Harnad, Stevan, Hanson, S.J. and Lubin, J.,
Categorical Perception and the Evolution of Supervised Learning in Neural Nets
Powers, D.W. and Reeker, L. (eds.)
At Proceedings of the AAAI Spring Symposium on Machine Learning of Natural Language and Ontology.
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.
Conference or Workshop Item
||In: Proceedings of the AAAI Spring Symposium on Machine Learning of Natural Language and Ontology (DW Powers & L Reeker, Eds.) Document D91-09, Deutsches Forschungszentrum fur Kuenstliche Intelligenz GmbH Kaiserslautern FRG, pp. 65-74. [Presented at Symposium on Symbol Grounding: Problems and Practice, Stanford University, March 1991] http://www.cogsci.soton.ac.uk/~harnad/Papers/Harnad/harnad91.cpnets.html
|Venue - Dates:
||Proceedings of the AAAI Spring Symposium on Machine Learning of Natural Language and Ontology, 1991-01-01
||Web & Internet Science
||26 May 2000
||17 Apr 2017 23:24
|Further Information:||Google Scholar|
Actions (login required)