Bayesian learning of visual chunks by human observers

doi:10.1073/pnas.0708424105

Bayesian learning of visual chunks by human observers

*Collegium Budapest Institute for Advanced Study, 2 Szentháromság utca, Budapest H-1014, Hungary;
^†Department of Psychology and Volen Center for Complex Systems, Brandeis University, 415 South Street, Waltham, MA 02454;
^‡Department of Brain and Cognitive Sciences, Center for Visual Science, Meliora 406, University of Rochester, Rochester, NY 14627;
^§Gatsby Computational Neuroscience Unit, University College London, Alexandra House, 17 Queen Square, London WC1N 3AR, United Kingdom; and
^¶Computational and Biological Learning Laboratory, Department of Engineering, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ, United Kingdom

Edited by James L. McClelland, Stanford University, Stanford, CA, and approved December 28, 2007 (received for review September 5, 2007)

Abstract

Efficient and versatile processing of any hierarchically structured information requires a learning mechanism that combines lower-level features into higher-level chunks. We investigated this chunking mechanism in humans with a visual pattern-learning paradigm. We developed an ideal learner based on Bayesian model comparison that extracts and stores only those chunks of information that are minimally sufficient to encode a set of visual scenes. Our ideal Bayesian chunk learner not only reproduced the results of a large set of previous empirical findings in the domain of human pattern learning but also made a key prediction that we confirmed experimentally. In accordance with Bayesian learning but contrary to associative learning, human performance was well above chance when pair-wise statistics in the exemplars contained no relevant information. Thus, humans extract chunks from complex visual patterns by generating accurate yet economical representations and not by encoding the full correlational structure of the input.

Footnotes

^‖To whom correspondence should be addressed. E-mail: lmate{at}gatsby.ucl.ac.uk
Author contributions: G.O., J.F., and M.L. designed research; G.O., J.F., and R.N.A. performed research; G.O., J.F., and M.L. analyzed data; and G.O., J.F., R.N.A., and M.L. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/cgi/content/full/0708424105/DC1.