An MCMC-based method of comparing connectionist models in cognitive science
Date
2004
Authors
Woojae, K.
Navarro, D.
Pitt, M.
Myung, J.
Editors
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Conference paper
Citation
Advances in neural information processing systems 16: proceedings of the 2003 conference / Sebastian Thrun, Lawrence K. Saul, and Bernhard Schölkopf (eds.): pp.937-944
Statement of Responsibility
Woojae Kim, Daniel J. Navarro, Mark A. Pitt, In Jae Myung
Conference Name
Conference on Neural Information Processing Systems (2003 : Vancouver, Canada)
Abstract
Despite the popularity of connectionist models in cognitive science, their performance can often be difficult to evaluate. Inspired by the geometric approach to statistical model selection, we introduce a conceptually similar method to examine the global behavior of a connectionist model, by counting the number and types of response patterns it can simulate. The Markov Chain Monte Carlo-based algorithm that we constructed nds these patterns efficiently. We demonstrate the approach using two localist network models of speech perception.