2022 roadmap on neuromorphic computing and engineering

Modern computation based on the von Neumann architecture is today a mature cutting-edge science. In the Von Neumann architecture, processing and memory units are implemented as separate blocks interchanging data intensively and continuously. This data transfer is responsible for a large part of the power consumption. The next generation computer technology is expected to solve problems at the exascale with 1018 calculations each second. Even though these future computers will be incredibly powerful, if they are based on von Neumann type architectures, they will consume between 20 and 30 megawatts of power and will not have intrinsic physically built-in capabilities to learn or deal with complex data as our brain does. These needs can be addressed by neuromorphic computing systems which are inspired by the biological concepts of the human brain. This new generation of computers has the potential to be used for the storage and processing of large amounts of digital information with much lower power consumption than conventional processors. Among their potential future applications, an important niche is moving the control from data centers to edge devices. The aim of this Roadmap is to present a snapshot of the present state of neuromorphic technology and provide an opinion on the challenges and opportunities that the future holds in the major areas of neuromorphic technology, namely materials, devices, neuromorphic circuits, neuromorphic algorithms, applications, and ethics. The Roadmap is a collection of perspectives where leading researchers in the neuromorphic community provide their own view about the current state and the future challenges for each research area. We hope that this Roadmap will be a useful resource by providing a concise yet comprehensive introduction to readers outside this field, for those who are just entering the field, as well as providing future perspectives for those who are well established in the neuromorphic computing community.

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.

Substrate-dependent differences in ferroelectric behavior and phase diagram of Si-doped hafnium oxide

Non-volatile memories based on ferroelectric hafnium oxide, especially the ferroelectric field-effect transistor (FeFET), have outstanding properties, e.g. for the application in neuromorphic circuits. However, material development has focused so far mainly on metal–ferroelectric–metal (MFM) capacitors, while FeFETs are based on metal–ferroelectric–insulator–semiconductor (MFIS) capacitors. Here, the influence of the interface properties, annealing temperature and Si-doping content are investigated. Antiferroelectric-like behavior is strongly suppressed with a thicker interface layer and high annealing temperature. In addition, high-k interface dielectrics allow for thicker interface layers without retention penalty. Moreover, the process window for ferroelectric behavior is much larger in MFIS capacitors compared to MFM-based films. This does not only highlight the substrate dependence of ferroelectric hafnium oxide films, but also gives evidence that the phase diagram of ferroelectric hafnium oxide is defined by the mechanical stress. Graphic Abstract

A Comprehensive Survey on Nanophotonics Neural Networks

: In the last years, materializations of neuromorphic circuits based on nanophotonic arrangements have been proposed, which contain complete optical circuits, laser, photodetectors, photonic crystals, optical fibers, flat waveguides, and other passive optical elements of nanostructured materials, which eliminate the time of simultaneous processing of big groups of data, taking advantage of the quantum perspective and thus highly increasing the potentials of contemporary intelligent computational systems. This article is an effort to record and study the research that has been conducted concerning the methods of development and materialization of neuromorphic circuits of Neural Networks of nanophotonic arrangements. In particular, an investigative study of the methods of developing nanophotonic neuromorphic processors, their originality in neuronic architectural structure, their training methods and their optimization has been realized along with the study of special issues such as optical activation functions and cost functions.

Analog Gaussian Function Circuit: Architectures, Operating Principles and Applications

This review paper explores existing architectures, operating principles, performance metrics and applications of analog Gaussian function circuits. Architectures based on the translinear principle, the bulk-controlled approach, the floating gate approach, the use of multiple differential pairs, compositions of different fundamental blocks and others are considered. Applications involving analog implementations of Machine Learning algorithms, neuromorphic circuits, smart sensor systems and fuzzy/neuro-fuzzy systems are discussed, focusing on the role of the Gaussian function circuit. Finally, a general discussion and concluding remarks are provided.

Electrospun bundled carbon nanofibers for skin-inspired tactile sensing, proprioception and gesture tracking applications

In this work, we report a class of wearable, stitchable, and sensitive carbon nanofiber (CNF)-polydimethylsiloxane (PDMS) composite-based piezoresistive sensors realized by carbonizing electrospun polyacrylonitrile (PAN) nanofibers and subsequently embedding in PDMS elastomeric thin films. Electro-mechanical tactile sensing characterization of the resulting piezoresistive strain sensors revealed a linear response with an average force sensitivity of ~1.82 kN−1 for normal forces up to 20 N. The real-time functionality of the CNF-PDMS composite sensors in wearable body sensor networks and advanced bionic skin applications was demonstrated through human motion and gesture monitoring experiments. A skin-inspired artificial soft sensor capable of demonstrating proprioceptive and tactile sensory perception utilizing CNF bundles has been shown. Furthermore, a 16-point pressure-sensitive flexible sensor array mimicking slow adapting low threshold mechanoreceptors of glabrous skin was demonstrated. Such devices in tandem with neuromorphic circuits can potentially recreate the sense of touch in robotic arms and restore somatosensory perception in amputees.

