Analog neuromorphic system based on multi input floating gate MOS neuron model

Ankit Tripathi, Mehdi Arabizadeh, Sourabh Khandelwal, Chetan Singh Thakur

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionResearchpeer-review

Abstract

This paper introduces a novel implementation of the low-power analog artificial neural network (ANN) using Multiple Input Floating Gate MOS (MIFGMOS) transistor for machine learning applications. The number of inputs to a neuron in an ANN is the major bottleneck in building a large scale analog system. The proposed MIFGMOS transistor enables to build a large scale system by combining multiple inputs in a single transistor with a small silicon footprint. Here, we show the MIFGMOS based implementation of the Extreme Learning Machine (ELM) architecture using the receptive field approach with transistor operating in the sub-threshold region. The MIFGMOS produces output current as a function of the weighted combination of the voltage applied to its gate terminals. In the ELM architecture, the weights between the input and the hidden layer are random and this allows exploiting the random device mismatch due to the fabrication process, for building Integrated Circuits (IC) based on ELM architecture. Thus, we use implicit random weights present due to device mismatch, and there is no need to store the input weights. We have verified our architecture using circuit simulations on regression and various classification problems such as on the MNIST data-set and a few UCI data-sets. The proposed MIFGMOS enables combining multiple inputs in a single transistor and will thus pave the way to build large scale deep learning neural networks.

LanguageEnglish
Title of host publication2019 IEEE International Symposium on Circuits and Systems (ISCAS)
Subtitle of host publicationproceedings
Place of PublicationPiscataway, NJ
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Number of pages5
ISBN (Print)9781728103976
DOIs
Publication statusPublished - 2019
Event2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019 - Sapporo, Japan
Duration: 26 May 201929 May 2019

Publication series

Name
ISSN (Print)2158-1525

Conference

Conference2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019
CountryJapan
CitySapporo
Period26/05/1929/05/19

Fingerprint

Neurons
Learning systems
Transistors
Neural networks
Circuit simulation
MOSFET devices
Integrated circuits
Large scale systems
Fabrication
Silicon
Electric potential

Cite this

Tripathi, A., Arabizadeh, M., Khandelwal, S., & Thakur, C. S. (2019). Analog neuromorphic system based on multi input floating gate MOS neuron model. In 2019 IEEE International Symposium on Circuits and Systems (ISCAS): proceedings Piscataway, NJ: Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/ISCAS.2019.8702492
Tripathi, Ankit ; Arabizadeh, Mehdi ; Khandelwal, Sourabh ; Thakur, Chetan Singh. / Analog neuromorphic system based on multi input floating gate MOS neuron model. 2019 IEEE International Symposium on Circuits and Systems (ISCAS): proceedings. Piscataway, NJ : Institute of Electrical and Electronics Engineers (IEEE), 2019.
@inproceedings{11147a27dfef47dfb9cf2fb8ca366ca0,
title = "Analog neuromorphic system based on multi input floating gate MOS neuron model",
abstract = "This paper introduces a novel implementation of the low-power analog artificial neural network (ANN) using Multiple Input Floating Gate MOS (MIFGMOS) transistor for machine learning applications. The number of inputs to a neuron in an ANN is the major bottleneck in building a large scale analog system. The proposed MIFGMOS transistor enables to build a large scale system by combining multiple inputs in a single transistor with a small silicon footprint. Here, we show the MIFGMOS based implementation of the Extreme Learning Machine (ELM) architecture using the receptive field approach with transistor operating in the sub-threshold region. The MIFGMOS produces output current as a function of the weighted combination of the voltage applied to its gate terminals. In the ELM architecture, the weights between the input and the hidden layer are random and this allows exploiting the random device mismatch due to the fabrication process, for building Integrated Circuits (IC) based on ELM architecture. Thus, we use implicit random weights present due to device mismatch, and there is no need to store the input weights. We have verified our architecture using circuit simulations on regression and various classification problems such as on the MNIST data-set and a few UCI data-sets. The proposed MIFGMOS enables combining multiple inputs in a single transistor and will thus pave the way to build large scale deep learning neural networks.",
author = "Ankit Tripathi and Mehdi Arabizadeh and Sourabh Khandelwal and Thakur, {Chetan Singh}",
year = "2019",
doi = "10.1109/ISCAS.2019.8702492",
language = "English",
isbn = "9781728103976",
publisher = "Institute of Electrical and Electronics Engineers (IEEE)",
booktitle = "2019 IEEE International Symposium on Circuits and Systems (ISCAS)",
address = "United States",

