A physics-based predictive model for pulse design to realize high-performance memristive neural networks
Memristive neural networks have extensively been investigated for their capability in handling various artificial intelligence tasks. The training performance of memristive neural networks depends on the pulse scheme applied to the constituent memristors. However, the design of the pulse scheme in m...
| Main Authors: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIP Publishing LLC
2023-12-01
|
| Series: | APL Machine Learning |
| Online Access: | http://dx.doi.org/10.1063/5.0180346 |
| _version_ | 1797367099681669120 |
|---|---|
| author | Haoyue Deng Zhen Fan Shuai Dong Zhiwei Chen Wenjie Li Yihong Chen Kun Liu Ruiqiang Tao Guo Tian Deyang Chen Minghui Qin Min Zeng Xubing Lu Guofu Zhou Xingsen Gao Jun-Ming Liu |
| author_facet | Haoyue Deng Zhen Fan Shuai Dong Zhiwei Chen Wenjie Li Yihong Chen Kun Liu Ruiqiang Tao Guo Tian Deyang Chen Minghui Qin Min Zeng Xubing Lu Guofu Zhou Xingsen Gao Jun-Ming Liu |
| author_sort | Haoyue Deng |
| collection | DOAJ |
| description | Memristive neural networks have extensively been investigated for their capability in handling various artificial intelligence tasks. The training performance of memristive neural networks depends on the pulse scheme applied to the constituent memristors. However, the design of the pulse scheme in most previous studies was approached in an empirical manner or through a trial-and-error method. Here, we choose ferroelectric tunnel junction (FTJ) as a model memristor and demonstrate a physics-based predictive model for the pulse design to achieve high training performance. This predictive model comprises a physical model for FTJ that can adequately describe the polarization switching and memristive switching behaviors of the FTJ and an FTJ-based neural network that uses the long-term potentiation (LTP)/long-term depression (LTD) characteristics of the FTJ for the weight update. Simulation results based on the predictive model demonstrate that the LTP/LTD characteristics with a good trade-off between ON/OFF ratio, nonlinearity, and asymmetry can lead to high training accuracies for the FTJ-based neural network. Moreover, it is revealed that an amplitude-increasing pulse scheme may be the most favorable pulse scheme as it offers the widest ranges of pulse amplitudes and widths for achieving high accuracies. This study may provide useful guidance for the pulse design in the experimental development of high-performance memristive neural networks. |
| first_indexed | 2024-03-08T17:12:27Z |
| format | Article |
| id | doaj.art-fe6298b6c2514f1ba000a639676bec1a |
| institution | Directory Open Access Journal |
| issn | 2770-9019 |
| language | English |
| last_indexed | 2024-03-08T17:12:27Z |
| publishDate | 2023-12-01 |
| publisher | AIP Publishing LLC |
| record_format | Article |
| series | APL Machine Learning |
| spelling | doaj.art-fe6298b6c2514f1ba000a639676bec1a2024-01-03T19:54:29ZengAIP Publishing LLCAPL Machine Learning2770-90192023-12-0114046110046110-1210.1063/5.0180346A physics-based predictive model for pulse design to realize high-performance memristive neural networksHaoyue Deng0Zhen Fan1Shuai Dong2Zhiwei Chen3Wenjie Li4Yihong Chen5Kun Liu6Ruiqiang Tao7Guo Tian8Deyang Chen9Minghui Qin10Min Zeng11Xubing Lu12Guofu Zhou13Xingsen Gao14Jun-Ming Liu15Institute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaNational Center for International Research on Green Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaInstitute for Advanced Materials and Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, ChinaMemristive neural networks have extensively been investigated for their capability in handling various artificial intelligence tasks. The training performance of memristive neural networks depends on the pulse scheme applied to the constituent memristors. However, the design of the pulse scheme in most previous studies was approached in an empirical manner or through a trial-and-error method. Here, we choose ferroelectric tunnel junction (FTJ) as a model memristor and demonstrate a physics-based predictive model for the pulse design to achieve high training performance. This predictive model comprises a physical model for FTJ that can adequately describe the polarization switching and memristive switching behaviors of the FTJ and an FTJ-based neural network that uses the long-term potentiation (LTP)/long-term depression (LTD) characteristics of the FTJ for the weight update. Simulation results based on the predictive model demonstrate that the LTP/LTD characteristics with a good trade-off between ON/OFF ratio, nonlinearity, and asymmetry can lead to high training accuracies for the FTJ-based neural network. Moreover, it is revealed that an amplitude-increasing pulse scheme may be the most favorable pulse scheme as it offers the widest ranges of pulse amplitudes and widths for achieving high accuracies. This study may provide useful guidance for the pulse design in the experimental development of high-performance memristive neural networks.http://dx.doi.org/10.1063/5.0180346 |
| spellingShingle | Haoyue Deng Zhen Fan Shuai Dong Zhiwei Chen Wenjie Li Yihong Chen Kun Liu Ruiqiang Tao Guo Tian Deyang Chen Minghui Qin Min Zeng Xubing Lu Guofu Zhou Xingsen Gao Jun-Ming Liu A physics-based predictive model for pulse design to realize high-performance memristive neural networks APL Machine Learning |
| title | A physics-based predictive model for pulse design to realize high-performance memristive neural networks |
| title_full | A physics-based predictive model for pulse design to realize high-performance memristive neural networks |
| title_fullStr | A physics-based predictive model for pulse design to realize high-performance memristive neural networks |
| title_full_unstemmed | A physics-based predictive model for pulse design to realize high-performance memristive neural networks |
| title_short | A physics-based predictive model for pulse design to realize high-performance memristive neural networks |
| title_sort | physics based predictive model for pulse design to realize high performance memristive neural networks |
| url | http://dx.doi.org/10.1063/5.0180346 |
| work_keys_str_mv | AT haoyuedeng aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT zhenfan aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT shuaidong aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT zhiweichen aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT wenjieli aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT yihongchen aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT kunliu aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT ruiqiangtao aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT guotian aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT deyangchen aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT minghuiqin aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT minzeng aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT xubinglu aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT guofuzhou aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT xingsengao aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT junmingliu aphysicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT haoyuedeng physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT zhenfan physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT shuaidong physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT zhiweichen physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT wenjieli physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT yihongchen physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT kunliu physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT ruiqiangtao physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT guotian physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT deyangchen physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT minghuiqin physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT minzeng physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT xubinglu physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT guofuzhou physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT xingsengao physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks AT junmingliu physicsbasedpredictivemodelforpulsedesigntorealizehighperformancememristiveneuralnetworks |