Artificial Neuron Research Articles

A predictive model for the estimation of industrial [formula omitted] emissions for IoT-based devices

The paper is devoted to solving the problem, associated with increasing the forecasting accuracy of pollutant concentration level based on PM2.5 dust. The LANN model (LANN – Lagrange powered Artificial Neuron Network), proposed in the paper, allows you to take into account the mutual influence of pollution sources on the concentration level at monitoring points, which, on the one hand, allows to manage the volume of emissions from industrial enterprises in such a way as to avoid penalties and, on the other hand, allows the regulatory agencies to determine points for environmental control, at which permissible standards will be most likely exceeded. The suggested solution makes it possible to use low-cost IoT-based air quality monitoring devices, which were primarily considered not applicable due to their low accuracy and exposure to the influence of short-term stochastic factors; besides, no models were available, which would be capable to provide the necessary accuracy and forecast horizon on the base of IoT data. At the same time, the obtained model has low requirements for computing resources compared to dispersion models and is dependent on the accuracy of weather information; besides, the model allows you to take into account trends and obtain better forecast values in terms of accuracy than the models, that use regression methods, neural network models and ML models, stochastic models and their ensembles. The application of the model improves well-known pollutant dispersion estimation calculation techniques through the use of measurement data flow and dynamic adaptation to changing conditions The model was trained on the dataset over the period 2022-2023 collected from 5 monitoring devices, which were installed in the catchment area of the group, incl. 13 industrial points of PM2.5 pollutant sources.

Programmable Optical Synaptic Linking of Neuromorphic Photonic-Electronic RTD Spiking Circuits.

Interconnectivity between functional building blocks (such as neurons and synapses) represents a fundamental functionality for realizing neuromorphic systems. However, in the domain of neuromorphic photonics, synaptic interlinking and cascadability of spiking optical artificial neurons remains challenging and mostly unexplored in experiments. In this work, we report an optical synaptic link between optoelectronic spiking artificial neurons based upon resonant tunneling diodes (RTDs) that allows for cascadable spike propagation. First, deterministic spiking is triggered using multimodal (electrical and optical) inputs in RTD-based spiking artificial neurons, which are optoelectronic (OE) circuits incorporating either micron-scale RTDs or photosensitive nanopillar-based RTDs. Second, feedforward linking with dynamical weighting of optical spiking signals between pre- and postsynaptic RTD artificial neurons is demonstrated, including cascaded spike activation. By dynamically weighting the amplitude of optical spikes, it is shown how the cascaded spike activation probability in the postsynaptic RTD node directly follows the amplitude of the weighted optical spikes. This work therefore provides the first experimental demonstration of programmable synaptic optical link and spike cascading between multiple fast and efficient RTD OE spiking artificial neurons, therefore providing a key functionality for photonic-electronic spiking neural networks and light-enabled neuromorphic hardware.

НЕЙРОМЕРЕЖЕВІ МАТЕМАТИЧНІ МОДЕЛІ ЛЮДИНОМАШИННОГО СПІЛКУВАННЯ ТА РОЗПІЗНАВАННЯ

The study is devoted to the neural network interpretation of the task of human-machine communication and recognition by multiple criteria, considered as a task of assignment. The main goal is to reduce this problem to a standard form, where the number of criteria groups is equal to the number of ranked documents. The study defined the architecture of a neural network and proposed the use of a network of binary neurons, which is a matrix of a certain dimension. The proposed ranking model is based on a neural network that contains arbitrary feedback. This allows the excitation to be transmitted back to the neuron, which contributes to the repeated performance of its function. However, in dynamic neural networks instability occurs, which is manifested in a random change in the states of neurons without reaching stationary states. The question of stability of the dynamics of such systems remains open. The considered discrete Hopfield neural network has the following characteristics: one layer of elements, each element is connected to all others, but not to itself; only one element is updated per stage; elements are updated in random order, but each is updated with the same frequency; the output function is binary (value "0" or "1"). A Hopfield neural network is recurrent: the output of the network is reused as input until a steady state is reached. After starting, the neural network changes its state, gradually moving to a stable mode, which allows identifying a plan for evaluating the process of human-machine communication according to a set of criteria. Random search procedures are used to refine the results. The proposed energy function is minimized to ensure that the constraints are met and the problem is solved. The constructed function reaches a minimum only in the states corresponding to the assignment plans. The definition of the network parameters is carried out by comparing the obtained functions with the energy function in general form. The practical implementation of the model demonstrated that Hopfield's neural network can be successfully applied to document ranking in human-machine communication and recognition systems, providing high accuracy and efficiency in solving ranking problems.

