Experimental Demonstration of Supervised Learning in Spiking Neural Networks with Phase-Change Memory Synapses

Strain-mediated multistate skyrmion for neuron devices

Magnetic skyrmions are potential candidates for neuromorphic computing because of their inherent topological stability, low drive current density and nanoscale size. However, an artificial neuron device based on current-driven skyrmion motion cannot satisfy the requirement of energy efficiency and integration density due to hundreds of millions of interconnected neurons and synapses present in the deep networks. Here, we present a compact and energy efficient skyrmion-based artificial neuron consisting of ferromagnetic/heavy metal/ferroelectric layers which uses strain-mediated voltage manipulation of skyrmion states to mimic the Integrate-and-Fire (IF) function of biological neurons. By implementation of a spiking neural network (SNN) based on the proposed skyrmionic neuronal devices, it can achieve a high accuracy of 95.08% on a modified National Institute of Standards and Technology (MNIST) handwritten digit dataset, as well as a low power consumption of ∼46.8 fJ per epoch per neuron. The present work suggests a novel way to realize energy-efficient and high-density neuromorphic computing.

Spike Optimization to Improve Properties of Ferroelectric Tunnel Junction Synaptic Devices for Neuromorphic Computing System Applications

The continuous advancement of Artificial Intelligence (AI) technology depends on the efficient processing of unstructured data, encompassing text, speech, and video. Traditional serial computing systems based on the von Neumann architecture, employed in information and communication technology development for decades, are not suitable for the concurrent processing of massive unstructured data tasks with relatively low-level operations. As a result, there arises a pressing need to develop novel parallel computing systems. Recently, there has been a burgeoning interest among developers in emulating the intricate operations of the human brain, which efficiently processes vast datasets with remarkable energy efficiency. This has led to the proposal of neuromorphic computing systems. Of these, Spiking Neural Networks (SNNs), designed to closely resemble the information processing mechanisms of biological neural networks, are subjects of intense research activity. Nevertheless, a comprehensive investigation into the relationship between spike shapes and Spike-Timing-Dependent Plasticity (STDP) to ensure efficient synaptic behavior remains insufficiently explored. In this study, we systematically explore various input spike types to optimize the resistive memory characteristics of Hafnium-based Ferroelectric Tunnel Junction (FTJ) devices. Among the various spike shapes investigated, the square-triangle (RT) spike exhibited good linearity and symmetry, and a wide range of weight values could be realized depending on the offset of the RT spike. These results indicate that the spike shape serves as a crucial indicator in the alteration of synaptic connections, representing the strength of the signals.

Spiking Dynamics in Dual Free Layer Perpendicular Magnetic Tunnel Junctions

Spintronic devices have recently attracted a lot of attention in the field of unconventional computing due to their non-volatility for short- and long-term memory, nonlinear fast response, and relatively small footprint. Here we demonstrate experimentally how voltage driven magnetization dynamics of dual free layer perpendicular magnetic tunnel junctions can emulate spiking neurons in hardware. The output spiking rate was controlled by varying the dc bias voltage across the device. The field-free operation of this two-terminal device and its robustness against an externally applied magnetic field make it a suitable candidate to mimic the neuron response in a dense neural network. The small energy consumption of the device (4-16 pJ/spike) and its scalability are important benefits for embedded applications. This compact perpendicular magnetic tunnel junction structure could finally bring spiking neural networks to sub-100 nm size elements.

Diffusive Memristors with Uniform and Tunable Relaxation Time for Spike Generation in Event-Based Pattern Recognition

A diffusive memristor is a promising building block for brain-inspired computing hardware. However, the randomness in the device relaxation dynamics limits the wide-range adoption of diffusive memristors in large arrays. In this work, the device stack is engineered to achieve a much-improved uniformity in the relaxation time (standard deviation σ reduced from ≈12 to ≈0.32 ms). The memristor is further connected with a resistor or a capacitor and the relaxation time is tuned between 1.13 µs and 1.25 ms, ranging from three orders of magnitude. The hierarchy of time surfaces (HOTS) algorithm, to utilize the tunable and uniform relaxation behavior for spike generation, is implemented. An accuracy of 77.3% is achieved in recognizing moving objects in the neuromorphic MNIST (N-MNIST) dataset. The work paves the way for building emerging neuromorphic computing hardware systems with ultralow power consumption.

