中国物理B ›› 2023, Vol. 32 ›› Issue (7): 78505-078505.doi: 10.1088/1674-1056/ac9cbc

• • 上一篇    下一篇

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)‡   

  1. College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
  • 收稿日期:2022-08-02 修回日期:2022-09-22 接受日期:2022-10-21 出版日期:2023-06-15 发布日期:2023-07-05
  • 通讯作者: Wei Wang, Qingjiang Li E-mail:wangwei_esss@nudt.edu.cn;qingjiangli@nudt.edu.cn
  • 基金资助:
    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).

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)‡   

  1. College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
  • Received:2022-08-02 Revised:2022-09-22 Accepted:2022-10-21 Online:2023-06-15 Published:2023-07-05
  • Contact: Wei Wang, Qingjiang Li E-mail:wangwei_esss@nudt.edu.cn;qingjiangli@nudt.edu.cn
  • Supported by:
    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).

RichHTML

0

PDF (PC)

217

赞

摘要： 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.

关键词: memristor, lightweight, robust, hybrid neural networks, depthwise separable convolution, bidirectional gate recurrent unit (BiGRU), one-transistor one-resistor (1T1R) arrays

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.

Key words: memristor, lightweight, robust, hybrid neural networks, depthwise separable convolution, bidirectional gate recurrent unit (BiGRU), one-transistor one-resistor (1T1R) arrays

中图分类号:  (Nanoelectronic devices)

  • 85.35.-p
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)