Variability Analysis of Memristor-based Sigmoid Function

Nursultan Kaiyrbekov, Olga Krestinskaya, Alex James Pappachen

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

Abstract

Activation functions are widely used in neural networks to decide the activation value of the neural unit based on linear combinations of the weighted inputs. The effective implementation of activation function is highly important to enhance he performance of a neural network. One of the most widely used activation functions is sigmoid. Therefore, there is a growing interest to enhance the performance of sigmoid circuits. In this paper, the main objective is to modify existing current mirror based sigmoid model by replacing CMOS transistors with memristive devices. We present the performance, variation of transistor sizes and temperature. The area, power and noise in the modified CMOS-memristive sigmoid circuit are shown. The application of memristors in the sigmoid circuit ensures the reduction of on-chip area, and power dissipation by 7%. The proposed sigmoid circuit was simulated in SPICE using TSMC 180nm CMOS design process.

Original languageEnglish
Title of host publicationProceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages206-209
Number of pages4
ISBN (Electronic)9781538659281
DOIs
Publication statusPublished - Sep 28 2018
Event2nd International Conference on Computing and Network Communications, CoCoNet 2018 - Astana, Kazakhstan
Duration: Aug 15 2018Aug 17 2018

Conference

Conference2nd International Conference on Computing and Network Communications, CoCoNet 2018
CountryKazakhstan
CityAstana
Period8/15/188/17/18

Fingerprint

Memristors
Chemical activation
Networks (circuits)
Transistors
Neural networks
SPICE
Energy dissipation
Mirrors
Temperature

Keywords

  • Artificial Neural Network
  • CMOS
  • Memristor
  • Sigmoid

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Computer Science Applications
  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

Kaiyrbekov, N., Krestinskaya, O., & James Pappachen, A. (2018). Variability Analysis of Memristor-based Sigmoid Function. In Proceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018 (pp. 206-209). [8476878] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CoCoNet.2018.8476878

Variability Analysis of Memristor-based Sigmoid Function. / Kaiyrbekov, Nursultan; Krestinskaya, Olga; James Pappachen, Alex.

Proceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 206-209 8476878.

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

Kaiyrbekov, N, Krestinskaya, O & James Pappachen, A 2018, Variability Analysis of Memristor-based Sigmoid Function. in Proceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018., 8476878, Institute of Electrical and Electronics Engineers Inc., pp. 206-209, 2nd International Conference on Computing and Network Communications, CoCoNet 2018, Astana, Kazakhstan, 8/15/18. https://doi.org/10.1109/CoCoNet.2018.8476878
Kaiyrbekov N, Krestinskaya O, James Pappachen A. Variability Analysis of Memristor-based Sigmoid Function. In Proceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018. Institute of Electrical and Electronics Engineers Inc. 2018. p. 206-209. 8476878 https://doi.org/10.1109/CoCoNet.2018.8476878
Kaiyrbekov, Nursultan ; Krestinskaya, Olga ; James Pappachen, Alex. / Variability Analysis of Memristor-based Sigmoid Function. Proceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 206-209
@inproceedings{24b49821d4694914a55ed672d92a413c,
title = "Variability Analysis of Memristor-based Sigmoid Function",
abstract = "Activation functions are widely used in neural networks to decide the activation value of the neural unit based on linear combinations of the weighted inputs. The effective implementation of activation function is highly important to enhance he performance of a neural network. One of the most widely used activation functions is sigmoid. Therefore, there is a growing interest to enhance the performance of sigmoid circuits. In this paper, the main objective is to modify existing current mirror based sigmoid model by replacing CMOS transistors with memristive devices. We present the performance, variation of transistor sizes and temperature. The area, power and noise in the modified CMOS-memristive sigmoid circuit are shown. The application of memristors in the sigmoid circuit ensures the reduction of on-chip area, and power dissipation by 7{\%}. The proposed sigmoid circuit was simulated in SPICE using TSMC 180nm CMOS design process.",
keywords = "Artificial Neural Network, CMOS, Memristor, Sigmoid",
author = "Nursultan Kaiyrbekov and Olga Krestinskaya and {James Pappachen}, Alex",
year = "2018",
month = "9",
day = "28",
doi = "10.1109/CoCoNet.2018.8476878",
language = "English",
pages = "206--209",
booktitle = "Proceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Variability Analysis of Memristor-based Sigmoid Function

AU - Kaiyrbekov, Nursultan

AU - Krestinskaya, Olga

AU - James Pappachen, Alex

PY - 2018/9/28

Y1 - 2018/9/28

N2 - Activation functions are widely used in neural networks to decide the activation value of the neural unit based on linear combinations of the weighted inputs. The effective implementation of activation function is highly important to enhance he performance of a neural network. One of the most widely used activation functions is sigmoid. Therefore, there is a growing interest to enhance the performance of sigmoid circuits. In this paper, the main objective is to modify existing current mirror based sigmoid model by replacing CMOS transistors with memristive devices. We present the performance, variation of transistor sizes and temperature. The area, power and noise in the modified CMOS-memristive sigmoid circuit are shown. The application of memristors in the sigmoid circuit ensures the reduction of on-chip area, and power dissipation by 7%. The proposed sigmoid circuit was simulated in SPICE using TSMC 180nm CMOS design process.

AB - Activation functions are widely used in neural networks to decide the activation value of the neural unit based on linear combinations of the weighted inputs. The effective implementation of activation function is highly important to enhance he performance of a neural network. One of the most widely used activation functions is sigmoid. Therefore, there is a growing interest to enhance the performance of sigmoid circuits. In this paper, the main objective is to modify existing current mirror based sigmoid model by replacing CMOS transistors with memristive devices. We present the performance, variation of transistor sizes and temperature. The area, power and noise in the modified CMOS-memristive sigmoid circuit are shown. The application of memristors in the sigmoid circuit ensures the reduction of on-chip area, and power dissipation by 7%. The proposed sigmoid circuit was simulated in SPICE using TSMC 180nm CMOS design process.

KW - Artificial Neural Network

KW - CMOS

KW - Memristor

KW - Sigmoid

UR - http://www.scopus.com/inward/record.url?scp=85055984891&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055984891&partnerID=8YFLogxK

U2 - 10.1109/CoCoNet.2018.8476878

DO - 10.1109/CoCoNet.2018.8476878

M3 - Conference contribution

SP - 206

EP - 209

BT - Proceedings of the 2nd International Conference on Computing and Network Communications, CoCoNet 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -