簡易檢索 / 詳目顯示

回結果列表
研究生： 許竣翔
Hsu, Chun-Hsiang
論文名稱： 電阻式記憶體類神經網路應用之模型開發
Modeling of Resistive Random-Access Memory in Neural Networks
指導教授： 江孟學
Chiang, Meng-Hsueh
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 奈米積體電路工程碩士博士學位學程
MS Degree/Ph.D. Program on Nano-Integrated-Circuit Engineering
論文出版年： 2019
畢業學年度： 108
語文別： 英文
論文頁數： 61
中文關鍵詞： 電阻式隨機記憶體 、精簡模型 、深度學習 、人工神經突觸 、Verilog-A 、HSPICE 、TensorFlow
外文關鍵詞： RRAM, Compact Model, Deep Learning, Artificial Synapses, Verilog-A, HSPICE, TensorFlow
相關次數： 點閱：240  下載：9 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 近年來，人工智慧迅速蓬勃發展，對於類神經網路的元件要求也勢必提高，小尺寸與低耗能的優點使電阻式記憶體(RRAM)為目前廣泛討論、使用。本篇論文旨在探討利用硬體描述語言Verilog-A建立電阻式記憶體(RRAM)模型，同時結合元件物理特性，輔以電路模擬軟體(HSPICE)測試電性；RRAM在操作上類似破壞性質，藉由輸入偏壓(或電流)改變元件電阻，透過內部絕緣層狀態改變使之呈現高低不同阻值。在模型中，模擬RRAM I-V曲線及提供導電絲間隙、溫度變化計算模組。
    另外，模型提供多階式儲存單元(MLC)功能，預期能提高記憶體元件電路設計效率；權重計算模組能提供不同電導值，模擬仿生網路之類比行為操作的能力，並利用深度學習軟體(TensorFlow)建立類神經網路，進行權重與演算法的訓練，能夠預測長期增強(LTP)、長期漸弱(LTD)等類比行為的功耗及達成手寫圖形資料辨識能力，並展示二進位制模型特性與類比行為模型在深度學習上差異。

    Artificial Intelligence (AI) has been developed prosperously in recent years; therefore, the improvement of devices has been demanded strictly. RRAM has become one of extraordinary candidates to be discussed because of small size and low power consumption. The thesis focuses on developing the RRAM compact model with Verilog-A describing physical characteristics with mathematic formula and obtaining electrical characteristics in the circuit with HSPICE. The operation of RRAM is to present different resistance states resulted from variation in the insulting layer by the applied bias voltage or current. We can simulate the I-V curve of RRAM and propose the gap between top electrode and conductive filaments (CFs) using gap calculation module and temperature calculation module.
    Furthermore, we can achieve multilevel cell (MLC) to improve the efficiency of programming strategies and different conductance can be acquired to simulate analog synaptic behaviors with the weight modulation module. Using TensorFlow which is the software for deep learning to construct neural networks and have the training of the algorithm, we can predict analog behaviors and power consumption in long-term potentiation (LTP) and long-term depression (LTD), achieve the hand-written pattern recognition and present differences between the binary model and the analog behavior model.

    摘要 I Abstract II 致謝 III Contents V Figure Captions VIII Chapter 1 Introduction 1 1.1 Research Motivation 1 1.2 Introduction of Software 1 1.3 Chapter Architecture 2 Chapter 2 Literature Review 3 2.1 Introduction of Resistive Random-Access Memory (RRAM) 3 2.1.1 Resistance-Transforming Device 3 2.1.2 Mechanism of the Transformation of Resistances 4 2.2 Introduction of the Resistance-Transforming Mechanism 4 2.2.1 Space Charge Limited Current 5 2.2.2 Ionic Drift Current Model [16] 6 2.3 Relevant Research 7 2.3.1 Electric Field Distribution Simulation 7 2.3.2 Ion Migration Model [18] 8 Chapter 3 RRAM Compact Model [3] 23 3.1 Current Calculation Module 24 3.1.1 Forming Process 24 3.1.2 SET Process 25 3.1.3 RESET Process 26 3.2 Memory Module 28 3.3 Simulation Results and Discussion 28 3.3.1 Physical Characteristic Simulation 28 3.3.2 Transient Pulse Simulation 29 Chapter 4 Updated Module 36 4.1 Research on Physical Mechanism 36 4.1.1 Joule Heating Effect 36 4.2 Improved Architecture of the Model 38 4.2.1 Temperature Calculation Module 38 4.2.2 Gap Calculation Module [25] 39 4.2.3 Multilevel Resistance 39 4.3 Simulation Results 40 Chapter 5 RRAM in Neural Networks 45 5.1 Linear and Nonlinear Weight Model [38] 45 5.2 Synaptic Characteristic Device 47 5.3 Neural Networks Implementation 47 5.3.1 Introduction of Neural Networks 47 5.3.2 Implementation in TensorFlow 48 5.4 Simulation Results 49 Chapter 6 Conclusion 55 References 56 Appendix 60

    [1]Accellera, “Verilog-AMS language reference manual version2.2, November 2004,” available online: http://www.accelera.org/home.
