Emerging Memristive Artificial Synapses and Neurons for Energy-Efficient Neuromorphic Computing

Field-Created Coordinate Cation Bridges Enable Conductance Modulation and Artificial Synapse within Metal Nanoparticles

When metal nanoparticles are functionalized with charged ligands, the movement of counterions and conduction electrons is coupled, which enables us to develop electronic devices, including diodes, transistors, and logic gates, but dynamically modulating the conductivity of a synaptic device within these materials has proved challenging. Here we show that an artificial synapse can be created from thin films of functionalized metal nanoparticles using an active silver electrode. The electric-field-injected Ag⁺ coordinates with carboxyl ligands that sets up a conduction bridge to increase the nanoparticle conductivity by reducing the electron tunneling/hopping energy barriers. The dynamic modulation of conductivity allows us to implement several important synaptic functions such as potentiation/depression, paired-pulse facilitation, learning behaviors including short-term to long-term memory transition, self-learning, and massed leaning vs spaced learning. Finally, based on the nonvolatile characteristics, the metal nanoparticle synapse is used to build a single-layer hardware spiking neural network (SNN) for pattern recognition.

Multi-Stimuli-Responsive Synapse Based on Vertical van der Waals Heterostructures

Brain-inspired intelligent systems demand diverse neuromorphic devices beyond simple functionalities. Merging biomimetic sensing with weight-updating capabilities in artificial synaptic devices represents one of the key research focuses. Here, we report a multiresponsive synapse device that integrates synaptic and optical-sensing functions. The device adopts vertically stacked graphene/h-BN/WSe₂ heterostructures, including an ultrahigh-mobility readout layer, a weight-control layer, and a dual-stimuli-responsive layer. The unique structure endows synapse devices with excellent synaptic plasticity, short response time (3 μs), and excellent optical responsivity (10⁵ A/W). To demonstrate the application in neuromorphic computing, handwritten digit recognition was simulated based on an unsupervised spiking neural network (SNN) with a precision of 90.89%, well comparable with the state-of-the-art results. Furthermore, multiterminal neuromorphic devices are demonstrated to mimic dendritic integration and photoswitching logic. Different from other synaptic devices, the research work validates multifunctional integration in synaptic devices, supporting the potential fusion of sensing and self-learning in neuromorphic networks.

Progressive and Stable Synaptic Plasticity with Femtojoule Energy Consumption by the Interface Engineering of a Metal/Ferroelectric/Semiconductor

In the era of "big data," the cognitive system of the human brain is being mimicked through hardware implementation of highly accurate neuromorphic computing by progressive weight update in synaptic electronics. Low-energy synaptic operation requires both low reading current and short operation time to be applicable to large-scale neuromorphic computing systems. In this study, an energy-efficient synaptic device is implemented comprising a Ni/Pb(Zr_0.52 Ti_0.48 )O₃ (PZT)/0.5 wt.% Nb-doped SrTiO₃ (Nb:STO) heterojunction with a low reading current of 10 nA and short operation time of 20-100 ns. Ultralow femtojoule operation below 9 fJ at a synaptic event, which is comparable to the energy required for synaptic events in the human brain (10 fJ), is achieved by adjusting the Schottky barrier between the top electrode and ferroelectric film. Moreover, progressive domain switching in ferroelectric PZT successfully induces both low nonlinearity/asymmetry and good stability of the weight update. The synaptic device developed here can facilitate the development of large-scale neuromorphic arrays for artificial neural networks with low energy consumption and high accuracy.

Neuro-inspired electronic skin for robots

Touch is a complex sensing modality owing to large number of receptors (mechano, thermal, pain) nonuniformly embedded in the soft skin all over the body. These receptors can gather and encode the large tactile data, allowing us to feel and perceive the real world. This efficient somatosensation far outperforms the touch-sensing capability of most of the state-of-the-art robots today and suggests the need for neural-like hardware for electronic skin (e-skin). This could be attained through either innovative schemes for developing distributed electronics or repurposing the neuromorphic circuits developed for other sensory modalities such as vision and audio. This Review highlights the hardware implementations of various computational building blocks for e-skin and the ways they can be integrated to potentially realize human skin-like or peripheral nervous system-like functionalities. The neural-like sensing and data processing are discussed along with various algorithms and hardware architectures. The integration of ultrathin neuromorphic chips for local computation and the printed electronics on soft substrate used for the development of e-skin over large areas are expected to advance robotic interaction as well as open new avenues for research in medical instrumentation, wearables, electronics, and neuroprosthetics.

