Search

In recent years, artificial neural networks have become the flagship algorithm of artificial intelligence 1 . In these systems, neuron activation functions are static, and computing is achieved through standard arithmetic operations. By contrast, a prominent branch of neuroinspired computing embraces the dynamical nature of the brain and proposes to endow each component of a neural network with dynamical functionality, such as oscillations, and to rely on emergent physical phenomena, such as synchronization 2-6 , for solving complex problems with small networks 7-11 . This approach is especially interesting for hardware implementations, because emerging nanoelectronic devices can provide compact and energy-efficient nonlinear auto-oscillators that mimic the periodic spiking activity of biological neurons 12-16 . The dynamical couplings between oscillators can then be used to mediate the synaptic communication between the artificial neurons. One challenge for using nanodevices in this way is to achieve learning, which requires fine control and tuning of their coupled oscillations 17 ; the dynamical features of nanodevices can be difficult to control and prone to noise and variability 18 . Here we show that the outstanding tunability of spintronic nano-oscillators-that is, the possibility of accurately controlling their frequency across a wide range, through electrical current and magnetic field-can be used to address this challenge. We successfully train a hardware network of four spin-torque nano-oscillators to recognize spoken vowels by tuning their frequencies according to an automatic real-time learning rule. We show that the high experimental recognition rates stem from the ability of these oscillators to synchronize. Our results demonstrate that non-trivial pattern classification tasks can be achieved with small hardware neural networks by endowing them with nonlinear dynamical features such as oscillations and synchronization.

With conventional transistor technologies reaching their limits, alternative computing schemes based on novel technologies are currently gaining considerable interest. Notably, promising computing approaches have proposed to leverage the complex dynamics emerging in networks of coupled oscillators based on nanotechnologies. The physical implementation of such architectures remains a true challenge, however, as most proposed ideas are not robust to nanotechnology devices' non-idealities. In this work, we propose and investigate the implementation of an oscillator-based architecture, which can be used to carry out pattern recognition tasks, and which is tailored to the specificities of nanotechnologies. This scheme relies on a weak coupling between oscillators, and does not require a fine tuning of the coupling values. After evaluating its reliability under the severe constraints associated to nanotechnologies, we explore the scalability of such an architecture, suggesting its potential to realize pattern recognition tasks using limited resources. We show that it is robust to issues like noise, variability and oscillator non-linearity. Defining network optimization design rules, we show that nano-oscillator networks could be used for efficient cognitive processing.

Multiple modern applications of electronics call for inexpensive chips that can perform complex operations on natural data with limited energy. A vision for accomplishing this is implementing hardware neural networks, which fuse computation and memory, with low cost organic electronics. A challenge, however, is the implementation of synapses (analog memories) composed of such materials. In this work, we introduce robust, fastly programmable, nonvolatile organic memristive nanodevices based on electrografted redox complexes that implement synapses thanks to a wide range of accessible intermediate conductivity states. We demonstrate experimentally an elementary neural network, capable of learning functions, which combines four pairs of organic memristors as synapses and conventional electronics as neurons. Our architecture is highly resilient to issues caused by imperfect devices. It tolerates inter-device variability and an adaptable learning rule offers immunity against asymmetries in device switching. Highly compliant with conventional fabrication processes, the system can be extended to larger computing systems capable of complex cognitive tasks, as demonstrated in complementary simulations.

The electronic structure and magnetic properties of α-CoV2O6 are investigated using density functional theory calculations including spin-orbit coupling and orbital polarization effects. These calculations reveal a strong magnetocrystalline anisotropy with a magnetization easy axis close to the c axis. The evaluation of magnetic couplings on the basis of broken-symmetry formalism suggests the occurrence of an antiferromagnetic ground-state order where ferromagnetic chains running along b are coupled antiferromagnetically to their nearest neighbors along a and c. Monte Carlo simulations are finally employed to explore the origins of the 1/3 plateau observed in the magnetization curves of this compound and to propose a structure for the corresponding state. [ABSTRACT FROM AUTHOR]

Recent work has exposed the idea that interesting quantum-like (QL) probability laws, including interference effects, can be manifest in classical systems. Here, we propose a model for QL states and QL bits. We suggest a way that huge, complex systems can host robust states that can process information in a QL fashion. Axioms that such states should satisfy are proposed. Specifically, it is shown that building blocks suited for QL states are networks, possibly very complex, that we defined based on k -regular random graphs. These networks can dynamically encode a lot of information that is distilled into the emergent states we can use for QL processing. Although the emergent states are classical, they have properties analogous to quantum states. Concrete examples of how QL functions are possible are given. The possibility of a 'QL advantage' for computing-type operations and the potential relevance for new kinds of function in the brain are discussed and left as open questions. [ABSTRACT FROM AUTHOR]

