Assessing Self-Repair on FPGAs with Biologically Realistic Astrocyte-Neuron Networks

Karim Shvan, Jim Harkin, LJ McDaid, Bryan Gardiner, Junxiu Liu, David Halliday, Andy Tyrrell, Jon Timmis, Alan Millard, Anju Johnson

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

This paper presents a hardware based implementation of a biologically-faithful astrocyte-based selfrepairing mechanism for Spiking Neural Networks. Spiking Astrocyte-neuron Networks (SANNs) are a new computing paradigm which capture the key mechanisms of how the human brain performs repairs. Using SANN in hardware affords the potential for realizing computing architecture that can self-repair. This paper demonstrates that Spiking Astrocyte Neural Network (SANN) in hardware have a resilience to significant levels of faults. The key novelty of the paper resides in implementing an SANN on FPGAs using fixed-point representation and demonstrating graceful performance degradation to different levels of injected faults via its self-repair capability. A fixed-point implementation of astrocyte, neurons and tripartite synapses are presented and compared against previous hardware floating-point and Matlab software implementations of SANN. All results are obtained from the SANN FPGA implementation and show how the reduced fixedpoint representation can maintain the biologically-realistic repair capability.
LanguageEnglish
Title of host publicationUnknown Host Publication
Pages421-426
Number of pages6
DOIs
Publication statusE-pub ahead of print - 20 Jul 2017
EventIEEE Computer Society Annual Symposium on VLSI (ISVLSI) - Bochum
Duration: 20 Jul 2017 → …

Conference

ConferenceIEEE Computer Society Annual Symposium on VLSI (ISVLSI)
Period20/07/17 → …

Fingerprint

Neurons
Field programmable gate arrays (FPGA)
Repair
Hardware
Neural networks
Astrocytes
Brain
Degradation

Keywords

  • Field programmable gate arrays
  • Hardware
  • Maintenance engineering
  • Mathematical model
  • Neurons
  • Astrocytes
  • Bio-inspired computing
  • FPGA
  • Self-repair
  • Spiking neural networks

Cite this

Shvan, Karim ; Harkin, Jim ; McDaid, LJ ; Gardiner, Bryan ; Liu, Junxiu ; Halliday, David ; Tyrrell, Andy ; Timmis, Jon ; Millard, Alan ; Johnson, Anju. / Assessing Self-Repair on FPGAs with Biologically Realistic Astrocyte-Neuron Networks. Unknown Host Publication. 2017. pp. 421-426
@inproceedings{36d7408952e14419a29d259ae93fa5dc,
title = "Assessing Self-Repair on FPGAs with Biologically Realistic Astrocyte-Neuron Networks",
abstract = "This paper presents a hardware based implementation of a biologically-faithful astrocyte-based selfrepairing mechanism for Spiking Neural Networks. Spiking Astrocyte-neuron Networks (SANNs) are a new computing paradigm which capture the key mechanisms of how the human brain performs repairs. Using SANN in hardware affords the potential for realizing computing architecture that can self-repair. This paper demonstrates that Spiking Astrocyte Neural Network (SANN) in hardware have a resilience to significant levels of faults. The key novelty of the paper resides in implementing an SANN on FPGAs using fixed-point representation and demonstrating graceful performance degradation to different levels of injected faults via its self-repair capability. A fixed-point implementation of astrocyte, neurons and tripartite synapses are presented and compared against previous hardware floating-point and Matlab software implementations of SANN. All results are obtained from the SANN FPGA implementation and show how the reduced fixedpoint representation can maintain the biologically-realistic repair capability.",
keywords = "Field programmable gate arrays, Hardware, Maintenance engineering, Mathematical model, Neurons, Astrocytes, Bio-inspired computing, FPGA, Self-repair, Spiking neural networks",
author = "Karim Shvan and Jim Harkin and LJ McDaid and Bryan Gardiner and Junxiu Liu and David Halliday and Andy Tyrrell and Jon Timmis and Alan Millard and Anju Johnson",
year = "2017",
month = "7",
day = "20",
doi = "10.1109/ISVLSI.2017.80",
language = "English",
isbn = "978-1-5090-6763-3",
pages = "421--426",
booktitle = "Unknown Host Publication",