    [2]Tutorials – TensorFlow, available online: https://www.tensorflow.org/tutorials.
    [3]許凱翔(2013)，“電阻式記憶體的SPICE模型分析與建立，國立宜蘭大學電子工程學系碩士論文.
    [4]Meng-Hsueh Chiang, Senior Member, IEEE, Kai-Hsiang Hsu, Wei-Wen Ding, and Bo-Ren Yang, “A Predictive Compact Model of Bipolar RRAM Cells for Circuit Simulations,” IEEE Transactions on electron devices, vol. 62, No. 7, July 2015.
    [5]P. D. Greene, E. L. Bush, and I. R. Rawlings, Thin Film Dielectrics, New York: Electrochemical Society, 1968.
    [6]A. Beck, J. G. Bednorz, Ch. Gerber, C. Rossel, and D. Widmer, “Reproducible switching effect in thin oxide films for memory applications,” Applied Physics Letters, vol. 77, pp.139-141, 2000.
    [7]S. Seo, M. J. Lee, D. H. Seo, E. J. Jeoung, D.-S. Suh, Y. S. Joung, I. K. Yoo, I. R. Hwang, S. H. Kim, I. S. Byun, J.-S. Kim, J. S. Choi, and B. H. Park, “Reproducible resistance switching in polycrystalline NiO films,” Applied Physics Letters, vol. 85, pp.5655-5657, 2004.
    [8]B. J. Choi, D. S. Jeong, S. K. Kim, C. Rohde, S. Choi, J. H. Oh, H. J. Kim, C. S. Hwang, K. Szot, R. Waser, B. Reichenberg, and S. Tiedke, “Resistive switching mechanism of TiO2 thin films grown by atomic-layer deposition,” Applied Physics Letters, vol. 98, pp. 033715 - 033715-10, 2005.
    [9]H.-S. P. Wong, H. Y. Lee, S. Yu, Y. S. Chen, Y. Wu, P. S. Chen, B. Lee, F. T. Chen, and M. J. Tsai, “Metal–Oxide RRAM,“ Proceedings of the IEEE, vol. 100, pp. 1951-1970, June 2012.
    [10]A. Sawa, “Resistive switching in transition metal oxides,” Materials Today, vol. 11, pp. 28-36, June 2008.
    [11]R. Waser, and M. Aono, “Nanoionics-based resistive switching memories,” Nature Materials, vol. 6, pp. 833-840, 2007.
    [12]熊昌鉑(2010)，“垂直與水平架構之導電絲電阻記憶元件的製作與導電絲成長暨記憶穩定性機制的探討”，國立清華大學材料科學與工程學系博士論文p.18.
    [13]M. A. Lampert, and P. Mark, Current Injection in Solids (Academic, New York, 1970).
    [14]H. Y. Lee, P. S. Chen, T. Y. Wu, Y. S. Chen, C. C. Wang, P. J. Tzeng, C. H. Lin, F. Chen, C. H. Lien, and M.-J. Tsai, “Low power and high speed bipolar switching with a thin reactive Ti buffer layer in robust HfO2 based RRAM,“ Tech. Dig. IEDM, 2008, pp. 1-4.
    [15]S. Haddad, W. Y. Ching, F. T. Ning, L. Zhida, S. Avanzino, S. Pangrle, M. Buynoski, M. Rathor, C. Wei, N. Tripsas, C. Bill, M. VanBuskirk, and M. Taguchi, “Non-Volatile resistive switching for advanced memory applications,” Tech. Dig. IEDM, Dec. 2005, pp. 746-749.