Pattern recognition using spiking antiferromagnetic neurons

Spintronic devices offer a promising avenue for the development of nanoscale, energy-efficient artificial neurons for neuromorphic computing. It has previously been shown that with antiferromagnetic (AFM) oscillators, ultra-fast spiking artificial neurons can be made that mimic many unique features of biological neurons. In this work, we train an artificial neural network of AFM neurons to perform pattern recognition. A simple machine learning algorithm called spike pattern association neuron (SPAN), which relies on the temporal position of neuron spikes, is used during training. In under a microsecond of physical time, the AFM neural network is trained to recognize symbols composed from a grid by producing a spike within a specified time window. We further achieve multi-symbol recognition with the addition of an output layer to suppress undesirable spikes. Through the utilization of AFM neurons and the SPAN algorithm, we create a neural network capable of high-accuracy recognition with overall power consumption on the order of picojoules.

Stress Engineering of Magnetization Fluctuation and Noise Spectra in Low-Barrier Nanomagnets Used as Analog and Binary Stochastic Neurons.

A single-domain nanomagnet, shaped like a thin elliptical disk with small eccentricity, has a double-well potential profile with two degenerate energy minima separated by a small barrier of a few kT (k = Boltzmann constant and T = absolute temperature). The two minima correspond to the magnetization pointing along the two mutually anti-parallel directions along the major axis of the ellipse. At room temperature, the magnetization fluctuates randomly between the two minima, mimicking telegraph noise. This makes the nanomagnet act as a "binary" stochastic neuron (BSN) with the neuronal state encoded in the magnetization orientation. If the nanomagnet is magnetostrictive, then the barrier can be depressed further by applying (electrically generated) uniaxial stress along the ellipse's major axis, thereby gradually eroding the double-well shape. When the barrier almost vanishes, the magnetization begins to randomly assume any arbitrary orientation (not just along the major axis), making the nanomagnet act as an "analog" stochastic neuron (ASN). The magnetization fluctuation then begins to increasingly resemble white noise. The full width at half maximum (FWHM) of the noise auto-correlation function decreases with increasing stress, as the fluctuation gradually transforms from telegraph noise to white noise. Consistent with this trend, the noise spectral density exhibits a 1/fβ spectrum (at high frequencies) with β decreasing from 2.00 to 1.88 with increasing stress. Stress can thus not only reconfigure a BSN to an ASN, which has its own applications, but it can also perform "noise engineering", i.e., tune the auto-correlation function and power spectral density, having applications in signal processing.

An energy efficient leaky integrate and fire neuron using Ge-source TFET for spiking neural network: simulation analysis

The basic building block of neural network is a device, which can mimic the neural behavior. The spiking neural network (SNN) is an efficient methodology in terms of power and area. Due to the excess energy consumption and larger area, various spintronic neural devices are unfit for neuron applications. In this article, we have implemented Ge source based Tunnel FET (TFET) for ultralow energy spike generation using TCAD simulator. It is seen that Ge source TFET has signature spiking frequency in THz range versus input voltage curve of an artificial biological neuron. The simulated device deploy the leaky integrate and fire (LIF) technique for generation of neurons. The simulation result highlights that the energy of device is 1.08 aJ/spike, which is several order less than existing neural based FET devices in literature.