Multifunctional Analog Resistance Switching of Si3N4-Based Memristors through Migration of Ag+ Ions and Formation of Si-Dangling Bonds

With forming-free, self-rectifying, and self-compliant properties, memristors can effectively prevent themselves from experiencing leakage currents and overshoot voltages without any additional circuitry. However, the implementation of all these features in a single memristor remains a challenge. Herein, a multifunctional Si3N4-based memristor with a structure of Ag/a-SiNx/p++-Si has been fabricated, and it was demonstrated, for the first time, that the device exhibits novel analog resistance switching behaviors, such as being forming-free, self-rectifying, and self-compliant, presenting well a coexistence of volatile and nonvolatile performance of resistance switching. The multifunctional analog resistance switching could be attributed to the formation of the Si-dangling bond channel and the migration of Ag+ ions inside the a-SiNx layer. Our current results might provide an insightful understanding of the resistance switching mechanism of Si3N4-based memristors, and the device with a large on/off ratio (>103) and robust retention (>103 s) and endurance (>103 cycles) shows potential for application in crossbar synaptic array devices.

Multimodal Tuning of Synaptic Plasticity Using Persistent Luminescent Memitters

Mimicking memory processes, including encoding, storing, and retrieving information, is critical for neuromorphic computing and artificial intelligence. Synaptic behavior simulations through electronic, magnetic, or photonic devices based on metal oxides, 2D materials, molecular complex and phase change materials, represent important strategies for performing computational tasks with enhanced power efficiency. Here, a special class of memristive materials based on persistent luminescent memitters (termed as a portmanteau of "memory" and "emitter") with optical characteristics closely resembling those of biological synapses is reported. The memory process and synaptic plasticity can be successfully emulated using such memitters under precisely controlled excitation frequency, wavelength, pulse number, and power density. The experimental and theoretical data suggest that electron-coupled trap nucleation and propagation through clustering in persistent luminescent memitters can explain experience-dependent plasticity. The use of persistent luminescent memitters for multichannel image memorization that allows direct visualization of subtle changes in luminescence intensity and realization of short-term and long-term memory is also demonstrated. These findings may promote the discovery of new functional materials as artificial synapses and enhance the understanding of memory mechanisms.

Energy-Efficient III-V Tunnel FET-Based Synaptic Device with Enhanced Charge Trapping Ability Utilizing Both Hot Hole and Hot Electron Injections for Analog Neuromorphic Computing

A charge trap device based on field-effect transistors (FET) is a promising candidate for artificial synapses because of its high reliability and mature fabrication technology. However, conventional MOSFET-based charge trap synapses require a strong stimulus for synaptic update because of their inefficient hot-carrier injection into the charge trapping layer, consequently causing a slow speed operation and large power consumption. Here, we propose a highly efficient charge trap synapse using III-V materials-based tunnel field-effect transistor (TFET). Our synaptic TFETs present superior subthreshold swing and improved charge trapping ability utilizing both carriers as charge trapping sources: hot holes created by impact ionization in the narrow bandgap InGaAs after being provided from the p⁺-source, and band-to-band tunneling hot electrons (BBHEs) generated at the abrupt p⁺n junctions in the TFETs. Thanks to these advances, our devices achieved outstanding efficiency in synaptic characteristics with a 5750 times faster synaptic update speed and 51 times lower sub-fJ/um² energy consumption per single synaptic update in comparison to the MOSFET-based synapse. An artificial neural network (ANN) simulation also confirmed a high recognition accuracy of handwritten digits up to ∼90% in a multilayer perceptron neural network based on our synaptic devices.

