AMSNet: Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network

Olga Krestinskaya, Adilya Bakambekova, Alex Pappachen James

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

1 Citation (Scopus)

Abstract

This work proposes analog hardware implementation of Mean-Pooling Convolutional Neural Network (CNN) with 50% random dropout backpropagation training. We illustrate the effect of variabilities of real memristive devices on the performance of CNN, and tolerance to the input noise. The classification accuracy of CNN is approximately 93% independent on memristor variabilities and input noise. On-chip area and power consumption of analog 180nm CMOS CNN with WOx memristors are 0.09338995mm2 and 3.3992W, respectively.

Original languageEnglish
Title of host publicationProceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages272-273
Number of pages2
ISBN (Electronic)9781538678848
DOIs
Publication statusPublished - Mar 1 2019
Event1st IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019 - Hsinchu, Taiwan
Duration: Mar 18 2019Mar 20 2019

Publication series

NameProceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019

Conference

Conference1st IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019
CountryTaiwan
CityHsinchu
Period3/18/193/20/19

Fingerprint

Memristors
Neural networks
Backpropagation
Electric power utilization
Hardware

Keywords

  • Analog Circuits
  • CNN
  • Memristors
  • Variability

ASJC Scopus subject areas

  • Artificial Intelligence
  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

Krestinskaya, O., Bakambekova, A., & James, A. P. (2019). AMSNet: Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network. In Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019 (pp. 272-273). [8771611] (Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/AICAS.2019.8771611

AMSNet : Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network. / Krestinskaya, Olga; Bakambekova, Adilya; James, Alex Pappachen.

Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 272-273 8771611 (Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019).

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

Krestinskaya, O, Bakambekova, A & James, AP 2019, AMSNet: Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network. in Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019., 8771611, Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019, Institute of Electrical and Electronics Engineers Inc., pp. 272-273, 1st IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019, Hsinchu, Taiwan, 3/18/19. https://doi.org/10.1109/AICAS.2019.8771611
Krestinskaya O, Bakambekova A, James AP. AMSNet: Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network. In Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 272-273. 8771611. (Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019). https://doi.org/10.1109/AICAS.2019.8771611
Krestinskaya, Olga ; Bakambekova, Adilya ; James, Alex Pappachen. / AMSNet : Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network. Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 272-273 (Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019).
@inproceedings{4ff9aff2f2824978a625c55ea832e479,
title = "AMSNet: Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network",
abstract = "This work proposes analog hardware implementation of Mean-Pooling Convolutional Neural Network (CNN) with 50{\%} random dropout backpropagation training. We illustrate the effect of variabilities of real memristive devices on the performance of CNN, and tolerance to the input noise. The classification accuracy of CNN is approximately 93{\%} independent on memristor variabilities and input noise. On-chip area and power consumption of analog 180nm CMOS CNN with WOx memristors are 0.09338995mm2 and 3.3992W, respectively.",
keywords = "Analog Circuits, CNN, Memristors, Variability",
author = "Olga Krestinskaya and Adilya Bakambekova and James, {Alex Pappachen}",
year = "2019",
month = "3",
day = "1",
doi = "10.1109/AICAS.2019.8771611",
language = "English",
series = "Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "272--273",
booktitle = "Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019",
address = "United States",

}

TY - GEN

T1 - AMSNet

T2 - Analog Memristive System Architecture for Mean-Pooling with Dropout Convolutional Neural Network

AU - Krestinskaya, Olga

AU - Bakambekova, Adilya

AU - James, Alex Pappachen

PY - 2019/3/1

Y1 - 2019/3/1

N2 - This work proposes analog hardware implementation of Mean-Pooling Convolutional Neural Network (CNN) with 50% random dropout backpropagation training. We illustrate the effect of variabilities of real memristive devices on the performance of CNN, and tolerance to the input noise. The classification accuracy of CNN is approximately 93% independent on memristor variabilities and input noise. On-chip area and power consumption of analog 180nm CMOS CNN with WOx memristors are 0.09338995mm2 and 3.3992W, respectively.

AB - This work proposes analog hardware implementation of Mean-Pooling Convolutional Neural Network (CNN) with 50% random dropout backpropagation training. We illustrate the effect of variabilities of real memristive devices on the performance of CNN, and tolerance to the input noise. The classification accuracy of CNN is approximately 93% independent on memristor variabilities and input noise. On-chip area and power consumption of analog 180nm CMOS CNN with WOx memristors are 0.09338995mm2 and 3.3992W, respectively.

KW - Analog Circuits

KW - CNN

KW - Memristors

KW - Variability

UR - http://www.scopus.com/inward/record.url?scp=85064770110&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064770110&partnerID=8YFLogxK

U2 - 10.1109/AICAS.2019.8771611

DO - 10.1109/AICAS.2019.8771611

M3 - Conference contribution

AN - SCOPUS:85064770110

T3 - Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019

SP - 272

EP - 273

BT - Proceedings 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2019

PB - Institute of Electrical and Electronics Engineers Inc.

ER -