Synchronization dynamics on the picosecond time scale in coupled Josephson junction neurons

Characteristics analysis of a single electromechanical arm driven by a functional neural circuit

From a biological viewpoint, the muscle tissue produces efficient gait behavior that can be adjusted by neural signals. From the physical viewpoint, the limb movement can be simulated by applying a neural circuit to control the artificial electromechanical arm (EA). In this paper, a functional neural circuit is used to excite a single EA, the load circuit attached to the moving beam is driven by a neural circuit, and the Ampere force is activated by the load circuit to control the artificial EA. The dynamic equations of the neural circuit are derived using Kirchhoff's theorem, while the energy and motion equations of the beam are computed through the application of mechanics and related theoretical principles. Furthermore, the dynamic characteristics of the functional neural circuit forced EA are analyzed. The results indicate that the beam movement can be controlled by the electrical activity of this functional neural circuit. This work will provide theoretical guidance to build the electromechanical device for complex gait behaviors.

A self-training spiking superconducting neuromorphic architecture

Neuromorphic computing takes biological inspiration to the device level aiming to improve computational efficiency and capabilities. One of the major issues that arises is the training of neuromorphic hardware systems. Typically training algorithms require global information and are thus inefficient to implement directly in hardware. In this paper we describe a set of reinforcement learning based, local weight update rules and their implementation in superconducting hardware. Using SPICE circuit simulations, we implement a small-scale neural network with a learning time of order one nanosecond per update. This network can be trained to learn new functions simply by changing the target output for a given set of inputs, without the need for any external adjustments to the network. Further, this architecture does not require programing explicit weight values in the network, alleviating a critical challenge with analog hardware implementations of neural networks.

Optimisation Challenge for a Superconducting Adiabatic Neural Network That Implements XOR and OR Boolean Functions

In this article, we consider designs of simple analog artificial neural networks based on adiabatic Josephson cells with a sigmoid activation function. A new approach based on the gradient descent method is developed to adjust the circuit parameters, allowing efficient signal transmission between the network layers. The proposed solution is demonstrated on the example of a system that implements XOR and OR logical operations.

Magnetic Flux Sensor Based on Spiking Neurons with Josephson Junctions

Josephson junctions (JJs) are superconductor-based devices used to build highly sensitive magnetic flux sensors called superconducting quantum interference devices (SQUIDs). These sensors may vary in design, being the radio frequency (RF) SQUID, direct current (DC) SQUID, and hybrid, such as D-SQUID. In addition, recently many of JJ's applications were found in spiking models of neurons exhibiting nearly biological behavior. In this study, we propose and investigate a new circuit model of a sensory neuron based on DC SQUID as part of the circuit. The dependence of the dynamics of the designed model on the external magnetic flux is demonstrated. The design of the circuit and derivation of the corresponding differential equations that describe the dynamics of the system are given. Numerical simulation is used for experimental evaluation. The experimental results confirm the applicability and good performance of the proposed magnetic-flux-sensitive neuron concept: the considered device can encode the magnetic flux in the form of neuronal dynamics with the linear section. Furthermore, some complex behavior was discovered in the model, namely the intermittent chaotic spiking and plateau bursting. The proposed design can be efficiently applied to developing the interfaces between circuitry and spiking neural networks. However, it should be noted that the proposed neuron design shares the main limitation of all the superconductor-based technologies, i.e., the need for a cryogenic and shielding system.

Reproduced neuron-like excitability and bursting synchronization of memristive Josephson junctions loaded inductor

Employing electronic component including memristor and Josephson junction to mimic biological neuron or synapse has elicited intense research in recent years. Neurons described by nonlinear oscillators can exhibit complex electrical activities. Josephson junctions are excellent candidates for neuron-inspired components because of their physical properties with low energy costs and high efficiency. In this paper, we revisit a prior work on memristive Josephson junction (MJJ) to identify the dynamical mechanisms to mimic neuron-like excitability and spiking. The inductive memristive Josephson junction (L-MJJ) model is further developed by adding an inductor with internal resistor. It is found that the L-MJJ model can reproduce square-wave bursting of the classical neuronal model from the neurodynamics point of view. The coupling L-MJJ oscillators can achieve in-phase and antiphase bursting synchronization similar with nonlinear coupling neurons. From the framework of nonlinear dynamics theory, this work aspires to build effective bridge between superconducting physics and theoretical neuroscience. Obtained results confirm the potential feasibility of this junction in designing a neuron-inspired computation to explore dynamics of larger-scale neuromorphic network.

