Realization of fault tolerance for spiking neural networks with particle swarm optimization

Ruibin Feng, Chi-Sing Leung*, Peter Tsang

*Corresponding author for this work

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

Abstract

The spiking neural network (SNN) model has been an important topic in the past two decades. Many training algorithms, such as SpikeProp, were designed and applied to various applications. However, the fault tolerant ability in SNNs was not fully understood. Based on our study, the SNN model with the classical training objective function cannot even handle the single fault situation, in which one of the hidden neurons is damage. To improve the fault tolerant ability, we design an objective function and utilize the particle swarm optimization approach to minimize it. Simulation results show that our approach is much better than the classical objective function. © Springer International Publishing Switzerland 2015
Original languageEnglish
Title of host publicationNeural Information Processing
Subtitle of host publication22nd International Conference, ICONIP 2015, Istanbul, Turkey, November 9-12, 2015, Proceedings
Editors Sabri Arik, Tingwen Huang, Weng Kin Lai, Qingshan Liu
Place of PublicationCham
PublisherSpringer 
Pages79-86
VolumePart II
ISBN (Electronic)978-3-319-26535-3
ISBN (Print)9783319265346
DOIs
Publication statusPublished - 2015
Event22nd International Conference on Neural Information Processing (ICONIP 2015) - Istanbul, Türkiye
Duration: 9 Nov 201512 Nov 2015

Publication series

NameLecture Notes in Computer Science
Volume9490
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on Neural Information Processing (ICONIP 2015)
PlaceTürkiye
CityIstanbul
Period9/11/1512/11/15

Research Keywords

  • Fault tolerance
  • Particle swarm optimization
  • Spiking neural networks

Fingerprint

Dive into the research topics of 'Realization of fault tolerance for spiking neural networks with particle swarm optimization'. Together they form a unique fingerprint.

Cite this