Pulsed Neural Networks Research Articles

Method of Data Preprocessing on the Basis of Pulse Neural Network to Improve the Accuracy of Water Level Forecast on the Example of Ufa City of the Republic of Bashkortostan

The developed method of data preprocessing based on impulse neural network is considered. The essence of this method is to improve the quality of the initial dataset by minimising data noise. The peculiarity is to improve the accuracy of the predicted values of the time series of water levels obtained at the output of the pulse neural network. Retrospective data were obtained from hydrological posts (river gauge) and automatic stations with the help of FSUE «Centre of Register and Cadastre» from 01.01.1997 to 30.06.2023. On the example of hydrological post 76289 (Ufa) the experiment on forecasting of water levels with the help of the developed system «Flood 2.0» was carried out. The experiment proves the efficiency of the data preprocessing method developed in this study to improve the accuracy of water level forecasts.

Synaptic and resistive switching behaviors of Sm‐doped HfO2 films for bio‐inspired neuromorphic calculations

AbstractArtificial neural network‐based computing is anticipated to surpass the von Neumann bottleneck of traditional computers, thus dramatically boosting computational efficiency and showing a wide range of promising applications. In this paper, sol−gel deposition was used to prepare thin films of samarium‐doped hafnium dioxide (Sm:HfO2). When Sm is doped at a concentration of 4%, it mimics biological synapses; meantime, by voltage scanning, an obvious mimicry of resistive switching can be detected, demonstrating that the technology may be applied to simulate biological synapse characteristics, including long‐term potentiation (depression), short‐term potentiation (depression), paired‐pulse facilitation, and learning rules of spike‐time‐dependent plasticity. Additionally, a pulsed neural network is built on the MNIST dataset to test the memristor's capacity to handle visual input. The findings show the possibility of synthetic synapses in artificial intelligence systems that integrate neuromorphic computing with synaptic brain activity.

Spiking neural network research based on brain-inspired computing

With the shortcomings of deep learning in training cost, generalization ability, interpretability, and reliability, increasing attention was concentrated on the novel neuromorphic devices. A spiking neural network can better simulate the information transmission mode of biological neurons, and has the characteristics of strong computing power and low power consumption. Therefore, this study provides an overview of the pulse neural network from the perspective of its fundamental structure and operating principle. In terms of structure optimization, there is a summary of the aspects of the encoding mode of a spiking neural network and topology structure. In addition, the deficiency and development of the spiking neural network are analyzed. Pulsed neurons work by integrating the mechanism of firing, exchanging information with the time of the signal, and transmitting signals to adjacent neurons only when the membrane potential reaches a threshold. Spiking neural network (SNN) still lags behind traditional Artificial neural network (ANN) in accuracy, has large computational capacity and is inefficiently trained due to the complexity of pulsed neuron models. Future developments lie in bionics, deeper studies of the complex dynamics of the human brain, and synaptic connections in the human brain.

Review of medical data analysis based on spiking neural networks

Medical data mainly includes various types of biomedical signals and medical images, which can be used by professional doctors to make judgments on patients' health conditions. However, the interpretation of medical data requires a lot of human cost and there may be misjudgments, so many scholars use neural networks and deep learning to classify and study medical data, which can improve the efficiency and accuracy of doctors and detect diseases early for early diagnosis, etc. Therefore, it has a wide range of application prospects. However, traditional neural networks have disadvantages such as high energy consumption and high latency (slow computation speed). This paper presents recent research on signal classification and disease diagnosis based on a third-generation neural network, the spiking neuron network, using medical data including EEG signals, ECG signals, EMG signals and MRI images. The advantages and disadvantages of pulsed neural networks compared with traditional networks are summarized and its development orientation in the future is prospected.

Tunable plasticity in functionalized honeycomb synaptic memristor for neurocomputing

Tunable plasticity is one of the important features in synapses and plays a key role in neuromorphic computing (neurocomputing, NC), maintaining a stable pulse neural network environment, and realizing accurate image recognition. Herein, a functionalized honeycomb-like synaptic memristor (HLSM) composed of porous silicon oxide incorporated with MoS2 quantum dots (QDs) is fabricated and the synaptic properties such as paired-pulse facilitation (PPF), post-tetanic potentiation (PTP), learning-forgetting behavior, and spike-timing-dependent plasticity (STDP) are simulated. A critical characteristic of short/long-term plasticity (SLTP) is observed from the HLSM device for the sliding frequency threshold between 200 μs and 10 ms and tunable synaptic plasticity is realized without gate voltage regulation. An artificial neural network (ANN) based on the potentiation/depression characteristics is designed theoretically and the recognition rate is observed to increase from 54.2% to 91.8% by simply adjusting the input frequency. The underlying mechanism of the tunable synaptic plasticity is proposed and discussed by taking advantage of the honeycomb structure on fluid buffering and acceleration. The results and theoretical understanding are expected to accelerate the application in NC to image recognition.