Reconfigurable Neuromorphic Computing: Materials, Devices, and Integration

Neuromorphic computing has been attracting ever-increasing attention due to superior energy efficiency, with great promise to promote the next wave of artificial general intelligence in the post-Moore era. Current approaches are, however, broadly designed for stationary and unitary assignments, thus encountering reluctant interconnections, power consumption, and data-intensive computing in that domain. Reconfigurable neuromorphic computing, an on-demand paradigm inspired by the inherent programmability of brain, can maximally reallocate finite resources to perform the proliferation of reproducibly brain-inspired functions, highlighting a disruptive framework for bridging the gap between different primitives. Although relevant research has flourished in diverse materials and devices with novel mechanisms and architectures, a precise overview remains blank and urgently desirable. Herein, the recent strides along this pursuit are systematically reviewed from material, device, and integration perspectives. At the material and device level, one comprehensively conclude the dominant mechanisms for reconfigurability, categorized into ion migration, carrier migration, phase transition, spintronics, and photonics. Integration-level developments for reconfigurable neuromorphic computing are also exhibited. Finally, a perspective on the future challenges for reconfigurable neuromorphic computing is discussed, definitely expanding its horizon for scientific communities.

Dynamical properties of neuromorphic Josephson junctions

Neuromorphic computing exploits the dynamical analogy between many physical systems and neuron biophysics. Superconductor systems, in particular, are excellent candidates for neuromorphic devices due to their capacity to operate at great speeds and with low energy dissipation compared to their silicon counterparts. In this paper, we revisit a prior work on Josephson Junction-based neurons to identify the exact dynamical mechanisms underlying the system's neuronlike properties and reveal complex behaviors which are relevant for neurocomputation and the design of superconducting neuromorphic devices. Our paper lies at the intersection of superconducting physics and theoretical neuroscience, both viewed under a common framework-that of nonlinear dynamics theory.

A superconducting adiabatic neuron in a quantum regime

We explore the dynamics of an adiabatic neural cell of a perceptron artificial neural network in a quantum regime. This mode of cell operation is assumed for a hybrid system of a classical neural network whose configuration is dynamically adjusted by a quantum co-processor. Analytical and numerical studies take into account non-adiabatic processes as well as dissipation, which leads to smoothing of quantum coherent oscillations. The obtained results indicate the conditions under which the neuron possesses the required sigmoid activation function.

Tunable superconducting neurons for networks based on radial basis functions

The hardware implementation of signal microprocessors based on superconducting technologies seems relevant for a number of niche tasks where performance and energy efficiency are critically important. In this paper, we consider the basic elements for superconducting neural networks on radial basis functions. We examine the static and dynamic activation functions of the proposed neuron. Special attention is paid to tuning the activation functions to a Gaussian form with relatively large amplitude. For the practical implementation of the required tunability, we proposed and investigated heterostructures designed for the implementation of adjustable inductors that consist of superconducting, ferromagnetic, and normal layers.

Superconducting Bio-Inspired Au-Nanowire-Based Neurons

High-performance modeling of neurophysiological processes is an urgent task that requires new approaches to information processing. In this context, two- and three-junction superconducting quantum interferometers with Josephson weak links based on gold nanowires are fabricated and investigated experimentally. The studied cells are proposed for the implementation of bio-inspired neurons—high-performance, energy-efficient, and compact elements of neuromorphic processor. The operation modes of an advanced artificial neuron capable of generating the burst firing activation patterns are explored theoretically. A comparison with the Izhikevich mathematical model of biological neurons is carried out.

