利用電阻保護氧化層場效電晶體雜訊電流產生器實現隨機性Izhikevich神經元電路

簡易檢索 / 詳目顯示

回結果列表

研究生：	詹凱鈞 Chan, Kai-Chun
論文名稱：	利用電阻保護氧化層場效電晶體雜訊電流產生器實現隨機性Izhikevich神經元電路 Implementation of Stochastic Izhikevich Neuron Circuit Based on Resist-Protection-Oxide Field-Effect Transistor Noise Current Generator
指導教授：	陳新 Chen, Hsin
口試委員:	金雅琴 King, Ya-Chin 彭盛裕 Peng, Sheng-Yu
學位類別：	碩士 Master
系所名稱：	電機資訊學院 - 電子工程研究所 Institute of Electronics Engineering
論文出版年：	2022
畢業學年度：	111
語文別：	中文
論文頁數：	71
中文關鍵詞：	電阻保護氧化層場效電晶體、Izhikevich 神經元
外文關鍵詞：	RPOFET, Izhikevich Neuron
相關次數：	點閱：1 下載：0
分享至:	分享至facebook 分享至twitter

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

由於人工智慧演算法需要極大的運算量，以致現行的硬體實現都具
有功耗較高的缺點。幸運地是，受到生物神經為低功耗系統的啟發，突
波式神經網路即是時下有話題性的研究方向之一，若能妥善的運用其長
處於演算法的硬體實現，想必是一強大的優勢。
我們發現神經元能夠透過雜訊進行機率型運算，而提高學習效率。
而在電路的實現上，可以透過電晶體本身的雜訊，模仿神經元的機率型
運算行為。
本論文旨在改良及設計以電阻保護氧化層場效電晶體(RPOFET)雜
訊電流產生器的方式是否可實現隨機性 Izhikevich 神經元模型。經由兩
次的晶片下線與量測驗證， RPOFET雜訊電流產生器皆未能使Izhikevich
神經元產生符合泊松分布的突波序列，主因為雜訊電流產生器電路設計
不良及高估 RPOFET 本身的雜訊能力。

Because the artificial intelligence algorithm requires a huge amount of
computation, the current hardware implementation has the disadvantage of
high power consumption. Fortunately, inspired by the fact that biological
nerves are low-power systems, the spurious neural network is one of the most
topical research directions. is a powerful advantage.
We found that neurons can perform probabilistic operations through
noise, which improves learning efficiency.
In the realization of the circuit, the probabilistic operation behavior of neurons
can be imitated through the noise of the transistor itself.
This paper aims to improve and design whether the random Izhikevich
neuron model can be realized by means of resistive protection oxide field
effect transistor (RPOFET) noise current generator. After two times of chip
off-line and measurement verification, the RPOFET noise current generator
failed to make the Izhikevich neuron generate a surge sequence conforming
to the Poisson distribution, mainly due to the poor circuit design of the noise
current generator and the overestimation of the RPOFET itself noise
capability

目錄
致謝..................................................................................................................i
摘要.................................................................................................................ii
Abstract ......................................................................................................... iii
目錄................................................................................................................iv
圖目錄............................................................................................................vi
表目錄......................................................................................................... viii
第一章 緒論...................................................................................................1
1.1 動機及貢獻 ......................................................................................1
1.2 論文大綱 ..........................................................................................1
第二章 文獻回顧...........................................................................................3
2.1 Izhikevich 神經元模型 ....................................................................3
2.2 電阻保護氧化層場效電晶體(RPOFET).........................................5
2.2.1 RPOFET 元件...............................................................................5
2.2.2 RPOFET 直流模型.......................................................................6
2.3 電阻保護氧化層場效電晶體雜訊電流產生器..............................8
2.4 隨機性 Izhikevich 神經元電路 .....................................................11
2.4.1 閃爍雜訊對 Izhikevich 神經元的影響......................................11
2.4.2 隨機性 Izhikevich 神經元應用於神經網路..............................12
2.4.3 隨機性 Izhikevich 神經元電路架構及推導..............................13
第三章 第一代隨機性 Izhikevich 神經元電路..........................................16
3.1 第一代電阻保護氧化層場效電晶體雜訊電流產生器................16
3.1.1 預計規格.....................................................................................16
3.1.2 電路架構.....................................................................................20
3.1.3 模擬及量測結果.........................................................................21
3.2 第一代隨機性 Izhikevich 神經元電路 .........................................26
3.2.1 預計規格.....................................................................................26v
3.2.2 電路架構.....................................................................................27
3.2.3 模擬及量測結果.........................................................................28
3.3 討論 ................................................................................................31
第四章 第二代隨機性 Izhikevich 神經元電路..........................................34
4.1 第二代電阻保護氧化層場效電晶體雜訊電流產生器................34
4.1.1 電路架構....................................................................................34
4.1.2 模擬及量測結果........................................................................38
4.2 第二代隨機性 Izhikevich 神經元電路 .........................................45
4.2.1 電路架構....................................................................................45
4.2.2 模擬及量測結果........................................................................51
4.3 討論 ................................................................................................55
第五章 隨機遞迴式突波神經網路晶片 ....................................................59
5.1 隨機遞迴式突波神經網路晶片....................................................59
5.1.1 系統架構....................................................................................59
5.1.2 佈局圖........................................................................................60
第六章 總結與未來展望.............................................................................67
參考文獻.......................................................................................................69
                                

