Dystal: a self-organizing ANN with pattern independent training time

G. Barbour; K. Blackwell; T. Busse; D. Alkon; T. Vogl

doi:10.1109/IJCNN.1992.227217

Back

Conference proceeding

Dystal: a self-organizing ANN with pattern independent training time

G. Barbour, K. Blackwell, T. Busse, D. Alkon and T. Vogl

[Proceedings 1992] IJCNN International Joint Conference on Neural Networks, Vol.4, pp.814-819 vol.4

1992

DOI: 10.1109/IJCNN.1992.227217

View Online

Abstract

As the difficulty of problems increase, artificial neural networks (ANNs) that use nonlinear optimization suffer from degraded execution speed, particularly with respect to learning time. Dystal is an ANN which does not suffer this degradation. Dystal is an ANN based on properties of associative learning found in biological neural networks. To verify these theoretical properties of Dystal, the authors implement Dystal on MasPar, a massively parallel machine. They show that the execution time is independent of both the separability of the patterns and the number of output units, and that the training time is linear with the number of patterns in the training data set. That is, the number of iterations through the training set to achieve learning is small and independent of pattern content or training set size.< >

Artificial neural networks

Biological neural networks

Biology computing

Computer networks

Degradation

Neurons

Parallel machines

Parallel processing

System testing

Training data

Details

Title: Subtitle: Dystal: a self-organizing ANN with pattern independent training time
Creators: G. Barbour - Environmental Research Institute of Michigan
K. Blackwell - Environmental Research Institute of Michigan
T. Busse
D. Alkon
T. Vogl
Resource Type: Conference proceeding
Publication Details: [Proceedings 1992] IJCNN International Joint Conference on Neural Networks, Vol.4, pp.814-819 vol.4
DOI: 10.1109/IJCNN.1992.227217
Publisher: IEEE
Language: English
Date published: 1992
Academic Unit: Roy J. Carver Department of Biomedical Engineering; Iowa Neuroscience Institute
Record Identifier: 9984446280802771

Metrics

62 Record Views