本文選題：平面型結(jié)構(gòu) + 晶體管��； 參考：《科學(xué)通報》2017年Z2期

【摘要】：在數(shù)據(jù)信息膨脹的大背景下,傳統(tǒng)的馮諾依曼體系結(jié)構(gòu)的計算機早已無法滿足靈活處理和存儲大量信息的需求.與之相比,大腦高度并行的非線性信息處理能力展現(xiàn)出了明顯的優(yōu)勢.突觸是大腦神經(jīng)元之間信息交換的連接渠道,突觸可塑性是生物學(xué)習(xí)和記憶的分子基礎(chǔ),為了模擬大腦,必須研究出具有生物突觸功能的物理器件,從而實現(xiàn)能夠在功能上模擬類腦行為的神經(jīng)形態(tài)電路.研究表明,對于場效應(yīng)晶體管,把具有可調(diào)憶阻特性的場效應(yīng)晶體管的傳導(dǎo)溝道和柵極作為信號傳輸和調(diào)節(jié)模塊,可以分別對應(yīng)于生物突觸中的離子傳導(dǎo)和神經(jīng)遞質(zhì)釋放的過程;對于電子型晶體管,可以利用電操作模式中的OFF(ON)狀態(tài)的弛豫傾向來模擬生物興奮(抑制)過程,因而在晶體管中可以實現(xiàn)對突觸功能的模擬.本文總結(jié)了近年來研究人員用平面晶體管結(jié)構(gòu)的突觸器件模擬的突觸可塑性功能,包括:短時程突觸可塑性(包含雙脈沖易化和雙脈沖抑制,動態(tài)過濾(高通濾波/低通濾波),處理時空關(guān)聯(lián)的脈沖,適應(yīng)等)、長時程突觸可塑性、短時程向長時程的轉(zhuǎn)變、放電時間依賴可塑性、分流抑制等,詳細說明了這些功能的特點及模擬方法并對其未來發(fā)展進行了展望.
[Abstract]:In the context of the expansion of data and information, the traditional Von Neumann architecture computer has long been unable to meet the need for flexible processing and storage of large amounts of information. Plasticity is the molecular basis of biological learning and memory. In order to simulate the brain, a physical device with a biological synaptic function must be studied to achieve a neuromorphological circuit capable of simulating the brain behavior of the brain. The research shows that the field effect transistors have the conduction channel and gate with a field effect transistor with adjustable memory resistance characteristics. As a signal transmission and adjustment module, it can correspond to the process of ion conduction and neurotransmitter release in the biological synapse. For electronic transistors, the relaxation tendency of the OFF (ON) state in the electrical operation mode can be used to simulate the biological excitation (inhibition) process, thus the synaptic function can be simulated in the transistor. This paper summarizes the synaptic plasticity functions of the researchers in recent years using a planar transistor structure to simulate synaptic plasticity, including short term synaptic plasticity (including double pulse facilitation and double pulse suppression, dynamic filtering (high pass filter / low pass filtering), processing of spatio-temporal pulses, adaption, etc.), long term synaptic plasticity, short distance direction length The change of time, the time dependent plasticity of the discharge time, and the inhibition of flow diversion, etc., are described in detail. The characteristics and simulation methods of these functions are described in detail, and their future development is prospected.
【作者單位】： 南京郵電大學(xué)信息材料與納米技術(shù)研究院有機電子與信息顯示國家重點實驗室培育基地江蘇省國家先進材料協(xié)同創(chuàng)新中心;南京工業(yè)大學(xué)先進材料研究院江蘇省柔性電子重點實驗室先進生物與化學(xué)制造協(xié)同創(chuàng)新中心;
【基金】：國家重點基礎(chǔ)研究發(fā)展計劃(2014CB648300,2015CB932200) 國家自然科學(xué)基金(61475074,61204095,61136003,61377019);國家自然科學(xué)基金優(yōu)秀青年科學(xué)基金(21322402) 江蘇省自然科學(xué)基金(BM2012010) 省級大學(xué)生創(chuàng)新訓(xùn)練計劃(SYB2016009) 有機電子與信息顯示協(xié)同創(chuàng)新中心 長江學(xué)者和創(chuàng)新團隊(IRT_15R37) 江蘇省教育委員會自然科學(xué)基金(14KJB510027) 江蘇省高校學(xué)術(shù)優(yōu)先發(fā)展計劃(PAPD)資助
【分類號】：TN32

【相似文獻】

相關(guān)期刊論文 前1條

1 鄭暢;;ROHM開發(fā)出比以往產(chǎn)品小50%的世界最小晶體管“VML0604”[J];半導(dǎo)體信息;2014年02期

，

本文編號：2005493