Designing Application-Specific Neural Networks Using the Genetic Algorithm

Steven A. Harp; Tariq Samad; Aloke Guha

Designing Application-Specific Neural Networks Using the Genetic Algorithm

Steven A. Harp
, Tariq Samad
, Aloke Guha

Technological Leadership Institute

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

60 Scopus citations

Abstract

We present a general and systematic method for neural network design based on the genetic algorithm. The technique works in conjunction with network learning rules, addressing aspects of the network's gross architecture, connectivity, and learning rule parameters. Networks can be optimized for various application-specific criteria, such as learning speed, generalization, robustness and connectivity. The approach is model-independent. We describe a prototype system, NeuroGENESYS, that employs the backpropagation learning rule. Experiments on several small problems have been conducted. In each case, NeuroGENESYS has produced networks that perform significantly better than the randomly generated networks of its initial population. The computational feasibility of our approach is discussed.

Original language	English (US)
Title of host publication	Advances in Neural Information Processing Systems 2, NIPS 1989
Editors	David S. Touretzky
Publisher	Neural information processing systems foundation
Pages	447-454
Number of pages	8
ISBN (Electronic)	1558601007, 9781558601000
State	Published - 1989
Event	2nd Advances in Neural Information Processing Systems, NIPS 1989 - Denver, United States Duration: Nov 27 1989 → Nov 30 1989

Publication series

Name	Advances in Neural Information Processing Systems
Volume	2
ISSN (Print)	1049-5258

Conference

Conference	2nd Advances in Neural Information Processing Systems, NIPS 1989
Country/Territory	United States
City	Denver
Period	11/27/89 → 11/30/89

Bibliographical note

Publisher Copyright:
© 1989 Neural information processing systems foundation. All rights reserved.

Find It

Find It at U of M Libraries

Cite this

Designing Application-Specific Neural Networks Using the Genetic Algorithm. / Harp, Steven A.; Samad, Tariq; Guha, Aloke.
Advances in Neural Information Processing Systems 2, NIPS 1989. ed. / David S. Touretzky. Neural information processing systems foundation, 1989. p. 447-454 (Advances in Neural Information Processing Systems; Vol. 2).

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

Harp, SA, Samad, T & Guha, A 1989, Designing Application-Specific Neural Networks Using the Genetic Algorithm. in DS Touretzky (ed.), Advances in Neural Information Processing Systems 2, NIPS 1989. Advances in Neural Information Processing Systems, vol. 2, Neural information processing systems foundation, pp. 447-454, 2nd Advances in Neural Information Processing Systems, NIPS 1989, Denver, United States, 11/27/89.

@inproceedings{172a6d1af94a45ffaf73e8b67d501809,

title = "Designing Application-Specific Neural Networks Using the Genetic Algorithm",

abstract = "We present a general and systematic method for neural network design based on the genetic algorithm. The technique works in conjunction with network learning rules, addressing aspects of the network's gross architecture, connectivity, and learning rule parameters. Networks can be optimized for various application-specific criteria, such as learning speed, generalization, robustness and connectivity. The approach is model-independent. We describe a prototype system, NeuroGENESYS, that employs the backpropagation learning rule. Experiments on several small problems have been conducted. In each case, NeuroGENESYS has produced networks that perform significantly better than the randomly generated networks of its initial population. The computational feasibility of our approach is discussed.",

author = "Harp, \{Steven A.\} and Tariq Samad and Aloke Guha",

note = "Publisher Copyright: {\textcopyright} 1989 Neural information processing systems foundation. All rights reserved.; 2nd Advances in Neural Information Processing Systems, NIPS 1989 ; Conference date: 27-11-1989 Through 30-11-1989",

year = "1989",

language = "English (US)",

series = "Advances in Neural Information Processing Systems",

publisher = "Neural information processing systems foundation",

pages = "447--454",

editor = "Touretzky, \{David S.\}",

booktitle = "Advances in Neural Information Processing Systems 2, NIPS 1989",

}

TY - GEN

T1 - Designing Application-Specific Neural Networks Using the Genetic Algorithm

AU - Harp, Steven A.

AU - Samad, Tariq

AU - Guha, Aloke

PY - 1989

Y1 - 1989

N2 - We present a general and systematic method for neural network design based on the genetic algorithm. The technique works in conjunction with network learning rules, addressing aspects of the network's gross architecture, connectivity, and learning rule parameters. Networks can be optimized for various application-specific criteria, such as learning speed, generalization, robustness and connectivity. The approach is model-independent. We describe a prototype system, NeuroGENESYS, that employs the backpropagation learning rule. Experiments on several small problems have been conducted. In each case, NeuroGENESYS has produced networks that perform significantly better than the randomly generated networks of its initial population. The computational feasibility of our approach is discussed.

AB - We present a general and systematic method for neural network design based on the genetic algorithm. The technique works in conjunction with network learning rules, addressing aspects of the network's gross architecture, connectivity, and learning rule parameters. Networks can be optimized for various application-specific criteria, such as learning speed, generalization, robustness and connectivity. The approach is model-independent. We describe a prototype system, NeuroGENESYS, that employs the backpropagation learning rule. Experiments on several small problems have been conducted. In each case, NeuroGENESYS has produced networks that perform significantly better than the randomly generated networks of its initial population. The computational feasibility of our approach is discussed.

UR - https://www.scopus.com/pages/publications/105021399167

UR - https://www.scopus.com/pages/publications/105021399167#tab=citedBy

M3 - Conference contribution

AN - SCOPUS:105021399167

T3 - Advances in Neural Information Processing Systems

SP - 447

EP - 454

BT - Advances in Neural Information Processing Systems 2, NIPS 1989

A2 - Touretzky, David S.

PB - Neural information processing systems foundation

T2 - 2nd Advances in Neural Information Processing Systems, NIPS 1989

Y2 - 27 November 1989 through 30 November 1989

ER -

Designing Application-Specific Neural Networks Using the Genetic Algorithm

Abstract

Publication series

Conference

Bibliographical note

Find It

Other files and links

Fingerprint

Cite this