Review of Electrochemically Synthesized Resistive Switching Devices: Memory Storage, Neuromorphic Computing, and Sensing Applications

Resistive-switching-based memory devices meet most of the requirements for use in next-generation information and communication technology applications, including standalone memory devices, neuromorphic hardware, and embedded sensing devices with on-chip storage, due to their low cost, excellent memory retention, compatibility with 3D integration, in-memory computing capabilities, and ease of fabrication. Electrochemical synthesis is the most widespread technique for the fabrication of state-of-the-art memory devices. The present review article summarizes the electrochemical approaches that have been proposed for the fabrication of switching, memristor, and memristive devices for memory storage, neuromorphic computing, and sensing applications, highlighting their various advantages and performance metrics. We also present the challenges and future research directions for this field in the concluding section.

The Effect of Carbon Doping on the Crystal Structure and Electrical Properties of Sb2Te3

As a new generation of non-volatile memory, phase change random access memory (PCRAM) has the potential to fill the hierarchical gap between DRAM and NAND FLASH in computer storage. Sb₂Te₃, one of the candidate materials for high-speed PCRAM, has high crystallization speed and poor thermal stability. In this work, we investigated the effect of carbon doping on Sb₂Te₃. It was found that the FCC phase of C-doped Sb₂Te₃ appeared at 200 °C and began to transform into the HEX phase at 25 °C, which is different from the previous reports where no FCC phase was observed in C-Sb₂Te₃. Based on the experimental observation and first-principles density functional theory calculation, it is found that the formation energy of FCC-Sb₂Te₃ structure decreases gradually with the increase in C doping concentration. Moreover, doped C atoms tend to form C molecular clusters in sp² hybridization at the grain boundary of Sb₂Te₃, which is similar to the layered structure of graphite. And after doping C atoms, the thermal stability of Sb₂Te₃ is improved. We have fabricated the PCRAM device cell array of a C-Sb₂Te₃ alloy, which has an operating speed of 5 ns, a high thermal stability (10-year data retention temperature 138.1 °C), a low device power consumption (0.57 pJ), a continuously adjustable resistance value, and a very low resistance drift coefficient.

A Supervised Learning Algorithm for Multilayer Spiking Neural Networks Based on Temporal Coding Toward Energy-Efficient VLSI Processor Design

Spiking neural networks (SNNs) are brain-inspired mathematical models with the ability to process information in the form of spikes. SNNs are expected to provide not only new machine-learning algorithms but also energy-efficient computational models when implemented in very-large-scale integration (VLSI) circuits. In this article, we propose a novel supervised learning algorithm for SNNs based on temporal coding. A spiking neuron in this algorithm is designed to facilitate analog VLSI implementations with analog resistive memory, by which ultrahigh energy efficiency can be achieved. We also propose several techniques to improve the performance on recognition tasks and show that the classification accuracy of the proposed algorithm is as high as that of the state-of-the-art temporal coding SNN algorithms on the MNIST and Fashion-MNIST datasets. Finally, we discuss the robustness of the proposed SNNs against variations that arise from the device manufacturing process and are unavoidable in analog VLSI implementation. We also propose a technique to suppress the effects of variations in the manufacturing process on the recognition performance.

Metaplastic and energy-efficient biocompatible graphene artificial synaptic transistors for enhanced accuracy neuromorphic computing

CMOS-based computing systems that employ the von Neumann architecture are relatively limited when it comes to parallel data storage and processing. In contrast, the human brain is a living computational signal processing unit that operates with extreme parallelism and energy efficiency. Although numerous neuromorphic electronic devices have emerged in the last decade, most of them are rigid or contain materials that are toxic to biological systems. In this work, we report on biocompatible bilayer graphene-based artificial synaptic transistors (BLAST) capable of mimicking synaptic behavior. The BLAST devices leverage a dry ion-selective membrane, enabling long-term potentiation, with ~50 aJ/µm² switching energy efficiency, at least an order of magnitude lower than previous reports on two-dimensional material-based artificial synapses. The devices show unique metaplasticity, a useful feature for generalizable deep neural networks, and we demonstrate that metaplastic BLASTs outperform ideal linear synapses in classic image classification tasks. With switching energy well below the 1 fJ energy estimated per biological synapse, the proposed devices are powerful candidates for bio-interfaced online learning, bridging the gap between artificial and biological neural networks.

