TY - GEN
T1 - Modeling unsupervised perceptual category learning
AU - Lake, Brenden M.
AU - Vallabha, Gautam K.
AU - McClelland, James L.
N1 - Funding Information:
This research was a component of grant 2003/051 "Developing tagging models and validating assumptions for estimating key fishery assessment parameters in rock lobster fisheries" which was supported by funding from the Fisheries Research and Development Corporation on behalf of the Australian Government. Craig Mackinnon undertook the aquarium trials and deployment of PIT tag scanners on vessels. We thank the external reviewers for their constructive comments. Paul Burch was supported by ajoint CSIRO-UTAS PhDscholarship in quantitative marine science (QMS).
PY - 2008
Y1 - 2008
N2 - During the learning of speech sounds and other perceptual categories, category labels are not provided, the number of categories is unknown, and the stimuli are encountered sequentially. These constraints provide a challenge for models, but they have been recently addressed in the Online Mixture Estimation model of unsupervised vowel category learning [1]. the model treats categories as Gaussian distributions, proposing both the number and parameters of the categories. while the model has been shown to successfully learn vowel categories, it has not been evaluated as a model of the learning process. We account for three results regarding the learning process: Infants' discrimination of speech sounds is better after exposure to a bimodal rather than unimodal distribution [2], infants' discrimination of vowels is affected by acoustic distance [3], and subjects place category centers near frequent stimuli in an unsupervised visual classification task [4].
AB - During the learning of speech sounds and other perceptual categories, category labels are not provided, the number of categories is unknown, and the stimuli are encountered sequentially. These constraints provide a challenge for models, but they have been recently addressed in the Online Mixture Estimation model of unsupervised vowel category learning [1]. the model treats categories as Gaussian distributions, proposing both the number and parameters of the categories. while the model has been shown to successfully learn vowel categories, it has not been evaluated as a model of the learning process. We account for three results regarding the learning process: Infants' discrimination of speech sounds is better after exposure to a bimodal rather than unimodal distribution [2], infants' discrimination of vowels is affected by acoustic distance [3], and subjects place category centers near frequent stimuli in an unsupervised visual classification task [4].
UR - http://www.scopus.com/inward/record.url?scp=67650122399&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=67650122399&partnerID=8YFLogxK
U2 - 10.1109/DEVLRN.2008.4640800
DO - 10.1109/DEVLRN.2008.4640800
M3 - Conference contribution
AN - SCOPUS:67650122399
SN - 9781424426621
T3 - 2008 IEEE 7th International Conference on Development and Learning, ICDL
SP - 25
EP - 30
BT - 2008 IEEE 7th International Conference on Development and Learning, ICDL
T2 - 2008 IEEE 7th International Conference on Development and Learning, ICDL
Y2 - 9 August 2008 through 12 August 2008
ER -