[1] Nessler B, Pfeiffer M, Buesing L, Maass W (2013) Bayesian Computation Emerges
in Generic Cortical Microcircuits through Spike-Timing-Dependent Plasticity. PLOS
Computational Biology 9(4): e1003037. https://doi.org/10.1371/journal.pcbi.1003037
[2] Diehl PU and Cook M (2015) Unsupervised learning of digit recognition using
spike-timing-dependent plasticity. Front. Comput. Neurosci. 9:99. doi:
10.3389/fncom.2015.00099
[3] E. Neftci, S. Das, B. Pedroni, K. Kreutz-Delgado, and G. Cauwenberghs,
“Event-driven contrastive divergence for spiking neuromorphic systems,”
Frontiers in Neuroscience, vol. 8, pp. 1–14, 2014
[4] Moreno-Bote R (2014) Poisson-Like Spiking in Circuits with Probabilistic
Synapses. PLoS Comput Biol 10(7): e1003522. doi:10.1371/journal.pcbi.1003522
[5] T.J. Chiu, Y.C. King, J. Gong, Y.H. Tsai, and H. Chen* “A ResistiveProtectiveOxide Transistor with Adaptable Low-frequency Noise for Stochastic
Neuromorphic Computation in VLSI” the IEEE Electron Device Letters, vol.32,
no.9, pp.1293-1295, 2011
[6] T.H. Yu “Stochastic Silicon Neuron Circuits Based on RPO-FET for
Stochastic Neuromorphic System”,2020
[7] E. M. Izhikevich, "Simple model of spiking neurons," in IEEE Transactions on
Neural Networks, vol. 14, no. 6, pp. 1569-1572, Nov. 2003, doi:
10.1109/TNN.2003.820440.
[8] E. M. Izhikevich, "Which model to use for cortical spiking neurons?," in IEEE
Transactions on Neural Networks, vol. 15, no. 5, pp. 1063-1070, Sept. 2004, doi:
10.1109/TNN.2004.832719.
[9] Jason Huang, "Noise Controlled Transistor Modeling and Application Circuit",2016.
[10] Yu-Cheng, Yao, “The Implementation of Silicon Neuron Circuits with Stochastic
Behavior”, 201870
[11] Y. Huang, T. Yang, S. Hsu, X. Chen and J. Chiou, "A novel pseudo resistor structure
for biomedical front-end amplifiers," 2015 37th Annual International Conference of the
IEEE Engineering in Medicine and Biology Society (EMBC), Milan, 2015, pp. 2713-
2716, doi: 10.1109/EMBC.2015.7318952.
[12] Wu-Shun Lai, “The Analysis and Implementation of Stochastic Spiking Neural
Network with on-chip learning capability”, 2020.
[13] Fatahi, M., Ahmadi, M., Shahsavari, M., Ahmadi, A., & Devienne, P. (2016).
evt_MNIST: A spike based version of traditional MNIST. arXiv preprint
arXiv:1604.06751
[14] C. Bartolozzi, S. Mitra and G. Indiveri, "An ultra low power current-mode filter
for neuromorphic systems and biomedical signal processing," 2006 IEEE Biomedical
Circuits and Systems Conference, London, 2006, pp. 130-133, doi:
10.1109/BIOCAS.2006.4600325.
[15] Y. Huang, T. Yang, S. Hsu, X. Chen and J. Chiou, "A novel pseudo resistor structure
for biomedical front-end amplifiers," 2015 37th Annual International Conference of the
IEEE Engineering in Medicine and Biology Society (EMBC), Milan, 2015, pp. 2713-
2716, doi: 10.1109/EMBC.2015.7318952.
[16] R. Brederlow, R. Prakash, C. Paulus, R. Thewes, “A Low-Power True Random
Number Generator using Random Telegraph Noise of Single OxideTraps,” ISSCC Dig.
Tech. Papers, pp. 536-537, Feb. 2006
[17] M. Matsumoto, S. Yasuda, R. Ohba, K. Ikegami, T. Tanamoto and S.
Fujita,"1200μm2 Physical Random-Number Generators Based on SiN MOSFET for
Secure Smart-Card Application," 2008 IEEE International Solid-State Circuits
Conference -Digest of Technical Papers, San Francisco, CA, 2008, pp. 414-624,
doi:10.1109/ISSCC.2008.4523233.
[18] C. Frenkel, J. Legat and D. Bol, "A compact phenomenological digital neuron
implementing the 20 Izhikevich behaviors," 2017 IEEE Biomedical Circuits and 57
Systems Conference (BioCAS), Turin, 2017, pp. 1-4,
doi:10.1109/BIOCAS.2017.8325231.71
[19] M.Ortmanns, A.Buhmann, Y.Manoli, "Comprehensive Microsystems" , 2008

簡易檢索 / 詳目顯示

相關論文