Abstract

The results of a set of simulation experiments conducted to quantify the effects of faults in a classification network implemented as a three-layered perception model are reported. The percentage of vectors misclassified by the classification network, the time taken for the network to stabilize, and the output values are measured. The results show that both transient and permanent faults have a significant impact on the performance of the network. Transient faults are also found to cause the network to be increasingly unstable as the duration of a transient is increased. The average percentage of the vectors misclassified is about 25%; after relearning, this is reduced to 10%. The impact of link faults is relatively insignificant in comparison with node faults (1% versus 19% misclassified after relearning). A study of the impact of hardware redundancy shows a linear increase in misclassifications with increasing hardware size.

Original languageEnglish (US)
Pages513-518
Number of pages6
StatePublished - Dec 1 1990
Event9th Digital Avionics Systems Conference - Virginia Beach, VA, USA
Duration: Oct 15 1990Oct 18 1990

Other

Other9th Digital Avionics Systems Conference
CityVirginia Beach, VA, USA
Period10/15/9010/18/90

Fingerprint

Neural networks
Hardware
Redundancy
Experiments

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Tan, C. H., & Iyer, R. K. (1990). Fault characterization of a multilayered perceptron network. 513-518. Paper presented at 9th Digital Avionics Systems Conference, Virginia Beach, VA, USA, .

Fault characterization of a multilayered perceptron network. / Tan, Chang H.; Iyer, Ravishankar K.

1990. 513-518 Paper presented at 9th Digital Avionics Systems Conference, Virginia Beach, VA, USA, .

Research output: Contribution to conferencePaper

Tan, CH & Iyer, RK 1990, 'Fault characterization of a multilayered perceptron network', Paper presented at 9th Digital Avionics Systems Conference, Virginia Beach, VA, USA, 10/15/90 - 10/18/90 pp. 513-518.
Tan CH, Iyer RK. Fault characterization of a multilayered perceptron network. 1990. Paper presented at 9th Digital Avionics Systems Conference, Virginia Beach, VA, USA, .
Tan, Chang H. ; Iyer, Ravishankar K. / Fault characterization of a multilayered perceptron network. Paper presented at 9th Digital Avionics Systems Conference, Virginia Beach, VA, USA, .6 p.
@conference{97b19720973b4970b4b579acba6e2987,
title = "Fault characterization of a multilayered perceptron network",
abstract = "The results of a set of simulation experiments conducted to quantify the effects of faults in a classification network implemented as a three-layered perception model are reported. The percentage of vectors misclassified by the classification network, the time taken for the network to stabilize, and the output values are measured. The results show that both transient and permanent faults have a significant impact on the performance of the network. Transient faults are also found to cause the network to be increasingly unstable as the duration of a transient is increased. The average percentage of the vectors misclassified is about 25{\%}; after relearning, this is reduced to 10{\%}. The impact of link faults is relatively insignificant in comparison with node faults (1{\%} versus 19{\%} misclassified after relearning). A study of the impact of hardware redundancy shows a linear increase in misclassifications with increasing hardware size.",
author = "Tan, {Chang H.} and Iyer, {Ravishankar K}",
year = "1990",
month = "12",
day = "1",
language = "English (US)",
pages = "513--518",
note = "9th Digital Avionics Systems Conference ; Conference date: 15-10-1990 Through 18-10-1990",

}

TY - CONF

T1 - Fault characterization of a multilayered perceptron network

AU - Tan, Chang H.

AU - Iyer, Ravishankar K

PY - 1990/12/1

Y1 - 1990/12/1

N2 - The results of a set of simulation experiments conducted to quantify the effects of faults in a classification network implemented as a three-layered perception model are reported. The percentage of vectors misclassified by the classification network, the time taken for the network to stabilize, and the output values are measured. The results show that both transient and permanent faults have a significant impact on the performance of the network. Transient faults are also found to cause the network to be increasingly unstable as the duration of a transient is increased. The average percentage of the vectors misclassified is about 25%; after relearning, this is reduced to 10%. The impact of link faults is relatively insignificant in comparison with node faults (1% versus 19% misclassified after relearning). A study of the impact of hardware redundancy shows a linear increase in misclassifications with increasing hardware size.

AB - The results of a set of simulation experiments conducted to quantify the effects of faults in a classification network implemented as a three-layered perception model are reported. The percentage of vectors misclassified by the classification network, the time taken for the network to stabilize, and the output values are measured. The results show that both transient and permanent faults have a significant impact on the performance of the network. Transient faults are also found to cause the network to be increasingly unstable as the duration of a transient is increased. The average percentage of the vectors misclassified is about 25%; after relearning, this is reduced to 10%. The impact of link faults is relatively insignificant in comparison with node faults (1% versus 19% misclassified after relearning). A study of the impact of hardware redundancy shows a linear increase in misclassifications with increasing hardware size.

UR - http://www.scopus.com/inward/record.url?scp=0025563774&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0025563774&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0025563774

SP - 513

EP - 518

ER -