On spike-timing-dependent-plasticity, memristive devices, and building a self-learning visual cortex

A Hybrid CMOS-Memristor Spiking Neural Network Supporting Multiple Learning Rules

Artificial intelligence (AI) is changing the way computing is performed to cope with real-world, ill-defined tasks for which traditional algorithms fail. AI requires significant memory access, thus running into the von Neumann bottleneck when implemented in standard computing platforms. In this respect, low-latency energy-efficient in-memory computing can be achieved by exploiting emerging memristive devices, given their ability to emulate synaptic plasticity, which provides a path to design large-scale brain-inspired spiking neural networks (SNNs). Several plasticity rules have been described in the brain and their coexistence in the same network largely expands the computational capabilities of a given circuit. In this work, starting from the electrical characterization and modeling of the memristor device, we propose a neuro-synaptic architecture that co-integrates in a unique platform with a single type of synaptic device to implement two distinct learning rules, namely, the spike-timing-dependent plasticity (STDP) and the Bienenstock-Cooper-Munro (BCM). This architecture, by exploiting the aforementioned learning rules, successfully addressed two different tasks of unsupervised learning.

Electrical Impedance Spectroscopy with Bacterial Biofilms: Neuronal-like Behavior

Negative capacitance at low frequencies for spiking neurons was first demonstrated in 1941 (K. S. Cole) by using extracellular electrodes. The phenomenon subsequently was explained by using the Hodgkin-Huxley model and is due to the activity of voltage-gated potassium ion channels. We show that Escherichia coli (E. coli) biofilms exhibit significant stable negative capacitances at low frequencies when they experience a small DC bias voltage in electrical impedance spectroscopy experiments. Using a frequency domain Hodgkin-Huxley model, we characterize the conditions for the emergence of this feature and demonstrate that the negative capacitance exists only in biofilms containing living cells. Furthermore, we establish the importance of the voltage-gated potassium ion channel, Kch, using knock-down mutants. The experiments provide further evidence for voltage-gated ion channels in E. coli and a new, low-cost method to probe biofilm electrophysiology, e.g., to understand the efficacy of antibiotics. We expect that the majority of bacterial biofilms will demonstrate negative capacitances.

Proton-Gated Synaptic Transistors, Based on an Electron-Beam Patterned Nafion Electrolyte

Neuromorphic processors using artificial neural networks are the center of attention for energy-efficient analog computing. Artificial synapses act as building blocks in such neural networks for parallel information processing and data storage. Herein we describe the fabrication of a proton-gated synaptic transistor using a Nafion electrolyte thin film, which is patterned by electron-beam lithography (EBL). The device has an active channel of indium-zinc-oxide (IZO) between the source and drain electrodes, which shows Ohmic behavior with a conductance level on the order of 100 μS. Under voltage applications to the gate electrode, the channel conductance is changed due to the injection and extraction of protons between the IZO channel and the Nafion electrolyte, emulating various synaptic functions with short-term and long-term plasticity. When positive (negative) gate voltage pulses are consecutively applied, the device exhibits long-term potentiation (depression) at the same number of steps as the number of input pulses. Based on these characteristics, an artificial neural network using this transistor shows ∼84% image recognition accuracy for handwritten digits. The subject transistor also successfully mimics paired-pulse facilitation and depression, Hebbian spike-timing-dependent plasticity, and Pavlovian associative learning followed by extinction activities. Finally, dynamical pattern image memorization is demonstrated in a 5 × 5 array of these synaptic transistors. The results indicate that EBL patternable Nafion electrolytes have great potential for use in the fabrication and circuit-level integration of synaptic devices for neuromorphic computing applications.

Ionic-electronic halide perovskite memdiodes enabling neuromorphic computing with a second-order complexity

With increasing computing demands, serial processing in von Neumann architectures built with zeroth-order complexity digital circuits is saturating in computational capacity and power, entailing research into alternative paradigms. Brain-inspired systems built with memristors are attractive owing to their large parallelism, low energy consumption, and high error tolerance. However, most demonstrations have thus far only mimicked primitive lower-order biological complexities using devices with first-order dynamics. Memristors with higher-order complexities are predicted to solve problems that would otherwise require increasingly elaborate circuits, but no generic design rules exist. Here, we present second-order dynamics in halide perovskite memristive diodes (memdiodes) that enable Bienenstock-Cooper-Munro learning rules capturing both timing- and rate-based plasticity. A triplet spike timing-dependent plasticity scheme exploiting ion migration, back diffusion, and modulable Schottky barriers establishes general design rules for realizing higher-order memristors. This higher order enables complex binocular orientation selectivity in neural networks exploiting the intrinsic physics of the devices, without the need for complicated circuitry.

A CMOS-memristor hybrid system for implementing stochastic binary spike timing-dependent plasticity

This paper describes a fully experimental hybrid system in which a [Formula: see text] memristive crossbar spiking neural network (SNN) was assembled using custom high-resistance state memristors with analogue CMOS neurons fabricated in 180 nm CMOS technology. The custom memristors used NMOS selector transistors, made available on a second 180 nm CMOS chip. One drawback is that memristors operate with currents in the micro-amperes range, while analogue CMOS neurons may need to operate with currents in the pico-amperes range. One possible solution was to use a compact circuit to scale the memristor-domain currents down to the analogue CMOS neuron domain currents by at least 5-6 orders of magnitude. Here, we proposed using an on-chip compact current splitter circuit based on MOS ladders to aggressively attenuate the currents by over 5 orders of magnitude. This circuit was added before each neuron. This paper describes the proper experimental operation of an SNN circuit using a [Formula: see text] 1T1R synaptic crossbar together with four post-synaptic CMOS circuits, each with a 5-decade current attenuator and an integrate-and-fire neuron. It also demonstrates one-shot winner-takes-all training and stochastic binary spike-timing-dependent-plasticity learning using this small system. This article is part of the theme issue 'Advanced neurotechnologies: translating innovation for health and well-being'.