Intelligent transportation systems (ITSs) make it possible for applications such as autonomous driving, active safety systems and smart cities. As the infrastructure of ITSs, vehicular ad-hoc network (VANET) plays a key role in ensuring traffic safety while improving driving experience and comfort. However, VANET faces many challenges as the network is exposed to the public and involves sensitive information such as vehicle control commands and driving records. Although blockchain technology can provide such features as decentralization, immutability, non-reliance on trust and traceability, most of the existing blockchain systems impose high storage and computing capacity requirements for participating nodes while terminals nodes in VANET, such as road-side units (RSUs), on-board units (OBUs) and sensors, usually have very limited storage and computing capacity. To solve this problem, this paper proposes a blockchain system that provides scalable storage capacity for VANETs. The proposed scheme uses network sharding and multi-consensus strategy to improve the topology and consensus process of the blockchain. Aimed at lowering the storage requirement on the RSU nodes that participate in the blockchain, a collaborative storage mechanism and a dynamic copy number strategy for the blockchain ledger is designed. Experimental results show that compared to existing schemes, the proposed scalable blockchain storage scheme can significantly lower the storage requirement for the blockchain nodes, thus making it possible for the nodes with limited storage capacity in VANET to participate in the maintenance of the blockchain to prevent the blockchain from drifting towards centralization, and support maximal sharing of road traffic information while ensuring the security, privacy and trustworthiness of information. [ABSTRACT FROM AUTHOR]

While digital computers rely on software-generated pseudo-random number generators, hardware-based true random number generators (TRNGs), which employ the natural physics of the underlying hardware, provide true stochasticity, and power and area efficiency. Research into TRNGs has extensively relied on the unpredictability in phase transitions, but such phase transitions are difficult to control given their often abrupt and narrow parameter ranges (e.g., occurring in a small temperature window). Here we demonstrate a TRNG based on self-oscillations in LaCoO3 that is electrically biased within its spin crossover regime. The LaCoO3 TRNG passes all standard tests of true stochasticity and uses only half the number of components compared to prior TRNGs. Assisted by phase field modeling, we show how spin crossovers are fundamentally better in producing true stochasticity compared to traditional phase transitions. As a validation, by probabilistically solving the NP-hard max-cut problem in a memristor crossbar array using our TRNG as a source of the required stochasticity, we demonstrate solution quality exceeding that using software-generated randomness. Probabilistic computing demands low power and high quality random number generation. Woo et al. demonstrate the use of a spin crossover in LaCoO3 to generate random numbers that outperform software-generated random numbers in probabilistic computing. [ABSTRACT FROM AUTHOR]

Phase-encoded oscillating neural networks offer compelling advantages over metal-oxide-semiconductor-based technology for tackling complex optimization problems, with promising potential for ultralow power consumption and exceptionally rapid computational performance. In this work, we investigate the ability of these networks to solve optimization problems belonging to the nondeterministic polynomial time complexity class using nanoscale vanadium-dioxide-based oscillators integrated onto a Silicon platform. Specifically, we demonstrate how the dynamic behavior of coupled vanadium dioxide devices can effectively solve combinatorial optimization problems, including Graph Coloring, Max-cut, and Max-3SAT problems. The electrical mappings of these problems are derived from the equivalent Ising Hamiltonian formulation to design circuits with up to nine crossbar vanadium dioxide oscillators. Using sub-harmonic injection locking techniques, we binarize the solution space provided by the oscillators and demonstrate that graphs with high connection density (η > 0.4) converge more easily towards the optimal solution due to the small spectral radius of the problem's equivalent adjacency matrix. Our findings indicate that these systems achieve stability within 25 oscillation cycles and exhibit power efficiency and potential for scaling that surpasses available commercial options and other technologies under study. These results pave the way for accelerated parallel computing enabled by large-scale networks of interconnected oscillators. Oscillating neural networks promise ultralow power consumption and rapid computation for tackling complex optimization problems. Here, the authors demonstrate VO2 oscillators to solve NP-complete problems with projected power consumption of 13 µW/oscillator. [ABSTRACT FROM AUTHOR]