    [16]D. B. Strukov, and R. S. Williams, “Exponential ionic drift: fast switching and low volatility of thin-film memristors,” Applied Physics A, vol. 94, pp. 515-519, 2009.
    [17]Q. Liu, C. Dou, Y. Wang, S. Long, W. Wang, M. Liu, M. Zhang, and Junning Chen, “Formation of multiple conductive filaments in the Cu/ZrO2 :Cu/Pt device,” Applied Physics Letters, vol. 95, pp. 1-3, 2009.
    [18]S. Larentis, F. Nardi, S. Balatti, D. C. Gilmer, and Daniele Ielmini, “Resistive switching by voltage-driven ion migration in bipolar RRAM—Part II: Modeling,” IEEE Transactions on Electron Devices, vol. 59, pp. 2468-2475, Sept. 2012
    [19]C. Sire, S. Blonkowski, M.J. Gordon, and T. Baron, “Statistics of electrical breakdown field in HfO2 and SiO2 films from millimeter to nanometer length scales,”Applied Physics Letters, vol. 91, pp. 1-3, 2007.
    [20]S. Yu, X. Guan, and H.-S. Philip Wong, “Conduction mechanism of TiN/HfOx/Pt resistive switching memory: A trap-assisted-tunneling model,” Applied Physics Letters, vol. 99, pp. 1-3,2001.
    [21]廖翊博(2008)， “運用 Verilog-A 實現的相變化記憶體溫度變化模型” ，國立宜蘭大學電子工程研究所碩士論文pp. 24-25.
    [22]X. Guan, S. Yu, and H.-S. Philip Wong, “On the switching parameter variation of metal-oxide RRAM—Part I: Physical modeling and simulation methodology,” IEEE Transactions on Electron Devices, vol. 59, pp. 1172-1182, April 2012.
    [23]X. Q. Wei, L. P. Shi, R. Walia, T. C. Chong, R. Zhao, X. S. Miao, and B. S. Quek “HSPICE macromodel of PCRAM for binary and multilevel storage,” IEEE Transactions on Electron Devices, vol. 53, pp. 56-62, 2005.
    [24]S. S. Sheu, M. F. Chang, K. F. Lin, C. W. Wu, Y. S. Chen, P. F. Chiu, C. C. Kuo, Y. S. Yang, P. C. Chiang, W. P. Lin, C. H. Lin, H. Y. Lee, P. Y. Gu, S. M. Wang, F. T. Chen, K. L. Su, C. H. Lien, K. H. Cheng, H. T. Wu, T. K. Ku, M. J. Kao, and M. J. Tsai “A 4Mb embedded SLC resistive-RAM macro with 7.2ns read-write random-access time and 160ns MLC-access capability,” ISSCC, Feb. 2011, pp. 200-202.
    [25]Pai-Yu Chen ; Shimeng Yu , “Compact Modeling of RRAM Devices and Its Applications in 1T1R and 1S1R Array Design,” IEEE Transactions on Electron Devices, Volume: 62 , Issue: 12 , Dec. 2015.
    [26]D. Ielmini, “Modeling the universal set/reset characteristics of bipolar RRAM by field- and temperature-driven filament growth,” IEEE Trans. Electron Devices, vol. 58, no. 12, pp. 4309–4317, Dec. 2011.
    [27]Shimeng Yu, “Neuro-inspired computing with emerging nonvolatile memorys, ” Proceedings of the IEEE, Volume: 106 , Issue: 2 , Feb. 2018.
    [28]Jiyong Woo, Andrea Padovani, Kibong Moon, Myounghun Kwak, Luca Larcher, and Hyunsang Hwang, “Linking Conductive Filament Properties and Evolution to Synaptic Behavior of RRAM Devices for Neuromorphic Application, ” IEEE Electron Device Letters, Vol 38, No 9, Sep. 2017.
    [29]Daniele Garbin, Elisa Vianello, Olivier Bichler, Quentin Rafhay, Christian Gamrat, Gérard Ghibaudo, Fellow, IEEE, Barbara DeSalvo, and Luca Perniola, “HfO2-Based OxRAM Devices as Synapses for Convolutional Neural Networks,” IEEE Transactions on Electron Devices vol. 62, Issue: 8, Aug. 2015.