Applications and Techniques for Fast Machine Learning in Science

In this community review report, we discuss applications and techniques for fast machine learning (ML) in science-the concept of integrating powerful ML methods into the real-time experimental data processing loop to accelerate scientific discovery. The material for the report builds on two workshops held by the Fast ML for Science community and covers three main areas: applications for fast ML across a number of scientific domains; techniques for training and implementing performant and resource-efficient ML algorithms; and computing architectures, platforms, and technologies for deploying these algorithms. We also present overlapping challenges across the multiple scientific domains where common solutions can be found. This community report is intended to give plenty of examples and inspiration for scientific discovery through integrated and accelerated ML solutions. This is followed by a high-level overview and organization of technical advances, including an abundance of pointers to source material, which can enable these breakthroughs.

Low-temperature emergent neuromorphic networks with correlated oxide devices

Neuromorphic computing-which aims to mimic the collective and emergent behavior of the brain's neurons, synapses, axons, and dendrites-offers an intriguing, potentially disruptive solution to society's ever-growing computational needs. Although much progress has been made in designing circuit elements that mimic the behavior of neurons and synapses, challenges remain in designing networks of elements that feature a collective response behavior. We present simulations of networks of circuits and devices based on superconducting and Mott-insulating oxides that display a multiplicity of emergent states that depend on the spatial configuration of the network. Our proposed network designs are based on experimentally known ways of tuning the properties of these oxides using light ions. We show how neuronal and synaptic behavior can be achieved with arrays of superconducting Josephson junction loops, all within the same device. We also show how a multiplicity of synaptic states could be achieved by designing arrays of devices based on hydrogenated rare earth nickelates. Together, our results demonstrate a research platform that utilizes the collective macroscopic properties of quantum materials to mimic the emergent behavior found in biological systems.

Non-volatile artificial synapse based on a vortex nano-oscillator

In this work, a new mechanism to combine a non-volatile behaviour with the spin diode detection of a vortex-based spin torque nano-oscillator (STVO) is presented. Experimentally, it is observed that the spin diode response of the oscillator depends on the vortex chirality. Consequently, fixing the frequency of the incoming signal and switching the vortex chirality results in a different rectified voltage. In this way, the chirality can be deterministically controlled via the application of electrical signals injected locally in the device, resulting in a non-volatile control of the output voltage for a given input frequency. Micromagnetic simulations corroborate the experimental results and show the main contribution of the Oersted field created by the input RF current density in defining two distinct spin diode detections for different chiralities. By using two non-identical STVOs, we show how these devices can be used as programmable non-volatile synapses in artificial neural networks.

Roadmap on emerging hardware and technology for machine learning

Recent progress in artificial intelligence is largely attributed to the rapid development of machine learning, especially in the algorithm and neural network models. However, it is the performance of the hardware, in particular the energy efficiency of a computing system that sets the fundamental limit of the capability of machine learning. Data-centric computing requires a revolution in hardware systems, since traditional digital computers based on transistors and the von Neumann architecture were not purposely designed for neuromorphic computing. A hardware platform based on emerging devices and new architecture is the hope for future computing with dramatically improved throughput and energy efficiency. Building such a system, nevertheless, faces a number of challenges, ranging from materials selection, device optimization, circuit fabrication and system integration, to name a few. The aim of this Roadmap is to present a snapshot of emerging hardware technologies that are potentially beneficial for machine learning, providing the Nanotechnology readers with a perspective of challenges and opportunities in this burgeoning field.

Synaptic weighting in single flux quantum neuromorphic computing

Josephson junctions act as a natural spiking neuron-like device for neuromorphic computing. By leveraging the advances recently demonstrated in digital single flux quantum (SFQ) circuits and using recently demonstrated magnetic Josephson junction (MJJ) synaptic circuits, there is potential to make rapid progress in SFQ-based neuromorphic computing. Here we demonstrate the basic functionality of a synaptic circuit design that takes advantage of the adjustable critical current demonstrated in MJJs and implement a synaptic weighting element. The devices were fabricated with a restively shunted Nb/AlO_x-Al/Nb process that did not include MJJs. Instead, the MJJ functionality was tested by making multiple circuits and varying the critical current, but not the external shunt resistance, of the oxide Josephson junction that represents the MJJ. Experimental measurements and simulations of the fabricated circuits are in good agreement.

