Quickly trained artificial neural network with single hidden layer Gaussian units

Goutam Chakraborty; Norio Shiratori; Shoichi Noguchi

Quickly trained artificial neural network with single hidden layer Gaussian units

Goutam Chakraborty, Norio Shiratori, Shoichi Noguchi

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Nonlinear radial basis functions at the single layer hidden units are effective in generating complex nonlinear mapping and at the same time facilitate fast linear learning. In this work, we propose a model and an algorithm to arrive at a near optimum initial configuration very quickly. Thus the position of the hidden units in the input space and the connection weights from the hidden units to the output units, instead of arbitrarily, are optimally set. Simulations on this initial configuration are performed. Different parameters are further trained and their effects experimented.

Original language	English
Title of host publication	1993 IEEE International Conference on Neural Networks
Editors	Anon
Place of Publication	Piscataway, NJ, United States
Publisher	Publ by IEEE
Pages	466-472
Number of pages	7
ISBN (Print)	0780312007
Publication status	Published - 1993
Event	1993 IEEE International Conference on Neural Networks - San Francisco, CA, USA Duration: 1993 Mar 28 → 1993 Apr 1

Other

Other	1993 IEEE International Conference on Neural Networks
City	San Francisco, CA, USA
Period	93/3/28 → 93/4/1

ASJC Scopus subject areas

Engineering(all)

Cite this

@inproceedings{899b6a6d8f5f4e0b87dbc535212b98b3,

title = "Quickly trained artificial neural network with single hidden layer Gaussian units",

abstract = "Nonlinear radial basis functions at the single layer hidden units are effective in generating complex nonlinear mapping and at the same time facilitate fast linear learning. In this work, we propose a model and an algorithm to arrive at a near optimum initial configuration very quickly. Thus the position of the hidden units in the input space and the connection weights from the hidden units to the output units, instead of arbitrarily, are optimally set. Simulations on this initial configuration are performed. Different parameters are further trained and their effects experimented.",

author = "Goutam Chakraborty and Norio Shiratori and Shoichi Noguchi",

year = "1993",

language = "English",

isbn = "0780312007",

pages = "466--472",

editor = "Anon",

booktitle = "1993 IEEE International Conference on Neural Networks",

publisher = "Publ by IEEE",

note = "1993 IEEE International Conference on Neural Networks ; Conference date: 28-03-1993 Through 01-04-1993",

}

TY - GEN

T1 - Quickly trained artificial neural network with single hidden layer Gaussian units

AU - Chakraborty, Goutam

AU - Shiratori, Norio

AU - Noguchi, Shoichi

PY - 1993

Y1 - 1993

N2 - Nonlinear radial basis functions at the single layer hidden units are effective in generating complex nonlinear mapping and at the same time facilitate fast linear learning. In this work, we propose a model and an algorithm to arrive at a near optimum initial configuration very quickly. Thus the position of the hidden units in the input space and the connection weights from the hidden units to the output units, instead of arbitrarily, are optimally set. Simulations on this initial configuration are performed. Different parameters are further trained and their effects experimented.

AB - Nonlinear radial basis functions at the single layer hidden units are effective in generating complex nonlinear mapping and at the same time facilitate fast linear learning. In this work, we propose a model and an algorithm to arrive at a near optimum initial configuration very quickly. Thus the position of the hidden units in the input space and the connection weights from the hidden units to the output units, instead of arbitrarily, are optimally set. Simulations on this initial configuration are performed. Different parameters are further trained and their effects experimented.

UR - http://www.scopus.com/inward/record.url?scp=0027306610&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0027306610&partnerID=8YFLogxK

M3 - Conference contribution

SN - 0780312007

SP - 466

EP - 472

BT - 1993 IEEE International Conference on Neural Networks

A2 - Anon, null

PB - Publ by IEEE

CY - Piscataway, NJ, United States

T2 - 1993 IEEE International Conference on Neural Networks

Y2 - 28 March 1993 through 1 April 1993

ER -

Quickly trained artificial neural network with single hidden layer Gaussian units

Abstract

Other

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this