Extending Moore's law by augmenting complementary-metal-oxide semiconductor (CMOS) transistors with emerging nanotechnologies (X) has become increasingly important. One important class of problems involve sampling-based Monte Carlo algorithms used in probabilistic machine learning, optimization, and quantum simulation. Here, we combine stochastic magnetic tunnel junction (sMTJ)-based probabilistic bits (p-bits) with Field Programmable Gate Arrays (FPGA) to create an energy-efficient CMOS + X (X = sMTJ) prototype. This setup shows how asynchronously driven CMOS circuits controlled by sMTJs can perform probabilistic inference and learning by leveraging the algorithmic update-order-invariance of Gibbs sampling. We show how the stochasticity of sMTJs can augment low-quality random number generators (RNG). Detailed transistor-level comparisons reveal that sMTJ-based p-bits can replace up to 10,000 CMOS transistors while dissipating two orders of magnitude less energy. Integrated versions of our approach can advance probabilistic computing involving deep Boltzmann machines and other energy-based learning algorithms with extremely high throughput and energy efficiency. Designing energy-efficient and scalable hardware capable of accelerating Monte Carlo algorithms is highly desirable for probabilistic computing. Here, Singh et al. combine stochastic magnetic tunnel junction-based probabilistic bits with versatile field programmable gate arrays to achieve this goa [ABSTRACT FROM AUTHOR]

Logical gates have been used in implementation of logic sequential and combinational circuits especially in computers, DSPs and microprocessors. They are mostly fabricated based on CMOS technology that provides a few nano-seconds of delay for each digital gate. Due to limitations in more scaling CMOS transistors and cause of short channel effects, some engineers and researchers believe that there is a huge demand for new devices and fabrication technologies to produce faster logic gates. In this paper, a new architecture for NAND gate is presented which operates based on mechanical resonance of a network of weakly coupled resonators that resonate in radio frequencies. This design is achieved by employing the associative memory property of Hopfield neural networks and the theory of weakly coupled resonators. The main advantage of the proposed design is in its capability to reach out delay times of the order of 1 nano-seconds or even less. One solution to decrease the delay time can be increasing the resonance frequency of resonators which are processing elements of resonators network. In this paper just the new idea of implementation NAND gate based on weakly coupled RF MEMS resonators is presented and evaluated. Other criteria like gate power consumption, effects of temperature, fan in, fan out and noise margin are not discussed. Regarding the rapid growth in MEMS technology, resonators with super high frequency (SHF 3–30 GHz) are now available and those with Extremely High frequency (EHF 30–300 GHz) will soon be in market which enables the presented design to achieve higher speeds. In addition, mechanical resonators are more fault tolerant than CMOS circuits when utilized in harsh environments which exposed to ionic beams or electron beam radiations. In space applications, the satellite and payload are exposed to huge bombardments of space electrons and ions beams. So, the proposed NAND gate can be a good solution for enhancing reliability of devices and systems exposed to space radiations. [ABSTRACT FROM AUTHOR]

Domain walls (DWs) in magnetic nanowires are promising candidates for a variety of applications including Boolean/unconventional logic, memories, in-memory computing as well as magnetic sensors and biomagnetic implementations. They show rich physical behaviour and are controllable using a number of methods including magnetic fields, charge and spin currents and spin-orbit torques. In this review, we detail types of DWs in ferromagnetic nanowires and describe processes of manipulating their state. We look at the state of the art of DW applications and give our take on the their current status, technological feasibility and challenges. [ABSTRACT FROM AUTHOR]

The synthesis of poly[N-(3-(9H-carbazol-9-yl)propyl)methacrylamide] (PCaPMA) is presented, along with its physical, photophysical, and electrical properties, revealing its promising potential for the application in bistable memory devices. The incorporation of the carbazole heterocycle, as a functional charge carrier transporting unit attached to a saturated polymethacryalamide backbone by a flexible alkyl chain, facilitates resistive switching by the applied voltage. Thin films of the polymer, sandwiched between Al or Au and ITO electrodes, exhibit rewriteable flash memory behavior with bistable conductivity with a setting voltage ranging from 2 to 4.5 V, achieving a current ON/OFF ratio exceeding 100. The device demonstrates a remarkable lifetime and remains persistent for more than 104 seconds under a static voltage of 0.5 V. The main physical mechanisms driving the resistive switching have been attributed to the electric field-induced reorientation of heterocycles, which modulates charge transport, and trapping/detrapping of charges in localized states within the bulk of the polymer. Memory persistence is strengthened by the physical crosslinking caused by hydrogen bonds between amide and carbonyl groups in the aliphatic side chains. This physical network further enhances the thermal and mechanical stability of the PCaPMA in comparison to similar polymers highlighting its potential as a suitable material for organic memory devices. [ABSTRACT FROM AUTHOR]