Proton-gated organic thin-film transistors for leaky integrate-and-fire convolutional spiking neural networks

Artificial spiking neurons, integral to the functionality of spiking neural networks, are designed to mimic the information transmission via discrete spikes in biological nervous systems. Traditional approaches that necessitate the charging of capacitors and the inclusion of discharge circuits for neuron membrane potential integration and leakage, present challenges in terms of cost and space efficiency. To overcome the challenges, this work proposes a hardware leaky integrate-and-fire neuron based on organic thin-film transistors. Under the electric field, the ion dynamics in the gate electrolyte can mimic the processes of membrane potential integration, leakage, and reset in spiking neurons. The convolutional spiking neural networks composed of such organic spiking neurons achieves excellent recognition rates (∼97.26 %) on the MNIST dataset. This indicates that the organic spiking neuron has enormous potential in next-generation non-von Neumann neuromorphic computing.

An emergent attractor network in a passive resistive switching circuit

Resistive memory devices feature drastic conductance change and fast switching dynamics. Particularly, nonvolatile bipolar switching events (set and reset) can be regarded as a unique nonlinear activation function characteristic of a hysteretic loop. Upon simultaneous activation of multiple rows in a crosspoint array, state change of one device may contribute to the conditional switching of others, suggesting an interactive network existing in the circuit. Here, we prove that a passive resistive switching circuit is essentially an attractor network, where the binary memory devices are artificial neurons while the pairwise voltage differences define an anti-symmetric weight matrix. An energy function is successfully constructed for this network, showing that every switching in the circuit would decrease the energy. Due to the nonvolatile hysteretic function, the energy change for bit flip in this network is thresholded, which is different from the classic Hopfield network. It allows more stable states stored in the circuit, thus representing a highly compact and efficient solution for associative memory. Network dynamics (towards stable states) and their modulations by external voltages have been demonstrated in experiment by 3-neuron and 4-neuron circuits.

Novel Two-Terminal Synapse/Neuron Based on an Antiferroelectric Hafnium Zirconium Oxide Device for Neuromorphic Computing.

Functionally diverse devices with artificial neuron and synapse properties are critical for neuromorphic systems. We present a two-terminal artificial leaky-integrate-fire (LIF) neuron based on 6 nm Hf0.1Zr0.9O2 (HZO) antiferroelectric (AFE) thin films and develop a synaptic device through work function (WF) engineering. LIF neuron characteristics, including integration, firing, and leakage, are achieved in W/HZO/W devices due to the accumulated polarization and spontaneous depolarization of AFE HZO films. By engineering the top electrode with asymmetric WFs, we found that Au/Ti/HZO/W devices exhibit synaptic weight plasticity, such as paired-pulse facilitation and long-term potentiation/depression, achieving >90% accuracy in digit recognition within constructed artificial neural network systems. These findings suggest that AFE HZO capacitor-based neurons and WF-engineered artificial synapses hold promise for constructing efficient spiking neuron networks and artificial neural networks, thereby advancing neuromorphic computing applications based on emerging AFE HZO devices.

Emerging Optoelectronic Devices for Brain‐Inspired Computing

AbstractBrain‐inspired neuromorphic computing is recognized as a promising technology for implementing human intelligence in hardware. Neuromorphic devices, including artificial synapses and neurons, are regarded as essential components for the construction of neuromorphic hardware systems. Recently, optoelectronic neuromorphic devices are increasingly highlighted due to their potential applications in next‐generation artificial visual systems, attributed to their integrated sensing, computing, and memory capabilities. In this review, recent advancements in optoelectronic synapses and neurons are examined, with an emphasis on their structural characteristics, operational principles, and the replication of neuromorphic functions. For optoelectronic synaptic devices, such as memristor‐ and transistor‐based ones, attention is given to the two primary weight update modes: the light‐electricity synergistic mode and the all‐optical mode. Optoelectronic neurons are discussed in terms of different device types, including threshold switch neurons and semiconductor laser neurons. Last, the challenges that impede the progress of optoelectronic neuromorphic devices are identified, and potential future directions are suggested.

