DL-ReSuMe: A Delay Learning-Based Remote Supervised Method for Spiking Neurons

Aboozar Taherkhani, Ammar Belatreche, Yuhua Li, Liam Maguire

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

Recent research has shown the potential capability of spiking neural networks (SNNs) to model complex information processing in the brain. There is biological evidence to prove the use of the precise timing of spikes for information coding. However, the exact learning mechanism in which the neuron is trained to fire at precise times remains an open problem. The majority of the existing learning methods for SNNs are based on weight adjustment. However, there is also biological evidence that the synaptic delay is not constant. In this paper, a learning method for spiking neurons, called delay learning remote supervised method (DL-ReSuMe), is proposed to merge the delay shift approach and ReSuMe-based weight adjustment to enhance the learning performance. DL-ReSuMe uses more biologically plausible properties, such as delay learning, and needs less weight adjustment than ReSuMe. Simulation results have shown that the proposed DL-ReSuMe approach achieves learning accuracy and learning speed improvements compared with ReSuMe.
LanguageEnglish
Pages3137-3149
JournalIEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
Volume26
Issue number12
DOIs
Publication statusPublished - Dec 2015

Fingerprint

Neurons
Neural networks
Brain
Fires

Keywords

  • Delay shift learning
  • spiking neuron
  • supervised learning
  • synaptic delay

Cite this

Taherkhani, Aboozar ; Belatreche, Ammar ; Li, Yuhua ; Maguire, Liam. / DL-ReSuMe: A Delay Learning-Based Remote Supervised Method for Spiking Neurons. 2015 ; Vol. 26, No. 12. pp. 3137-3149.
@article{62f929bbde7f4ec486f1d6c7a5ea77d7,
title = "DL-ReSuMe: A Delay Learning-Based Remote Supervised Method for Spiking Neurons",
abstract = "Recent research has shown the potential capability of spiking neural networks (SNNs) to model complex information processing in the brain. There is biological evidence to prove the use of the precise timing of spikes for information coding. However, the exact learning mechanism in which the neuron is trained to fire at precise times remains an open problem. The majority of the existing learning methods for SNNs are based on weight adjustment. However, there is also biological evidence that the synaptic delay is not constant. In this paper, a learning method for spiking neurons, called delay learning remote supervised method (DL-ReSuMe), is proposed to merge the delay shift approach and ReSuMe-based weight adjustment to enhance the learning performance. DL-ReSuMe uses more biologically plausible properties, such as delay learning, and needs less weight adjustment than ReSuMe. Simulation results have shown that the proposed DL-ReSuMe approach achieves learning accuracy and learning speed improvements compared with ReSuMe.",
keywords = "Delay shift learning, spiking neuron, supervised learning, synaptic delay",
author = "Aboozar Taherkhani and Ammar Belatreche and Yuhua Li and Liam Maguire",
year = "2015",
month = "12",
doi = "10.1109/TNNLS.2015.2404938",
language = "English",
volume = "26",
pages = "3137--3149",
number = "12",

}

DL-ReSuMe: A Delay Learning-Based Remote Supervised Method for Spiking Neurons. / Taherkhani, Aboozar; Belatreche, Ammar; Li, Yuhua; Maguire, Liam.

Vol. 26, No. 12, 12.2015, p. 3137-3149.

Research output: Contribution to journalArticle

TY - JOUR

T1 - DL-ReSuMe: A Delay Learning-Based Remote Supervised Method for Spiking Neurons

AU - Taherkhani, Aboozar

AU - Belatreche, Ammar

AU - Li, Yuhua

AU - Maguire, Liam

PY - 2015/12

Y1 - 2015/12

N2 - Recent research has shown the potential capability of spiking neural networks (SNNs) to model complex information processing in the brain. There is biological evidence to prove the use of the precise timing of spikes for information coding. However, the exact learning mechanism in which the neuron is trained to fire at precise times remains an open problem. The majority of the existing learning methods for SNNs are based on weight adjustment. However, there is also biological evidence that the synaptic delay is not constant. In this paper, a learning method for spiking neurons, called delay learning remote supervised method (DL-ReSuMe), is proposed to merge the delay shift approach and ReSuMe-based weight adjustment to enhance the learning performance. DL-ReSuMe uses more biologically plausible properties, such as delay learning, and needs less weight adjustment than ReSuMe. Simulation results have shown that the proposed DL-ReSuMe approach achieves learning accuracy and learning speed improvements compared with ReSuMe.

AB - Recent research has shown the potential capability of spiking neural networks (SNNs) to model complex information processing in the brain. There is biological evidence to prove the use of the precise timing of spikes for information coding. However, the exact learning mechanism in which the neuron is trained to fire at precise times remains an open problem. The majority of the existing learning methods for SNNs are based on weight adjustment. However, there is also biological evidence that the synaptic delay is not constant. In this paper, a learning method for spiking neurons, called delay learning remote supervised method (DL-ReSuMe), is proposed to merge the delay shift approach and ReSuMe-based weight adjustment to enhance the learning performance. DL-ReSuMe uses more biologically plausible properties, such as delay learning, and needs less weight adjustment than ReSuMe. Simulation results have shown that the proposed DL-ReSuMe approach achieves learning accuracy and learning speed improvements compared with ReSuMe.

KW - Delay shift learning

KW - spiking neuron

KW - supervised learning

KW - synaptic delay

U2 - 10.1109/TNNLS.2015.2404938

DO - 10.1109/TNNLS.2015.2404938

M3 - Article

VL - 26

SP - 3137

EP - 3149

IS - 12

ER -