Synaptic Emulation via Ferroelectric P(VDF-TrFE) Reinforced Charge Trapping/Detrapping in Zinc-Tin Oxide Transistor

Brain inspired artificial synapses are highly desirable for neuromorphic computing and are an alternative to a conventional computing system. Here, we report a simple and cost-effective ferroelectric capacitively coupled zinc-tin oxide (ZTO) thin-film transistor (TFT) topped with ferroelectric copolymer poly(vinylidene fluoride-trifluoroethylene) (P(VDF-TrFE)) for artificial synaptic devices. Ferroelectric dipoles enhance the charge trapping/detrapping effect in ZTO TFT, as confirmed by the transfer curve (I_D-V_G) analysis. This substantiates superior artificial synapse responses in ferroelectric-coupled ZTO TFT because the current potentiation and depression are individually improved. The ferroelectric-coupled ZTO TFT successfully emulates the essential features of the artificial synapse, including pair-pulsed facilitation (PPF) and potentiation/depression (P/D) characteristics. In addition, the device also mimics the memory consolidation behavior through intensified stimulation. This work demonstrates that the ferroelectric-coupled ZTO synaptic transistor possesses great potential as a hardware candidate for neuromorphic computing.

Engineering Spiking Neurons Using Threshold Switching Devices for High-Efficient Neuromorphic Computing

Inspired by the human brain, the spike-based neuromorphic system has attracted strong research enthusiasm because of the high energy efficiency and powerful computational capability, in which the spiking neurons and plastic synapses are two fundamental building blocks. Recently, two-terminal threshold switching (TS) devices have been regarded as promising candidates for building spiking neurons in hardware. However, how circuit parameters affect the spiking behavior of TS-based neurons is still an open question. Here, based on a leaky integrate-and-fire (LIF) neuron circuit, we systematically study the effect of both the extrinsic and intrinsic factors of NbO x -based TS neurons on their spiking behaviors. The extrinsic influence factors contain input intensities, connected synaptic weights, and parallel capacitances. To illustrate the effect of intrinsic factors, including the threshold voltage, holding voltage, and high/low resistance states of NbO x devices, we propose an empirical model of the fabricated NbO x devices, fitting well with the experimental results. The results indicate that with enhancing the input intensity, the spiking frequency increases first then decreases after reaching a peak value. Except for the connected synaptic weights, all other parameters can modulate the spiking peak frequency under high enough input intensity. Also, the relationship between energy consumption per spike and frequency of the neuron cell is further studied, leading guidance to design neuron circuits in a system to obtain the lowest energy consumption. At last, to demonstrate the practical applications of TS-based neurons, we construct a spiking neural network (SNN) to control the cart-pole using reinforcement learning, obtaining a reward score up to 450. This work provides valuable guidance on building compact LIF neurons based on TS devices and further bolsters the construction of high-efficiency neuromorphic systems.

Antiferromagnetic skyrmion repulsion based artificial neuron device

Magnetic skyrmions are potential candidates for neuromorphic computing due to their inherent topologically stable particle-like behavior, low driving current density, and nanoscale size. Antiferromagnetic skyrmions are favored as they can be driven parallel to in-plane electrical currents as opposed to ferromagnetic skyrmions which exhibit the skyrmion Hall effect and eventually cause their annihilation at the edge of nanotracks. In this paper, an antiferromagnetic skyrmion based artificial neuron device consisting of a magnetic anisotropy barrier on a nanotrack is proposed. It exploits inter-skyrmion repulsion, mimicking the integrate-fire (IF) functionality of a biological neuron. The device threshold represented by the maximum number of skyrmions that can be pinned by the barrier can be tuned based on the particular current density employed on the nanotrack. The corresponding neuron spiking event occurs when a skyrmion overcomes the barrier. By raising the device threshold, lowering the barrier width and height, the operating current density of the device can be decreased to further enhance its energy efficiency. The proposed device paves the way for developing energy-efficient neuromorphic computing in antiferromagnetic spintronics.

Novel hardware and concepts for unconventional computing

Neuromorphic systems are currently experiencing a rapid upswing due to the fact that today's CMOS (complementary metal oxide silicon) based technologies are increasingly approaching their limits. In particular, for the area of machine learning, energy consumption of today's electronics is an important limitation, that also contributes toward the ever-increasing impact of digitalization on our climate. Thus, in order to better meet the special requirements of unconventional computing, new physical substrates for bio-inspired computing schemes are extensively exploited. The aim of this Guest Edited Collection is to provide a platform for interdisciplinary research along three main lines: memristive materials and devices, emulation of cellular learning (neurons and synapses), and unconventional computing and network schemes.