Step forward to map fully parallel energy efficient cortical columns on field programmable gate arrays

Arfan Ghani, Chan H. See, Syed M. Usman Ali

Research output: Contribution to journalArticle

2 Citations (Scopus)
15 Downloads (Pure)

Abstract

This study presents energy and area-efficient hardware architectures to map fully parallel cortical columns on reconfigurable platform - field programmable gate arrays (FPGAs). An area-efficient architecture is proposed at the system level and benchmarked with a speech recognition application. Owing to the spatio-temporal nature of spiking neurons it is more suitable to map such architectures on FPGAs where signals can be represented in binary form and communication can be performed through the use of spikes. The viability of implementing multiple recurrent neural reservoirs is demonstrated with a novel multiplier-less reconfigurable architectures and a design strategy is devised for its implementation.

Original languageEnglish
Pages (from-to)432-440
Number of pages9
JournalIET Science, Measurement and Technology
Volume8
Issue number6
DOIs
Publication statusPublished - 1 Nov 2014
Externally publishedYes

Fingerprint

field-programmable gate arrays
Field programmable gate arrays (FPGA)
Reconfigurable architectures
Speech recognition
Neurons
Hardware
spiking
energy
Communication
multipliers
speech recognition
neurons
viability
spikes
hardware
platforms
communication

Bibliographical note

© 2014 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Keywords

  • speech recognition
  • field programmable gate arrays
  • neurophysiology
  • reconfigurable architectures
  • recurrent neural nets
  • fully parallel energy efficient cortical columns
  • multiplier-less reconfigurable architectures
  • multiple recurrent neural reservoirs
  • spiking neurons
  • speech recognition application
  • FPGAs
  • reconfigurable platform
  • area-efficient hardware architectures

ASJC Scopus subject areas

  • Atomic and Molecular Physics, and Optics
  • Electrical and Electronic Engineering

Cite this

Step forward to map fully parallel energy efficient cortical columns on field programmable gate arrays. / Ghani, Arfan; See, Chan H.; Usman Ali, Syed M.

In: IET Science, Measurement and Technology, Vol. 8, No. 6, 01.11.2014, p. 432-440.

Research output: Contribution to journalArticle

@article{7fe3473c0ada4be1a53f8ac0870fdcaf,
title = "Step forward to map fully parallel energy efficient cortical columns on field programmable gate arrays",
abstract = "This study presents energy and area-efficient hardware architectures to map fully parallel cortical columns on reconfigurable platform - field programmable gate arrays (FPGAs). An area-efficient architecture is proposed at the system level and benchmarked with a speech recognition application. Owing to the spatio-temporal nature of spiking neurons it is more suitable to map such architectures on FPGAs where signals can be represented in binary form and communication can be performed through the use of spikes. The viability of implementing multiple recurrent neural reservoirs is demonstrated with a novel multiplier-less reconfigurable architectures and a design strategy is devised for its implementation.",
keywords = "speech recognition, field programmable gate arrays, neurophysiology, reconfigurable architectures, recurrent neural nets , fully parallel energy efficient cortical columns, multiplier-less reconfigurable architectures, multiple recurrent neural reservoirs, spiking neurons, speech recognition application, FPGAs, reconfigurable platform, area-efficient hardware architectures",
author = "Arfan Ghani and See, {Chan H.} and {Usman Ali}, {Syed M.}",
note = "{\circledC} 2014 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.",
year = "2014",
month = "11",
day = "1",
doi = "10.1049/iet-smt.2014.0004",
language = "English",
volume = "8",
pages = "432--440",
journal = "IET Science, Measurement and Technology",
issn = "1751-8822",
publisher = "Institution of Engineering and Technology",
number = "6",

}

TY - JOUR

T1 - Step forward to map fully parallel energy efficient cortical columns on field programmable gate arrays

AU - Ghani, Arfan

AU - See, Chan H.

AU - Usman Ali, Syed M.

N1 - © 2014 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2014/11/1

Y1 - 2014/11/1

N2 - This study presents energy and area-efficient hardware architectures to map fully parallel cortical columns on reconfigurable platform - field programmable gate arrays (FPGAs). An area-efficient architecture is proposed at the system level and benchmarked with a speech recognition application. Owing to the spatio-temporal nature of spiking neurons it is more suitable to map such architectures on FPGAs where signals can be represented in binary form and communication can be performed through the use of spikes. The viability of implementing multiple recurrent neural reservoirs is demonstrated with a novel multiplier-less reconfigurable architectures and a design strategy is devised for its implementation.

AB - This study presents energy and area-efficient hardware architectures to map fully parallel cortical columns on reconfigurable platform - field programmable gate arrays (FPGAs). An area-efficient architecture is proposed at the system level and benchmarked with a speech recognition application. Owing to the spatio-temporal nature of spiking neurons it is more suitable to map such architectures on FPGAs where signals can be represented in binary form and communication can be performed through the use of spikes. The viability of implementing multiple recurrent neural reservoirs is demonstrated with a novel multiplier-less reconfigurable architectures and a design strategy is devised for its implementation.

KW - speech recognition

KW - field programmable gate arrays

KW - neurophysiology

KW - reconfigurable architectures

KW - recurrent neural nets

KW - fully parallel energy efficient cortical columns

KW - multiplier-less reconfigurable architectures

KW - multiple recurrent neural reservoirs

KW - spiking neurons

KW - speech recognition application

KW - FPGAs

KW - reconfigurable platform

KW - area-efficient hardware architectures

U2 - 10.1049/iet-smt.2014.0004

DO - 10.1049/iet-smt.2014.0004

M3 - Article

VL - 8

SP - 432

EP - 440

JO - IET Science, Measurement and Technology

JF - IET Science, Measurement and Technology

SN - 1751-8822

IS - 6

ER -