| INTERDISCIPLINARY PHYSICS AND RELATED AREAS OF SCIENCE AND TECHNOLOGY |
Prev
Next
|
|
|
Lightweight and highly robust memristor-based hybrid neural networks for electroencephalogram signal processing |
| Peiwen Tong(童霈文), Hui Xu(徐晖), Yi Sun(孙毅), Yongzhou Wang(汪泳州), Jie Peng(彭杰),Cen Liao(廖岑), Wei Wang(王伟)†, and Qingjiang Li(李清江)‡ |
| College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China |
|
|
|
|
Abstract Memristor-based neuromorphic computing shows great potential for high-speed and high-throughput signal processing applications, such as electroencephalogram (EEG) signal processing. Nonetheless, the size of one-transistor one-resistor (1T1R) memristor arrays is limited by the non-ideality of the devices, which prevents the hardware implementation of large and complex networks. In this work, we propose the depthwise separable convolution and bidirectional gate recurrent unit (DSC-BiGRU) network, a lightweight and highly robust hybrid neural network based on 1T1R arrays that enables efficient processing of EEG signals in the temporal, frequency and spatial domains by hybridizing DSC and BiGRU blocks. The network size is reduced and the network robustness is improved while ensuring the network classification accuracy. In the simulation, the measured non-idealities of the 1T1R array are brought into the network through statistical analysis. Compared with traditional convolutional networks, the network parameters are reduced by 95% and the network classification accuracy is improved by 21% at a 95% array yield rate and 5% tolerable error. This work demonstrates that lightweight and highly robust networks based on memristor arrays hold great promise for applications that rely on low consumption and high efficiency.
|
Received: 02 August 2022
Revised: 22 September 2022
Accepted manuscript online: 21 October 2022
|
|
PACS:
|
85.35.-p
|
(Nanoelectronic devices)
|
| |
84.37.+q
|
(Measurements in electric variables (including voltage, current, resistance, capacitance, inductance, impedance, and admittance, etc.))
|
| |
07.05.Mh
|
(Neural networks, fuzzy logic, artificial intelligence)
|
| |
87.19.lv
|
(Learning and memory)
|
|
| Fund: Project supported by the National Key Research and Development Program of China (Grant No. 2019YFB2205102) and the National Natural Science Foundation of China (Grant Nos. 61974164, 62074166, 61804181, 62004219, 62004220, and 62104256). |
Corresponding Authors:
Wei Wang, Qingjiang Li
E-mail: wangwei_esss@nudt.edu.cn;qingjiangli@nudt.edu.cn
|
Cite this article:
Peiwen Tong(童霈文), Hui Xu(徐晖), Yi Sun(孙毅), Yongzhou Wang(汪泳州), Jie Peng(彭杰),Cen Liao(廖岑), Wei Wang(王伟), and Qingjiang Li(李清江) Lightweight and highly robust memristor-based hybrid neural networks for electroencephalogram signal processing 2023 Chin. Phys. B 32 078505
|
[1] Shanechi M M 2019 Nat. Neurosci. 22 1554
[2] Kao J C, Stavisky S D, Sussillo D, Nuyujukian P and Shenoy K V 2014 Proc. IEEE 102 666
[3] Hong G and Lieber C M 2019 Nat. Rev. Neurosci. 20 330
[4] Maharbiz M M, Muller R, Alon E, Rabaey J M and Carmena J M 2017 Proc. IEEE 105 73
[5] Brandman D M, Cash S S and Hochberg L R 2017 IEEE Trans. Neural Syst. Rehabil. Eng. 25 1687
[6] Liu X, Zhang M, Xiong T, Richardson A G, Lucas T H, Chin P S, Etienne-Cummings R, Tran T D and Van Der Spiegel J 2016 IEEE Trans. Biomed. Circuits Syst. 10 874
[7] Cheng C H, Tsai P Y, Yang T Y, Cheng W H, Yen T Y, Luo Z, Qian X H, Chen Z X, Lin T H, Chen W H, Chen W M, Liang S F, Shaw F Z, Chang C S, Hsin Y L, Lee C Y, Ker M D and Wu C Y 2018 IEEE J. Solid-State Circuits 53 3314
[8] Merkel C, Saleh Q, Donahue C and Kudithipudi D 2014 Procedia Comput. Sci. 41 249
[9] Park S, Chu M, Kim J, Noh J, Jeon M, Hun Lee B, Hwang H, Lee B and Lee B G 2015 Sci. Rep. 5 1
[10] Zhang J, Xu H, Long H, Hou K, Li Z, Yang Z, Wang Y and Li Q 2021 IEEE 4th Int. Conf. Electron. Technol. ICET 2021 851
[11] Liu Z, Tang J, Hong B, Wu H and Gao B 2020 Nat. Commun. 11 4234
[12] Liu J, Ye F and Xiong H 2021 J. Neural Eng. 18
[13] Cai H 2018 ACS Appl. Mater. Interfaces 10 22408
[14] Lawhern V J, Solon A J, Waytowich N R, Gordon S M, Hung C P and Lance B J 2018 J. Neural Eng. 15 056013
[15] Schirrmeister R T, Springenberg J T, Fiederer L D J, Glasstetter M, Eggensperger K, Tangermann M, Hutter F, Burgard W and Ball T 2017 Hum. Brain Mapp. 38 5391
[16] Ang K K, Chin Z Y, Zhang H and Guan C 2008 Proc. Int. Jt. Conf. Neural Networks 2390
[17] Gramfort A, Luessi M, Larson E, Engemann D A, Strohmeier D, Brodbeck C, Parkkonen L and Hämäläinen M S 2014 Neuroimage 86 446
[18] Gramfort A, Luessi M, Larson E, Engemann D A, Strohmeier D, Brodbeck C, Goj R, Jas M, Brooks T, Parkkonen L and Hämäläinen M 2013 Front. Neurosci. 7
[19] Chen J, Wu H, Gao B, Tang J, Zhang W and Qian H 2019 IEEE Int. Work. Futur. Comput. IWOFC 2019 2019 |
| No Suggested Reading articles found! |
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
Altmetric
|
|
blogs
Facebook pages
Wikipedia page
Google+ users
|
Online attention
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.
View more on Altmetrics
|
|
|