A Photoelectric Spiking Neuron for Visual Depth Perception

The biological visual system encodes optical information into spikes and processes them by the neural network, which enables the perception with high throughput of visual processing with ultralow energy budget. This has inspired a wide spectrum of devices to imitate such neural process, while precise mimicking such procedure is still highly required. Here, a highly bio-realistic photoelectric spiking neuron for visual depth perception is presented. The firing spikes generated by the TaO_X memristive spiking encoders have a biologically similar frequency range of 1-200 Hz and sub-micro watts power. Such spiking encoder is integrated with a photodetector and a network of neuromorphic transistors, for information collection and recognition tasks, respectively. The distance-dependent response and eye fatigue of biological visual systems have been mimicked based on such photoelectric spiking neuron. The simulated depth perception shows a recognition improvement by adapting to sights at different distances. The results can advance the technologies in bioinspired or robotic systems that may be endowed with depth perception and power efficiency at the same time.

Filament Engineering of Two-Dimensional h-BN for a Self-Power Mechano-Nociceptor System

The switching variability caused by intrinsic stochasticity of the ionic/atomic motions during the conductive filaments (CFs) formation process largely limits the applications of diffusive memristors (DMs), including artificial neurons, neuromorphic computing and artificial sensory systems. In this study, a DM device with improved device uniformity based on well-crystallized two-dimensional (2D) h-BN, which can restrict the CFs formation from three to two dimensions due to the high migration barrier of Ag⁺ between h-BN interlayer, is developed. The BN-DM has potential arrayable feature with high device yield of 88%, which can be applied for building a reservoir computing system for digital pattern recognition with high accuracy rate of 96%, and used as an artificial nociceptor to sense the external noxious stimuli and mimic the important biological nociceptor properties. By connecting the BN-DM to a self-made triboelectric nanogenerator (TENG), a self-power mechano-nociceptor system, which can successfully mimic the important nociceptor features of "threshold", "relaxation" and "allodynia" is designed.

Flexible Neural Network Realized by the Probabilistic SiOx Memristive Synaptic Array for Energy-Efficient Image Learning

The human brain's neural networks are sparsely connected via tunable and probabilistic synapses, which may be essential for performing energy-efficient cognitive and intellectual functions. In this sense, the implementation of a flexible neural network with probabilistic synapses is a first step toward realizing the ultimate energy-efficient computing framework. Here, inspired by the efficient threshold-tunable and probabilistic rod-to-rod bipolar synapses in the human visual system, a 16 × 16 crossbar array comprising the vertical form of gate-tunable probabilistic SiO_x memristive synaptic barristor utilizing the Si/graphene heterojunction is designed and fabricated. Controllable stochastic switching dynamics in this array are achieved via various input voltage pulse schemes. In particular, the threshold tunability via electrostatic gating enables the efficient in situ alteration of the probabilistic switching activation (P_Act ) from 0 to 1.0, and can even modulate the degree of the P_Act change. A drop-connected algorithm based on the P_Act is constructed and used to successfully classify the shapes of several fashion items. The suggested approach can decrease the learning energy by up to ≈2,116 times relative to that of the conventional all-to-all connected network while exhibiting a high recognition accuracy of ≈93 %.

The Image Identification Application with HfO2-Based Replaceable 1T1R Neural Networks

This paper mainly studies the hardware implementation of a fully connected neural network based on the 1T1R (one-transistor-one-resistor) array and its application in handwritten digital image recognition. The 1T1R arrays are prepared by connecting the memristor and nMOSFET in series, and a single-layer and a double-layer fully connected neural network are established. The recognition accuracy of 8 × 8 handwritten digital images reaches 95.19%. By randomly replacing the devices with failed devices, it is found that the stuck-off devices have little effect on the accuracy of the network, but the stuck-on devices will cause a sharp reduction of accuracy. By using the measured conductivity adjustment range and precision data of the memristor, the relationship between the recognition accuracy of the network and the number of hidden neurons is simulated. The simulation results match the experimental results. Compared with the neural network based on the precision of 32-bit floating point, the difference is lower than 1%.