Comparative Analysis of Reconfigurable Platforms for Memristor Emulation

The memristor is the fourth fundamental element in the electronic circuit field, whose memory and resistance properties make it unique. Although there are no electronic solutions based on the memristor, interest in application development has increased significantly. Nevertheless, there are only numerical Matlab or Spice models that can be used for simulating memristor systems, and designing is limited to using memristor emulators only. A memristor emulator is an electronic circuit that mimics a memristor. In this way, a research approach is to build discrete-component emulators of memristors for its study without using the actual models. In this work, two reconfigurable hardware architectures have been proposed for use in the prototyping of a non-linearity memristor emulator: the FPAA (Field Programing Analog Arrays) and the FPGA (Field Programming Gate Array). The easy programming and reprogramming of the first architecture and the performance, high area density, and parallelism of the second one allow the implementation of this type of system. In addition, a detailed comparison is shown to underline the main differences between the two approaches. These platforms could be used in more complex analog and/or digital systems, such as neural networks, CNN, digital circuits, etc.

Toward Reflective Spiking Neural Networks Exploiting Memristive Devices

The design of modern convolutional artificial neural networks (ANNs) composed of formal neurons copies the architecture of the visual cortex. Signals proceed through a hierarchy, where receptive fields become increasingly more complex and coding sparse. Nowadays, ANNs outperform humans in controlled pattern recognition tasks yet remain far behind in cognition. In part, it happens due to limited knowledge about the higher echelons of the brain hierarchy, where neurons actively generate predictions about what will happen next, i.e., the information processing jumps from reflex to reflection. In this study, we forecast that spiking neural networks (SNNs) can achieve the next qualitative leap. Reflective SNNs may take advantage of their intrinsic dynamics and mimic complex, not reflex-based, brain actions. They also enable a significant reduction in energy consumption. However, the training of SNNs is a challenging problem, strongly limiting their deployment. We then briefly overview new insights provided by the concept of a high-dimensional brain, which has been put forward to explain the potential power of single neurons in higher brain stations and deep SNN layers. Finally, we discuss the prospect of implementing neural networks in memristive systems. Such systems can densely pack on a chip 2D or 3D arrays of plastic synaptic contacts directly processing analog information. Thus, memristive devices are a good candidate for implementing in-memory and in-sensor computing. Then, memristive SNNs can diverge from the development of ANNs and build their niche, cognitive, or reflective computations.

Neuromorphic Signal Filter for Robot Sensoring

Noise management associated with input signals in sensor devices arises as one of the main problems limiting robot control performance. This article introduces a novel neuromorphic filter model based on a leaky integrate and fire (LIF) neural model cell, which encodes the primary information from a noisy input signal and delivers an output signal with a significant noise reduction in practically real-time with energy-efficient consumption. A new approach for neural decoding based on the neuron-cell spiking frequency is introduced to recover the primary signal information. The simulations conducted on the neuromorphic filter demonstrate an outstanding performance of white noise rejecting while preserving the original noiseless signal with a low information loss. The proposed filter model is compatible with the CMOS technology design methodologies for implementing low consumption smart sensors with applications in various fields such as robotics and the automotive industry demanded by Industry 4.0.

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.

R-STDP Spiking Neural Network Architecture for Motion Control on a Changing Friction Joint Robotic Arm

Neuromorphic computing is a recent class of brain-inspired high-performance computer platforms and algorithms involving biologically-inspired models adopting hardware implementation in integrated circuits. The neuromorphic computing applications have provoked the rise of highly connected neurons and synapses in analog circuit systems that can be used to solve today's challenging machine learning problems. In conjunction with biologically plausible learning rules, such as the Hebbian learning and memristive devices, biologically-inspired spiking neural networks are considered the next-generation neuromorphic hardware construction blocks that will enable the deployment of new analog in situ learning capable and energetic efficient brain-like devices. These features are envisioned for modern mobile robotic implementations, currently challenging to overcome the pervasive von Neumann computer architecture. This study proposes a new neural architecture using the spike-time-dependent plasticity learning method and step-forward encoding algorithm for a self tuning neural control of motion in a joint robotic arm subjected to dynamic modifications. Simulations were conducted to demonstrate the proposed neural architecture's feasibility as the network successfully compensates for changing dynamics at each simulation run.

Chemical Inductor

A multitude of chemical, biological, and material systems present an inductive behavior that is not electromagnetic in origin. Here, it is termed a chemical inductor. We show that the structure of the chemical inductor consists of a two-dimensional system that couples a fast conduction mode and a slowing down element. Therefore, it is generally defined in dynamical terms rather than by a specific physicochemical mechanism. The chemical inductor produces many familiar features in electrochemical reactions, including catalytic, electrodeposition, and corrosion reactions in batteries and fuel cells, and in solid-state semiconductor devices such as solar cells, organic light-emitting diodes, and memristors. It generates the widespread phenomenon of negative capacitance, it causes negative spikes in voltage transient measurements, and it creates inverted hysteresis effects in current–voltage curves and cyclic voltammetry. Furthermore, it determines stability, bifurcations, and chaotic properties associated to self-sustained oscillations in biological neurons and electrochemical systems. As these properties emerge in different types of measurement techniques such as impedance spectroscopy and time-transient decays, the chemical inductor becomes a useful framework for the interpretation of the electrical, optoelectronic, and electrochemical responses in a wide variety of systems. In the paper, we describe the general dynamical structure of the chemical inductor and we comment on a broad range of examples from different research areas.

Mem-fractive properties of mushrooms

Memristors close the loop forI-Vcharacteristics of the traditional, passive, semi-conductor devices. A memristor is a physical realisation of the material implication and thus is a universal logical element. Memristors are getting particular interest in the field of bioelectronics. Electrical properties of living substrates are not binary and there is nearly a continuous transitions from being non-memristive to mem-fractive (exhibiting a combination of passive memory) to ideally memristive. In laboratory experiments we show that living oyster mushroomsPleurotus ostreatusexhibit mem-fractive properties. We offer a piece-wise polynomial approximation of theI-Vbehaviour of the oyster mushrooms. We also report spiking activity, oscillations in conduced current of the oyster mushrooms.

