Principle Of Neural Network Research Articles

Teaching the basic principles of neural networks using online services

Interest in neural networks is huge today. Neural networks are used by schoolchildren, students, and teachers to solve their problems. In accordance with the federal project "Artificial Intelligence", many universities include disciplines on artificial intelligence and neural networks in their educational programs, organize advanced training courses for teachers on this theme. The article discusses the main approaches to teaching students of pedagogical specialties neural networks. An approach based on the study of the principles of neural networks using online services for training artificial intelligence models is proposed. A comparison of services by functionality and ease of use is given. The experience of students studying the training program 44.03.05 "Pedagogical education (with two training profiles)" in the process of teaching neural networks using the online service Teachable Machine is described.The approach to the study of neural networks described in the article may be of interest to teachers for organizing educational and extracurricular activities to introduce schoolchildren to the work of neural networks and organizing project activities of students using neural networks.

Applications and Progress of Machine Learning Techniques in the Ladle Furnace Refining Process: A Review

Machine learning techniques have now become key technologies in optimizing the ladle furnace (LF) refining process, significantly improving prediction accuracy and operational efficiency in various aspects such as temperature control, slag optimization, and alloying processes. This article first reviews the principles of neural networks, utilizing their significant advantages in data processing capability and model generalization to explore the current applications and development trends of machine learning techniques in the LF refining process. It analyzes the advantages of hybrid dynamic models and ensemble learning methods in improving prediction accuracy and handling complex data. Additionally, this article explains the optimization of the LF refining process by combining machine learning techniques with traditional methods to reduce costs and improve production efficiency. It demonstrates the application effects and potential value of machine learning techniques in actual production, discusses the strategies and challenges in real‐time control and prediction accuracy, and envisions the future development direction of intelligent models for the LF refining process, providing a scientific basis and technical reference for the subsequent development of LF refining technology.

Medical image recognition based on multilayer neural network

As a common accidental injury, accurate identification of burns and scalded injuries is of great significance to the development of treatment programs and prognosis assessment. Traditional identification methods are subjective and inaccurate, this paper proposes a burns and scalds grade identification technique based on multilayer neural network. The three-degree classification standard of burns and scalds is described, and the structure and principle of multilayer neural network are introduced, including the network structure composed of input layer, hidden layer and output layer, forward propagation, back propagation and training process. The implementation steps of burns and scald grade recognition technology based on multi-layer neural network are discussed in detail, such as data acquisition, image pre-processing, feature extraction, classifier design, model training and evaluation. Through experiments, the model is tested using a dataset containing 362 burns and scalds images, and the model achieves more than 90% accuracy on the test set with high accuracy and reliability. The multi-layer neural network model is used to classify the burn and scald data, which improves the diagnostic accuracy, shortens the delay time of the disease, reduces the pressure of professional doctors, and helps patients have a preliminary understanding of the degree of burn.

Hybrid Neuro-Fuzzy Based Improved Control Design of Electro-Pneumatic Clutch Actuation Control System for Heavy Duty Vehicles

The application of hybrid Neuro-Fuzzy principles of clutch actuation control in enhancing the performance of electro-pneumatic clutch actuation system for heavy duty vehicles is the aim of this presentation. Neuro-Fuzzy is a hybrid design that accommodates the principles of fuzzy Logic and Neural Network in a manner that take advantage of their positive sides. The inability of some heavy-duty vehicles to operate optimally on hilly terrains and which often results in road mishaps are traceable to the inadequacies in conventional clutch actuation control mechanisms. These conventional actuation control techniques provide for routine calibration of clutch actuators after maintenance. Often times, this important requirement of calibration is neglected with attendant ugly consequences. To stem this tide of manual calibration and its obvious defects, an intelligent method of clutch actuation modelled in a hybrid Neuro-Fuzzy control is implemented. Conventional data obtained for errors, speed, torque and power from Mercedes Benz Actros Truck model MP 2, 2031 provided the reference points. These data were embedded into a Simulink block and cascaded into a designed Neuro-Fuzzy Simulink model. Both conventional and Neuro-Fuzzy Simulink model controllers were also simulated. Different percentages of improvements were recorded for piston error, angular speed, engine torque and power respectively. The level and percentage of improvements stood at 0.4821mm or 33.04% decrease for error, increases of 334.1 RPM or 33% for angular speed, 0.0594NM or 33.26% for torque and 2.79W or 16.53% for power respectively.

Modeling Brain Functional Networks Using Graph Neural Networks: A Review and Clinical Application

The integration of graph neural networks (GNNs) with brain functional network analysis is an emerging field that combines neuroscience and machine learning to enhance our understanding of complex brain dynamics. We first briefly introduce the fundamentals of brain functional networks, followed by an overview of Graph Neural Network principles and architectures. The review then focuses on the applications of these networks and address current challenges in the field, such as the need for interpretable models and effective integration of multi-modal neuroimaging data. We also highlight the potential of GNNs in clinical perimenopausal areas such as perimenopausal depression research, demonstrating the broad applicability of this approach. The review concludes by outlining future research directions, including the development of more sophisticated architectures for large-scale, heterogeneous brain graphs, and the exploration of causal inference in brain networks. By synthesizing recent advances and identifying key research directions, this review aims to summarize the focal points of brain functional network analysis and GNNs, explore the potential of their integration, and provide a reference for advancing this interdisciplinary field.

Enhancing Melanoma Diagnosis with Advanced Deep Learning Models Focusing on Vision Transformer, Swin Transformer, and ConvNeXt.