In-Memory Computation Based Mapping of Keccak-f Hash Function

Cryptographic hash functions play a central role in data security for applications such as message authentication, data verification, and detecting malicious or illegal modification of data. However, such functions typically require intensive computations with high volume of memory accesses. Novel computing architectures such as logic-in-memory (LIM)/in-memory computing (IMC) have been investigated in the literature to address the limitations of intense compute and memory bottleneck. In this work, we present an implementation of Keccak-f (a state-of-the-art secure hash algorithm) using a variant of simultaneous logic-in-memory (SLIM) that utilizes emerging non-volatile memory (NVM) devices. Detailed operation and instruction mapping on SLIM-based digital gates is presented. Through simulations, we benchmark the proposed approach using LIM cells based on four different emerging NVM devices (OxRAM, CBRAM, PCM, and FeRAM). The proposed mapping strategy when used with state-of-the-art emerging NVM devices offers EDP savings of up to 300× compared to conventional methods.

Covalent Coupling of Porphyrins with Monolayer Graphene for Low-Voltage Synaptic Transistors

Synaptic devices emulating biological synapses are a key building component of artificial neural networks. Porphyrins and graphene, as two kinds of emerging electronic materials, have attracted extensive attention in the research of photoelectric devices due to their excellent structural and functional properties. Herein, we present a photonic synaptic transistor based on porphyrin-graphene covalent hybrids utilizing 5,10,15,20-tetrakis (4-aminophenyl)-21H,23H-porphine and monolayer graphene linked through the diazo addition reaction. The photonic synaptic device successfully simulates several essential biological functions, and the synaptic plasticity can be regulated by adjusting the parameters of light spikes and gate voltages of the device. Moreover, learning and memory behaviors under different wavelengths are studied to imitate the learning efficiency of humans in diverse emotional states. It is worth noting that all the synaptic functions can be realized at a low operating voltage of -10 mV, which is much lower than that required by most reported photonic synaptic devices. These results indicate that covalent coupling products of porphyrins with graphene have broad prospects in the construction of synaptic transistors and may arouse new research advances in neuromorphic devices with ultralow operating voltage and low energy consumption.

Memristive Behaviors Dominated by Reversible Nucleation Dynamics of Phase-Change Nanoclusters

One of the important steps for realizing artificial intelligence is identifying elementary units that are beneficial for neural network construction. A type of memristive behavior in which phase-change nanoclusters nucleate adaptively in two adjacent dielectric layers with distinct distribution patterns is demonstrated. This memristive system responds in potentiation to increased stimulation strength and fire action potential after threshold stimulation. Reversible nucleation of phase-change nanoclusters is confirmed after both in situ and ex situ examinations using high-resolution transmission electron microscopy. The dynamics at the nanoscale level dominates the actions of the two dielectric layers. The oscillation response over a long period is due to the competition between crystalline and amorphous phases in the layer near the bottom electrode. Weight mutation, that is, action potential firing, is caused by the blockage of the filament in the layer near the top electrode. The memristive system is compact and able to execute complicated functions of a complete neuron and performs an important role in neuromorphic computing.

Low-Dimensional Layered Light-Sensitive Memristive Structures for Energy-Efficient Machine Vision

Layered two-dimensional (2D) and quasi-zero-dimensional (0D) materials effectively absorb radiation in the wide ultraviolet, visible, infrared, and terahertz ranges. Photomemristive structures made of such low-dimensional materials are of great interest for creating optoelectronic platforms for energy-efficient storage and processing of data and optical signals in real time. Here, photosensor and memristor structures based on graphene, graphene oxide, bismuth oxyselenide, and transition metal dichalcogenides are reviewed from the point of view of application in broadband image recognition in artificial intelligence systems for autonomous unmanned vehicles, as well as the compatibility of the formation of layered neuromorphic structures with CMOS technology.