EVOLUTIONARY METHOD FOR SYNTHESIS SPIKING NEURAL NETWORKS USING THE NEUROPATTHERN MECHANISM

Context. The problem of synthesizing pulsed neural networks based on an evolutionary approach to the synthesis of artificial neural networks using a neuropathic mechanism for constructing diagnostic models with a high level of accuracy is considered. The object of research is the process of synthesis of pulsed neural networks using an evolutionary approach and a neuropathic mechanism.
 Objective of the work is to develop a method for synthesizing pulsed neural networks based on an evolutionary approach using a neuropathic mechanism to build diagnostic models with a high level of accuracy of work.
 Method. A method for synthesizing pulsed neural networks based on an evolutionary approach is proposed. At the beginning, a population of pulsed neural networks is generated, and a neuropathic mechanism is used for their encoding and further development, which consists in separate encoding of neurons with different activation functions that are determined beforehand. So each pattern with multiple entry points can define the relationship between a pair of points. In the future, this simplifies the evolutionary development of networks. To decipher a pulsed neural network from a pattern, the coordinates for a pair of neurons are passed to the network that creates the pattern. The network output determines the weight and delay of the connection between two neurons in a pulsed neural network. After that, you can evaluate each neuromodel after evolutionary changes and check the criteria for stopping synthesis. This method allows you to reduce the resource intensity during network synthesis by abstracting the evolutionary changes of the network pattern from itself.
 Results. The developed method is implemented and investigated on the example of the synthesis of a pulsed neural network for use as a model for technical diagnostics. Using the developed method to increase the accuracy of the neuromodel with a test sample by 20%, depending on the computing resources used.
 Conclusions. The conducted experiments confirmed the operability of the proposed mathematical software and allow us to recommend it for use in practice in the synthesis of pulsed neural networks as the basis of diagnostic models for further automation of tasks of diagnostics, forecasting, evaluation and pattern recognition using big data. Prospects for further research may lie in the use of a neuropathic mechanism for indirect encoding of pulsed neural networks, which will provide even more compact data storage and speed up the synthesis process.

Analysis of Teaching Effect of English Classroom Mind Map Based on a Logistic Regression Model

In order to improve the effect of English classroom teaching, this paper combines the pulse neural network to identify the teaching process of English classroom and conducts real-time supervision of classroom dynamics in combination with the actual situation. Moreover, this paper constructs an intelligent system, uses the recognition method based on the spiking neural network image to collect the real-time dynamics of classroom teaching, and combines the logistic regression model to evaluate the teaching effect. After constructing the system framework, the system process analysis is carried out. This paper combines experimental research to verify the model of this paper. The experimental research results show that the analysis model of teaching effect of mind map in English classroom based on the logistic regression model proposed in this paper has good results, which also verifies that the mind map teaching can play an important role in English classroom teaching.

Image Manipulation Detection Through Laterally Linked Pixels and Kernel Algorithms

In this paper, copy-move forgery in image is detected for single image with multiple manipulations such as blurring, noise addition, gray scale conversion, brightness modifications, rotation, Hu adjustment, color adjustment, contrast changes and JPEG Compression. However, traditional algorithms detect only copy-move attacks in image and never for different manipulation in single image. The proposed LLP (Laterally linked pixel) algorithm has two dimensional arrays and single layer is obtained through unit linking pulsed neural network for detection of copied region and kernel tricks is applied for detection of multiple manipulations in single forged image. LLP algorithm consists of two channels such as feeding component (F-Channel) and linking component (L channel) for linking pixels. LLP algorithm linking pixels detects image with multiple manipulation and copy-move forgery due to one-to-one correspondence between pixel and neuron, where each pixel’s intensity is taken as input for F channel of neuron and connected for forgery identification. Furthermore, neuron is connected with neighboring field of neuron by L channel for detecting forged images with multiple manipulations in the image along with copy-move, through kernel trick classifier (KTC). From experimental results, proposed LLP algorithm performs better than traditional algorithms for multiple manipulated copy and paste images. The accuracy obtained through LLP algorithm is about 90% and further forgery detection is improved based on optimized kernel selections in classification algorithm.

