Title: Modeling strategies and individual differences in learning a diagrammatic reasoning task　 
Speaker: Professor Frank Ritter (Pennsylvania State University)

Time: 15:00, Thursday, Aug 9th, 2012
Venue: Lecture Room, 3rd Floor, Building 5#, State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences

Abstract:
The Diag model learns in multiple ways while finding faults in a simple control panel device. The model accounts very well for measures such as problem-solving strategy, the relative difficulty of faults, and average fault-finding time. Because the model learns and transfers its learning across problems, it also accounts for the faster problem-solving times due to learning when examined across participants, across faults, and across the series of 20 trials on an individual participant basis.  The model shows how learning while problem solving can lead to more recognition-based performance, and helps explain how noise in the learning curve might arise through differential
transfer. This talk introduces new work on modeling different strategies on this task using a high level language for creating Soar cognitive models.

Related reading:
Ritter, F. E., & Bibby, P. A. (2008). Modeling how, when, and what learning happens in a diagrammatic reasoning task. Cognitive Science, 32, 862-892.
http://acs.ist.psu.edu/papers/ritterB08.pdf

Friedrich, M. B., & Frank E. Ritter, F. E. (2009).  Reimplementing a diagrammatic reasoning model in Herbal.  In Proceedings of ICCM - 2009- Ninth International Conference on Cognitive Modeling. 438-439. Manchester, England.
http://acs.ist.psu.edu/papers/friedrichR09.pdf

Bio:
Frank Ritter is the prof. of Information Sciences and Technology in the College of Information Sciences and Technology, the prof. of Psychology, and the prof. of Computer Science and Engineering. He obtained his PhD degree and MS degree at Carnegie-Mellon University in 1992 and 1989 seperately. He is interested in using cognitive modeling within unified theories of cognition to improve human-computer interaction and to test theories of learning, interaction, and behavior moderators. He has built several models that help explain how people learn and sets of tools to make model building, protocol analysis, and statistical analysis easier. He is also interested in developing stochastic learning and optimization algorithms to model behavior and to improve other analyses.  We are currently applying many of these techniques to a long-term attempt to study and model the effects of stress on cognition, and is exploring how to use high-performance computing to understand and optimize the fit of data to models.