Donor-acceptor-type poly[chalcogenoviologen-alt-triphenylamine] for synaptic biomimicking and neuromorphic computing

Polymer memristors are preeminent candidates for low-power edge computing paradigms. Poly[chalcogenoviologen-alt-triphenylamine] (PCVTPA) has been synthesized by direct coupling of chalcogeno-viologen as electron acceptor and 4-(bromomethyl)-N-(4-(bromo-methyl)phenyl)-N-phenylaniline as electron donor. The introduction of chalcogen atoms (S, Se, Te) into viologen scaffolds can greatly improve electrical conductive, electrochemical, and electrochromic properties of the materials when compared with the conventional viologens. Taking PTeVTPA as an example, the as-fabricated electronic device with a configuration of Al/PTeVTPA/ITO exhibits excellent multilevel storage and history-dependent memristive switching performance. Associated with the unique memristive behavior, the PTeVTPA-based device can not only be used to emulate the synaptic potentiation/depression, the human's learning and memorizing functions, and the transition from short-term synaptic plasticity to long-term plasticity but also carry out decimal arithmetic operations as well. This work will be expected to offer a train of new thought for constructing high-performance synaptic biomimicking and neuromorphic computing system in the near future.

Controllable SiOx Nanorod Memristive Neuron for Probabilistic Bayesian Inference

Modern artificial neural network technology using a deterministic computing framework is faced with a critical challenge in dealing with massive data that are largely unstructured and ambiguous. This challenge demands the advances of an elementary physical device for tackling these uncertainties. Here, we designed and fabricated a SiO_x nanorod memristive device by employing the glancing angle deposition (GLAD) technique, suggesting a controllable stochastic artificial neuron that can mimic the fundamental integrate-and-fire signaling and stochastic dynamics of a biological neuron. The nanorod structure provides the random distribution of multiple nanopores all across the active area, capable of forming a multitude of Si filaments at many SiO_x nanorod edges after the electromigration process, leading to a stochastic switching event with very high dynamic range (≈5.15 × 10¹⁰ ) and low energy (≈4.06 pJ). Different probabilistic activation (ProbAct) functions in a sigmoid form are implemented, showing its controllability with low variation by manufacturing and electrical programming schemes. Furthermore, as an application prospect, based on the suggested memristive neuron, we demonstrated the self-resting neural operation with the local circuit configuration and revealed probabilistic Bayesian inferences for genetic regulatory networks with low normalized mean squared errors (≈2.41 × 10^-2 ) and its robustness to the ProbAct variation.

A Framework for Ultra Low-Power Hardware Accelerators Using NNs for Embedded Time Series Classification

In embedded applications that use neural networks (NNs) for classification tasks, it is important to not only minimize the power consumption of the NN calculation, but of the whole system. Optimization approaches for individual parts exist, such as quantization of the NN or analog calculation of arithmetic operations. However, there is no holistic approach for a complete embedded system design that is generic enough in the design process to be used for different applications, but specific in the hardware implementation to waste no energy for a given application. Therefore, we present a novel framework that allows an end-to-end ASIC implementation of a low-power hardware for time series classification using NNs. This includes a neural architecture search (NAS), which optimizes the NN configuration for accuracy and energy efficiency at the same time. This optimization targets a custom designed hardware architecture that is derived from the key properties of time series classification tasks. Additionally, a hardware generation tool is used that creates a complete system from the definition of the NN. This system uses local multi-level RRAM memory as weight and bias storage to avoid external memory access. Exploiting the non-volatility of these devices, such a system can use a power-down mode to save significant energy during the data acquisition process. Detection of atrial fibrillation (AFib) in electrocardiogram (ECG) data is used as an example for evaluation of the framework. It is shown that a reduction of more than 95% of the energy consumption compared to state-of-the-art solutions is achieved.

