NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care

Hanrui Li; Fengshi Tian; Jie Yang; Mohamad Sawan; Nazek El-Atab

doi:10.1109/BioCAS58349.2023.10388528

NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care

Hanrui Li, Fengshi Tian, Jie Yang, Mohamad Sawan, Nazek El-Atab^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Energy-efficient and noisy-adaptive signal processing system are in high demand of edge biomedical applications. In this paper, we present a Noise-Adaptive Memristive Bayesian Neural Network (NMBNN) architecture for various biosignal applications. The memristor has the inherent physical property of exhibiting variability in resistance, which makes it a promising candidate of uncertainty weight in Bayesian Neural Networks (BNN). The NMBNN architecture combines the noise-resilient attributes of BNN with the implementation of an energy-efficient RRAM array. By utilizing BNN's probabilistic predictions and implementation with the conductance fluctuations of memristors, NMBNN offers a robust and energy-efficient solution adept at processing biosignals in noisy environments. In order to evaluate the network robustness, we conduct the experiments to introduce multiple types of noise as adversarial sample. The experimental results indicate that the proposed NMBNN approach has the advantages of being both noise-adaptive and energy-efficient.

Original language	English (US)
Title of host publication	BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350300260
DOIs	https://doi.org/10.1109/BioCAS58349.2023.10388528
State	Published - 2023
Event	2023 IEEE Biomedical Circuits and Systems Conference, BioCAS 2023 - Toronto, Canada Duration: Oct 19 2023 → Oct 21 2023

Publication series

Name	BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings

Conference

Conference	2023 IEEE Biomedical Circuits and Systems Conference, BioCAS 2023
Country/Territory	Canada
City	Toronto
Period	10/19/23 → 10/21/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Keywords

bayesian neural network
memristor
network robustness
signal processing

ASJC Scopus subject areas

Signal Processing
Biomedical Engineering
Electrical and Electronic Engineering
Clinical Neurology
Neurology

Access to Document

10.1109/BioCAS58349.2023.10388528

Cite this

Li, H., Tian, F., Yang, J., Sawan, M., & El-Atab, N. (2023). NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care. In BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings (BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/BioCAS58349.2023.10388528

Li, Hanrui ; Tian, Fengshi ; Yang, Jie et al. / NMBNN : Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care. BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings. Institute of Electrical and Electronics Engineers Inc., 2023. (BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings).

@inproceedings{7754eae2a4fd4f82b42d6e14f1484d08,

title = "NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care",

abstract = "Energy-efficient and noisy-adaptive signal processing system are in high demand of edge biomedical applications. In this paper, we present a Noise-Adaptive Memristive Bayesian Neural Network (NMBNN) architecture for various biosignal applications. The memristor has the inherent physical property of exhibiting variability in resistance, which makes it a promising candidate of uncertainty weight in Bayesian Neural Networks (BNN). The NMBNN architecture combines the noise-resilient attributes of BNN with the implementation of an energy-efficient RRAM array. By utilizing BNN's probabilistic predictions and implementation with the conductance fluctuations of memristors, NMBNN offers a robust and energy-efficient solution adept at processing biosignals in noisy environments. In order to evaluate the network robustness, we conduct the experiments to introduce multiple types of noise as adversarial sample. The experimental results indicate that the proposed NMBNN approach has the advantages of being both noise-adaptive and energy-efficient.",

keywords = "bayesian neural network, memristor, network robustness, signal processing",

author = "Hanrui Li and Fengshi Tian and Jie Yang and Mohamad Sawan and Nazek El-Atab",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE Biomedical Circuits and Systems Conference, BioCAS 2023 ; Conference date: 19-10-2023 Through 21-10-2023",

year = "2023",

doi = "10.1109/BioCAS58349.2023.10388528",

language = "English (US)",

series = "BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings",

address = "United States",

}

Li, H, Tian, F, Yang, J, Sawan, M & El-Atab, N 2023, NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care. in BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings. BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings, Institute of Electrical and Electronics Engineers Inc., 2023 IEEE Biomedical Circuits and Systems Conference, BioCAS 2023, Toronto, Canada, 10/19/23. https://doi.org/10.1109/BioCAS58349.2023.10388528

NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care. / Li, Hanrui; Tian, Fengshi; Yang, Jie et al.
BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings. Institute of Electrical and Electronics Engineers Inc., 2023. (BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - NMBNN

T2 - 2023 IEEE Biomedical Circuits and Systems Conference, BioCAS 2023

AU - Li, Hanrui

AU - Tian, Fengshi

AU - Yang, Jie

AU - Sawan, Mohamad

AU - El-Atab, Nazek

PY - 2023

Y1 - 2023

N2 - Energy-efficient and noisy-adaptive signal processing system are in high demand of edge biomedical applications. In this paper, we present a Noise-Adaptive Memristive Bayesian Neural Network (NMBNN) architecture for various biosignal applications. The memristor has the inherent physical property of exhibiting variability in resistance, which makes it a promising candidate of uncertainty weight in Bayesian Neural Networks (BNN). The NMBNN architecture combines the noise-resilient attributes of BNN with the implementation of an energy-efficient RRAM array. By utilizing BNN's probabilistic predictions and implementation with the conductance fluctuations of memristors, NMBNN offers a robust and energy-efficient solution adept at processing biosignals in noisy environments. In order to evaluate the network robustness, we conduct the experiments to introduce multiple types of noise as adversarial sample. The experimental results indicate that the proposed NMBNN approach has the advantages of being both noise-adaptive and energy-efficient.

AB - Energy-efficient and noisy-adaptive signal processing system are in high demand of edge biomedical applications. In this paper, we present a Noise-Adaptive Memristive Bayesian Neural Network (NMBNN) architecture for various biosignal applications. The memristor has the inherent physical property of exhibiting variability in resistance, which makes it a promising candidate of uncertainty weight in Bayesian Neural Networks (BNN). The NMBNN architecture combines the noise-resilient attributes of BNN with the implementation of an energy-efficient RRAM array. By utilizing BNN's probabilistic predictions and implementation with the conductance fluctuations of memristors, NMBNN offers a robust and energy-efficient solution adept at processing biosignals in noisy environments. In order to evaluate the network robustness, we conduct the experiments to introduce multiple types of noise as adversarial sample. The experimental results indicate that the proposed NMBNN approach has the advantages of being both noise-adaptive and energy-efficient.

KW - bayesian neural network

KW - memristor

KW - network robustness

KW - signal processing

UR - http://www.scopus.com/inward/record.url?scp=85184925986&partnerID=8YFLogxK

U2 - 10.1109/BioCAS58349.2023.10388528

DO - 10.1109/BioCAS58349.2023.10388528

M3 - Conference contribution

AN - SCOPUS:85184925986

T3 - BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings

BT - BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 19 October 2023 through 21 October 2023

ER -

Li H, Tian F, Yang J, Sawan M, El-Atab N. NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care. In BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings. Institute of Electrical and Electronics Engineers Inc. 2023. (BioCAS 2023 - 2023 IEEE Biomedical Circuits and Systems Conference, Conference Proceedings). doi: 10.1109/BioCAS58349.2023.10388528

NMBNN: Noise-Adaptive Memristive Bayesian Neural Network for Energy-Efficient Edge Health Care

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this