Categorical Perception and the Evolution of Supervised Learning in Neural Nets

Harnad, Stevan and Hanson, S.J. and Lubin, J. (1991) Categorical Perception and the Evolution of Supervised Learning in Neural Nets. [Book Chapter]

Full text available as:

HTML
28Kb

Abstract

Some of the features of animal and human categorical perception (CP) for color, pitch and speech are exhibited by neural net simulations of CP with one-dimensional inputs: When a backprop net is trained to discriminate and then categorize a set of stimuli, the second task is accomplished by "warping" the similarity space (compressing within-category distances and expanding between-category distances). This natural side-effect also occurs in humans and animals. Such CP categories, consisting of named, bounded regions of similarity space, may be the ground level out of which higher-order categories are constructed; nets are one possible candidate for the mechanism that learns the sensorimotor invariants that connect arbitrary names (elementary symbols?) to the nonarbitrary shapes of objects. This paper examines how and why such compression/expansion effects occur in neural nets.

Item Type:	Book Chapter
Keywords:	categorical perception, neural nets, learning
Subjects:	Computer Science > Neural Nets Psychology > Perceptual Cognitive Psychology
ID Code:	1579
Deposited By:	Harnad, Stevan
Deposited On:	18 Jun 2001
Last Modified:	11 Mar 2011 08:54

References in Article

Select the SEEK icon to attempt to find the referenced article. If it does not appear to be in cogprints you will be forwarded to the paracite service. Poorly formated references will probably not work.

Metadata

Repository Staff Only: item control page