Activating Silent Synapses in Sulfurized Indium Selenide for Neuromorphic Computing

The transformation from silent to functional synapses is accompanied by the evolutionary process of human brain development and is essential to hardware implementation of the evolutionary artificial neural network but remains a challenge for mimicking silent to functional synapse activation. Here, we developed a simple approach to successfully realize activation of silent to functional synapses by controlled sulfurization of chemical vapor deposition-grown indium selenide crystals. The underlying mechanism is attributed to the migration of sulfur anions introduced by sulfurization. One of our most important findings is that the functional synaptic behaviors can be modulated by the degree of sulfurization and temperature. In addition, the essential synaptic behaviors including potentiation/depression, paired-pulse facilitation, and spike-rate-dependent plasticity are successfully implemented in the partially sulfurized functional synaptic device. The developed simple approach of introducing sulfur anions in layered selenide opens an effective new avenue to realize activation of silent synapses for application in evolutionary artificial neural networks.

Resistance Switching Behavior of a Perhydropolysilazane-Derived SiOx-Based Memristor

SiO_x is an important dielectric material layer for resistive switching memory due to its compatibility with complementary metal-oxide semiconductor (CMOS) technology. Here we propose a solution process for a SiO_x dielectric layer based on perhydropolysilazane (PHPS). A series of SiO_x layers with different compositions are prepared by controlling the conversion process from PHPS, then the resistance switching behaviors of typical Ag/SiO_x/Au memristors are analyzed. The effect of oxygen vacancies and Si-OH groups on the formation and rupture of Ag conducting filaments (CFs) in the SiO_x layer was thoroughly investigated. Ultimately, we achieved a high-performance memristor with a coefficient of variation (σ/μ) as low as 0.16 ± 0.08 and an on/off ratio as high as 10⁶, which can rival the performance of the SiO_x memristors based on the high-vacuum and high-cost vapor deposition methods. These findings demonstrate the high promise of the PHPS-derived SiO_x dielectric layer in the field of memristors.

Analog Synaptic Transistor with Al-Doped HfO2 Ferroelectric Thin Film

Neuromorphic computing has garnered significant attention because it can overcome the limitations of the current von-Neumann computing system. Analog synaptic devices are essential for realizing hardware-based artificial neuromorphic devices; however, only a few systematic studies in terms of both synaptic materials and device structures have been conducted so far, and thus, further research is required in this direction. In this study, we demonstrate the synaptic characteristics of a ferroelectric material-based thin-film transistor (FeTFT) that uses partial switching of ferroelectric polarization to implement analog conductance modulation. For a ferroelectric material, an aluminum-doped hafnium oxide (Al-doped HfO2) thin film was prepared by atomic layer deposition. As an analog synaptic device, our FeTFT successfully emulated short-term plasticity and long-term plasticity characteristics, such as paired-pulse facilitation and spike timing-dependent plasticity. In addition, we obtained potentiation/depression weight updates with high linearity, an on/off ratio, and low cycle-to-cycle variation by adjusting the amplitude and number of input pulses. In the simulation trained with optimized potentiation/depression conditions, we achieved a pattern recognition accuracy of approximately 90% for the Modified National Institute of Standard and Technology (MNIST) handwritten data set. Our results indicated that ferroelectric transistors can be used as an alternative artificial synapse.

Stimuli-Responsive Memristive Materials for Artificial Synapses and Neuromorphic Computing

Neuromorphic computing holds promise for building next-generation intelligent systems in a more energy-efficient way than the conventional von Neumann computing architecture. Memristive hardware, which mimics biological neurons and synapses, offers high-speed operation and low power consumption, enabling energy- and area-efficient, brain-inspired computing. Here, recent advances in memristive materials and strategies that emulate synaptic functions for neuromorphic computing are highlighted. The working principles and characteristics of biological neurons and synapses, which can be mimicked by memristive devices, are presented. Besides device structures and operation with different external stimuli such as electric, magnetic, and optical fields, how memristive materials with a rich variety of underlying physical mechanisms can allow fast, reliable, and low-power neuromorphic applications is also discussed. Finally, device requirements are examined and a perspective on challenges in developing memristive materials for device engineering and computing science is given.

