Journal of the Semiconductor & Display Technology (반도체디스플레이기술학회지)

The Korean Society Of Semiconductor & Display Technology (한국반도체디스플레이기술학회)
권호 표지

Improved Accuracy in Neuromorphic Computing Based on IGZO Memristor Devices

IGZO 멤리스터 소자기반 뉴로모픽 컴퓨팅 정확도 향상

  • Seojin Choi (Department of System Semiconductor Engineering, Sangmyung University) ;
  • Kyoungjin Min (Department of System Semiconductor Engineering, Sangmyung University) ;
  • Jonghwan Lee (Department of System Semiconductor Engineering, Sangmyung University)
  • 최서진 (상명대학교 시스템반도체공학과) ;
  • 민경진 (상명대학교 시스템반도체공학과) ;
  • 이종환 (상명대학교 시스템반도체공학과)
  • Received : 2023.12.05
  • Accepted : 2023.12.19
  • Published : 2023.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

This paper presents the synaptic characteristics of IGZO memristors in neuromorphic computing, using MATLAB/Simulink and NeuroSim. In order to investigate the variations in the conductivity of IGZO memristor and the corresponding changes in the hidden layer, simulations are conducted by using the MNIST dataset. It was observed from simulation results that the recognition accuracy could be dependent on various parameters of IGZO memristor, along with the experimental exploration. Moreover, we identified optimal parameters to achieve high accuracy, showing an outstanding accuracy of 96.83% in image classification.

Keywords

Acknowledgement

This work is funded by a 2023 research Grant from Sangmyung University.

References

  1. Park, Geon-Woo, et al. "A Review of RRAM-based Synaptic Device to Improve Neuromorphic Systems." Journal of The Korean Society of Semiconductor & Display Technology, Vol. 21, pp. 50-56, 2022.
  2. Lu, Tian, et al. "Optimal Weight Models for Ferroelectric Synapses Toward Neuromorphic Computing." IEEE Transactions on Electron Devices (2023).
  3. Chen, Wenbin, et al. "Essential Characteristics of Memristors for Neuromorphic Computing." Advanced Electronic Materials 9.2 (2023): 2200833.
  4. Agarwal, Sapan, et al. "Resistive memory device requirements for a neural algorithm accelerator." 2016 International Joint Conference on Neural Networks (IJCNN). IEEE, 2016.
  5. Zhang, Li, et al. "Resistive switching performance improvement of InGaZnO-based memory device by nitrogen plasma treatment." Journal of Materials Science & Technology 49 (2020): 1-6.
  6. Pereira, Maria Elias, et al. "Tailoring the synaptic properties of a-IGZO memristors for artificial deep neural networks." APL Materials 10.1 (2022).
  7. Choi, Hyun-Woong, et al. "Zinc oxide and indium-gallium-zinc-oxide bi-layer synaptic device with highly linear long-term potentiation and depression characteristics." Scientific reports 12.1 (2022): 1259.
  8. Agarwal, Sapan, et al. "Resistive memory device requirements for a neural algorithm accelerator." 2016 International Joint Conference on Neural Networks (IJCNN). IEEE, 2016.
  9. Chen, Pai-Yu, Xiaochen Peng, and Shimeng Yu. "NeuroSim: A circuit-level macro model for benchmarking neuro-inspired architectures in online learning." IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems 37.12 (2018): 3067-3080. https://doi.org/10.1109/TCAD.2018.2789723
  10. Liu, Huan, Wei, Min and Chen, Yuzhong. "Optimization of non-linear conductance modulation based on metal oxide memristors" Nanotechnology Reviews, vol. 7, no. 5, 2018, pp. 443-468. https://doi.org/10.1515/ntrev-2018-0045
  11. Karsoliya, Saurabh. "Approximating number of hidden layer neurons in multiple hidden layer BPNN architecture." International Journal of Engineering Trends and Technology 3.6 (2012): 714-717.
  12. Lee, Yong-Hwan and Kim, Heung-Jun. "Implementation of Fish Detection Based on Convolutional Neural Networks." Journal of The Korean Society of Semiconductor & Display Technology, Vol. 20, pp. 124-129, 2020.
  13. Ying, Xue. "An overview of overfitting and its solutions." Journal of physics: Conference series. Vol. 1168. IOP Publishing, 2019.