Emulation of synaptic functions with low voltage organic memtransistor for hardware oriented neuromorphic computing

Here, various synaptic functions and neural network simulation based pattern-recognition using novel, solution-processed organic memtransistors (memTs) with an unconventional redox-gating mechanism are demonstrated. Our synaptic memT device using conjugated polymer thin-film and redox-active solid electrolyte as the gate dielectric can be routinely operated at gate voltages (V_GS) below − 1.5 V, subthreshold-swings (S) smaller than 120 mV/dec, and ON/OFF current ratio larger than 10⁸. Large hysteresis in transfer curves depicts the signature of non-volatile resistive switching (RS) property with ON/OFF ratio as high as 10⁵. In addition, our memT device also shows many synaptic functions, including the availability of many conducting-states (> 500) that are used for efficient pattern recognition using the simplest neural network simulation model with training and test accuracy higher than 90%. Overall, the presented approach opens a new and promising way to fabricate high-performance artificial synapses and their arrays for the implementation of hardware-oriented neural network.

An Artificial Reflex Arc That Perceives Afferent Visual and Tactile Information and Controls Efferent Muscular Actions

Neural perception and action-inspired electronics is becoming important for interactive human-machine interfaces and intelligent robots. A system that implements neuromorphic environmental information coding, synaptic signal processing, and motion control is desired. We report a neuroinspired artificial reflex arc that possesses visual and somatosensory dual afferent nerve paths and an efferent nerve path to control artificial muscles. A self-powered photoelectric synapse between the afferent and efferent nerves was used as the key information processor. The artificial reflex arc successfully responds to external visual and tactile information and controls the actions of artificial muscle in response to these external stimuli and thus emulates reflex activities through a full reflex arc. The visual and somatosensory information is encoded as impulse spikes, the frequency of which exhibited a sublinear dependence on the obstacle proximity or pressure stimuli. The artificial reflex arc suggests a promising strategy toward developing soft neurorobotic systems and prostheses.

High-Speed Switching and Giant Electroresistance in an Epitaxial Hf0.5Zr0.5O2-Based Ferroelectric Tunnel Junction Memristor

HfO₂-based ferroelectric materials are good candidates for constructing next-generation nonvolatile memories and high-performance electronic synapses and have attracted extensive attention from both academia and industry. Here, a Hf_0.5Zr_0.5O₂-based ferroelectric tunnel junction (FTJ) memristor is successfully fabricated by epitaxially growing a Hf_0.5Zr_0.5O₂ film on a 0.7 wt % Nb-doped SrTiO₃ (001) substrate with a buffer layer of La_2/3Sr_1/3MnO₃ (∼1 u.c.). The FTJ shows a high switching speed of 20 ns, a giant electroresistance ratio of ∼834, and multiple states (eight states or three bits) with good retention >10⁴ s. As a solid synaptic device, tunable synapse functions have also been obtained, including long-term potentiation, long-term depression, and spike-timing-dependent plasticity. These results highlight the promising applications of Hf_0.5Zr_0.5O₂-based FTJ in ultrafast-speed and high-density nonvolatile memories and artificial synapses.

Monolayer MoS2 Synaptic Transistors for High-Temperature Neuromorphic Applications

As essential units in an artificial neural network (ANN), artificial synapses have to adapt to various environments. In particular, the development of synaptic transistors that can work above 125 °C is desirable. However, it is challenging due to the failure of materials or mechanisms at high temperatures. Here, we report a synaptic transistor working at hundreds of degrees Celsius. It employs monolayer MoS₂ as the channel and Na⁺-diffused SiO₂ as the ionic gate medium. A large on/off ratio of 10⁶ can be achieved at 350 °C, 5 orders of magnitude higher than that of a normal MoS₂ transistor in the same range of gate voltage. The short-term plasticity has a synaptic transistor function as an excellent low-pass dynamic filter. Long-term potentiation/depression and spike-timing-dependent plasticity are demonstrated at 150 °C. An ANN can be simulated, with the recognition accuracy reaching 90%. Our work provides promising strategies for high-temperature neuromorphic applications.

Toward Learning in Neuromorphic Circuits Based on Quantum Phase Slip Junctions

We explore the use of superconducting quantum phase slip junctions (QPSJs), an electromagnetic dual to Josephson Junctions (JJs), in neuromorphic circuits. These small circuits could serve as the building blocks of neuromorphic circuits for machine learning applications because they exhibit desirable properties such as inherent ultra-low energy per operation, high speed, dense integration, negligible loss, and natural spiking responses. In addition, they have a relatively straight-forward micro/nano fabrication, which shows promise for implementation of an enormous number of lossless interconnections that are required to realize complex neuromorphic systems. We simulate QPSJ-only, as well as hybrid QPSJ + JJ circuits for application in neuromorphic circuits including artificial synapses and neurons, as well as fan-in and fan-out circuits. We also design and simulate learning circuits, where a simplified spike timing dependent plasticity rule is realized to provide potential learning mechanisms. We also take an alternative approach, which shows potential to overcome some of the expected challenges of QPSJ-based neuromorphic circuits, via QPSJ-based charge islands coupled together to generate non-linear charge dynamics that result in a large number of programmable weights or non-volatile memory states. Notably, we show that these weights are a function of the timing and frequency of the input spiking signals and can be programmed using a small number of DC voltage bias signals, therefore exhibiting spike-timing and rate dependent plasticity, which are mechanisms to realize learning in neuromorphic circuits.

Stochastic binary synapses having sigmoidal cumulative distribution functions for unsupervised learning with spike timing-dependent plasticity

Spike timing-dependent plasticity (STDP), which is widely studied as a fundamental synaptic update rule for neuromorphic hardware, requires precise control of continuous weights. From the viewpoint of hardware implementation, a simplified update rule is desirable. Although simplified STDP with stochastic binary synapses was proposed previously, we find that it leads to degradation of memory maintenance during learning, which is unfavourable for unsupervised online learning. In this work, we propose a stochastic binary synaptic model where the cumulative probability of the weight change evolves in a sigmoidal fashion with potentiation or depression trials, which can be implemented using a pair of switching devices consisting of serially connected multiple binary memristors. As a benchmark test we perform simulations of unsupervised learning of MNIST images with a two-layer network and show that simplified STDP in combination with this model can outperform conventional rules with continuous weights not only in memory maintenance but also in recognition accuracy. Our method achieves 97.3% in recognition accuracy, which is higher than that reported with standard STDP in the same framework. We also show that the high performance of our learning rule is robust against device-to-device variability of the memristor's probabilistic behaviour.

