Simultaneous emulation of synaptic and intrinsic plasticity using a memristive synapse

Abstract

Neuromorphic computing targets the hardware embodiment of neural network, and device implementation of individual neuron and synapse has attracted considerable attention. The emulation of synaptic plasticity has shown promising results after the advent of memristors. However, neuronal intrinsic plasticity, which involves in learning process through interactions with synaptic plasticity, has been rarely demonstrated. Synaptic and intrinsic plasticity occur concomitantly in learning process, suggesting the need of the simultaneous implementation. Here, we report a neurosynaptic device that mimics synaptic and intrinsic plasticity concomitantly in a single cell. Threshold switch and phase change memory are merged in threshold switch-phase change memory device. Neuronal intrinsic plasticity is demonstrated based on bottom threshold switch layer, which resembles the modulation of firing frequency in biological neuron. Synaptic plasticity is also introduced through the nonvolatile switching of top phase change layer. Intrinsic and synaptic plasticity are simultaneously emulated in a single cell to establish the positive feedback between them. A positive feedback learning loop which mimics the retraining process in biological system is implemented in threshold switch-phase change memory array for accelerated training.

Highly Reliable Threshold Switching Characteristics of Surface-Modulated Diffusive Memristors Immune to Atmospheric Changes

“…The crossbar array using the above 1S1R device with 1/2 V operation scheme ,− can be scaled up to approximately 1000 × 1000 for a 10% readout margin, according to a calculation based on the one-bit line pull-up method and extracted parameters (see Figure S15 and Table S2). − Stable threshold switching behavior and fast switching are suitable for mimicking the action potential of biological neurons. − The leaky LIF model, a popular neuronal firing model, is well manifested by using a diffusive memristor with a multicomponent circuit designed with resistors and capacitors. A neuron generates an output spike when the membrane potential, which is the difference in potential between the interior and exterior of the neuron, is integrated through the input stimulus and reaches a particular threshold voltage.…”

Synapse-Mimetic Hardware-Implemented Resistive Random-Access Memory for Artificial Neural Network

“…A memristor is a two-terminal electrical component in which an active material is sandwiched between a top electrode (TE) and a bottom electrode (BE) [ 57 , 58 , 59 , 60 ]. Memristive behavior entails the functionalized hysteresis of electrical resistance and can be quantified by a current value corresponding to a voltage sweep [ 61 , 62 , 63 , 64 ]. As shown in Figure 1 , a typical I–V curve of a memristor exhibits an identical “butterfly curve” shape by changing its resistance.…”

“…Sung et al, demonstrated a heterostructured memristor that simultaneously mimicked both neurons and synapses [ 64 ]. Figure 4 a shows a structure in which a silver-filament-based memristor is combined with a phase-change memory (PCM) material using the formed silver filament as the BE, corresponding to the signal-processing roles of biological neurons and synapses, respectively.…”

“…( g ) Excited signal output of TS layer indicating emulated intrinsic plasticity; adapted with permission from ref. [ 64 ], copyright 2022 Springer Nature.…”

Temperature Regulated Artificial Neuron Based on Memristor

“…Artificial neurons based on the implementation of a single device play an important role in neuromorphic computing. In contrast to traditional CMOS-based approaches, which have complex architectures and energy inefficiencies, devices with individual neuron are energy efficient and scalable [4], [5], [6], [7], [8], [9]. Therefore, the artificial neurons base on a single memristor have attracted extensive attention in recent years, and artificial neuron devices are significant for developing neuromorphic intelligent computers with advanced cognitive functions, developing brain-like computing and advancing artificial intelligence [10], [11], [12], 13].…”