Mathematical Modeling of an Analogue Self-Learning Neural Network Based on Memristive Elements Taking into Account Stochastic Switching Dynamics

In this paper, we consider the issues of modeling an analog self-learning pulsed neural network based on memristive elements. One of the important properties of a memristor is the stochastic switching dynamics, which is mainly due to the stochastic processes of generation/recombination and the movement of ions (or oxygen vacancies) in a dielectric film under the effect of an electric field. The stochastic features are taken into account by adding a term responsible for the additive (Gaussian) noise to the memristor-state equation. The effect of noise on the functioning of an element is demonstrated using the dynamic model of a memristor as an example. The switching of a memristor from the high-resistance state to low-resistance state and vice versa is shown to occur from cycle to cycle in different ways, which is consistent with experimental data. We formulate a stochastic model that describes the hardware analog implementation of a pulsed neural network with memristive elements as synaptic weights and a learning mechanism based on the spike timing dependent plasticity (STDP) method. The operation of two neural networks consisting of one neuron with 64 synapses and two neurons with 128 synapses, respectively, is modeled. The recognition of 8 × 8 images is performed. The stochastic component in the memristor model is shown to have an effect on the fact that the templates from implementation to implementation are distributed among the neurons in different ways and the adaptation of weights (network training) occurs at different rates. In both cases, the network successfully learns to recognize the specified images.

Magnetic Resonance Imaging Image Segmentation and Brain Tumour Detection Using Pulse-Coupled Neural Networks

Image processing can be defined as a functional structure to correct and change the images viewed and their interpretation. One of the applications of digital image processing is using image processing techniques in the component and image segmentation. One of these techniques is magnetic resonance imaging (MRI) in the medical world. In this article, a brain tumour detection system and various anomalies and abnormalities are presented where image pre-processing and preparation include image enhancement, filtering and noise reduction. Then image segmentation is done by a pulse neural network. Next, the image features are extracted and finally, the tumour and abnormal area are separated from the normal area by the algorithms. In this research, the feature selection and integration method are used and the most important statistical features of brain MRI images are used to improve brain tumour detection. Along with the studies done and the implementation of tumour detection systems, the following suggestions can be provided for future researches and the tumour detection system will work more efficiently. The pulse-coupled neural network (PCNN) can be used for image segmentation in the pre-processing stage, especially in the image filtering.

Fabrication technology and electrophysical properties of a composite memristor-diode crossbar used as a basis for hardware implementation of a biomorphic neuroprocessor

In order to build a biomorphic neuroprocessor hardware, a laboratory TiN/Ti0.93Al0.07Ox/p-Si/n-Si/W composite memristor-diode crossbar intended to serve as a basis for a memory matrix and a logic matrix, was manufactured. For this purpose, the materials and the fabrication nanotechnology for semiconductor layers of the Zener diode and a memristive layer, which ensured the required characteristics of the diode and of the memristors, were selected. It was demonstrated that the magnetron sputtering technology was unified both for the fabrication of the diodes and of the memristors. Therefore, all the layers in a composite memristor-diode crossbar, including conductive paths, can be fabricated in one technological cycle.Electrophysical properties of fabricated memristor-diode crossbar were investigated. Current-voltage curves of the memristor, of the diode and of the memristor-diode cell were measured. Signal processing was executed by the following procedures: output neuron pulses routing to other neurons' synapses; matrix-vector product, which is performed in the memory matrix while weighting and summing the signals; associative self-learning. It was the first time that the generation of a new association (new knowledge) in a fabricated memristor-diode crossbar of an artificial pulsed neural network was demonstrated, as opposed to associative self-learning in the existing hardware-based neural networks with synapses based on discrete memristors. The obtained experimental findings were compared with the SPICE simulation of the signal processing routines in a memristor-diode crossbar. The memristor parameter scatter at which the crossbar demonstrates stable operation was identified. The cells SPICE model was constructed from experimentally obtained i-v curve. The change in the output current of the cell, associated with the parasitic currents passing through the adjacent cells, was measured. It was demonstrated that the Zener diode parameters can reduce power consumption of the composite crossbar.The experimental findings highlighted that the composite memristor-diode crossbar intended to fabricate the memory and logic matrices of the neuroprocessor is essentially workable.

Mechanisms of Invariant Anti-Noise Encoding in Pulse Neural Networks

Nonlinear dynamics, physical processes, and information processing in selective pulse neurons are studied. A way of encoding input information, in which a piece of information is the entire pulse sequence and the intensity of input signals is coded according to the asynchronous change in the mean frequency of pulse sequences, is considered. The proposed means of encoding pulse sequences is tested. The invariance of positional recognition is demonstrated with respect to changes in the input signal intensity, which, according to the Weber–Fechner law, results in a synchronous change in the frequency of pulse sequences.