Safety-critical sensory applications, like medical diagnosis, demand accurate decisions from limited, noisy data. Bayesian neural networks excel at such tasks, offering predictive uncertainty assessment. However, because of their probabilistic nature, they are computationally intensive. An innovative solution utilizes memristors' inherent probabilistic nature to implement Bayesian neural networks. However, when using memristors, statistical effects follow the laws of device physics, whereas in Bayesian neural networks, those effects can take arbitrary shapes. This work overcome this difficulty by adopting a variational inference training augmented by a "technological loss", incorporating memristor physics. This technique enabled programming a Bayesian neural network on 75 crossbar arrays of 1,024 memristors, incorporating CMOS periphery for in-memory computing. The experimental neural network classified heartbeats with high accuracy, and estimated the certainty of its predictions. The results reveal orders-of-magnitude improvement in inference energy efficiency compared to a microcontroller or an embedded graphics processing unit performing the same task. Bayesian networks gain importance in safety-critical applications. Authors conducted experiments with a memristor-based Bayesian network trained with variational inference with technological loss, achieving accurate heartbeats classification and prediction certainty. [ABSTRACT FROM AUTHOR]

Arrays of coupled nanoelectromechanical resonators are a promising foundation for implementing large-scale network applications, such as mechanical-based information processing and computing, but their practical realization remains an outstanding challenge. In this work, we demonstrate a scalable platform of suspended graphene resonators, such that neighboring resonators are persistently coupled mechanically. We provide evidence of strong coupling between neighboring resonators using two different tuning methods. Additionally, we provide evidence of inter-resonator coupling of higher-order modes, demonstrating the rich dynamics that can be accessed with this platform. Our results establish this platform as a viable option for realizing large-scale programmable networks, enabling applications such as phononic circuits, tunable waveguides, and reconfigurable metamaterials. [ABSTRACT FROM AUTHOR]

Over the decades, the spin dynamics of a large set of lanthanide complexes have been explored. Lanthanide-based molecular nanomagnets are bistable spin systems, generally conceptualised as classical bits, but many lanthanide complexes have also been presented as candidate quantum bits (qubits). Here, we offer a third alternative and model them as probabilistic bits (p-bits), where their stochastic behaviour constitutes a computational resource instead of a limitation. Employing an ad-hoc modelling tool for molecular spin p-bits and molecular nanomagnets, we simulate a minimal p-bit network under realistic conditions. Finally, we go back to a recently published dataset and screen the best lanthanide complexes for p-bit behaviour, lay out the performance of the different lanthanide ions and chemical families and offer some chemical design considerations. [ABSTRACT FROM AUTHOR]

Future brain–computer interfaces will require local and highly individualized signal processing of fully integrated electronic circuits within the nervous system and other living tissue. New devices will need to be developed that can receive data from a sensor array, process these data into meaningful information, and translate that information into a format that can be interpreted by living systems. Here, the first example of interfacing a hardware‐based pattern classifier with a biological nerve is reported. The classifier implements the Widrow–Hoff learning algorithm on an array of evolvable organic electrochemical transistors (EOECTs). The EOECTs' channel conductance is modulated in situ by electropolymerizing the semiconductor material within the channel, allowing for low voltage operation, high reproducibility, and an improvement in state retention by two orders of magnitude over state‐of‐the‐art OECT devices. The organic classifier is interfaced with a biological nerve using an organic electrochemical spiking neuron to translate the classifier's output to a simulated action potential. The latter is then used to stimulate muscle contraction selectively based on the input pattern, thus paving the way for the development of adaptive neural interfaces for closed‐loop therapeutic systems. [ABSTRACT FROM AUTHOR]

