TY - JOUR
T1 - Predicting shifts in generalization gradients with perceptrons
AU - Wisniewski, Matthew
AU - Radell, Milen
AU - Guillette, Lauren
AU - Sturdy, Christopher
AU - Mercado III, Eduardo
PY - 2012
Y1 - 2012
N2 - Perceptron models have been used extensively to model perceptual learning and the effects of discrimination training on generalization, as well as to explore natural classification mechanisms. Here, we assess the ability of existing models to account for the time course of generalization shifts that occur when individuals learn to distinguish sounds. A set of simulations demonstrates that commonly used single-layer and multilayer perceptron networks do not predict transitory shifts in generalization over the course of training but that such dynamics can be accounted for when the output functions of these networks are modified to mimic the properties of cortical tuning curves. The simulations further suggest that prudent selection of stimuli and training criteria can allow for more precise predictions of learning-related shifts in generalization gradients in behavioral experiments. In particular, the simulations predict that individuals will show maximal peak shift after different numbers of trials, that easier contrasts will lead to slower development of shifted peaks, and that whether generalization shifts persist or dissipate will depend on which stimulus dimensions individuals use to distinguish stimuli and how those dimensions are neurally encoded.
AB - Perceptron models have been used extensively to model perceptual learning and the effects of discrimination training on generalization, as well as to explore natural classification mechanisms. Here, we assess the ability of existing models to account for the time course of generalization shifts that occur when individuals learn to distinguish sounds. A set of simulations demonstrates that commonly used single-layer and multilayer perceptron networks do not predict transitory shifts in generalization over the course of training but that such dynamics can be accounted for when the output functions of these networks are modified to mimic the properties of cortical tuning curves. The simulations further suggest that prudent selection of stimuli and training criteria can allow for more precise predictions of learning-related shifts in generalization gradients in behavioral experiments. In particular, the simulations predict that individuals will show maximal peak shift after different numbers of trials, that easier contrasts will lead to slower development of shifted peaks, and that whether generalization shifts persist or dissipate will depend on which stimulus dimensions individuals use to distinguish stimuli and how those dimensions are neurally encoded.
UR - http://www.ncbi.nlm.nih.gov/pubmed/21983938
U2 - 10.3758/s13420-011-0050-6
DO - 10.3758/s13420-011-0050-6
M3 - Article
SN - 1543-4494
VL - 40
SP - 128
EP - 144
JO - Learning and Behavior
JF - Learning and Behavior
ER -