Experimental Demonstration of Ferroelectric Spiking Neurons for Unsupervised Clustering
Experimental Demonstration of Ferroelectric Spiking Neurons for Unsupervised Clustering
复制标题
DOI：
10.1109/iedm.2018.8614586
复制
发表时间：
2018-12
期刊：
影响因子：
--
通讯作者：
Z. Wang;Brian Crafton;Jorge Gomez;R. Xu;Aileen Luo;Z. Krivokapic;L. Martin;S. Datta;A. Raychowdhury;A. Khan
中科院分区：
文献类型：
--
作者：
Z. Wang;Brian Crafton;Jorge Gomez;R. Xu;Aileen Luo;Z. Krivokapic;L. Martin;S. Datta;A. Raychowdhury;A. Khan
我们报告了第一个基于铁电场效应晶体管（FEFET）的脉冲神经元的实验演示。本文所证明的铁电（FE）神经元的一个独特特征是在紧凑的1T-1FEFET结构中存在兴奋性和抑制性输入连接，这也是任何神经元实现的第一次报道。这种双神经元功能是仿生神经网络的关键要求，代表了第三代峰值神经网络（snn）实现的突破——本文也首次报道了对真实世界数据的无监督学习和聚类。我们演示的关键是仔细设计两个重要的器件级特征：(1)在没有稳定状态的情况下，ffet的突然迟滞转变，以及(2)通过偏置条件允许抑制功能的ffet迟滞的动态可调性。实验校准，基于多域Preisach的FEFET模型用于精确模拟FE神经元并在缩放节点上预测其性能。我们还实现了一个SNN用于无监督聚类，并对模拟CMOS和新兴技术的网络性能进行了基准测试，并观察到(1)兴奋性和抑制性神经连接的统一，(2)基于STDP的学习，(3)分类过程中最低的报告功率（3.6nW），以及(4)分类准确率为93%。
We report the first experimental demonstration of ferroelectric field-effect transistor (FEFET) based spiking neurons. A unique feature of the ferroelectric (FE) neuron demonstrated herein is the availability of both excitatory and inhibitory input connections in the compact 1T-1FEFET structure, which is also reported for the first time for any neuron implementations. Such dual neuron functionality is a key requirement for bio-mimetic neural networks and represents a breakthrough for implementation of the third generation spiking neural networks (SNNs)—also reported herein for unsupervised learning and clustering on real world data for the first time. The key to our demonstration is the careful design of two important device level features: (1) abrupt hysteretic transitions of the FEFET with no stable states therein, and (2) the dynamic tunability of the FEFET hysteresis by bias conditions which allows for the inhibition functionality. Experimentally calibrated, multi-domain Preisach based FEFET models were used to accurately simulate the FE neurons and project their performance at scaled nodes. We also implement an SNN for unsupervised clustering and benchmark the network performance across analog CMOS and emerging technologies and observe (1) unification of excitatory and inhibitory neural connections, (2) STDP based learning, (3) lowest reported power (3.6nW) during classification, and (4) a classification accuracy of 93%.