A neuromorphic visual system using RRAM synaptic devices with Sub-pJ energy and tolerance to variability

Experimental characterization and large-scale modeling

Shimeng Yu, Bin Gao, Zheng Fang, Hongyu Yu, Jinfeng Kang, H. S Philip Wong

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

63 Citations (Scopus)

Abstract

We report the use of metal oxide resistive switching memory (RRAM) as synaptic devices for a neuromorphic visual system. At the device level, we experimentally characterized the gradual resistance modulation of RRAM by hundreds of identical pulses. As compared with phase change memory (PCM) reported recently in [1,2], >100×-1000× energy consumption reduction was achieved in RRAM as synaptic devices (<1 pJ per spike). Based on the experimental results, we developed a stochastic model to quantify the device switching dynamics. At the system level, we simulated the performance of image orientation selectivity on a neuromorphic visual system which consists of 1,024 CMOS neuron circuits and 16,348 RRAM synaptic devices. It was found that the system can tolerate the temporal and spatial variability which are commonly present in RRAM devices, suggesting the feasibility of large-scale hardware implementation of neuromorphic system using RRAM synaptic devices.

Original languageEnglish (US)
Title of host publicationTechnical Digest - International Electron Devices Meeting, IEDM
DOIs
StatePublished - 2012
Externally publishedYes
Event2012 IEEE International Electron Devices Meeting, IEDM 2012 - San Francisco, CA, United States
Duration: Dec 10 2012Dec 13 2012

Other

Other2012 IEEE International Electron Devices Meeting, IEDM 2012
CountryUnited States
CitySan Francisco, CA
Period12/10/1212/13/12

Fingerprint

energy
Phase change memory
Stochastic models
Oxides
energy consumption
Neurons
RRAM
neurons
spikes
Energy utilization
metal oxides
Metals
Modulation
CMOS
hardware
Hardware
Data storage equipment
selectivity
Networks (circuits)
modulation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics
  • Electronic, Optical and Magnetic Materials
  • Materials Chemistry

Cite this

Yu, S., Gao, B., Fang, Z., Yu, H., Kang, J., & Wong, H. S. P. (2012). A neuromorphic visual system using RRAM synaptic devices with Sub-pJ energy and tolerance to variability: Experimental characterization and large-scale modeling. In Technical Digest - International Electron Devices Meeting, IEDM [6479018] https://doi.org/10.1109/IEDM.2012.6479018

A neuromorphic visual system using RRAM synaptic devices with Sub-pJ energy and tolerance to variability : Experimental characterization and large-scale modeling. / Yu, Shimeng; Gao, Bin; Fang, Zheng; Yu, Hongyu; Kang, Jinfeng; Wong, H. S Philip.

Technical Digest - International Electron Devices Meeting, IEDM. 2012. 6479018.

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

Yu, S, Gao, B, Fang, Z, Yu, H, Kang, J & Wong, HSP 2012, A neuromorphic visual system using RRAM synaptic devices with Sub-pJ energy and tolerance to variability: Experimental characterization and large-scale modeling. in Technical Digest - International Electron Devices Meeting, IEDM., 6479018, 2012 IEEE International Electron Devices Meeting, IEDM 2012, San Francisco, CA, United States, 12/10/12. https://doi.org/10.1109/IEDM.2012.6479018
Yu, Shimeng ; Gao, Bin ; Fang, Zheng ; Yu, Hongyu ; Kang, Jinfeng ; Wong, H. S Philip. / A neuromorphic visual system using RRAM synaptic devices with Sub-pJ energy and tolerance to variability : Experimental characterization and large-scale modeling. Technical Digest - International Electron Devices Meeting, IEDM. 2012.
@inproceedings{c1df923f31fa4bee93759e88ae498119,
title = "A neuromorphic visual system using RRAM synaptic devices with Sub-pJ energy and tolerance to variability: Experimental characterization and large-scale modeling",
abstract = "We report the use of metal oxide resistive switching memory (RRAM) as synaptic devices for a neuromorphic visual system. At the device level, we experimentally characterized the gradual resistance modulation of RRAM by hundreds of identical pulses. As compared with phase change memory (PCM) reported recently in [1,2], >100×-1000× energy consumption reduction was achieved in RRAM as synaptic devices (<1 pJ per spike). Based on the experimental results, we developed a stochastic model to quantify the device switching dynamics. At the system level, we simulated the performance of image orientation selectivity on a neuromorphic visual system which consists of 1,024 CMOS neuron circuits and 16,348 RRAM synaptic devices. It was found that the system can tolerate the temporal and spatial variability which are commonly present in RRAM devices, suggesting the feasibility of large-scale hardware implementation of neuromorphic system using RRAM synaptic devices.",
author = "Shimeng Yu and Bin Gao and Zheng Fang and Hongyu Yu and Jinfeng Kang and Wong, {H. S Philip}",
year = "2012",
doi = "10.1109/IEDM.2012.6479018",
language = "English (US)",
isbn = "9781467348706",
booktitle = "Technical Digest - International Electron Devices Meeting, IEDM",

}

TY - GEN

T1 - A neuromorphic visual system using RRAM synaptic devices with Sub-pJ energy and tolerance to variability

T2 - Experimental characterization and large-scale modeling

AU - Yu, Shimeng

AU - Gao, Bin

AU - Fang, Zheng

AU - Yu, Hongyu

AU - Kang, Jinfeng

AU - Wong, H. S Philip

PY - 2012

Y1 - 2012

N2 - We report the use of metal oxide resistive switching memory (RRAM) as synaptic devices for a neuromorphic visual system. At the device level, we experimentally characterized the gradual resistance modulation of RRAM by hundreds of identical pulses. As compared with phase change memory (PCM) reported recently in [1,2], >100×-1000× energy consumption reduction was achieved in RRAM as synaptic devices (<1 pJ per spike). Based on the experimental results, we developed a stochastic model to quantify the device switching dynamics. At the system level, we simulated the performance of image orientation selectivity on a neuromorphic visual system which consists of 1,024 CMOS neuron circuits and 16,348 RRAM synaptic devices. It was found that the system can tolerate the temporal and spatial variability which are commonly present in RRAM devices, suggesting the feasibility of large-scale hardware implementation of neuromorphic system using RRAM synaptic devices.

AB - We report the use of metal oxide resistive switching memory (RRAM) as synaptic devices for a neuromorphic visual system. At the device level, we experimentally characterized the gradual resistance modulation of RRAM by hundreds of identical pulses. As compared with phase change memory (PCM) reported recently in [1,2], >100×-1000× energy consumption reduction was achieved in RRAM as synaptic devices (<1 pJ per spike). Based on the experimental results, we developed a stochastic model to quantify the device switching dynamics. At the system level, we simulated the performance of image orientation selectivity on a neuromorphic visual system which consists of 1,024 CMOS neuron circuits and 16,348 RRAM synaptic devices. It was found that the system can tolerate the temporal and spatial variability which are commonly present in RRAM devices, suggesting the feasibility of large-scale hardware implementation of neuromorphic system using RRAM synaptic devices.

UR - http://www.scopus.com/inward/record.url?scp=84876153660&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876153660&partnerID=8YFLogxK

U2 - 10.1109/IEDM.2012.6479018

DO - 10.1109/IEDM.2012.6479018

M3 - Conference contribution

SN - 9781467348706

BT - Technical Digest - International Electron Devices Meeting, IEDM

ER -