The conventional computing method based on the von Neumann architecture is limited by a series of problems such as high energy consumption, finite data exchange bandwidth between processors and storage media, etc., and it is difficult to achieve higher computing efficiency. A more efficient unconventional computing architecture is urgently needed to overcome these problems. Neuromorphic computing and stochastic computing have been considered to be two competitive candidates for unconventional computing, due to their extraordinary potential for energy-efficient and high-performance computing. Although conventional electronic devices can mimic the topology of the human brain, these require high power consumption and large area. Spintronic devices represented by magnetic tunnel junctions (MTJs) exhibit remarkable high-energy efficiency, non-volatility, and similarity to biological nervous systems, making them one of the promising candidates for unconventional computing. In this work, we review the fundamentals of MTJs as well as the development of MTJ-based neurons, synapses, and probabilistic-bit. In the section on neuromorphic computing, we review a variety of neural networks composed of MTJ-based neurons and synapses, including multilayer perceptrons, convolutional neural networks, recurrent neural networks, and spiking neural networks, which are the closest to the biological neural system. In the section on stochastic computing, we review the applications of MTJ-based p-bits, including Boltzmann machines, Ising machines, and Bayesian networks. Furthermore, the challenges to developing these novel technologies are briefly discussed at the end of each section. [ABSTRACT FROM AUTHOR]

Using experimental and numerical investigation, we demonstrate laser-controlled propagation and interaction of spin waves in an irregular magnetic structure in the geometry of the Mach–Zehnder interferometer. It is shown that the use of laser radiation for heating one of the interferometer arms leads to controlled interference of a spin-wave signal in the output section. The yttrium–iron garnet film heating under the action of laser radiation is measured experimentally. Using micromagnetic modeling, the evolution of the spin-wave interference pattern under the action of laser heating of one of the interferometer arm is demonstrated. The results of this study ensure a simple solution for developing tunable spin-wave interferometers for the paradigm of the magnonic logics. [ABSTRACT FROM AUTHOR]

Pattern classification architectures leveraging the physics of coupled nano-oscillators have been demonstrated as promising alternative computing approaches but lack effective learning algorithms. In this work, we propose a nano-oscillator based classification architecture where the natural frequencies of the oscillators are learned linear combinations of the inputs and define an offline learning algorithm based on gradient back-propagation. Our results show significant classification improvements over a related approach with online learning. We also compare our architecture with a standard neural network on a simple machine learning case, which suggests that our approach is economical in terms of the number of adjustable parameters. The introduced architecture is also compatible with existing nano-technologies: the architecture does not require changes in the coupling between nano-oscillators, and it is tolerant to oscillator phase noise. [ABSTRACT FROM AUTHOR]

Neuromorphic computing and engineering has been the focus of intense research efforts that have been intensified recently by the mutation of Information and Communication Technologies. In fact, new computing solutions and new hardware platforms are expected to emerge to answer to the new needs and challenges of our societies. In this revolution, lots of candidates' technologies are explored and will require leveraging of the pros and cons. In this perspective paper belonging to the special issue on neuromorphic engineering of Journal of Applied Physics, we focus on the current achievements in the field of organic electronics and the potentialities and specificities of this research field. We highlight how unique material features available through organic materials can be used to engineer useful and promising bio-inspired devices and circuits. We also discuss the opportunities that organic electronics offer for future research directions in the neuromorphic engineering field. [ABSTRACT FROM AUTHOR]

Bayesian networks (BNs) find widespread application in many real-world probabilistic problems including diagnostics, forecasting, computer vision, etc. The basic computing primitive for BNs is a stochastic bit (s-bit) generator that can control the probability of obtaining '1' in a binary bit-stream. While silicon-based complementary metal-oxide-semiconductor (CMOS) technology can be used for hardware implementation of BNs, the lack of inherent stochasticity makes it area and energy inefficient. On the other hand, memristors and spintronic devices offer inherent stochasticity but lack computing ability beyond simple vector matrix multiplication due to their two-terminal nature and rely on extensive CMOS peripherals for BN implementation, which limits area and energy efficiency. Here, we circumvent these challenges by introducing a hardware platform based on 2D memtransistors. First, we experimentally demonstrate a low-power and compact s-bit generator circuit that exploits cycle-to-cycle fluctuation in the post-programmed conductance state of 2D memtransistors. Next, the s-bit generators are monolithically integrated with 2D memtransistor-based logic gates to implement BNs. Our findings highlight the potential for 2D memtransistor-based integrated circuits for non-von Neumann computing applications. Bayesian networks are applied to resolve several types of probabilistic problems. Here, Das et al. develop a stochastic computing hardware platform using two-dimensional memtransistors for the implementation of Bayesian network with high accuracy. [ABSTRACT FROM AUTHOR]