Hydrogel-Based Artificial Synapses for Sustainable Neuromorphic Electronics.

Hydrogels find widespread applications in biomedicine because of their outstanding biocompatibility, biodegradability, and tunable material properties. Hydrogels can be chemically functionalized or reinforced to respond to physical or chemical stimulation, which opens up new possibilities in the emerging field of intelligent bioelectronics. Here, the state-of-the-art in functional hydrogel-based transistors and memristors is reviewed as potential artificial synapses. Within these systems, hydrogels can serve as semisolid dielectric electrolytes in transistors and as switching layers in memristors. These synaptic devices with volatile and non-volatile resistive switching show good adaptability to external stimuli for short-term and long-term synaptic memory effects, some of which are integrated into synaptic arrays as artificial neurons; although, there are discrepancies in switching performance and efficacy. By comparing different hydrogels and their respective properties, an outlook is provided on a new range of biocompatible, environment-friendly, and sustainable neuromorphic hardware. How potential energy-efficient information storage and processing can be achieved using artificial neural networks with brain-inspired architecture for neuromorphic computing is described. The development of hydrogel-based artificial synapses can significantly impact the fields of neuromorphic bionics, biometrics, and biosensing.

One-Transistor One-Memristor Based Universal Oscillating Units for Spike-Encoding Artificial Sensory Neuron

One-Transistor One-Memristor Based Universal Oscillating Units for Spike-Encoding Artificial Sensory Neuron

Développement et validation d’un outil de dépistage systématique des fragilités sociales des patients atteints de cancer : l’outil DEFCO

Literature suggests that patients from deprived backgrounds are less likely to adhere to their treatments, continue to expose themselves to risk factors and, as a result, have poorer health outcomes. It is therefore crucial to identify these vulnerable populations early on, in order to provide them with tailored and reinforced care. The primary aim of this research is to construct and validate a systematic screening tool for identifying patients at highest risk of social vulnerability due to deprivation, through the use of psychometric techniques. This tool is intended to be easily used by healthcare professionals, to provide tailored and targeted care throughout the patient's journey. This study involves the development and assessment of a screening tool, along with a self-questionnaire and a decision support tool incorporating an artificial neural network. It is a prospective, monocentric, 2-stage psychometric validation study. This study has demonstrated the successful development of the self-questionnaire using psychometric methodology. The tool was found a good performance in screening social vulnerabilities. This validated self-questionnaire is an easy-to-use tool, allowing systematic screening for social vulnerabilities for cancer patients. This early identification allows to reinforce patient's pathway in order to avoid disruption. The integration of the tool in an artificial neuron network system allows to automate and disseminate this method of deprived patients' detection, while limiting the workload for the staff.

Investigating green hydrogen production operated by redundant energy on a solar PV mini-grid through matlab simulation and artificial neural network

Green hydrogen is an important part of the transition to a net zero economy, due to its potential to deliver high amounts of energy without pollution. One important opportunity to produce green hydrogen is the use of redundant energy on solar PV mini-grids. Redundant energy is the potential generation of a mini-grid that is never utilized due to low demand and the mismatch of peak solar PV generation and peak demand. In this study, a typical rural community solar PV mini-grid is simulated using Matlab Simulink and incorporated with a water electrolyser for hydrogen production. Having run the simulation for 365 days of the year, it was revealed that the mini-grid can produce a daily average of 1.18–2.16 kg, a monthly average of 36.59–64.84 kg and a yearly total of 609.26 kg of hydrogen. Subsequently, an artificial neuron network was trained with data acquired from the simulation. With varying configurations of the number of neurons and hidden layers, it was found that the model successfully predicts hydrogen production on the mini-grid with the highest performance (RMSE = 39.85 g and R2 = 0.9979) seen when the number of hidden layers is 40 and the number of neurons per hidden layer is 3.