UV light modulated synaptic behavior of MoTe2/BN heterostructure

Electrical synaptic devices are the basic components for the hardware based neuromorphic computational systems, which are expected to break the bottleneck of current von Neumann architecture. So far, synaptic devices based on three-terminal transistors are considered to provide the most stable performance, which usually use gate pulses to modulate the channel conductance through a floating gate and/or charge trapping layer. Herein, we report a three-terminal synaptic device based on a two-dimensional molybdenum ditelluride (MoTe₂)/hexagonal boron nitride (hBN) heterostructure. This structure enables stable and prominent conductance modulation of the MoTe₂channel by the photo-induced doping method through electron migration between the MoTe₂channel and ultraviolet (UV) light excited mid-gap defect states in hBN. Therefore, it is free of the floating gate and charge trapping layer to reduce the thickness and simplify the fabrication/design of the device. Moreover, since UV illumination is indispensable for stable doping in MoTe₂channel, the device can realize both short- (without UV illumination) and long- (with UV illumination) term plasticity. Meanwhile, the introduction of UV light allows additional tunability on the MoTe₂channel conductance through the wavelength and power intensity of incident UV, which may be important to mimic advanced synaptic functions. In addition, the photo-induced doping method can bidirectionally dope MoTe₂channel, which not only leads to large high/low resistance ratio for potential multi-level storage, but also implement both potentiation (n-doping) and depression (p-doping) of synaptic weight. This work explores alternative three-terminal synaptic configuration without floating gate and charge trapping layer, which may inspire researches on novel electrical synapse mechanisms.

Artificial Neuron and Synapse Devices Based on 2D Materials

Neuromorphic systems, which emulate neural functionalities of a human brain, are considered to be an attractive next-generation computing approach, with advantages of high energy efficiency and fast computing speed. After these neuromorphic systems are proposed, it is demonstrated that artificial synapses and neurons can mimic neural functions of biological synapses and neurons. However, since the neuromorphic functionalities are highly related to the surface properties of materials, bulk material-based neuromorphic devices suffer from uncontrollable defects at surfaces and strong scattering caused by dangling bonds. Therefore, 2D materials which have dangling-bond-free surfaces and excellent crystallinity have emerged as promising candidates for neuromorphic computing hardware. First, the fundamental synaptic behavior is reviewed, such as synaptic plasticity and learning rule, and requirements of artificial synapses to emulate biological synapses. In addition, an overview of recent advances on 2D materials-based synaptic devices is summarized by categorizing these into various working principles of artificial synapses. Second, the compulsory behavior and requirements of artificial neurons such as the all-or-nothing law and refractory periods to simulate a spike neural network are described, and the implementation of 2D materials-based artificial neurons to date is reviewed. Finally, future challenges and outlooks of 2D materials-based neuromorphic devices are discussed.

Engineering Method for Tailoring Electrical Characteristics in TiN/TiOx/HfOx/Au Bi-Layer Oxide Memristive Devices

Memristive devices have led to an increased interest in neuromorphic systems. However, different device requirements are needed for the multitude of computation schemes used there. While linear and time-independent conductance modulation is required for machine learning, non-linear and time-dependent properties are necessary for neurobiologically realistic learning schemes. In this context, an adaptation of the resistance switching characteristic is necessary with regard to the desired application. Recently, bi-layer oxide memristive systems have proven to be a suitable device structure for this purpose, as they combine the possibility of a tailored memristive characteristic with low power consumption and uniformity of the device performance. However, this requires technological solutions that allow for precise adjustment of layer thicknesses, defect densities in the oxide layers, and suitable area sizes of the active part of the devices. For this purpose, we have investigated the bi-layer oxide system TiN/TiOx/HfOx/Au with respect to tailored I-V non-linearity, the number of resistance states, electroforming, and operating voltages. Therefore, a 4-inch full device wafer process was used. This process allows a systematic investigation, i.e., the variation of physical device parameters across the wafer as well as a statistical evaluation of the electrical properties with regard to the variability from device to device and from cycle to cycle. For the investigation, the thickness of the HfOx layer was varied between 2 and 8 nm, and the size of the active area of devices was changed between 100 and 2,500 µm2. Furthermore, the influence of the HfOx deposition condition was investigated, which influences the conduction mechanisms from a volume-based, filamentary to an interface-based resistive switching mechanism. Our experimental results are supported by numerical simulations that show the contribution of the HfOx film in the bi-layer memristive system and guide the development of a targeting device.

Effects of memristive synapse radiation interactions on learning in spiking neural networks

This study uses advanced modeling and simulation to explore the effects of external events such as radiation interactions on the synaptic devices in an electronic spiking neural network. Specifically, the networks are trained using the spike-timing-dependent plasticity (STDP) learning rule to recognize spatio-temporal patterns (STPs) representing 25 and 100-pixel characters. Memristive synapses based on a TiO2 non-linear drift model designed in Verilog-A are utilized, with STDP learning behavior achieved through bi-phasic pre- and post-synaptic action potentials. The models are modified to include experimentally observed state-altering and ionizing radiation effects on the device. It is found that radiation interactions tend to make the connection between afferents stronger by increasing the conductance of synapses overall, subsequently distorting the STDP learning curve. In the absence of consistent STPs, these effects accumulate over time and make the synaptic weight evolutions unstable. With STPs at lower flux intensities, the network can recover and relearn with constant training. However, higher flux can overwhelm the leaky integrate-and-fire post-synaptic neuron circuits and reduce stability of the network.

EqSpike: spike-driven equilibrium propagation for neuromorphic implementations

Finding spike-based learning algorithms that can be implemented within the local constraints of neuromorphic systems, while achieving high accuracy, remains a formidable challenge. Equilibrium propagation is a promising alternative to backpropagation as it only involves local computations, but hardware-oriented studies have so far focused on rate-based networks. In this work, we develop a spiking neural network algorithm called EqSpike, compatible with neuromorphic systems, which learns by equilibrium propagation. Through simulations, we obtain a test recognition accuracy of 97.6% on the MNIST handwritten digits dataset (Mixed National Institute of Standards and Technology), similar to rate-based equilibrium propagation, and comparing favorably to alternative learning techniques for spiking neural networks. We show that EqSpike implemented in silicon neuromorphic technology could reduce the energy consumption of inference and training, respectively, by three orders and two orders of magnitude compared to graphics processing units. Finally, we also show that during learning, EqSpike weight updates exhibit a form of spike-timing-dependent plasticity, highlighting a possible connection with biology.