Inductively coupled burst oscillators in neural network information processing systems

In the current study, we present a self-generating circuit based on an LC oscillatory circuit with a switching element having an S-type current-voltage characteristic. Using the example of a VO2 switch, we demonstrate that such circuits generate burst oscillations and have a specific mode of bursts synchronization using inductive coupling between generators. Such generators of burst oscillations (of a neural type) can find application in the tasks of cognitive technologies, including the implementation of pulsed neural networks.

A demonstration of using the model reference principle to develop the function-oriented adaptive pulse-coded neural network

How can one design an adaptive pulsed neural network that is based on psycho-phenomenological foundations? In other words, how can one migrate the adaptive capability of a psychologically modeled neural network to a pulsed network? Neural networks that model psychological phenomena are at a larger scale than physiological models. There is a common presumption that pulse-coded neural network analogs to non-pulsing networks can be obtained by a simple mapping and scaling process of some sort. But the actual in vivo environment of pulse-coded neural network systems produces a much more diverse set of firing patterns. Thus, functional mapping from traditional neural network systems to pulse-coded neural network systems is much more challenging than has been presumed. This paper demonstrates that the employment of model reference adaptation as a method for applying scientific reduction is a powerful design tool for the development of a function-oriented adaptive pulse-coded neural network. The performance surface is empirically obtained by comparing the performance of the pulsed network to the non-pulsing network. Based on this surface, the adaptive algorithm is a combination of gain scheduling and steepest-descent method. Therefore, the adaptive property of the pulse-coded neural network is built upon a psycho-physiological foundation.

Discerning Functional Connections in the Pulsed Neural Networks with the Dynamic Bayesian Network Structure Search Method Based on a Genetic Algorithm.

It is important to explore potential structural characteristics of biological networks and regulatory mechanisms of network behaviors at the system level. In this study, a dynamic Bayesian network structure search method (DBNSSM) based on a genetic algorithm is employed to infer and locate functional connections in pulsed neural networks (PNNs) as typical artificial neural networks. In the process of network structure searching, a minimum description length score is calculated for each candidate network structure. The score indicates two characteristics of the network structure: (1) the likelihood based on network dynamic response data and (2) the complexity. Both should be considered together on selecting network structures. The DBNSSM is applied to analyze time-series data from PNNs, thereby discerns functional connections showing network structures collectively. It is feasible to analyze multichannel electrophysiological data of biological neural networks using the DBNSSM.

Chaotic Pattern Alternations Can Reproduce Properties of Dominance Durations in Multistable Perception.

We propose a pulse neural network that exhibits chaotic pattern alternations among stored patterns as a model of multistable perception, which is reflected in phenomena such as binocular rivalry and perceptual ambiguity. When we regard the mixed state of patterns as a part of each pattern, the durations of the retrieved pattern obey unimodal distributions. We confirmed that no chaotic properties are observed in the time series of durations, consistent with the findings of previous psychological studies. Moreover, it is shown that our model also reproduces two properties of multistable perception that characterize the relationship between the contrast of inputs and the durations.

Real-time multimodal transport path planning based on a pulse neural network model

Real-time multimodal transport path planning based on a pulse neural network model

Real-time multimodal transport path planning based on a pulse neural network model

A modified pulse-coupled neural network (MPCNN) model is designed for real-time collision-free path planning of multimodal transport choice in stationary or non-stationary environments. The proposed neural network is topologically organised with only local lateral connections among neurons. The optimisation networks model consists of transport distance, transport time, transit costs and transit time and other factors, and then all factors compose to weight of the networks to realise the transformation to solve the shortest path problem. The effectiveness and efficiency of the proposed approach is demonstrated through simulation and comparison studies.

Identification of Functional Connections in Biological Neural Networks Using Dynamic Bayesian Networks

Abstract: Investigation of the underlying structural characteristics and network properties of biological networks is crucial to understanding the system-level regulatory mechanism of network behaviors. A Dynamic Bayesian Network (DBN) identification method is developed based on the Minimum Description Length (MDL) to identify and locate functional connections among Pulsed Neural Networks (PNN), which are typical in synthetic biological networks. A score of MDL is evaluated for each candidate network structure which includes two factors: i) the complexity of the network; and ii) the likelihood of the network structure based on network dynamic response data. These two factors are combined together to determine the network structure. The DBN is then used to analyze the time-series data from the PNNs, thereby discerning causal connections which collectively show the network structures. Numerical studies on PNN with different number of nodes illustrate the effectiveness of the proposed strategy in network structure identification.

リカレントパルスニューラルネットワークを用いた自律移動ロボットの時系列処理

リカレントパルスニューラルネットワークを用いた自律移動ロボットの時系列処理