}

Tripathi, A, Arabizadeh, M, Khandelwal, S & Thakur, CS 2019, Analog neuromorphic system based on multi input floating gate MOS neuron model. in 2019 IEEE International Symposium on Circuits and Systems (ISCAS): proceedings. Institute of Electrical and Electronics Engineers (IEEE), Piscataway, NJ, 2019 IEEE International Symposium on Circuits and Systems, ISCAS 2019, Sapporo, Japan, 26/05/19. https://doi.org/10.1109/ISCAS.2019.8702492

Analog neuromorphic system based on multi input floating gate MOS neuron model. / Tripathi, Ankit; Arabizadeh, Mehdi; Khandelwal, Sourabh; Thakur, Chetan Singh.

2019 IEEE International Symposium on Circuits and Systems (ISCAS): proceedings. Piscataway, NJ : Institute of Electrical and Electronics Engineers (IEEE), 2019.

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionResearchpeer-review

TY - GEN

T1 - Analog neuromorphic system based on multi input floating gate MOS neuron model

AU - Tripathi, Ankit

AU - Arabizadeh, Mehdi

AU - Khandelwal, Sourabh

AU - Thakur, Chetan Singh

PY - 2019

Y1 - 2019

N2 - This paper introduces a novel implementation of the low-power analog artificial neural network (ANN) using Multiple Input Floating Gate MOS (MIFGMOS) transistor for machine learning applications. The number of inputs to a neuron in an ANN is the major bottleneck in building a large scale analog system. The proposed MIFGMOS transistor enables to build a large scale system by combining multiple inputs in a single transistor with a small silicon footprint. Here, we show the MIFGMOS based implementation of the Extreme Learning Machine (ELM) architecture using the receptive field approach with transistor operating in the sub-threshold region. The MIFGMOS produces output current as a function of the weighted combination of the voltage applied to its gate terminals. In the ELM architecture, the weights between the input and the hidden layer are random and this allows exploiting the random device mismatch due to the fabrication process, for building Integrated Circuits (IC) based on ELM architecture. Thus, we use implicit random weights present due to device mismatch, and there is no need to store the input weights. We have verified our architecture using circuit simulations on regression and various classification problems such as on the MNIST data-set and a few UCI data-sets. The proposed MIFGMOS enables combining multiple inputs in a single transistor and will thus pave the way to build large scale deep learning neural networks.

AB - This paper introduces a novel implementation of the low-power analog artificial neural network (ANN) using Multiple Input Floating Gate MOS (MIFGMOS) transistor for machine learning applications. The number of inputs to a neuron in an ANN is the major bottleneck in building a large scale analog system. The proposed MIFGMOS transistor enables to build a large scale system by combining multiple inputs in a single transistor with a small silicon footprint. Here, we show the MIFGMOS based implementation of the Extreme Learning Machine (ELM) architecture using the receptive field approach with transistor operating in the sub-threshold region. The MIFGMOS produces output current as a function of the weighted combination of the voltage applied to its gate terminals. In the ELM architecture, the weights between the input and the hidden layer are random and this allows exploiting the random device mismatch due to the fabrication process, for building Integrated Circuits (IC) based on ELM architecture. Thus, we use implicit random weights present due to device mismatch, and there is no need to store the input weights. We have verified our architecture using circuit simulations on regression and various classification problems such as on the MNIST data-set and a few UCI data-sets. The proposed MIFGMOS enables combining multiple inputs in a single transistor and will thus pave the way to build large scale deep learning neural networks.

UR - http://www.scopus.com/inward/record.url?scp=85066808063&partnerID=8YFLogxK

U2 - 10.1109/ISCAS.2019.8702492

DO - 10.1109/ISCAS.2019.8702492

M3 - Conference proceeding contribution

SN - 9781728103976

BT - 2019 IEEE International Symposium on Circuits and Systems (ISCAS)

PB - Institute of Electrical and Electronics Engineers (IEEE)

CY - Piscataway, NJ

ER -

Tripathi A, Arabizadeh M, Khandelwal S, Thakur CS. Analog neuromorphic system based on multi input floating gate MOS neuron model. In 2019 IEEE International Symposium on Circuits and Systems (ISCAS): proceedings. Piscataway, NJ: Institute of Electrical and Electronics Engineers (IEEE). 2019 https://doi.org/10.1109/ISCAS.2019.8702492