Mimicking Neurotransmitter Activity and Realizing Algebraic Arithmetic on Flexible Protein-Gated Oxide Neuromorphic Transistors

Recently, flexible neuromorphic devices have attracted extensive attention for the construction of perception cognitive systems with the ultimate objective to achieve robust computation, efficient learning, and adaptability to evolutionary changes. In particular, the design of flexible neuromorphic devices with data processing and arithmetic capabilities is highly desirable for wearable cognitive platforms. Here, an albumen-based protein-gated flexible indium tin oxide (ITO) ionotronic neuromorphic transistor was proposed. First, the transistor demonstrates excellent mechanical robustness against bending stress. Moreover, spike-duration-dependent synaptic plasticity and spike-amplitude-dependent synaptic plasticity behaviors are not affected by bending stress. With the unique protonic gating behaviors, neurotransmission processes in biological synapses are emulated, exhibiting three characteristics in neurotransmitter release, including quantal release, stochastic release, and excitatory or inhibitory release. In addition, three types of spike-timing-dependent plasticity learning rules are mimicked on the ITO ionotronic neuromorphic transistor. Most interestingly, algebraic arithmetic operations, including addition, subtraction, multiplication, and division, are implemented on the protein gated neuromorphic transistor for the first time. The present work would open a promising biorealistic avenue to the scientific community to control and design wearable "green" cognitive platforms, with potential applications including but not limited to intelligent humanoid robots and replacement neuroprosthetics.

Plasticity and Adaptation in Neuromorphic Biohybrid Systems

Neuromorphic systems take inspiration from the principles of biological information processing to form hardware platforms that enable the large-scale implementation of neural networks. The recent years have seen both advances in the theoretical aspects of spiking neural networks for their use in classification and control tasks and a progress in electrophysiological methods that is pushing the frontiers of intelligent neural interfacing and signal processing technologies. At the forefront of these new technologies, artificial and biological neural networks are tightly coupled, offering a novel "biohybrid" experimental framework for engineers and neurophysiologists. Indeed, biohybrid systems can constitute a new class of neuroprostheses opening important perspectives in the treatment of neurological disorders. Moreover, the use of biologically plausible learning rules allows forming an overall fault-tolerant system of co-developing subsystems. To identify opportunities and challenges in neuromorphic biohybrid systems, we discuss the field from the perspectives of neurobiology, computational neuroscience, and neuromorphic engineering.

Toward Scalable, Efficient, and Accurate Deep Spiking Neural Networks With Backward Residual Connections, Stochastic Softmax, and Hybridization

Spiking Neural Networks (SNNs) may offer an energy-efficient alternative for implementing deep learning applications. In recent years, there have been several proposals focused on supervised (conversion, spike-based gradient descent) and unsupervised (spike timing dependent plasticity) training methods to improve the accuracy of SNNs on large-scale tasks. However, each of these methods suffer from scalability, latency, and accuracy limitations. In this paper, we propose novel algorithmic techniques of modifying the SNN configuration with backward residual connections, stochastic softmax, and hybrid artificial-and-spiking neuronal activations to improve the learning ability of the training methodologies to yield competitive accuracy, while, yielding large efficiency gains over their artificial counterparts. Note, artificial counterparts refer to conventional deep learning/artificial neural networks. Our techniques apply to VGG/Residual architectures, and are compatible with all forms of training methodologies. Our analysis reveals that the proposed solutions yield near state-of-the-art accuracy with significant energy-efficiency and reduced parameter overhead translating to hardware improvements on complex visual recognition tasks, such as, CIFAR10, Imagenet datatsets.

Photoinduced Persistent Electron Accumulation and Depletion in LaAlO_{3}/SrTiO_{3} Quantum Wells

Persistent photoconductance is a phenomenon found in many semiconductors, by which light induces long-lived excitations in electronic states. Commonly, persistent photoexcitation leads to an increase of carriers (accumulation), though occasionally it can be negative (depletion). Here, we present the quantum well at the LaAlO_{3}/SrTiO_{3} interface, where in addition to photoinduced accumulation, a secondary photoexcitation enables carrier depletion. The balance between both processes is wavelength dependent, and allows tunable accumulation or depletion in an asymmetric manner, depending on the relative arrival time of photons of different frequencies. We use Green's function formalism to describe this unconventional photoexcitation, which paves the way to an optical implementation of neurobiologically inspired spike-timing-dependent plasticity.

Flexible Neuromorphic Electronics for Computing, Soft Robotics, and Neuroprosthetics

Flexible neuromorphic electronics that emulate biological neuronal systems constitute a promising candidate for next-generation wearable computing, soft robotics, and neuroprosthetics. For realization, with the achievement of simple synaptic behaviors in a single device, the construction of artificial synapses with various functions of sensing and responding and integrated systems to mimic complicated computing, sensing, and responding in biological systems is a prerequisite. Artificial synapses that have learning ability can perceive and react to events in the real world; these abilities expand the neuromorphic applications toward health monitoring and cybernetic devices in the future Internet of Things. To demonstrate the flexible neuromorphic systems successfully, it is essential to develop artificial synapses and nerves replicating the functionalities of the biological counterparts and satisfying the requirements for constructing the elements and the integrated systems such as flexibility, low power consumption, high-density integration, and biocompatibility. Here, the progress of flexible neuromorphic electronics is addressed, from basic backgrounds including synaptic characteristics, device structures, and mechanisms of artificial synapses and nerves, to applications for computing, soft robotics, and neuroprosthetics. Finally, future research directions toward wearable artificial neuromorphic systems are suggested for this emerging area.

