A Weight evolution algorithm with deterministic perturbation

Sin-Chun Ng; Shu Hung Leung; Andrew Luk

doi:10.1109/IJCNN.2000.857834

A Weight evolution algorithm with deterministic perturbation

Sin-Chun Ng
, Shu Hung Leung
, Andrew Luk

Department of Electrical Engineering

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

1 Citation (Scopus)

Abstract

This paper introduces a new learning algorithm for multi-layered feedforward network - weight evolution algorithm with deterministic perturbation. During the learning phase of the gradient algorithm (such as back-propagation), the network weights are adjusted intentionally in order to have an improvement in system performance. The intention is to reduce the overall system error after every weight update. By looking at the error component, it is possible to adjust some of the network weights deterministically so as to have an overall reduction in system error. Using the deterministic perturbation, it is found that the weight evolution between the hidden and output layer can accelerate the convergence speed, whereas the weight evolution between the input layer and the hidden layer can assist in solving the local minima problem. © 2000 IEEE

Original language	English
Title of host publication	Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks
Subtitle of host publication	IJCNN 2000
Publisher	IEEE
Pages	185-190
Volume	1
ISBN (Print)	0-7695-0619-4
DOIs	https://doi.org/10.1109/IJCNN.2000.857834
Publication status	Published - Jul 2000
Event	International Joint Conference on Neural Networks (IJCNN'2000) - Como, Italy Duration: 24 Jul 2000 → 27 Jul 2000 https://ieeexplore.ieee.org/xpl/tocresult.jsp?isnumber=18621

Publication series

Name
ISSN (Print)	1098-7576

Conference

Conference	International Joint Conference on Neural Networks (IJCNN'2000)
Place	Italy
City	Como
Period	24/07/00 → 27/07/00
Internet address	https://ieeexplore.ieee.org/xpl/tocresult.jsp?isnumber=18621

Access Output

10.1109/IJCNN.2000.857834

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{a0195140ef2c4095a8d7869e5d44bc3e,

title = "A Weight evolution algorithm with deterministic perturbation",

abstract = "This paper introduces a new learning algorithm for multi-layered feedforward network - weight evolution algorithm with deterministic perturbation. During the learning phase of the gradient algorithm (such as back-propagation), the network weights are adjusted intentionally in order to have an improvement in system performance. The intention is to reduce the overall system error after every weight update. By looking at the error component, it is possible to adjust some of the network weights deterministically so as to have an overall reduction in system error. Using the deterministic perturbation, it is found that the weight evolution between the hidden and output layer can accelerate the convergence speed, whereas the weight evolution between the input layer and the hidden layer can assist in solving the local minima problem. {\textcopyright} 2000 IEEE",

author = "Sin-Chun Ng and Leung, \{Shu Hung\} and Andrew Luk",

year = "2000",

month = jul,

doi = "10.1109/IJCNN.2000.857834",

language = "English",

isbn = "0-7695-0619-4",

volume = "1",

publisher = "IEEE",

pages = "185--190",

booktitle = "Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks",

address = "United States",

note = "International Joint Conference on Neural Networks (IJCNN'2000) ; Conference date: 24-07-2000 Through 27-07-2000",

url = "https://ieeexplore.ieee.org/xpl/tocresult.jsp?isnumber=18621",

}

A Weight evolution algorithm with deterministic perturbation. / Ng, Sin-Chun; Leung, Shu Hung; Luk, Andrew.
Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks: IJCNN 2000. Vol. 1 IEEE, 2000. p. 185-190.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - A Weight evolution algorithm with deterministic perturbation

AU - Ng, Sin-Chun

AU - Leung, Shu Hung

AU - Luk, Andrew

PY - 2000/7

Y1 - 2000/7

N2 - This paper introduces a new learning algorithm for multi-layered feedforward network - weight evolution algorithm with deterministic perturbation. During the learning phase of the gradient algorithm (such as back-propagation), the network weights are adjusted intentionally in order to have an improvement in system performance. The intention is to reduce the overall system error after every weight update. By looking at the error component, it is possible to adjust some of the network weights deterministically so as to have an overall reduction in system error. Using the deterministic perturbation, it is found that the weight evolution between the hidden and output layer can accelerate the convergence speed, whereas the weight evolution between the input layer and the hidden layer can assist in solving the local minima problem. © 2000 IEEE

AB - This paper introduces a new learning algorithm for multi-layered feedforward network - weight evolution algorithm with deterministic perturbation. During the learning phase of the gradient algorithm (such as back-propagation), the network weights are adjusted intentionally in order to have an improvement in system performance. The intention is to reduce the overall system error after every weight update. By looking at the error component, it is possible to adjust some of the network weights deterministically so as to have an overall reduction in system error. Using the deterministic perturbation, it is found that the weight evolution between the hidden and output layer can accelerate the convergence speed, whereas the weight evolution between the input layer and the hidden layer can assist in solving the local minima problem. © 2000 IEEE

UR - https://www.scopus.com/pages/publications/0033712526

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-0033712526&origin=recordpage

U2 - 10.1109/IJCNN.2000.857834

DO - 10.1109/IJCNN.2000.857834

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 0-7695-0619-4

VL - 1

SP - 185

EP - 190

BT - Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks

PB - IEEE

T2 - International Joint Conference on Neural Networks (IJCNN'2000)

Y2 - 24 July 2000 through 27 July 2000

ER -

A Weight evolution algorithm with deterministic perturbation

Abstract

Publication series

Conference

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this