Two-dimensional mutually synchronized spin Hall nano-oscillator arrays for neuromorphic computing

In spin Hall nano-oscillators (SHNOs), pure spin currents drive local regions of magnetic films and nanostructures into auto-oscillating precession. If such regions are placed in close proximity to each other they can interact and may mutually synchronize. Here, we demonstrate robust mutual synchronization of two-dimensional SHNO arrays ranging from 2 × 2 to 8 × 8 nano-constrictions, observed both electrically and using micro-Brillouin light scattering microscopy. On short time scales, where the auto-oscillation linewidth [Formula: see text] is governed by white noise, the signal quality factor, [Formula: see text], increases linearly with the number of mutually synchronized nano-constrictions (N), reaching 170,000 in the largest arrays. We also show that SHNO arrays exposed to two independently tuned microwave frequencies exhibit the same synchronization maps as can be used for neuromorphic vowel recognition. Our demonstrations may hence enable the use of SHNO arrays in two-dimensional oscillator networks for high-quality microwave signal generation and ultra-fast neuromorphic computing.

Between phase and amplitude oscillators

We analyze an intermediate collective regime where amplitude oscillators distribute themselves along a closed, smooth, time-dependent curve C, thereby maintaining the typical ordering of (identical) phase oscillators. This is achieved by developing a general formalism based on two partial differential equations, which describe the evolution of the probability density along C and of the shape of C itself. The formalism is specifically developed for Stuart-Landau oscillators, but it is general enough to apply to other classes of amplitude oscillators. The main achievements consist in (i) identification and characterization of a transition to self-consistent partial synchrony (SCPS), which confirms the crucial role played by higher Fourier harmonics in the coupling function; (ii) an analytical treatment of SCPS, including a detailed stability analysis; and (iii) the discovery of a different form of collective chaos, which can be seen as a generalization of SCPS and characterized by a multifractal probability density. Finally, we are able to describe given dynamical regimes at both the macroscopic and the microscopic level, thereby shedding additional light on the relationship between the two different levels of description.

Ultralow power artificial synapses using nanotextured magnetic Josephson junctions

Neuromorphic computing promises to markedly improve the efficiency of certain computational tasks, such as perception and decision-making. Although software and specialized hardware implementations of neural networks have made tremendous accomplishments, both implementations are still many orders of magnitude less energy efficient than the human brain. We demonstrate a new form of artificial synapse based on dynamically reconfigurable superconducting Josephson junctions with magnetic nanoclusters in the barrier. The spiking energy per pulse varies with the magnetic configuration, but in our demonstration devices, the spiking energy is always less than 1 aJ. This compares very favorably with the roughly 10 fJ per synaptic event in the human brain. Each artificial synapse is composed of a Si barrier containing Mn nanoclusters with superconducting Nb electrodes. The critical current of each synapse junction, which is analogous to the synaptic weight, can be tuned using input voltage spikes that change the spin alignment of Mn nanoclusters. We demonstrate synaptic weight training with electrical pulses as small as 3 aJ. Further, the Josephson plasma frequencies of the devices, which determine the dynamical time scales, all exceed 100 GHz. These new artificial synapses provide a significant step toward a neuromorphic platform that is faster, more energy-efficient, and thus can attain far greater complexity than has been demonstrated with other technologies.

Neuromorphic computing with nanoscale spintronic oscillators

Neurons in the brain behave as nonlinear oscillators, which develop rhythmic activity and interact to process information. Taking inspiration from this behaviour to realize high-density, low-power neuromorphic computing will require very large numbers of nanoscale nonlinear oscillators. A simple estimation indicates that to fit 10⁸ oscillators organized in a two-dimensional array inside a chip the size of a thumb, the lateral dimension of each oscillator must be smaller than one micrometre. However, nanoscale devices tend to be noisy and to lack the stability that is required to process data in a reliable way. For this reason, despite multiple theoretical proposals and several candidates, including memristive and superconducting oscillators, a proof of concept of neuromorphic computing using nanoscale oscillators has yet to be demonstrated. Here we show experimentally that a nanoscale spintronic oscillator (a magnetic tunnel junction) can be used to achieve spoken-digit recognition with an accuracy similar to that of state-of-the-art neural networks. We also determine the regime of magnetization dynamics that leads to the greatest performance. These results, combined with the ability of the spintronic oscillators to interact with each other, and their long lifetime and low energy consumption, open up a path to fast, parallel, on-chip computation based on networks of oscillators.