Neuron circuit made of a single locally-active memristor

Neuromorphic computing, inspired by the human brain’s architecture, is expected to break the physical limits of transistors and von Neumann bottleneck. The multiple internal state variables of higher-order memristors (second-order or above) possess dynamic complexity and adaptability, enabling them to mimick the characteristics of biological neurons, which are very important building blocks for neuromorphic computing. This paper presents a simple neuron circuit containing a single second-order current-controlled locally-active memristor (LAM). The pinched hysteresis loop and DC V–I curve of the proposed second-order LAM show good odd symmetry. Applying small signal analysis method, we obtain the small-signal equivalent circuit of the neuron circuit, showing a [Formula: see text] parallel structure and an edge of chaos kernel in its locally-active domain. Also, we draw a parameter classification of the neuron circuit, showing four symmetrical edge of chaos domains, which plays an important role in biphasic action potentials. Finally, we demonstrate that the simple neuron circuit can produce monophasic action potentials, biphasic action potentials and co-existing neuromorphic phenomena via subcritical Hopf bifurcation with different input, verifying the simple circuit is suitable as artificial neurons.

REVOLUTIONIZING POWER TRANSFORMER FAULT DIAGNOSIS THROUGH COGNITIVE ARTIFICIAL INTELLIGENCE AND DISSOLVED GAS ANALYSIS INTEGRATION

The research introduces a cognitive artificial intelligence (CAI) model that leverages dissolved gas analysis (DGA) to investigate power transformer faults. Conventional fault interpretation methods using DGA are limited in accuracy and uncertainty. In response, the proposed CAI model utilizes cognitive learning and direct interaction to achieve remarkably accurate fault identification without the need for supervised training. By extracting fault features through key gas ratio limitations. However, the CAI model also has a gap in data perception due to the information sensory challenges. Using gas ratios based on the conventional fault interpretation methods in the latest study still limited data perception of the CAI model to only three or four gas ratios. Thus, this study aims to increase data perception by extracting fault features through ten gas ratio limitations. The proposed CAI model's performance is validated, outperforming traditional methods like the Duval triangle method, Duval pentagon method, Doernenburg ratio method, and Roger ratio method, as well as common AI approaches including artificial neuron network, long short-term memory, nearest neighbor classifiers, support vector machine, ensemble classifiers, and decision trees. Notably, the CAI model's success rate in fault type identification stands at an impressive 98.04%. A distinctive trait of the CAI model is its autonomous knowledge accumulation and enhancement, enabled by inferring-fusion information and sensor-based knowledge integration. This intrinsic learning ability further contributes to its exceptional fault diagnosis accuracy. The proposed CAI model showcases promising potential for revolutionizing power transformer fault investigation and diagnosis, mitigating unplanned outages, and ultimately bolstering power system reliability.

A Comparative Investigation of Hybrid MPPT Methods for Enhancing Solar Power Generation in Renewable Energy Systems

Photovoltaic (PV) systems are among the types of renewable energy that are frequently employed. Since the characteristics of the solar cell depend on the amount of insolation and temperature, it is necessary to use MPPT “Maximum Power Point Tracking” to move the operating voltage close to the maximum power point under changing weather conditions. This article aims to design a photovoltaic energy system based on boost converter control to obtain maximum power using a hybrid algorithm based on artificial neurons (ANN). Additionally included is a proportional-integral (PI) controller, which improves the performance of the ANN-MPPT controller; this method is quick and precise for tracking the maximum power point (MPP) in the face of variations in temperature and solar radiation. The efficiency of the tracking algorithm was calculated and compared to one of the traditional methods, the incremental conductance (INC) method, in addition to comparing it to other hybrid methods and by simulating the system using MATLAB/Simulation and analyzing the results. This study unequivocally proves the superiority of the hybrid ANN+PI strategy; the efficiency reached 99.91%. This approach excels at tracking maximum power accuracy by leveraging the adaptive learning capabilities of neural networks, ensuring maximum power even in the face of changing environmental conditions.

Audio Signal-Stimulated Multilayered HfOx/TiOy Spiking Neuron Network for Neuromorphic Computing.

