A Reconfigurable and Biologically Inspired Paradigm for Computation using Networks-on-chip and Spiking Neural Networks

Jim Harkin, F Morgan, Liam McDaid, S Hall, B McGinely, S Cawley

Research output: Contribution to journalArticle

Abstract

FPGA devices have emerged as a popular platform for the rapid prototyping of biological Spiking Neural Networks (SNNs) applications, offering the key requirement of reconfigurability. However, FPGAs do not efficiently realise the biologically plausible neuron and synaptic models of SNNs, and current FPGA routing structures cannot accommodate the high levels of interneuron connectivity inherent in complex SNNs. This paper highlights and discusses the current challenges of implementing scalable SNNs on reconfigurable FPGAs. The paper proposes a novel field programmable neural network architecture (EMBRACE), incorporating low-power analogue spiking neurons, interconnected using a Network-on-Chip architecture. Results on the evaluation of the EMBRACE architecture using the XOR benchmark problem are presented, and the performance of the architecture is discussed. The paper also discusses the adaptability of the EMBRACE architecture in supporting fault tolerant computing.
LanguageEnglish
Article number908740
Number of pages14
JournalInternational Journal of Reconfigurable Computing
Volume7
Issue number2
DOIs
Publication statusPublished - 1 Jun 2009

Fingerprint

Field programmable gate arrays (FPGA)
Neural networks
Neurons
Fault tolerant computer systems
Rapid prototyping
Network architecture
Network-on-chip

Cite this

@article{5f6782d806a14dc7a6f6af4fe649bcb1,
title = "A Reconfigurable and Biologically Inspired Paradigm for Computation using Networks-on-chip and Spiking Neural Networks",
abstract = "FPGA devices have emerged as a popular platform for the rapid prototyping of biological Spiking Neural Networks (SNNs) applications, offering the key requirement of reconfigurability. However, FPGAs do not efficiently realise the biologically plausible neuron and synaptic models of SNNs, and current FPGA routing structures cannot accommodate the high levels of interneuron connectivity inherent in complex SNNs. This paper highlights and discusses the current challenges of implementing scalable SNNs on reconfigurable FPGAs. The paper proposes a novel field programmable neural network architecture (EMBRACE), incorporating low-power analogue spiking neurons, interconnected using a Network-on-Chip architecture. Results on the evaluation of the EMBRACE architecture using the XOR benchmark problem are presented, and the performance of the architecture is discussed. The paper also discusses the adaptability of the EMBRACE architecture in supporting fault tolerant computing.",
author = "Jim Harkin and F Morgan and Liam McDaid and S Hall and B McGinely and S Cawley",
year = "2009",
month = "6",
day = "1",
doi = "10.1155/2009/908740",
language = "English",
volume = "7",
journal = "International Journal of Reconfigurable Computing",
issn = "1687-7195",
number = "2",

}

A Reconfigurable and Biologically Inspired Paradigm for Computation using Networks-on-chip and Spiking Neural Networks. / Harkin, Jim; Morgan, F; McDaid, Liam; Hall, S; McGinely, B; Cawley, S.

In: International Journal of Reconfigurable Computing, Vol. 7, No. 2, 908740, 01.06.2009.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A Reconfigurable and Biologically Inspired Paradigm for Computation using Networks-on-chip and Spiking Neural Networks

AU - Harkin, Jim

AU - Morgan, F

AU - McDaid, Liam

AU - Hall, S

AU - McGinely, B

AU - Cawley, S

PY - 2009/6/1

Y1 - 2009/6/1

N2 - FPGA devices have emerged as a popular platform for the rapid prototyping of biological Spiking Neural Networks (SNNs) applications, offering the key requirement of reconfigurability. However, FPGAs do not efficiently realise the biologically plausible neuron and synaptic models of SNNs, and current FPGA routing structures cannot accommodate the high levels of interneuron connectivity inherent in complex SNNs. This paper highlights and discusses the current challenges of implementing scalable SNNs on reconfigurable FPGAs. The paper proposes a novel field programmable neural network architecture (EMBRACE), incorporating low-power analogue spiking neurons, interconnected using a Network-on-Chip architecture. Results on the evaluation of the EMBRACE architecture using the XOR benchmark problem are presented, and the performance of the architecture is discussed. The paper also discusses the adaptability of the EMBRACE architecture in supporting fault tolerant computing.

AB - FPGA devices have emerged as a popular platform for the rapid prototyping of biological Spiking Neural Networks (SNNs) applications, offering the key requirement of reconfigurability. However, FPGAs do not efficiently realise the biologically plausible neuron and synaptic models of SNNs, and current FPGA routing structures cannot accommodate the high levels of interneuron connectivity inherent in complex SNNs. This paper highlights and discusses the current challenges of implementing scalable SNNs on reconfigurable FPGAs. The paper proposes a novel field programmable neural network architecture (EMBRACE), incorporating low-power analogue spiking neurons, interconnected using a Network-on-Chip architecture. Results on the evaluation of the EMBRACE architecture using the XOR benchmark problem are presented, and the performance of the architecture is discussed. The paper also discusses the adaptability of the EMBRACE architecture in supporting fault tolerant computing.

U2 - 10.1155/2009/908740

DO - 10.1155/2009/908740

M3 - Article

VL - 7

JO - International Journal of Reconfigurable Computing

T2 - International Journal of Reconfigurable Computing

JF - International Journal of Reconfigurable Computing

SN - 1687-7195

IS - 2

M1 - 908740

ER -