    [30]D. Kuzum, S. Yu, and H.-S. P. Wong, “Synaptic electronics: Materials, devices and applications,” Nanotechnology, vol. 24, no. 38, p. 382001, 2013.
    [31]Jia-Wei Lee, Chun-Hsiang Hsu, and Meng-Hsueh Chiang, “A Predictive Resistive RAM Compact Model with Synaptic Behavior for Circuit Simulations,” Informatics, Electronics and Microsystems: TechConnect Briefs, vol. 4, May. 2018.
    [32]Po-An Chen, Rui-Jing Ge, Jia-Wei Lee, Chun-Hsiang Hsu, Wei-Chou Hsu, Deji Akinwande, and Meng-Hsueh Chiang, “An RRAM with a 2D Material Embedded Double Switching Layer for Neuromorphic Computing,” Oct. 2018 IEEE Nanotechnology Materials and Devices Conference (NMDC).
    [33]Ligang Gao, Pai-Yu Chen, and Shimeng Yu, “Weight Tuning of Resistive Memories and Convolution Kernel Operation on Cross-Point Array for Neuro-Inspired Computing,” Oct. 2016 IEEE International Conference on Solid-State and Integrated Circuit Technology (ICSICT).
    [34]Ligang Gao, Pai-Yu Chen, and Shimeng Yu, “Demonstration of Convolution Kernel Operation on Resistive Cross-Point Array,” IEEE Electron Device Letters, Vol 37, No 7, Jul. 2016.
    [35]D. Garbin, O. Bichler, E. Vianello, Q. Rafhay, C. Gamrat, L.Perniola, G. Ghibaudo, B. DeSalvo, “Variability-tolerant Convolutional Neural Network for Pattern Recognition Applications based on OxRAM Synapses,” Dec. 2014 IEEE International Electron Devices Meeting (IEDM).
    [36]Park S, Chu M, Kim J, Noh J, Jeon M, Hun Lee B, Hwang H, Lee B, Lee BG, “Electronic system with memristive synapses for pattern recognition,” Sci Rep. 2015 May 5;5:10123. doi: 10.1038/srep10123.
    [37]Chih-Cheng Chang, Jen-Chieh Liu, Yu-Lin Shen, Teyuh Chou, Pin-Chun Chen, I.-Ting Wang, Chih-Chun Su, Ming-Hong Wu, Boris Hudec, Che-Chia Chang, Chia-Ming Tsai, Tian-Sheuan Chang, H.-S. Philip Wong, Tuo-Hung Hou, “Challenges and opportunities toward online training acceleration using RRAM-based hardware neural network,” Dec. 2017 IEEE International Electron Devices Meeting (IEDM).
    [38]Ligang Gao, I-Ting Wang, Pai-Yu Chen, Sarma Vrudhula, Jae-sun Seo, Yu Cao, Tuo-Hung Hou and Shimeng Yu, “Fully parallel write/read in resistive synaptic array for accelerating on-chip learning,” Nanotechnology 26 (2015) 455204 (9pp).
    [39]K. Moon, S. Lim, J. Park, C. Sung, S. Oh, J. Woo, J. Lee and H. Wang, “RRAM-based synapse devices for neuromorphic systems,” Faraday Discussions, Vol 213, 2019.
    [40]Sohyeon Kim, Yawar Abbas, Yu-Rim Jeon, Andrey Sergeevich Sokolov, Boncheeol Ku and Changhwan Choi, “Engineering synaptic characteristics of TaOx/HfO2 bi-layered resistive switching device,” Nanotechnology 29 (2018) 415204 (8pp)
    [41]Shimeng Yu, “Neuro-inspired computing with emerging nonvolatile memorys,” Proceedings of the IEEE, Vol 106, Issue 2, Feb. 2018.
    [42]洪瑋辰(2019)，類比非揮發記憶體仿生電路之深度學習模擬平台，國立成功大學微電子所碩士論文.
    [43]Gong, N., et al., “Signal and extraction from analog memory elements for neuromorphic computing,” Nature Communication, 2018.9 (I): p2012.

    無法下載圖示 校內：2023-01-01公開
    校外：不公開
    電子論文尚未授權公開，紙本請查館藏目錄
    QR CODE