As the key hardware of a brain-like chip based on a spiking neuron network (SNN), memristor has attracted more attention due to its similarity with biological neurons and synapses to deal with the audio signal. However, designing stable artificial neurons and synapse devices with a controllable switching pathway to form a hardware network is a challenge. For the first time, we report that artificial neurons and synapses based on multilayered HfOx/TiOy memristor crossbar arrays can be used for the SNN training of audio signals, which display the tunable threshold switching and memory switching characteristics. It is found that tunable volatile and nonvolatile switching from the multilayered HfOx/TiOy memristor is induced by the size-controlled atomic oxygen vacancy pathway, which depends on the atomic sublayer in the multilayered structure. The successful emulation of the biological neuron's integrate-and-fire function can be achieved through the utilization of the tunable threshold switching characteristic. Based on the stable performance of the multilayered HfOx/TiOy neuron and synapse, we constructed a hardware SNN architecture for processing audio signals, which provides a base for the recognition of audio signals through the function of integration and firing. Our design of an atomic conductive pathway by using a multilayered TiOy/HfOx memristor supplies a new method for the construction of an artificial neuron and synapse in the same matrix, which can reduce the cost of integration in an AI chip. The implementation of synaptic functionalities by the hardware of SNNs paves the way for novel neuromorphic computing paradigms in the AI era.

Integration of Ag-based threshold switching devices in silicon microchips

Threshold-type resistive switching (RS) is an essential electronic behavior in many types of integrated circuits and can be exploited in multiple applications, such as leaky integrate-and-fire neurons for artificial neural networks (ANNs). Many research articles have shown that metal/insulator/metal (MIM) devices using Ag electrodes exhibit stable threshold-type RS, but all of them presented large devices (>1 µm2) fabricated on unfunctional SiO2/Si substrates. In this article, for the first time we integrate Ag-based threshold-type RS devices at the back-end-of-line interconnections of silicon microchips. The insulator used is multilayer hexagonal boron nitride (h-BN), and the size of the devices is ∼0.05 µm2. The devices can switch between a high resistive state (HRS) and a low resistive state (LRS) without the need of any forming process, and we observe a high endurance over 1 million cycles over multiple devices. By performing circuit simulation using SPICE software, we confirm that this electrical behavior is suitable for being used as leaky integrate-and-fire electronic neuron in spiking neural networks for image recognition, and the h-BN artificial neurons operate correctly for 94 % of the images presented. Our study represents a significant advancement towards the integration of Ag-based threshold-type RS devices in silicon microchips.

Network Attack Classification with a Shallow Neural Network for Internet and Internet of Things (IoT) Traffic

In recent years, there has been a tremendous increase in the use of connected devices as part of the so-called Internet of Things (IoT), both in private spaces and the industry. Integrated distributed systems have shown many benefits compared to isolated devices. However, exposing industrial infrastructure to the global Internet also generates security challenges that need to be addressed to benefit from tighter systems integration and reduced reaction times. Machine learning algorithms have demonstrated their capacity to detect sophisticated cyber attack patterns. However, they often consume significant amounts of memory, computing resources, and scarce energy. Furthermore, their training relies on the availability of datasets that accurately represent real-world data traffic subject to cyber attacks. Network attacks are relatively rare events, as is reflected in the distribution of typical training datasets. Such imbalanced datasets can bias the training of a neural network and prevent it from successfully detecting underrepresented attack samples, generally known as the problem of imbalanced learning. This paper presents a shallow neural network comprising only 110 ReLU-activated artificial neurons capable of detecting representative attacks observed on a communication network. To enable the training of such small neural networks, we propose an improved attack-sharing loss function to cope with imbalanced learning. We demonstrate that our proposed solution can detect network attacks with an F1 score above 99% for various attacks found in current intrusion detection system datasets, focusing on IoT device communication. We further show that our solution can reduce the false negative detection rate of our proposed shallow network and thus further improve network security while enabling processing at line rate in low-complexity network intrusion systems.