Sub-nanosecond memristor based on ferroelectric tunnel junction

Next-generation non-volatile memories with ultrafast speed, low power consumption, and high density are highly desired in the era of big data. Here, we report a high performance memristor based on a Ag/BaTiO₃/Nb:SrTiO₃ ferroelectric tunnel junction (FTJ) with the fastest operation speed (600 ps) and the highest number of states (32 states or 5 bits) per cell among the reported FTJs. The sub-nanosecond resistive switching maintains up to 358 K, and the write current density is as low as 4 × 10³ A cm⁻². The functionality of spike-timing-dependent plasticity served as a solid synaptic device is also obtained with ultrafast operation. Furthermore, it is demonstrated that a Nb:SrTiO₃ electrode with a higher carrier concentration and a metal electrode with lower work function tend to improve the operation speed. These results may throw light on the way for overcoming the storage performance gap between different levels of the memory hierarchy and developing ultrafast neuromorphic computing systems.

Memristor devices based on ferroelectric tunnel junctions are promising, but suffer from quite slow switching times. Here, the authors report on ultrafast switching times at and above room temperature of 600ps in Ag/BaTiO3/Nb:SrTiO3 based ferroelectric tunnel junctions.

Spatial Properties of STDP in a Self-Learning Spiking Neural Network Enable Controlling a Mobile Robot

Development of spiking neural networks (SNNs) controlling mobile robots is one of the modern challenges in computational neuroscience and artificial intelligence. Such networks, being replicas of biological ones, are expected to have a higher computational potential than traditional artificial neural networks (ANNs). The critical problem is in the design of robust learning algorithms aimed at building a “living computer” based on SNNs. Here, we propose a simple SNN equipped with a Hebbian rule in the form of spike-timing-dependent plasticity (STDP). The SNN implements associative learning by exploiting the spatial properties of STDP. We show that a LEGO robot controlled by the SNN can exhibit classical and operant conditioning. Competition of spike-conducting pathways in the SNN plays a fundamental role in establishing associations of neural connections. It replaces the irrelevant associations by new ones in response to a change in stimuli. Thus, the robot gets the ability to relearn when the environment changes. The proposed SNN and the stimulation protocol can be further enhanced and tested in developing neuronal cultures, and also admit the use of memristive devices for hardware implementation.

Emerging neuromorphic devices

Artificial intelligence (AI) has the ability of revolutionizing our lives and society in a radical way, by enabling machine learning in the industry, business, health, transportation, and many other fields. The ability to recognize objects, faces, and speech, requires, however, exceptional computational power and time, which is conflicting with the current difficulties in transistor scaling due to physical and architectural limitations. As a result, to accelerate the progress of AI, it is necessary to develop materials, devices, and systems that closely mimic the human brain. In this work, we review the current status and challenges on the emerging neuromorphic devices for brain-inspired computing. First, we provide an overview of the memory device technologies which have been proposed for synapse and neuron circuits in neuromorphic systems. Then, we describe the implementation of synaptic learning in the two main types of neural networks, namely the deep neural network and the spiking neural network (SNN). Bio-inspired learning, such as the spike-timing dependent plasticity scheme, is shown to enable unsupervised learning processes which are typical of the human brain. Hardware implementations of SNNs for the recognition of spatial and spatio-temporal patterns are also shown to support the cognitive computation in silico. Finally, we explore the recent advances in reproducing bio-neural processes via device physics, such as insulating-metal transitions, nanoionics drift/diffusion, and magnetization flipping in spintronic devices. By harnessing the device physics in emerging materials, neuromorphic engineering with advanced functionality, higher density and better energy efficiency can be developed.

Memristive and Memory Impedance Behavior in a Photo-Annealed ZnO–rGO Thin-Film Device

An oxygen-rich ZnO-reduced graphene oxide (rGO) thin film was synthesized using a photo-annealing technique from zinc precursor (ZnO)–graphene oxide (GO) sol–gel solution. X-ray diffraction (XRD) results show a clear characteristic peak corresponding to rGO. The scanning electron microscope (SEM) image of the prepared thin film shows an evenly distributed wrinkled surface structure. Transition Metal Oxide (TMO)-based memristive devices are nominees for beyond CMOS Non-Volatile Memory (NVRAM) devices. The two-terminal Metal–TMO (Insulator)–Metal (MIM) memristive device is fabricated using a synthesized ZnO–rGO as an active layer on fluorine-doped tin oxide (FTO)-coated glass substrate. Aluminum (Al) is deposited as a top metal contact on the ZnO–rGO active layer to complete the device. Photo annealing was used to reduce the GO to rGO to make the proposed method suitable for fabricating ZnO–rGO thin-film devices on flexible substrates. The electrical characterization of the Al–ZnO–rGO–FTO device confirms the coexistence of memristive and memimpedance characteristics. The coexistence of memory resistance and memory impedance in the same device could be valuable for developing novel programmable analog filters and self-resonating circuits and systems.

Spike-shape dependence of the spike-timing dependent synaptic plasticity in ferroelectric-tunnel-junction synapses

Resistive switching (RS) devices have attracted increasing attention for artificial synapse applications in neural networks because of their nonvolatile and analogue resistance changes. Among the neural networks, a spiking neural network (SNN) based on spike-timing-dependent plasticity (STDP) is highly energy efficient. To implement STDP in resistive switching devices, several types of voltage spikes have been proposed to date, but there have been few reports on the relationship between the STDP characteristics and spike types. Here, we report the STDP characteristics implemented in ferroelectric tunnel junctions (FTJs) by several types of spikes. Based on simulated time evolutions of superimposed spikes and taking the nonlinear current-voltage (I-V) characteristics of FTJs into account, we propose equations for simulating the STDP curve parameters of a magnitude of the conductance change (ΔG_max) and a time window (τ_C) from the spike parameters of a peak amplitude (V_peak) and time durations (t_p and t_d) for three spike types: triangle-triangle, rectangular-triangle, and rectangular-rectangular. The power consumption experiments of the STDP revealed that the power consumption under the inactive-synapse condition (spike timing |Δt| > τ_C) was as large as 50–82% of that under the active-synapse condition (|Δt| < τ_C). This finding indicates that the power consumption under the inactive-synapse condition should be reduced to minimize the total power consumption of an SNN implemented by using FTJs as synapses.