A robust model of Stimulus-Specific Adaptation validated on neuromorphic hardware

Stimulus-Specific Adaptation (SSA) to repetitive stimulation is a phenomenon that has been observed across many different species and in several brain sensory areas. It has been proposed as a computational mechanism, responsible for separating behaviorally relevant information from the continuous stream of sensory information. Although SSA can be induced and measured reliably in a wide variety of conditions, the network details and intracellular mechanisms giving rise to SSA still remain unclear. Recent computational studies proposed that SSA could be associated with a fast and synchronous neuronal firing phenomenon called Population Spikes (PS). Here, we test this hypothesis using a mean-field rate model and corroborate it using a neuromorphic hardware. As the neuromorphic circuits used in this study operate in real-time with biologically realistic time constants, they can reproduce the same dynamics observed in biological systems, together with the exploration of different connectivity schemes, with complete control of the system parameter settings. Besides, the hardware permits the iteration of multiple experiments over many trials, for extended amounts of time and without losing the networks and individual neural processes being studied. Following this “neuromorphic engineering” approach, we therefore study the PS hypothesis in a biophysically inspired recurrent networks of spiking neurons and evaluate the role of different linear and non-linear dynamic computational primitives such as spike-frequency adaptation or short-term depression (STD). We compare both the theoretical mean-field model of SSA and PS to previously obtained experimental results in the area of novelty detection and observe its behavior on its neuromorphic physical equivalent model. We show how the approach proposed can be extended to other computational neuroscience modelling efforts for understanding high-level phenomena in mechanistic models.

Lévy Noise-induced Self-induced Stochastic Resonance in a Memristive Neuron

Self-induced stochastic resonance (SISR) is a subtle resonance mechanism requiring a nontrivial scaling limit between the stochastic and the deterministic timescales of an excitable system, leading to the emergence of a limit cycle behavior which is absent without noise. All previous studies on SISR in neural systems have only considered the idealized Gaussian white noise. Moreover, these studies have ignored one electrophysiological aspect of the nerve cell: its memristive properties. In this paper, first, we show that in the excitable regime, the asymptotic matching of the mean escape timescale of an α-stable Lévy process (with value increasing as a power σ-α of the noise amplitude σ, unlike the mean escape timescale of a Gaussian process with the value increasing as in Kramers' law) and the deterministic timescale (controlled by the singular parameter) can also induce a strong SISR. In addition, it is shown that the degree of SISR induced by Lévy noise is not always higher than that of Gaussian noise. Second, we show that, for both types of noises, the two memristive properties of the neuron have opposite effects on the degree of SISR: the stronger the feedback gain parameter that controls the modulation of the membrane potential with the magnetic flux and the weaker the feedback gain parameter that controls the saturation of the magnetic flux, the higher the degree of SISR. Finally, we show that, for both types of noises, the degree of SISR in the memristive neuron is always higher than in the non-memristive neuron. Our results could find applications in designing neuromorphic circuits operating in noisy regimes.

Reliability-Aware Design of Spike-Event Neuromorphic Circuits

Very Large Scale Integration (VLSI) based neuromorphic circuits also known as Silicon Neurons (SiNs) emulate the electrophysiological behavior of biological neurons. With the advancement in technology, neuromorphic systems also lead to various reliability issues and hence making their study important. Bias Temperature Instability (BTI) and Hot Carrier Injection (HCI) are the two major reliability issues present in VLSI circuits. In this work, we have investigated the combined effect of BTI and HCI on the two types of integrate-and-fire based SiNs namely (a) Axon-Hillock and (b) Simplified Leaky integrate-and-fire circuits using their key performance parameters. Novel reliability-aware AH and SLIF circuits are proposed to mitigate the reliability issues. Proposed reliability-aware designs show negligible deviation in performance parameters after aging. The time-zero process variability analysis is also carried out for proposed reliability-aware SiNs. The power consumption of existing and proposed reliability-aware neuron circuits is analyzed and compared.<br>