Hybrid spin-CMOS stochastic spiking neuron for high-speed emulation of In vivo neuron dynamics
Hybrid spin-CMOS stochastic spiking neuron for high-speed emulation of In vivo neuron dynamics
复制标题
DOI：
10.1049/iet-cdt.2017.0145
复制
发表时间：
2018-02
期刊：
影响因子：
--
通讯作者：
Steven D. Pyle;Kerem Y Çamsarı;R. Demara
中科院分区：
文献类型：
--
作者：
Steven D. Pyle;Kerem Y Çamsarı;R. Demara
本文所开发的自旋电子随机脉冲神经元（S3N）实现了在体皮质神经元中观察到的仿生随机脉冲特性，同时其运行速度快几个数量级，并且具有良好的能量特性。这项工作利用了一种新型的概率性自旋电子开关元件器件，该器件在紧凑、低能耗和高速的封装中提供了热驱动和电流控制的可调随机性。为了形成闭环，作者使用了具有可变权重控制的二阶互补金属氧化物半导体（CMOS）突触，它将传入的脉冲累积为二阶瞬态电流信号，这些信号类似于生物神经元中的兴奋性突触后电位，并且可用于驱动突触后的S3N。集成电路专用模拟程序（SPICE）模拟结果表明，相当于1秒的在体神经元脉冲特性可以在纳秒级产生，这使得对皮质信息处理的未来统计模型进行极快速的在体神经元行为模拟成为可能。他们的结果还表明，根据脉冲频率的不同，S3N可以在十皮秒级产生脉冲，而功耗仅为0.6 - 9.6微瓦。此外，他们证明了S3N可以实现感知机功能，例如基于与门和或门的逻辑处理，并为更先进的随机神经形态架构提供了该工作未来的扩展方向。
The spintronic stochastic spiking neuron (S3N) developed herein realises biologically mimetic stochastic spiking characteristics observed within in vivo cortical neurons, while operating several orders of magnitude more rapidly and exhibiting a favourable energy profile. This work leverages a novel probabilistic spintronic switching element device that provides thermally-driven and current-controlled tunable stochasticity in a compact, low-energy, and high-speed package. In order to close the loop, the authors utilise a second-order complementary metal-oxide-semiconductor (CMOS) synapse with variable weight control that accumulates incoming spikes into second-order transient current signals, which resemble the excitatory post-synaptic potentials found in biological neurons, and can be used to drive post-synaptic S3Ns. Simulation program with integrated circuit emphasis (SPICE) simulation results indicate that the equivalent of 1 s of in vivo neuronal spiking characteristics can be generated on the order of nanoseconds, enabling the feasibility of extremely rapid emulation of in vivo neuronal behaviours for future statistical models of cortical information processing. Their results also indicate that the S3N can generate spikes on the order of ten picoseconds while dissipating only 0.6-9.6 μW, depending on the spiking rate. Additionally, they demonstrate that an S3N can implement perceptron functionality, such as AND-gate- and OR-gate-based logic processing, and provide future extensions of the work to more advanced stochastic neuromorphic architectures.