Artificial Neurons Based on Ag/V2C/W Threshold Switching Memristors

Artificial synapses and neurons are two critical, fundamental bricks for constructing hardware neural networks. Owing to its high-density integration, outstanding nonlinearity, and modulated plasticity, memristors have attracted emerging attention on emulating biological synapses and neurons. However, fabricating a low-power and robust memristor-based artificial neuron without extra electrical components is still a challenge for brain-inspired systems. In this work, we demonstrate a single two-dimensional (2D) MXene(V₂C)-based threshold switching (TS) memristor to emulate a leaky integrate-and-fire (LIF) neuron without auxiliary circuits, originating from the Ag diffusion-based filamentary mechanism. Moreover, our V₂C-based artificial neurons faithfully achieve multiple neural functions including leaky integration, threshold-driven fire, self-relaxation, and linear strength-modulated spike frequency characteristics. This work demonstrates that three-atom-type MXene (e.g., V₂C) memristors may provide an efficient method to construct the hardware neuromorphic computing systems.

A Low-Power RRAM Memory Block for Embedded, Multi-Level Weight and Bias Storage in Artificial Neural Networks

Pattern recognition as a computing task is very well suited for machine learning algorithms utilizing artificial neural networks (ANNs). Computing systems using ANNs usually require some sort of data storage to store the weights and bias values for the processing elements of the individual neurons. This paper introduces a memory block using resistive memory cells (RRAM) to realize this weight and bias storage in an embedded and distributed way while also offering programming and multi-level ability. By implementing power gating, overall power consumption is decreased significantly without data loss by taking advantage of the non-volatility of the RRAM technology. Due to the versatility of the peripheral circuitry, the presented memory concept can be adapted to different applications and RRAM technologies.

Reconfigurable 2D WSe2 -Based Memtransistor for Mimicking Homosynaptic and Heterosynaptic Plasticity

The mimicking of both homosynaptic and heterosynaptic plasticity using a high-performance synaptic device is important for developing human-brain-like neuromorphic computing systems to overcome the ever-increasing challenges caused by the conventional von Neumann architecture. However, the commonly used synaptic devices (e.g., memristors and transistors) require an extra modulate terminal to mimic heterosynaptic plasticity, and their capability of synaptic plasticity simulation is limited by the low weight adjustability. In this study, a WSe₂ -based memtransistor for mimicking both homosynaptic and heterosynaptic plasticity is fabricated. By applying spikes on either the drain or gate terminal, the memtransistor can mimic common homosynaptic plasticity, including spiking rate dependent plasticity, paired pulse facilitation/depression, synaptic potentiation/depression, and filtering. Benefitting from the multi-terminal input and high adjustability, the resistance state number and linearity of the memtransistor can be improved by optimizing the conditions of the two inputs. Moreover, the device can successfully mimic heterosynaptic plasticity without introducing an extra terminal and can simultaneously offer versatile reconfigurability of excitatory and inhibitory plasticity. These highly adjustable and reconfigurable characteristics offer memtransistors more freedom of choice for tuning synaptic weight, optimizing circuit design, and building artificial neuromorphic computing systems.

Free-Standing Multilayer Molybdenum Disulfide Memristor for Brain-Inspired Neuromorphic Applications

Recently, atomically thin two-dimensional (2D) transition-metal dichalcogenides (TMDs) have attracted great interest in electronic and opto-electronic devices for high-integration-density applications such as data storage due to their small vertical dimension and high data storage capability. Here, we report a memristor based on free-standing multilayer molybdenum disulfide (MoS₂) with a high current on/off ratio of ∼10³ and a stable retention for at least 3000 s. Through light modulation of the carrier density in the suspended MoS₂ channel, the on/off ratio can be further increased to ∼10⁵. Moreover, the essential photosynaptic functions with short- and long-term memory (STM and LTM) behaviors are successfully mimicked by such devices. These results also indicate that STM can be transferred to LTM by increasing the light stimuli power, pulse duration, and number of pulses. The electrical measurements performed under vacuum and ambient air conditions propose that the observed resistive switching is due to adsorbed oxygen and water molecules on both sides of the MoS₂ channel. Thus, our free-standing 2D multilayer MoS₂-based memristors propose a simple approach for fabrication of a low-power-consumption and reliable resistive switching device for neuromorphic applications.

