Lisa C. Kaczmarczyk
University of Texas at Austin

The Acquisition of Intellectual Expertise: A Computational Theory

Intellectual expertise is a state of cognitive development in which a person can rapidly grasp subtleties of complex problems, and produce very high quality solutions. One of the most important reasons for wanting to better understand the nature of expertise is so that we can improve educational strategies. As a result of many research studies about experts and expertise, we have established some important information. For example, we know that intellectual experts have a great deal of knowledge in their fields, and have either large or highly efficient memories. But there is more to expertise than having an efficient memory. Otherwise it would be easy to create experts, whether human or computational, but it is clearly not easy. One reason for this difficulty is that experts can identify subtle patterns and intuit solution strategies that novices cannot. Understanding how to acquire intuition has proven particularly elusive for educators, students and psychologists alike.

A primary goal of my work is to increase understanding of the process by which humans become intellectual experts, i.e. how do people develop the intuitive ability to identify high quality solutions to problems just by looking at them? The second main goal is to understand this process in the context of formal adult education; specifically, how does the order in which material is delivered to the learner, affect their learning and conceptual development? I am testing the hypothesis that the acquisition of intellectual expertise takes place within the human mind via a statistical learning process that can be best understood using a distributed computational model. In my talk at the Doctoral Consortium, I will present results from preliminary experiments examining how different delivery methods influence learning and conceptual development. I will also present results showing how learning and conceptual development change as concepts become more difficult. These experiments use a real-world adult educational problem: the ability to identify correct solution strategies for calculus integration problems.

Current Status of My Dissertation Studies

I am working on an Interdisciplinary PhD in Computer Science Education. I have completed all the required course work and passed my Qualifying Exams. My dissertation committee will be made up of faculty from many departments here at the University of Texas. I am in the process of putting together this group of people, who will reside in Computer Science, Science/Math Education, and Psychology. I have completed the preliminary research necessary for my Proposal. As of this writing (11/02) I am conducting some additional experiments that will be included in my formal Proposal presentation. I expect to make my Proposal presentation early in the Spring semester.

What I Hope to Gain from the DC

One of the most difficult aspects of doing interdisciplinary research, is learning how to address multiple audiences. In my case I need to address the concerns of the Artificial Intelligence/Machine Learning community, as well as people whose background is Psychology. In addition, my audiences will contain CS Educators from a variety of subdisciplines. I need more experience putting together and formally presenting my work so that the focus is clear and the results convincing. I plan to use this opportunity to test out some new ways of effectively communicating my work. One of the nicest parts of the DC is that it brings together a wide range of CS people - they will ask the kind of questions and provide the suggestions that I need.

Bibliographic References

Classic Studies of Expertise:

Chi, M. T. H., Glaser, R., & Farr, M. J. (Eds.). (1988). The Nature of Expertise. Hillsdale, New Jersey: Lawrence Erlbaum Assoc.

deGroot, A. D. (1965). Thought and choice in chess. New York: Basic Books.

deGroot, A. D. (1966). Perception and Memory versus thought: Some ideas and recent findings. In B. Kleinmuntz (Ed.), Problem Solving: Research, methods and theory (pp. 19-50). New York: John Wiley.

Adult Calculus Learning:

Kaput, J. J., & Dubinsky, E. E. (1994). Research Issues in Undergraduate Mathematics Learning (MAA Notes 33). Washington, DC: Mathematical Association of America.

Selden, J., Selden, A., & Mason, A. (1994). Even Good Calculus Students Can't Solve Nonroutine Problems Research Issues in Undergraduate Mathematics Learning, MAA Notes Number 33, 19-26.

Complex Conceptual Learning with Neural Networks:

Cottrell, G. W., & Tseng, F.-S. (1991). Learning Simple Arithmetic Procedures. In J. A. Barnden & J. B. Pollack (Eds.), High-Level Connectionist Models (Vol. 1, pp. 305-321). Norwood: NJ: Ablex Publishing.

Taraban, R., & Palacios, J. M. (1993). Exemplar Models and Weighted Cue Models in Category Learning. In G. V. Nakamura, D. L. Medin, & R. Taraban (Eds.), Categorization by Humans and Machines (Vol. 29, pp. 91-127). San Diego: Academic Press, Inc.

Viscuso, S. R., Anderson, J. A., & Spoehr, K. T. (1989). Representing simple arithmetic in neural networks. In G. Tiberghien (Ed.), Advances in Cognitive Science (Vol. 2, pp. 141-164). NY and Chichester: Ellis Horwood John Wiley & Sons.

Last Revised: 11/25/02