Oscillator networks rapidly become one of the promising vehicles for energy-efficient computing due to their intrinsic parallelism of execution. The criticality property of the oscillator-based networks is regarded to be essential for performing complex tasks. There are numerous bio-inspired synaptic and structural plasticity mechanisms available, especially for spiking neural networks, which can drive the network towards the criticality. However, there is no solid connection between these self-adaption mechanisms and the task performance, and it is not clear how and why particular self-adaptation mechanisms contribute to the solution of the task, although their relation to criticality is understood. Here we propose an evolutionary approach for the structural plasticity that relies solely on the task performance and does not contain any task-independent adaptation mechanisms, which usually contribute towards the criticality of the network. As a driver for the structural plasticity, we use a direct binary search guided by the performance of the classification task that can be interpreted as an interaction of the network with the environment. Remarkably, such interaction with the environment brings the network to criticality, although this property was not a part of the objectives of the employed structural plasticity mechanism. This observation confirms a duality of criticality and task performance, and legitimizes internal activity-dependent plasticity mechanisms from the viewpoint of evolution as mechanisms contributing to the task performance, but following the dual route. Finally, we analyze the trained network against task-independent information-theoretic measures and identify the interconnection graph's entropy to be an essential ingredient for the classification task performance and network's criticality. [ABSTRACT FROM AUTHOR]

Radial vortex-based spin torque nano-oscillators (RV-STNOs) have attracted extensive attention as potential nano microwave signal generators due to their advantages over other topological states, such as their higher oscillation, higher microwave power, and lower power consumption. However, the current driving the oscillation frequency of the STNOs must be limited in a small range of adjustment, which means less data transmission channels. In this paper, a new RV-STNO system is proposed with a multiferroic nanostructure, which consists of an ultrathin magnetic multilayer and a piezoelectric layer. Phase diagrams of oscillation frequency and amplitude with respect to piezostrain and current are obtained through micromagnetic simulation. The results show that the threshold current density of −4000-ppm compressive strain-assisted RV-STNOs is reduced from 2 × 109 A/m2 to 2 × 108 A/m2, showing one order of magnitude lower than that of conventional current-driven nano-oscillators. Meanwhile, the range of oscillation frequency adjustment is significantly enhanced, and there is an increased amplitude at the low oscillation point. Moreover, a promising digital binary frequency-shift key (BFSK) and binary amplitude-shift key (BASK) modulation technique is proposed under the combined action of current pulse and piezostrain pulse. They can transmit bit signals and show good modulation characteristics with a minimal transient state. These results provide a reference for developing the next generation of spintronic nano-oscillators with a wide frequency range and low power consumption, showing potential for future wireless communication applications. [ABSTRACT FROM AUTHOR]

Oscillatory neural network (ONN)-based classification of clustered data relies on frequency synchronization to injected signals representing input data, showing a more efficient structure than a conventional deep neural network. A frequency tunable oscillator is a core component of the network, requiring energy-efficient, and areascalable characteristics for large-scale hardware implementation. From a hardware viewpoint, insulator-metal transition (IMT) device-based oscillators are attractive owing to their simple structure and low power consumption. Furthermore, by introducing non-volatile analog memory, non-volatile frequency programmability can be obtained. However, the required device characteristics of the oscillator for high performance of coupled oscillator have not been identified. In this article, we investigated the effect of device parameters of IMT oscillator with non-volatile analog memory on coupled oscillators network for classification of clustered data. We confirmed that linear conductance response with identical pulses is crucial to accurate training. In addition, considering dispersed clustered inputs, a wide synchronization window achieved by controlling the hold voltage of the IMT shows resilient classification. As an oscillator that satisfies the requirements, we evaluated the NbO2-based IMT oscillator with non-volatile Li-based electrochemical random access memory (Li-ECRAM). Finally, we demonstrated a coupled oscillator network for classifying spoken vowels, achieving an accuracy of 85%, higher than that of a ring oscillator-based system. Our results show that an NbO2-based oscillator with Li-ECRAM has the potential for an area-scalable and energy-efficient network with high performance. [ABSTRACT FROM AUTHOR]

