-1333
http://www.LNK.com


It has been conclusively demonstrated that, other things being equal, individualized tutoring, on average, provides a "2 Sigma" improvement in student performance over classroom instruction (Bloom, 1986). Individualized tutoring is prohibitively expensive in classroom settings in which there are more than a very few students. software systems, based on Pattern Recognition, Artificial Intelligence software systems, based on Pattern Recognition, Artificial Intelligence e.g., Case-Based Reasoning, and e e.g., Case-Based Reasoning, and technologies such as Fuzzy Logic, Neural Networks, Genetic Algorithms and Bayesian Belief Networks, are emerging. Some technologies such as Fuzzy Logic, Neural Networks, Genetic Algorithms and Bayesian Belief Networks, are emerging. Some of these software systems can greatly assist teachers and learners in enhancing the learning experience.

Current assessment practices are limited to performance on a few tests and often do not provide students with feedback until several lessons are completed. This type of "summative" feedback may have uses for administrators but does not typically help teaching and learning, while the process is underway. It has proven to be of little help in bringing up the performance of low performing students. The current slogan of "high standards and leave no child behind" is clearly an oxymoron and will remain an empty promise unless a radical new approach to instruction, learning and assessment is used.

Dynamic Assessment is an important area of research. The term is used for learner assessment based on regular intervention during a student's learning, by a group of professionals, possibly including psychologists, using a broad cognitive view of the learner rather than just test performance at the end of a course. This approach is not yet used widely in education and training because it is expensive and at best can only serve a few students. While no machine can replace a caring teacher individually tutoring a few students with support from a cadre of professionals, we can provide a feasible approach to personalized tutoring and dynamic assessment using machine intelligence.

A learner's performance assessment depends not only on the learner but on an entire chain of elements from the Analysis stage -definition of Learning Objectives, Task Analysis etc, to the Content Authoring and Media design stage, to the Delivery stage and finally the Testing stage. Thus a multidimensional approach to performance assessment is warranted. DAMSeL™ (Dynamic, Adaptive Machines for Smart e-Learning, also Student e-Learning and Situated e-Learning), is a domain independent authoring and tutoring system, which represents a multidimensional approach. DAMSeL dynamically assesses performance and provides individualized, adaptive sequencing of presentations to each learner, in all modes of instruction: instruction, practice, testing, diagnosis, and remediation. DAMSeL™ can be used to ensure mastery of learning objectives at desired levels and also alleviate much of the burden on classroom teachers in order to diagnose student learning gaps and get students certified or ready for standardized tests. For additional comments and references on the above or other current and emerging research on assessment please contact Dr. Kanal at kanal@LNK.com, or kanal@cs.umd.edu.

________________________________________________________________________

Laveen Kanal, Ph.D., Professor Emeritus of Computer Science, University of Maryland at College Park, and President of L N K Corporation, Inc. He was awarded his Ph.D. in E.E. from the University of Pennsylvania in 1960. His dissertation (in E.E.), on the analysis of a nonlinear stochastic model for Individual Choice Behavior was guided by the mathematical psychologists, Robert R. Bush and R. Duncan Luce and was published as two papers in the journal, Psychometrika. Dr. Kanal joined the University of Maryland in July 1970 where he taught courses in AI and Pattern Recognition and served as principal investigator on numerous research grants for twenty-five years. Prior to that, he provided technical leadership on many R & D contracts from the Department of Defense agencies at Philco-Ford and General Dynamics/Electronics. While at Philco-Ford, Dr. Kanal was also a Visiting Professor at the Univ. of Pennsylvania's Wharton Graduate School of Business and in the E.E. Dept. of Lehigh Univ. He was elected a Fellow of the IEEE, the AAAS, the American Association for AI and the International Association for Pattern Recognition (IAPR). In 1992 he received the highest award of IAPR for his contributions to pattern recognition, and in 1996 he received the "Contributions to Science" award of UMD's chapter of the honor research society Sigma Xi. He has served on Board of Directors of the Maryland High-Technology Council, the Collonade Society of the Univ. of Maryland, the Board of Directors of the University of Maryland Friends of the Libraries, and is currently an education program evaluator for ABET, the Accreditation Board for Engineering and Technology programs at US engineering colleges and institutes of technology.. His biography appears in Who's Who in America, Who's Who in Engineering, American Men and Women of Science and other national and international biographical sources.