Enhancing Security of Memristor Computing System Through Secure Weight Mapping

Minhui Zou; Junlong Zhou; Xiaotong Cui; Wei Wang; Shahar Kvatinsky

doi:10.1109/ISVLSI54635.2022.00044

Enhancing Security of Memristor Computing System Through Secure Weight Mapping

Minhui Zou
, Junlong Zhou
, Xiaotong Cui
, Wei Wang
, Shahar Kvatinsky

Technion - Israel Institute of Technology

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

Abstract

Emerging memristor computing systems have demonstrated great promise in improving the energy efficiency of neural network (NN) algorithms. The NN weights stored in memristor crossbars, however, may face potential theft attacks due to the nonvolatility of the memristor devices. In this paper, we propose to protect the NN weights by mapping selected columns of them in the form of 1's complements and leaving the other columns in their original form, preventing the adversary from knowing the exact representation of each weight. The results show that compared with prior work, our method achieves effectiveness comparable to the best of them and reduces the hardware overhead by more than 18X.

Original language	English GB
Title of host publication	Proceedings - 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022
Pages	182-187
Number of pages	6
ISBN (Electronic)	9781665466059
DOIs	https://doi.org/10.1109/ISVLSI54635.2022.00044
State	Published - 2022
Event	2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022 - Pafos, Cyprus Duration: 4 Jul 2022 → 6 Jul 2022

Publication series

Name	Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI
Volume	2022-July

Conference

Conference	2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022
Country/Territory	Cyprus
City	Pafos
Period	4/07/22 → 6/07/22

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Keywords

Memristor
Neural network
Security

ASJC Scopus subject areas

Hardware and Architecture
Control and Systems Engineering
Electrical and Electronic Engineering

Access to Document

10.1109/ISVLSI54635.2022.00044

Cite this

Zou, M., Zhou, J., Cui, X., Wang, W., & Kvatinsky, S. (2022). Enhancing Security of Memristor Computing System Through Secure Weight Mapping. In Proceedings - 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022 (pp. 182-187). (Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI; Vol. 2022-July). https://doi.org/10.1109/ISVLSI54635.2022.00044

@inproceedings{cff48595ee66401cb59caf740ba84e34,

title = "Enhancing Security of Memristor Computing System Through Secure Weight Mapping",

abstract = "Emerging memristor computing systems have demonstrated great promise in improving the energy efficiency of neural network (NN) algorithms. The NN weights stored in memristor crossbars, however, may face potential theft attacks due to the nonvolatility of the memristor devices. In this paper, we propose to protect the NN weights by mapping selected columns of them in the form of 1's complements and leaving the other columns in their original form, preventing the adversary from knowing the exact representation of each weight. The results show that compared with prior work, our method achieves effectiveness comparable to the best of them and reduces the hardware overhead by more than 18X.",

keywords = "Memristor, Neural network, Security",

author = "Minhui Zou and Junlong Zhou and Xiaotong Cui and Wei Wang and Shahar Kvatinsky",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022 ; Conference date: 04-07-2022 Through 06-07-2022",

year = "2022",

doi = "10.1109/ISVLSI54635.2022.00044",

language = "الإنجليزيّة",

series = "Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI",

pages = "182--187",

booktitle = "Proceedings - 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022",

}

Zou, M, Zhou, J, Cui, X, Wang, W & Kvatinsky, S 2022, Enhancing Security of Memristor Computing System Through Secure Weight Mapping. in Proceedings - 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022. Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI, vol. 2022-July, pp. 182-187, 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022, Pafos, Cyprus, 4/07/22. https://doi.org/10.1109/ISVLSI54635.2022.00044

Enhancing Security of Memristor Computing System Through Secure Weight Mapping. / Zou, Minhui; Zhou, Junlong; Cui, Xiaotong et al.
Proceedings - 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022. 2022. p. 182-187 (Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI; Vol. 2022-July).

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

TY - GEN

T1 - Enhancing Security of Memristor Computing System Through Secure Weight Mapping

AU - Zou, Minhui

AU - Zhou, Junlong

AU - Cui, Xiaotong

AU - Wang, Wei

AU - Kvatinsky, Shahar

PY - 2022

Y1 - 2022

N2 - Emerging memristor computing systems have demonstrated great promise in improving the energy efficiency of neural network (NN) algorithms. The NN weights stored in memristor crossbars, however, may face potential theft attacks due to the nonvolatility of the memristor devices. In this paper, we propose to protect the NN weights by mapping selected columns of them in the form of 1's complements and leaving the other columns in their original form, preventing the adversary from knowing the exact representation of each weight. The results show that compared with prior work, our method achieves effectiveness comparable to the best of them and reduces the hardware overhead by more than 18X.

AB - Emerging memristor computing systems have demonstrated great promise in improving the energy efficiency of neural network (NN) algorithms. The NN weights stored in memristor crossbars, however, may face potential theft attacks due to the nonvolatility of the memristor devices. In this paper, we propose to protect the NN weights by mapping selected columns of them in the form of 1's complements and leaving the other columns in their original form, preventing the adversary from knowing the exact representation of each weight. The results show that compared with prior work, our method achieves effectiveness comparable to the best of them and reduces the hardware overhead by more than 18X.

KW - Memristor

KW - Neural network

KW - Security

UR - https://www.scopus.com/pages/publications/85140892970

U2 - 10.1109/ISVLSI54635.2022.00044

DO - 10.1109/ISVLSI54635.2022.00044

M3 - منشور من مؤتمر

T3 - Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI

SP - 182

EP - 187

BT - Proceedings - 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022

T2 - 2022 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2022

Y2 - 4 July 2022 through 6 July 2022

ER -

Enhancing Security of Memristor Computing System Through Secure Weight Mapping

Abstract

Publication series

Conference

UN SDGs

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this