A nanoimprinted artificial engram device

At present, mainstream neuromorphic hardware is based on artificial synapses; however, an engram, instead of a synapse, has recently been confirmed as the basic unit of memory, which verifies the engram theory proposed by Richard Semon in 1904. Here, we demonstrate an artificial engram device based on a nanoimprinted curable resin. The variation in the relative diffraction efficiency based on the asymmetric reversible topological change of the nanoimprinted resin enables the device to meet all the requirements for artificial engrams, including synaptic plasticity, long memory storage time, asymmetric memorizing-forgetting behaviour and measurable changes and responses. On this basis, we demonstrate the concept of realizing memory formation, memory manipulation and implantation, and memory consolidation using our artificial engram device in comparison with its biological counterpart.

Emerging Artificial Neuron Devices for Probabilistic Computing

In recent decades, artificial intelligence has been successively employed in the fields of finance, commerce, and other industries. However, imitating high-level brain functions, such as imagination and inference, pose several challenges as they are relevant to a particular type of noise in a biological neuron network. Probabilistic computing algorithms based on restricted Boltzmann machine and Bayesian inference that use silicon electronics have progressed significantly in terms of mimicking probabilistic inference. However, the quasi-random noise generated from additional circuits or algorithms presents a major challenge for silicon electronics to realize the true stochasticity of biological neuron systems. Artificial neurons based on emerging devices, such as memristors and ferroelectric field-effect transistors with inherent stochasticity can produce uncertain non-linear output spikes, which may be the key to make machine learning closer to the human brain. In this article, we present a comprehensive review of the recent advances in the emerging stochastic artificial neurons (SANs) in terms of probabilistic computing. We briefly introduce the biological neurons, neuron models, and silicon neurons before presenting the detailed working mechanisms of various SANs. Finally, the merits and demerits of silicon-based and emerging neurons are discussed, and the outlook for SANs is presented.

Artificial Synapses Based on Ferroelectric Schottky Barrier Field-Effect Transistors for Neuromorphic Applications

Artificial synapses based on ferroelectric Schottky barrier field-effect transistors (FE-SBFETs) are experimentally demonstrated. The FE-SBFETs employ single-crystalline NiSi₂ contacts with an atomically flat interface to Si and Hf_0.5Zr_0.5O₂ ferroelectric layers on silicon-on-insulator substrates. The ferroelectric polarization switching dynamics gradually modulate the NiSi₂/Si Schottky barriers and the potential of the channel, thus programming the device conductance with input voltage pulses. The short-term synaptic plasticity is characterized in terms of excitatory/inhibitory post-synaptic current (EPSC) and paired-pulse facilitation/depression. The EPSC amplitude shows a linear response to the amplitude of the pre-synaptic spike. Very low energy/spike consumption as small as ∼2 fJ is achieved, demonstrating high energy efficiency. Long-term potentiation/depression results show very high endurance and very small cycle-to-cycle variations (∼1%) after 10⁵ pulse measurements. Furthermore, spike-timing-dependent plasticity is also emulated using the gate voltage pulse as the pre-synaptic spike and the drain voltage pulse as the post-synaptic spikes. These findings indicate that FE-SBFET synapses have high potential for future neuromorphic computing applications.

Self-Powered Memristive Systems for Storage and Neuromorphic Computing

A neuromorphic computing chip that can imitate the human brain’s ability to process multiple types of data simultaneously could fundamentally innovate and improve the von-neumann computer architecture, which has been criticized. Memristive devices are among the best hardware units for building neuromorphic intelligence systems due to the fact that they operate at an inherent low voltage, use multi-bit storage, and are cost-effective to manufacture. However, as a passive device, the memristor cell needs external energy to operate, resulting in high power consumption and complicated circuit structure. Recently, an emerging self-powered memristive system, which mainly consists of a memristor and an electric nanogenerator, had the potential to perfectly solve the above problems. It has attracted great interest due to the advantages of its power-free operations. In this review, we give a systematic description of self-powered memristive systems from storage to neuromorphic computing. The review also proves a perspective on the application of artificial intelligence with the self-powered memristive system.