Spin Hall nano-oscillators (SHNOs) exploiting current-driven magnetization auto-oscillation have recently received much attention because of their potential for neuromorphic computing. Widespread applications of neuromorphic devices with SHNOs require an energy-efficient method of tuning oscillation frequency over broad ranges and storing trained frequencies in SHNOs without the need for additional memory circuitry. While the voltage-driven frequency tuning of SHNOs has been demonstrated, it was volatile and limited to megahertz ranges. Here, we show that the frequency of SHNOs is controlled up to 2.1 GHz by an electric field of 1.25 MV/cm. The large frequency tuning is attributed to the voltage-controlled magnetic anisotropy (VCMA) in a perpendicularly magnetized Ta/Pt/[Co/Ni]n/Co/AlOx structure. Moreover, the non-volatile VCMA effect enables cumulative control of the frequency using repetitive voltage pulses which mimic the potentiation and depression functions of biological synapses. Our results suggest that the voltage-driven frequency tuning of SHNOs facilitates the development of energy-efficient neuromorphic devices. Spin-hall nano-oscillators are a variation on spin-torque nano-oscillators, where the spin-Hall effect is used to drive oscillations, however past examples have had limited frequency tunability. Here, Choi et al demonstrate spin-hall oscillators with wide voltage controlled frequency tunability. [ABSTRACT FROM AUTHOR]

Wave-based analog signal processing holds the promise of extremely fast, on-the-fly, power-efficient data processing, occurring as a wave propagates through an artificially engineered medium. Yet, due to the fundamentally weak non-linearities of traditional electromagnetic materials, such analog processors have been so far largely confined to simple linear projections such as image edge detection or matrix multiplications. Complex neuromorphic computing tasks, which inherently require strong non-linearities, have so far remained out-of-reach of wave-based solutions, with a few attempts that implemented non-linearities on the digital front, or used weak and inflexible non-linear sensors, restraining the learning performance. Here, we tackle this issue by demonstrating the relevance of time-Floquet physics to induce a strong non-linear entanglement between signal inputs at different frequencies, enabling a power-efficient and versatile wave platform for analog extreme deep learning involving a single, uniformly modulated dielectric layer and a scattering medium. We prove the efficiency of the method for extreme learning machines and reservoir computing to solve a range of challenging learning tasks, from forecasting chaotic time series to the simultaneous classification of distinct datasets. Our results open the way for optical wave-based machine learning with high energy efficiency, speed and scalability. Wave-based analog signal processing has been challenging for complex nonlinear operations such as data forecasting or classification. The authors propose here an analog neuromorphic platform for optical wave-based machine learning characterized by energy efficiency, speed and scalability. [ABSTRACT FROM AUTHOR]

Composed of computation units each containing a parallel, dual-gate (DG) thin-film transistor (TFT) and a memory capacitor addressed by an access TFT, a neuromorphic circuit implementing fundamental AND and OR logic functions is realized and characterized. The exceptionally low leakage current of a TFT built on indium-gallium-zinc oxide (IGZO) allows the capacitor to be used as a quasi-static memory element. Appropriate weight signals for the realization of the two logic functions are obtained using a gradient-descent training algorithm. The circuit can be configured to execute one of the two functions by the setting of the weight signals. [ABSTRACT FROM AUTHOR]

Biomimicry at the hardware level is expected to overcome at least some of the challenges, including high power consumption, large footprint, two‐dimensionality, and limited functionality, which arise as the field of artificial intelligence matures. One of the main attributes that allow biological systems to thrive is the successful interpretation of and response to environmental signals. Taking inspiration from these systems, the first demonstration of using multiple environmental inputs to trigger the formation and control the growth of an evolvable synaptic transistor is reported here. The resulting transistor exhibits long‐term changes in the channel conductance at a fixed gate voltage. Biomimetic logic circuits are investigated based on this evolvable transistor that implement temperature and pressure inputs to achieve higher order processes like self‐regulation of synaptic strength and coincidence detection. [ABSTRACT FROM AUTHOR]