}

Shvan, K, Harkin, J, McDaid, LJ, Gardiner, B, Liu, J, Halliday, D, Tyrrell, A, Timmis, J, Millard, A & Johnson, A 2017, Assessing Self-Repair on FPGAs with Biologically Realistic Astrocyte-Neuron Networks. in Unknown Host Publication. pp. 421-426, IEEE Computer Society Annual Symposium on VLSI (ISVLSI), 20/07/17. https://doi.org/10.1109/ISVLSI.2017.80

Assessing Self-Repair on FPGAs with Biologically Realistic Astrocyte-Neuron Networks. / Shvan, Karim; Harkin, Jim; McDaid, LJ; Gardiner, Bryan; Liu, Junxiu; Halliday, David; Tyrrell, Andy; Timmis, Jon; Millard, Alan; Johnson, Anju.

Unknown Host Publication. 2017. p. 421-426.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Assessing Self-Repair on FPGAs with Biologically Realistic Astrocyte-Neuron Networks

AU - Shvan, Karim

AU - Harkin, Jim

AU - McDaid, LJ

AU - Gardiner, Bryan

AU - Liu, Junxiu

AU - Halliday, David

AU - Tyrrell, Andy

AU - Timmis, Jon

AU - Millard, Alan

AU - Johnson, Anju

PY - 2017/7/20

Y1 - 2017/7/20

N2 - This paper presents a hardware based implementation of a biologically-faithful astrocyte-based selfrepairing mechanism for Spiking Neural Networks. Spiking Astrocyte-neuron Networks (SANNs) are a new computing paradigm which capture the key mechanisms of how the human brain performs repairs. Using SANN in hardware affords the potential for realizing computing architecture that can self-repair. This paper demonstrates that Spiking Astrocyte Neural Network (SANN) in hardware have a resilience to significant levels of faults. The key novelty of the paper resides in implementing an SANN on FPGAs using fixed-point representation and demonstrating graceful performance degradation to different levels of injected faults via its self-repair capability. A fixed-point implementation of astrocyte, neurons and tripartite synapses are presented and compared against previous hardware floating-point and Matlab software implementations of SANN. All results are obtained from the SANN FPGA implementation and show how the reduced fixedpoint representation can maintain the biologically-realistic repair capability.

AB - This paper presents a hardware based implementation of a biologically-faithful astrocyte-based selfrepairing mechanism for Spiking Neural Networks. Spiking Astrocyte-neuron Networks (SANNs) are a new computing paradigm which capture the key mechanisms of how the human brain performs repairs. Using SANN in hardware affords the potential for realizing computing architecture that can self-repair. This paper demonstrates that Spiking Astrocyte Neural Network (SANN) in hardware have a resilience to significant levels of faults. The key novelty of the paper resides in implementing an SANN on FPGAs using fixed-point representation and demonstrating graceful performance degradation to different levels of injected faults via its self-repair capability. A fixed-point implementation of astrocyte, neurons and tripartite synapses are presented and compared against previous hardware floating-point and Matlab software implementations of SANN. All results are obtained from the SANN FPGA implementation and show how the reduced fixedpoint representation can maintain the biologically-realistic repair capability.

KW - Field programmable gate arrays

KW - Hardware

KW - Maintenance engineering

KW - Mathematical model

KW - Neurons

KW - Astrocytes

KW - Bio-inspired computing

KW - FPGA

KW - Self-repair

KW - Spiking neural networks

U2 - 10.1109/ISVLSI.2017.80

DO - 10.1109/ISVLSI.2017.80

M3 - Conference contribution

SN - 978-1-5090-6763-3

SP - 421

EP - 426

BT - Unknown Host Publication

ER -