Low Cost Interconnected Architecture for the Hardware Spiking Neural Networks

Yuling Luo, Lei Wan, Junxiu Liu, Jim Harkin, LJ McDaid, Yi Cao, Xuemei Ding

Research output: Contribution to journalArticle

Abstract

A novel low cost interconnected architecture (LCIA) is proposed in this paper, which is an efficient solution for the neuron interconnections for the hardware spiking neural networks (SNNs). It is based on an all-to-all connection that takes each paired input and output nodes of multi-layer SNNs as the source and destination of connections. The aim is to maintain an efficient routing performance under low hardware overhead. A Networks-on-Chip (NoC) router is proposed as the fundamental component of the LCIA, where an effective scheduler is designed to address the traffic challenge due to irregular spikes. The router can find requests rapidly, make the arbitration decision promptly, and provide equal services to different network traffic requests. Experimental results show that the LCIA can manage the intercommunication of the multi-layer neural networks efficiently and have a low hardware overhead which can maintain the scalability of hardware SNNs.
LanguageEnglish
Article number857
Pages1-14
Number of pages14
JournalFrontiers in Neurosciences
Volume12
DOIs
Publication statusPublished - 21 Nov 2018

Fingerprint

Costs and Cost Analysis
Negotiating
Neurons

Keywords

  • interconnected architecture
  • Spiking Neural Networks
  • Networks-on-Chip
  • system scalability
  • arbitration scheme

Cite this

@article{15111c7b6c3d44a990477d0f39afb668,
title = "Low Cost Interconnected Architecture for the Hardware Spiking Neural Networks",
abstract = "A novel low cost interconnected architecture (LCIA) is proposed in this paper, which is an efficient solution for the neuron interconnections for the hardware spiking neural networks (SNNs). It is based on an all-to-all connection that takes each paired input and output nodes of multi-layer SNNs as the source and destination of connections. The aim is to maintain an efficient routing performance under low hardware overhead. A Networks-on-Chip (NoC) router is proposed as the fundamental component of the LCIA, where an effective scheduler is designed to address the traffic challenge due to irregular spikes. The router can find requests rapidly, make the arbitration decision promptly, and provide equal services to different network traffic requests. Experimental results show that the LCIA can manage the intercommunication of the multi-layer neural networks efficiently and have a low hardware overhead which can maintain the scalability of hardware SNNs.",
keywords = "interconnected architecture, Spiking Neural Networks, Networks-on-Chip, system scalability, arbitration scheme",
author = "Yuling Luo and Lei Wan and Junxiu Liu and Jim Harkin and LJ McDaid and Yi Cao and Xuemei Ding",
year = "2018",
month = "11",
day = "21",
doi = "10.3389/fnins.2018.00857",
language = "English",
volume = "12",
pages = "1--14",
journal = "Frontiers in Neuroscience",
issn = "1662-4548",

}

Low Cost Interconnected Architecture for the Hardware Spiking Neural Networks. / Luo, Yuling; Wan, Lei; Liu, Junxiu; Harkin, Jim; McDaid, LJ; Cao, Yi; Ding, Xuemei.

In: Frontiers in Neurosciences, Vol. 12, 857, 21.11.2018, p. 1-14.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Low Cost Interconnected Architecture for the Hardware Spiking Neural Networks

AU - Luo, Yuling

AU - Wan, Lei

AU - Liu, Junxiu

AU - Harkin, Jim

AU - McDaid, LJ

AU - Cao, Yi

AU - Ding, Xuemei

PY - 2018/11/21

Y1 - 2018/11/21

N2 - A novel low cost interconnected architecture (LCIA) is proposed in this paper, which is an efficient solution for the neuron interconnections for the hardware spiking neural networks (SNNs). It is based on an all-to-all connection that takes each paired input and output nodes of multi-layer SNNs as the source and destination of connections. The aim is to maintain an efficient routing performance under low hardware overhead. A Networks-on-Chip (NoC) router is proposed as the fundamental component of the LCIA, where an effective scheduler is designed to address the traffic challenge due to irregular spikes. The router can find requests rapidly, make the arbitration decision promptly, and provide equal services to different network traffic requests. Experimental results show that the LCIA can manage the intercommunication of the multi-layer neural networks efficiently and have a low hardware overhead which can maintain the scalability of hardware SNNs.

AB - A novel low cost interconnected architecture (LCIA) is proposed in this paper, which is an efficient solution for the neuron interconnections for the hardware spiking neural networks (SNNs). It is based on an all-to-all connection that takes each paired input and output nodes of multi-layer SNNs as the source and destination of connections. The aim is to maintain an efficient routing performance under low hardware overhead. A Networks-on-Chip (NoC) router is proposed as the fundamental component of the LCIA, where an effective scheduler is designed to address the traffic challenge due to irregular spikes. The router can find requests rapidly, make the arbitration decision promptly, and provide equal services to different network traffic requests. Experimental results show that the LCIA can manage the intercommunication of the multi-layer neural networks efficiently and have a low hardware overhead which can maintain the scalability of hardware SNNs.

KW - interconnected architecture

KW - Spiking Neural Networks

KW - Networks-on-Chip

KW - system scalability

KW - arbitration scheme

U2 - 10.3389/fnins.2018.00857

DO - 10.3389/fnins.2018.00857

M3 - Article

VL - 12

SP - 1

EP - 14

JO - Frontiers in Neuroscience

T2 - Frontiers in Neuroscience

JF - Frontiers in Neuroscience

SN - 1662-4548

M1 - 857

ER -