The modern era of digital computing systems is deeply embedded around memory devices, which store an ocean of information accessible by an electronic gadget. Such electronic circuitry can be mimicked by molecules that became known to the scientific community over the last two decades. However, their mechanistic studies, fast switching events, reproducibility, and scaling-up approaches are crucial points for such molecular devices to have practical applications. Among the well-explored systems, redox-based resistive switching memories (RRAM) are the most prosperous research domain, expressly for technological augmentation. In situ stabilization of electrical charges on the molecule for the subsequent reduction of the frontier molecular orbital's energy gap, can be an excellent method to overcome the slow switching speed of prototype devices. The recent extension of resistive memory to memristors has also offered enormous charge storage capacity and processing ability to overcome Moore's limitation law and the Von Neumann bottleneck issue. Thereby, the central objective of the present review is to showcase the recent progress in the fabrication of devices incorporating inorganic, organic, and redox-active molecules, either in self-assembled monolayers or many molecules assembled between two electrical contacts followed by stimuli-responsive current–voltage (I–V) analysis. Solid-state molecular memory devices as a 'proof-of-concept' might be prolific in future applications. [ABSTRACT FROM AUTHOR]

Neuromorphic computing using spintronic devices, such as spin-torque oscillators (STOs), has been intensively studied for energy-efficient data processing. One of the critical issues in this application is stochasticity in magnetization dynamics, which limits the accuracy of computation. Such stochastic behavior, however, plays a key role in stochastic computing and machine learning. It is therefore important to develop methods for both suppressing and enhancing stochastic response in spintronic devices. We report on experimental investigations on control of stochastic quantity, such as the width of a distribution of transient time in magnetization dynamics in vortex-type STO. The spin-transfer effect can suppress stochasticity in transient dynamics from a non-oscillating to oscillating state, whereas an application of a radio-frequency magnetic field is effective in reducing stochasticity on the time evolution of the oscillating state. [ABSTRACT FROM AUTHOR]

Chemical mixtures can be leveraged to store large amounts of data in a highly compact form and have the potential for massive scalability owing to the use of large-scale molecular libraries. With the parallelism that comes from having many species available, chemical-based memory can also provide the physical substrate for computation with increased throughput. Here, we represent non-binary matrices in chemical solutions and perform multiple matrix multiplications and additions, in parallel, using chemical reactions. As a case study, we demonstrate image processing, in which small greyscale images are encoded in chemical mixtures and kernel-based convolutions are performed using phenol acetylation reactions. In these experiments, we use the measured concentrations of reaction products (phenyl acetates) to reconstruct the output image. In addition, we establish the chemical criteria required to realize chemical image processing and validate reaction-based multiplication. Most importantly, this work shows that fundamental arithmetic operations can be reliably carried out with chemical reactions. Our approach could serve as a basis for developing more advanced chemical computing architectures. [ABSTRACT FROM AUTHOR]

In-memory computing based on non-volatile resistive memory can significantly improve the energy efficiency of artificial neural networks. However, accurate in situ training has been challenging due to the nonlinear and stochastic switching of the resistive memory elements. One promising analog memory is the electrochemical random-access memory (ECRAM), also known as the redox transistor. Its low write currents and linear switching properties across hundreds of analog states enable accurate and massively parallel updates of a full crossbar array, which yield rapid and energy-efficient training. While simulations predict that ECRAM based neural networks achieve high training accuracy at significantly higher energy efficiency than digital implementations, these predictions have not been experimentally achieved. In this work, we train a 3 × 3 array of ECRAM devices that learns to discriminate several elementary logic gates (AND, OR, NAND). We record the evolution of the network's synaptic weights during parallel in situ (on-line) training, with outer product updates. Due to linear and reproducible device switching characteristics, our crossbar simulations not only accurately simulate the epochs to convergence, but also quantitatively capture the evolution of weights in individual devices. The implementation of the first in situ parallel training together with strong agreement with simulation results provides a significant advance toward developing ECRAM into larger crossbar arrays for artificial neural network accelerators, which could enable orders of magnitude improvements in energy efficiency of deep neural networks. [ABSTRACT FROM AUTHOR]

Spintronic technology is emerging as a direction for the hardware implementation of neurons and synapses of neuromorphic architectures. In particular, a single spintronic device can be used to implement the nonlinear activation function of neurons. Here, we present how to implement spintronic neurons with sigmoidal and rectified linear unit (ReLU)-like activation functions. We then perform a numerical experiment showing the reliability of neural networks made by spintronic neurons, all having different activation functions to emulate device-to-device variations in a possible hardware implementation of the network. Therefore, we consider a “vanilla'' neural network implemented to recognize the categories of the Mixed National Institute of Standards and Technology database, and we show an average accuracy of 98.87% in the test dataset, which is very close to 98.89% as obtained for the ideal case (all neurons have the same sigmoid activation function). Similar results are obtained with neurons having a ReLU-like activation function. [ABSTRACT FROM AUTHOR]