Time and rate dependent synaptic learning in neuro-mimicking resistive memories

Memristors have demonstrated immense potential as building blocks in future adaptive neuromorphic architectures. Recently, there has been focus on emulating specific synaptic functions of the mammalian nervous system by either tailoring the functional oxides or engineering the external programming hardware. However, high device-to-device variability in memristors induced by the electroforming process and complicated programming hardware are among the key challenges that hinder achieving biomimetic neuromorphic networks. Here, a simple hybrid complementary metal oxide semiconductor (CMOS)-memristor approach is reported to implement different synaptic learning rules by utilizing a CMOS-compatible memristor based on oxygen-deficient SrTiO3-x (STOx). The potential of such hybrid CMOS-memristor approach is demonstrated by successfully imitating time-dependent (pair and triplet spike-time-dependent-plasticity) and rate-dependent (Bienenstosk-Cooper-Munro) synaptic learning rules. Experimental results are benchmarked against in-vitro measurements from hippocampal and visual cortices with good agreement. The scalability of synaptic devices and their programming through a CMOS drive circuitry elaborates the potential of such an approach in realizing adaptive neuromorphic computation and networks.

Unsupervised online learning of temporal information in spiking neural network using thin-film transistor-type NOR flash memory devices

Brain-inspired analog neuromorphic systems based on the synaptic arrays have attracted large attention due to low-power computing. Spike-timing-dependent plasticity (STDP) algorithm is considered as one of the appropriate neuro-inspired techniques to be applied for on-chip learning. The aim of this study is to investigate the methodology of unsupervised STDP based learning in temporal encoding systems. The system-level simulation was performed based on the measurement results of thin-film transistor-type asymmetric floating-gate NOR flash memory. With proposed learning methods, 91.53% of recognition accuracy is obtained in inferencing MNIST standard dataset with 200 output neurons. Moreover, temporal encoding rules showed that the number of input pulses and the computing power can be compressed without significant loss of recognition accuracy compared to the conventional rate encoding scheme. In addition, temporal computing in a multi-layer network is suitable for learning data sequences, suggesting the possibility of applying to real-world tasks such as classifying direction of moving objects.

Self-Organizing Neural Networks Based on OxRAM Devices under a Fully Unsupervised Training Scheme

A fully-unsupervised learning algorithm for reaching self-organization in neuromorphic architectures is provided in this work. We experimentally demonstrate spike-timing dependent plasticity (STDP) in Oxide-based Resistive Random Access Memory (OxRAM) devices, and propose a set of waveforms in order to induce symmetric conductivity changes. An empirical model is used to describe the observed plasticity. A neuromorphic system based on the tested devices is simulated, where the developed learning algorithm is tested, involving STDP as the local learning rule. The design of the system and learning scheme permits to concatenate multiple neuromorphic layers, where autonomous hierarchical computing can be performed.

Impact of Synaptic Device Variations on Classification Accuracy in a Binarized Neural Network

Brain-inspired neuromorphic systems (hardware neural networks) are expected to be an energy-efficient computing architecture for solving cognitive tasks, which critically depend on the development of reliable synaptic weight storage (i.e., synaptic device). Although various nanoelectronic devices have successfully reproduced the learning rules of biological synapses through their internal analog conductance states, the sustainability of such devices is still in doubt due to the variability common to all nanoelectronic devices. Alternatively, a neuromorphic system based on a relatively more reliable digital-type switching device has been recently demonstrated, i.e., a binarized neural network (BNN). The synaptic device is a more mature digital-type switching device, and the training/recognition algorithm developed for the BNN enables the task of facial image classification with a supervised training scheme. Here, we quantitatively investigate the effects of device parameter variations on the classification accuracy; the parameters include the number of weight states (N_state), the weight update margin (ΔG), and the weight update variation (G_var). This analysis demonstrates the feasibility of the BNN and introduces a practical neuromorphic system based on mature, conventional digital device technologies.

Synaptic learning behavior of a TiO2 nanowire memristor

TiO₂ nanowire memristors were fabricated by dielectrophoresis. The responding current of the memristor continuously increases and decreases with the consecutive positive and negative sweep voltage, which is similar to the nonlinear transmission characteristics of biological synapses. Spike-rate-dependent plasticity and learning behaviors of TiO₂ memristor were studied by applying programmed pulses. The pulses with higher amplitude, bigger width and smaller interval cause a larger excitatory postsynaptic current. The number of relearning pulses is decreased with the learning experience, and a deepening memory will be consolidated by the repeated learning process. A mechanism based on the oxygen vacancy migration is proposed for the learning behavior. Excess oxygen vacancies are generated during the learning process and the conducting pathways are formed by the vacancy drift under the applied voltage. After removing the voltage at the forgetting process, back diffusion and electron trapping of the oxygen vacancies dominate the relaxation time, and the metastable atoms are formed with the involvement of the oxygen atoms. However, weak chemical bonding among the metastable atoms leads to the migration of the regenerated oxygen vacancies again, contributing to the enhanced current in the relearning process.

Nanoscale resistive switching memory devices: a review

In this review the different concepts of nanoscale resistive switching memory devices are described and classified according to their I-V behaviour and the underlying physical switching mechanisms. By means of the most important representative devices, the current state of electrical performance characteristics is illuminated in-depth. Moreover, the ability of resistive switching devices to be integrated into state-of-the-art CMOS circuits under the additional consideration with a suitable selector device for memory array operation is assessed. From this analysis, and by factoring in the maturity of the different concepts, a ranking methodology for application of the nanoscale resistive switching memory devices in the memory landscape is derived. Finally, the suitability of the different device concepts for beyond pure memory applications, such as brain inspired and neuromorphic computational or logic in memory applications that strive to overcome the vanNeumann bottleneck, is discussed.

Neuromorphic Spiking Neural Networks and Their Memristor-CMOS Hardware Implementations

Inspired by biology, neuromorphic systems have been trying to emulate the human brain for decades, taking advantage of its massive parallelism and sparse information coding. Recently, several large-scale hardware projects have demonstrated the outstanding capabilities of this paradigm for applications related to sensory information processing. These systems allow for the implementation of massive neural networks with millions of neurons and billions of synapses. However, the realization of learning strategies in these systems consumes an important proportion of resources in terms of area and power. The recent development of nanoscale memristors that can be integrated with Complementary Metal-Oxide-Semiconductor (CMOS) technology opens a very promising solution to emulate the behavior of biological synapses. Therefore, hybrid memristor-CMOS approaches have been proposed to implement large-scale neural networks with learning capabilities, offering a scalable and lower-cost alternative to existing CMOS systems.

