Low-energy robust neuromorphic computation using synaptic devices

Duygu Kuzum, Rakesh Gnana David Jeyasingh, Shimeng Yu, H. S Philip Wong

Research output: Contribution to journalArticle

43 Citations (Scopus)

Abstract

Brain-inspired computing is an emerging field, which aims to reach brainlike performance in real-time processing of sensory data. The challenges that need to be addressed toward reaching such a computational system include building a compact massively parallel architecture with scalable interconnection devices, ultralow-power consumption, and robust neuromorphic computational schemes for implementation of learning in hardware. In this paper, we discuss programming strategies, material characteristics, and spike schemes, which enable implementation of symmetric and asymmetric synaptic plasticity with devices using phase-change materials. We demonstrate that energy consumption can be optimized by tuning the device operation regime and the spike scheme. Our simulations illustrate that a crossbar array consisting of synaptic devices and neurons can achieve hippocampus-like associative learning with symmetric synapses and sequence learning with asymmetric synapses. Pattern completion for patterns with 50% missing elements is achieved via associative learning with symmetric plasticity. Robustness of learning against input noise, variation in sensory data, and device resistance variation are investigated through simulations.

Original languageEnglish (US)
Article number6340321
Pages (from-to)3489-3494
Number of pages6
JournalIEEE Transactions on Electron Devices
Volume59
Issue number12
DOIs
StatePublished - 2012
Externally publishedYes

Fingerprint

Plasticity
Phase change materials
Parallel architectures
Neurons
Brain
Electric power utilization
Energy utilization
Tuning
Hardware
Processing

Keywords

  • Hopfield network
  • neuromorphic
  • phase-change materials
  • plasticity
  • spike-timing-dependent plasticity (STDP)
  • synaptic device

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Electronic, Optical and Magnetic Materials

Cite this

Kuzum, D., Jeyasingh, R. G. D., Yu, S., & Wong, H. S. P. (2012). Low-energy robust neuromorphic computation using synaptic devices. IEEE Transactions on Electron Devices, 59(12), 3489-3494. [6340321]. https://doi.org/10.1109/TED.2012.2217146

Low-energy robust neuromorphic computation using synaptic devices. / Kuzum, Duygu; Jeyasingh, Rakesh Gnana David; Yu, Shimeng; Wong, H. S Philip.

In: IEEE Transactions on Electron Devices, Vol. 59, No. 12, 6340321, 2012, p. 3489-3494.

Research output: Contribution to journalArticle

Kuzum, D, Jeyasingh, RGD, Yu, S & Wong, HSP 2012, 'Low-energy robust neuromorphic computation using synaptic devices', IEEE Transactions on Electron Devices, vol. 59, no. 12, 6340321, pp. 3489-3494. https://doi.org/10.1109/TED.2012.2217146
Kuzum, Duygu ; Jeyasingh, Rakesh Gnana David ; Yu, Shimeng ; Wong, H. S Philip. / Low-energy robust neuromorphic computation using synaptic devices. In: IEEE Transactions on Electron Devices. 2012 ; Vol. 59, No. 12. pp. 3489-3494.
@article{d5525e2a5e084ddcacb7760108782359,
title = "Low-energy robust neuromorphic computation using synaptic devices",
abstract = "Brain-inspired computing is an emerging field, which aims to reach brainlike performance in real-time processing of sensory data. The challenges that need to be addressed toward reaching such a computational system include building a compact massively parallel architecture with scalable interconnection devices, ultralow-power consumption, and robust neuromorphic computational schemes for implementation of learning in hardware. In this paper, we discuss programming strategies, material characteristics, and spike schemes, which enable implementation of symmetric and asymmetric synaptic plasticity with devices using phase-change materials. We demonstrate that energy consumption can be optimized by tuning the device operation regime and the spike scheme. Our simulations illustrate that a crossbar array consisting of synaptic devices and neurons can achieve hippocampus-like associative learning with symmetric synapses and sequence learning with asymmetric synapses. Pattern completion for patterns with 50{\%} missing elements is achieved via associative learning with symmetric plasticity. Robustness of learning against input noise, variation in sensory data, and device resistance variation are investigated through simulations.",
keywords = "Hopfield network, neuromorphic, phase-change materials, plasticity, spike-timing-dependent plasticity (STDP), synaptic device",
author = "Duygu Kuzum and Jeyasingh, {Rakesh Gnana David} and Shimeng Yu and Wong, {H. S Philip}",
year = "2012",
doi = "10.1109/TED.2012.2217146",
language = "English (US)",
volume = "59",
pages = "3489--3494",
journal = "IEEE Transactions on Electron Devices",
issn = "0018-9383",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "12",

}

TY - JOUR

T1 - Low-energy robust neuromorphic computation using synaptic devices

AU - Kuzum, Duygu

AU - Jeyasingh, Rakesh Gnana David

AU - Yu, Shimeng

AU - Wong, H. S Philip

PY - 2012

Y1 - 2012

N2 - Brain-inspired computing is an emerging field, which aims to reach brainlike performance in real-time processing of sensory data. The challenges that need to be addressed toward reaching such a computational system include building a compact massively parallel architecture with scalable interconnection devices, ultralow-power consumption, and robust neuromorphic computational schemes for implementation of learning in hardware. In this paper, we discuss programming strategies, material characteristics, and spike schemes, which enable implementation of symmetric and asymmetric synaptic plasticity with devices using phase-change materials. We demonstrate that energy consumption can be optimized by tuning the device operation regime and the spike scheme. Our simulations illustrate that a crossbar array consisting of synaptic devices and neurons can achieve hippocampus-like associative learning with symmetric synapses and sequence learning with asymmetric synapses. Pattern completion for patterns with 50% missing elements is achieved via associative learning with symmetric plasticity. Robustness of learning against input noise, variation in sensory data, and device resistance variation are investigated through simulations.

AB - Brain-inspired computing is an emerging field, which aims to reach brainlike performance in real-time processing of sensory data. The challenges that need to be addressed toward reaching such a computational system include building a compact massively parallel architecture with scalable interconnection devices, ultralow-power consumption, and robust neuromorphic computational schemes for implementation of learning in hardware. In this paper, we discuss programming strategies, material characteristics, and spike schemes, which enable implementation of symmetric and asymmetric synaptic plasticity with devices using phase-change materials. We demonstrate that energy consumption can be optimized by tuning the device operation regime and the spike scheme. Our simulations illustrate that a crossbar array consisting of synaptic devices and neurons can achieve hippocampus-like associative learning with symmetric synapses and sequence learning with asymmetric synapses. Pattern completion for patterns with 50% missing elements is achieved via associative learning with symmetric plasticity. Robustness of learning against input noise, variation in sensory data, and device resistance variation are investigated through simulations.

KW - Hopfield network

KW - neuromorphic

KW - phase-change materials

KW - plasticity

KW - spike-timing-dependent plasticity (STDP)

KW - synaptic device

UR - http://www.scopus.com/inward/record.url?scp=84870294722&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84870294722&partnerID=8YFLogxK

U2 - 10.1109/TED.2012.2217146

DO - 10.1109/TED.2012.2217146

M3 - Article

AN - SCOPUS:84870294722

VL - 59

SP - 3489

EP - 3494

JO - IEEE Transactions on Electron Devices

JF - IEEE Transactions on Electron Devices

SN - 0018-9383

IS - 12

M1 - 6340321

ER -