A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing

Siddhaling Urolagin, K. V. Prema, N. V.Subba Reddy

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

Abstract

The artificial neural network is typically trained from initial weight/bias position. As training progresses the network parameters such as weights and biases are updated according to learning algorithm to reduce the performance index. Not all the network parameters are equally learning the input-output mapping. Some parameters would hold more discriminating capability while others are not so effective. We propose a novel method of measuring the learning capability of a network parameter. The learning capability for a parameter we call it as learnability is contribution of that parameter to reduce performance index as the network is training. The proposed method of measuring learnability is applied on network parameters freezing on feedforward neural network. Our method is validated on MNIST handwritten numeral database using backpropagation learning algorithm.

Original languageEnglish
Title of host publicationProceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007
Pages242-249
Number of pages8
Volume1
DOIs
Publication statusPublished - 31-03-2008
EventInternational Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007 - Sivakasi, Tamil Nadu, India
Duration: 13-12-200715-12-2007

Conference

ConferenceInternational Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007
CountryIndia
CitySivakasi, Tamil Nadu
Period13-12-0715-12-07

Fingerprint

Freezing
Learning algorithms
Neural networks
Backpropagation algorithms
Feedforward neural networks

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Electrical and Electronic Engineering
  • Media Technology

Cite this

Urolagin, S., Prema, K. V., & Reddy, N. V. S. (2008). A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing. In Proceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007 (Vol. 1, pp. 242-249). [4426587] https://doi.org/10.1109/ICCIMA.2007.35
Urolagin, Siddhaling ; Prema, K. V. ; Reddy, N. V.Subba. / A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing. Proceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007. Vol. 1 2008. pp. 242-249
@inproceedings{aa8cbfad114948fb9a28d29f106e5f96,
title = "A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing",
abstract = "The artificial neural network is typically trained from initial weight/bias position. As training progresses the network parameters such as weights and biases are updated according to learning algorithm to reduce the performance index. Not all the network parameters are equally learning the input-output mapping. Some parameters would hold more discriminating capability while others are not so effective. We propose a novel method of measuring the learning capability of a network parameter. The learning capability for a parameter we call it as learnability is contribution of that parameter to reduce performance index as the network is training. The proposed method of measuring learnability is applied on network parameters freezing on feedforward neural network. Our method is validated on MNIST handwritten numeral database using backpropagation learning algorithm.",
author = "Siddhaling Urolagin and Prema, {K. V.} and Reddy, {N. V.Subba}",
year = "2008",
month = "3",
day = "31",
doi = "10.1109/ICCIMA.2007.35",
language = "English",
isbn = "0769530508",
volume = "1",
pages = "242--249",
booktitle = "Proceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007",

}

Urolagin, S, Prema, KV & Reddy, NVS 2008, A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing. in Proceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007. vol. 1, 4426587, pp. 242-249, International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007, Sivakasi, Tamil Nadu, India, 13-12-07. https://doi.org/10.1109/ICCIMA.2007.35

A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing. / Urolagin, Siddhaling; Prema, K. V.; Reddy, N. V.Subba.

Proceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007. Vol. 1 2008. p. 242-249 4426587.

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

TY - GEN

T1 - A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing

AU - Urolagin, Siddhaling

AU - Prema, K. V.

AU - Reddy, N. V.Subba

PY - 2008/3/31

Y1 - 2008/3/31

N2 - The artificial neural network is typically trained from initial weight/bias position. As training progresses the network parameters such as weights and biases are updated according to learning algorithm to reduce the performance index. Not all the network parameters are equally learning the input-output mapping. Some parameters would hold more discriminating capability while others are not so effective. We propose a novel method of measuring the learning capability of a network parameter. The learning capability for a parameter we call it as learnability is contribution of that parameter to reduce performance index as the network is training. The proposed method of measuring learnability is applied on network parameters freezing on feedforward neural network. Our method is validated on MNIST handwritten numeral database using backpropagation learning algorithm.

AB - The artificial neural network is typically trained from initial weight/bias position. As training progresses the network parameters such as weights and biases are updated according to learning algorithm to reduce the performance index. Not all the network parameters are equally learning the input-output mapping. Some parameters would hold more discriminating capability while others are not so effective. We propose a novel method of measuring the learning capability of a network parameter. The learning capability for a parameter we call it as learnability is contribution of that parameter to reduce performance index as the network is training. The proposed method of measuring learnability is applied on network parameters freezing on feedforward neural network. Our method is validated on MNIST handwritten numeral database using backpropagation learning algorithm.

UR - http://www.scopus.com/inward/record.url?scp=41149149273&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=41149149273&partnerID=8YFLogxK

U2 - 10.1109/ICCIMA.2007.35

DO - 10.1109/ICCIMA.2007.35

M3 - Conference contribution

SN - 0769530508

SN - 9780769530505

VL - 1

SP - 242

EP - 249

BT - Proceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007

ER -

Urolagin S, Prema KV, Reddy NVS. A novel method to measure the learning capability of a parameter in artificial neural network with application to network freezing. In Proceedings - International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2007. Vol. 1. 2008. p. 242-249. 4426587 https://doi.org/10.1109/ICCIMA.2007.35