Dynamic topology representing networks

Siming Lin, Jennie Si

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In the present paper, we propose a new algorithm named dynamic topology representing Networks (DTRN) for learning both topology and clustering information from input data. In contrast to other models with adaptive architecture of this kind, the DTRN algorithm adaptively grows the number of output nodes by applying a vigilance test. The clustering procedure is based on a winner-take-quota learning strategy in conjunction with an annealing process in order to minimize the associated mean square error. A competitive Hebbian rule is applied to learn the global topology information concurrently with the clustering process. The topology information learned is also utilized for dynamically deleting the nodes and for the annealing process. The specific properties of this new algorithm are illustrated by some analyses and simulation examples.

Original languageEnglish (US)
Title of host publicationIEEE International Conference on Neural Networks - Conference Proceedings
Editors Anon
Place of PublicationPiscataway, NJ, United States
PublisherIEEE
Pages353-358
Number of pages6
Volume1
StatePublished - 1998
EventProceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3) - Anchorage, AK, USA
Duration: May 4 1998May 9 1998

Other

OtherProceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3)
CityAnchorage, AK, USA
Period5/4/985/9/98

Fingerprint

Topology
Annealing
Mean square error

ASJC Scopus subject areas

  • Software

Cite this

Lin, S., & Si, J. (1998). Dynamic topology representing networks. In Anon (Ed.), IEEE International Conference on Neural Networks - Conference Proceedings (Vol. 1, pp. 353-358). Piscataway, NJ, United States: IEEE.

Dynamic topology representing networks. / Lin, Siming; Si, Jennie.

IEEE International Conference on Neural Networks - Conference Proceedings. ed. / Anon. Vol. 1 Piscataway, NJ, United States : IEEE, 1998. p. 353-358.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Lin, S & Si, J 1998, Dynamic topology representing networks. in Anon (ed.), IEEE International Conference on Neural Networks - Conference Proceedings. vol. 1, IEEE, Piscataway, NJ, United States, pp. 353-358, Proceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3), Anchorage, AK, USA, 5/4/98.
Lin S, Si J. Dynamic topology representing networks. In Anon, editor, IEEE International Conference on Neural Networks - Conference Proceedings. Vol. 1. Piscataway, NJ, United States: IEEE. 1998. p. 353-358
Lin, Siming ; Si, Jennie. / Dynamic topology representing networks. IEEE International Conference on Neural Networks - Conference Proceedings. editor / Anon. Vol. 1 Piscataway, NJ, United States : IEEE, 1998. pp. 353-358
@inproceedings{3f2ecdc673a3404b9dcd05e7542223da,
title = "Dynamic topology representing networks",
abstract = "In the present paper, we propose a new algorithm named dynamic topology representing Networks (DTRN) for learning both topology and clustering information from input data. In contrast to other models with adaptive architecture of this kind, the DTRN algorithm adaptively grows the number of output nodes by applying a vigilance test. The clustering procedure is based on a winner-take-quota learning strategy in conjunction with an annealing process in order to minimize the associated mean square error. A competitive Hebbian rule is applied to learn the global topology information concurrently with the clustering process. The topology information learned is also utilized for dynamically deleting the nodes and for the annealing process. The specific properties of this new algorithm are illustrated by some analyses and simulation examples.",
author = "Siming Lin and Jennie Si",
year = "1998",
language = "English (US)",
volume = "1",
pages = "353--358",
editor = "Anon",
booktitle = "IEEE International Conference on Neural Networks - Conference Proceedings",
publisher = "IEEE",

}

TY - GEN

T1 - Dynamic topology representing networks

AU - Lin, Siming

AU - Si, Jennie

PY - 1998

Y1 - 1998

N2 - In the present paper, we propose a new algorithm named dynamic topology representing Networks (DTRN) for learning both topology and clustering information from input data. In contrast to other models with adaptive architecture of this kind, the DTRN algorithm adaptively grows the number of output nodes by applying a vigilance test. The clustering procedure is based on a winner-take-quota learning strategy in conjunction with an annealing process in order to minimize the associated mean square error. A competitive Hebbian rule is applied to learn the global topology information concurrently with the clustering process. The topology information learned is also utilized for dynamically deleting the nodes and for the annealing process. The specific properties of this new algorithm are illustrated by some analyses and simulation examples.

AB - In the present paper, we propose a new algorithm named dynamic topology representing Networks (DTRN) for learning both topology and clustering information from input data. In contrast to other models with adaptive architecture of this kind, the DTRN algorithm adaptively grows the number of output nodes by applying a vigilance test. The clustering procedure is based on a winner-take-quota learning strategy in conjunction with an annealing process in order to minimize the associated mean square error. A competitive Hebbian rule is applied to learn the global topology information concurrently with the clustering process. The topology information learned is also utilized for dynamically deleting the nodes and for the annealing process. The specific properties of this new algorithm are illustrated by some analyses and simulation examples.

UR - http://www.scopus.com/inward/record.url?scp=0031628740&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031628740&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0031628740

VL - 1

SP - 353

EP - 358

BT - IEEE International Conference on Neural Networks - Conference Proceedings

A2 - Anon, null

PB - IEEE

CY - Piscataway, NJ, United States

ER -