On the Application of a Diffusive Memristor Compact Model to Neuromorphic Circuits

Memristive devices have found application in both random access memory and neuromorphic circuits. In particular, it is known that their behavior resembles that of neuronal synapses. However, it is not simple to come by samples of memristors and adjusting their parameters to change their response requires a laborious fabrication process. Moreover, sample to sample variability makes experimentation with memristor-based synapses even harder. The usual alternatives are to either simulate or emulate the memristive systems under study. Both methodologies require the use of accurate modeling equations. In this paper, we present a diffusive compact model of memristive behavior that has already been experimentally validated. Furthermore, we implement an emulation architecture that enables us to freely explore the synapse-like characteristics of memristors. The main advantage of emulation over simulation is that the former allows us to work with real-world circuits. Our results can give some insight into the desirable characteristics of the memristors for neuromorphic applications.

Fast and Accurate Sparse Coding of Visual Stimuli With a Simple, Ultralow-Energy Spiking Architecture

Memristive crossbars have become a popular means for realizing unsupervised and supervised learning techniques. In previous neuromorphic architectures with leaky integrate-and-fire neurons, the crossbar itself has been separated from the neuron capacitors to preserve mathematical rigor. In this paper, we sought to design a simplified sparse coding circuit without this restriction, resulting in a fast circuit that approximated a sparse coding operation at a minimal loss in accuracy. We showed that connecting the neurons directly to the crossbar resulted in a more energy-efficient sparse coding architecture and alleviated the need to prenormalize receptive fields. This paper provides derivations for the design of such a network, named the simple spiking locally competitive algorithm, as well as CMOS designs and results on the CIFAR and MNIST data sets. Compared to a nonspiking, nonapproximate model which scored 33% on CIFAR-10 with a single-layer classifier, this hardware scored 32% accuracy. When used with a state-of-the-art deep learning classifier, the nonspiking model achieved 82% and our simplified, spiking model achieved 80% while compressing the input data by 92%. Compared to a previously proposed spiking model, our proposed hardware consumed 99% less energy to do the same work at 21 × the throughput. Accuracy held out with online learning to a write variance of 3%, suitable for the often reported 4-bit resolution required for neuromorphic algorithms, with offline learning to a write variance of 27%, and with read variance to 40%. The proposed architecture's excellent accuracy, throughput, and significantly lower energy usage demonstrate the utility of our innovations.

Artificial Neuron and Synapse Realized in an Antiferromagnet/Ferromagnet Heterostructure Using Dynamics of Spin-Orbit Torque Switching

Efficient information processing in the human brain is achieved by dynamics of neurons and synapses, motivating effective implementation of artificial spiking neural networks. Here, the dynamics of spin-orbit torque switching in antiferromagnet/ferromagnet heterostructures is studied to show the capability of the material system to form artificial neurons and synapses for asynchronous spiking neural networks. The magnetization switching, driven by a single current pulse or trains of pulses, is examined as a function of the pulse width (1 s to 1 ns), amplitude, number, and pulse-to-pulse interval. Based on this dynamics and the unique ability of the system to exhibit binary or analog behavior depending on the device size, key functionalities of a synapse (spike-timing-dependent plasticity) and a neuron (leaky integrate-and-fire) are reproduced in the same material and on the basis of the same working principle. These results open a way toward spintronics-based neuromorphic hardware that executes cognitive tasks with the efficiency of the human brain.

Emerging memory technologies for neuromorphic computing

In this paper, we reviewed the recent trends on neuromorphic computing using emerging memory technologies. Two representative learning algorithms used to implement a hardware-based neural network are described as a bio-inspired learning algorithm and software-based learning algorithm, in particular back-propagation. The requirements of the synaptic device to apply each algorithm were analyzed. Then, we reviewed the research trends of synaptic devices to implement an artificial neural network.

Synaptic plasticity, metaplasticity and memory effects in hybrid organic-inorganic bismuth-based materials

Since the discovery of memristors, their application in computing systems utilizing multivalued logic and a neuromimetic approach is of great interest. A thin film device made of methylammonium bismuth iodide exhibits a wide variety of neuromorphic effects simultaneously, and is thus able to mimic synaptic behaviour and learning phenomena. Standard learning protocols, such as spike-timing dependent plasticity and spike-rate dependent plasticity might be further modulated via metaplasticity in order to amplify or alter changes in the synaptic weight. Moreover, transfer of information from short-term to long-term memory is observed. These effects show that the diversity of functions of memristive devices can be strongly affected by the pre-treatment of the sample. Modulation of the resistive switching amplitude is of great importance for the application of memristive elements in computational applications, as additional sub-states might be utilized in multi-valued logic systems and metaplasticity and memory consolidation will contribute to the development of more efficient bioinspired computational schemes.

Neuromorphic computation with spiking memristors: habituation, experimental instantiation of logic gates and a novel sequence-sensitive perceptron model

This paper presents rules based on the physical behaviour of the device to instantiate logic gates for further computation and a method of understanding the memristor’s operation as a type of non-linear, sequence-sensitive perceptron.

A multi-level memristor based on atomic layer deposition of iron oxide

This work reports the fabrication of memristive devices based on iron oxide (Fe₂O₃) thin films grown by atomic layer deposition (ALD) using ferrocene as iron precursor and ozone as oxidant. An excellent control of the ALD process was achieved by using an experimental procedure based on a sequence of micro-pulses, which provided long residence time and homogeneous diffusion of precursors, allowing ALD of thin films with smooth morphology and crystallinity which was found to increase with layer thickness, at temperatures as low as 250 °C. The resistive switching of symmetric Pt/Fe₂O₃/Pt thin film devices exhibited bipolar mode with good stability and endurance. Multi-level switching was achieved via current and voltage control. It was proved that the ON state regime can be tuned by changing the current compliance while the OFF state can be changed to intermediate levels by decreasing the maximum voltage during RESET. The structural analysis of the switched oxide layer revealed the presence of nano-sized crystalline domains corresponding to different iron oxide phases, suggesting that Joule heating effects during I-V cycling are responsible for a crystallization process of the pristine amorphous layer.