Skin tumors, especially melanoma, which is highly aggressive and progresses quickly to other sites, are an issue in various parts of the world. Nevertheless, the one and only way to save lives is to detect it at its initial stages. This study explores the application of advanced deep learning models for classifying benign and malignant melanoma using dermoscopic images. The aim of the study is to enhance the accuracy and efficiency of melanoma diagnosis with the ConvNeXt, Vision Transformer (ViT) Base-16, and Swin Transformer V2 Small (Swin V2 S) deep learning models. The ConvNeXt model, which integrates principles of both convolutional neural networks and transformers, demonstrated superior performance, with balanced precision and recall metrics. The dataset, sourced from Kaggle, comprises 13,900 uniformly sized images, preprocessed to standardize the inputs for the models. Experimental results revealed that ConvNeXt achieved the highest diagnostic accuracy among the tested models. Experimental results revealed that ConvNeXt achieved an accuracy of 91.5%, with balanced precision and recall rates of 90.45% and 92.8% for benign cases, and 92.61% and 90.2% for malignant cases, respectively. The F1-scores for ConvNeXt were 91.61% for benign cases and 91.39% for malignant cases. This research points out the potential of hybrid deep learning architectures in medical image analysis, particularly for early melanoma detection.

Sound source localization based on data and neural network model

Abstract In this study, the integration of sound source localization with neural network techniques is proposed. Firstly, the development history and basic principle of neural networks are introduced. Combined with the characteristics of sound source location, the residual neural network is analyzed, and the residual neural network is used as the neural network model. The sound signal received by the microphone array is computed by PHAT-SCOT and the output is trained as the input feature of ResNet. Finally, the results of indoor sound source location are obtained in the test set. Experimental findings demonstrate that the neural network-based localization algorithm exhibits robust localization accuracy across diverse environmental conditions.

Finite-time adaptive NN dynamic surface control for nonstrict nonlinear FOSs subject to input dead-zone and full-states constraints

This article focuses on a kind of nonstrict nonlinear fractional-order systems (FOSs) suffering from state constraints and dead-zone input. Meanwhile, a finite-time adaptive dynamic surface control (DSC) approach based on backstepping technology and approximation principle of radial basis function neural network (RBFNN) is developed. To overcome the problem of inherent computational complexity, a fractional-order filter is applied to approach the virtual controller and its fractional-order derivative in each step of the backstepping procedure. The barrier Lyapunov function (BLF) is employed to handle the state constraints, and finite-time stability criteria on the basis of fractional-order Lyapunov method are introduced to prove the finite-time convergence of the tracking error into a small region around the origin. It is shown that all the solutions of the closed-loop system are bounded, while the state constraints are satisfied within a predetermined finite time. Finally, two examples are provided to demonstrate the effectiveness of the presented control scheme.

Neuroscience inspired neural operator for partial differential equations

We propose, in this paper, a Variable Spiking Wavelet Neural Operator (VS-WNO), which aims to bridge the gap between theoretical and practical implementation of Artificial Intelligence (AI) algorithms for mechanics applications. With recent developments like the introduction of neural operators, AI's potential for being used in mechanics applications has increased significantly. However, AI's immense energy and resource requirements are a hurdle in its practical field use case. The proposed VS-WNO is based on the principles of spiking neural networks, which have shown promise in reducing the energy requirements of the neural networks. This makes possible the use of such algorithms in edge computing. The proposed VS-WNO utilizes variable spiking neurons, which promote sparse communication, thus conserving energy, and its use is further supported by its ability to tackle regression tasks, often faced in the field of mechanics. Various examples dealing with partial differential equations, like Burger's equation, Allen Cahn's equation, and Darcy's equation, have been shown. Comparisons have been shown against wavelet neural operator utilizing leaky integrate and fire neurons (direct and encoded inputs) and vanilla wavelet neural operator utilizing artificial neurons. The results produced illustrate the ability of the proposed VS-WNO to converge to ground truth while promoting sparse communication.

PROSPECTS AND LIMITS OF ARTIFICIAL INTELLIGENCE IN THE CONTEXT OF NATURAL INTELLIGENCE (philosophical and methodological aspects) / ПЕРСПЕКТИВЫ И ПРЕДЕЛЫ ВОЗМОЖНОСТЕЙ ИСКУССТВЕННОГО ИНТЕЛЛЕКТА В КОНТЕКСТЕ ЕСТЕСТВЕННОГО ИНТЕЛЛЕКТА (философско-методологические аспекты)

Artificial intelligence is characterized as non-living intelligence, although attempts are being made to pass from the principles of biological neural networks to artificial living neural networks. The human brain is assumed to be a complex system of physicochemical neurophysiological biochemical integration of neural networks, which encodes a specific type of biological information, cognitive information. Intelligence is a specific quality of the process of thinking as an expression based on the neurophysiological mechanism of prediction. Nowadays, people more and more often speak about the anthropological, axiological crisis, modification and transformation of human nature, the creation of biological hybrids that compete with natural ones. It will hardly be possible to formalize the complexities of the human brain and mind into digital code in artificial intelligence.

Research on Kalman Filter Fusion Navigation Algorithm Assisted by CNN-LSTM Neural Network

In response to the challenge of single navigation methods failing to meet the high precision requirements for unmanned aerial vehicle (UAV) navigation in complex environments, a novel algorithm that integrates Global Navigation Satellite System/Inertial Navigation System (GNSS/INS) navigation information is proposed to enhance the positioning accuracy and robustness of UAV navigation systems. First, the fundamental principles of Kalman filtering and its application in navigation are introduced. Second, the basic principles of Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks and their applications in the navigation domain are elaborated. Subsequently, an algorithm based on a CNN and LSTM-assisted Kalman filtering fusion navigation is proposed. Finally, the feasibility and effectiveness of the proposed algorithm are validated through experiments. Experimental results demonstrate that the Kalman filtering fusion navigation algorithm assisted by a CNN and LSTM significantly improves the positioning accuracy and robustness of UAV navigation systems in highly interfered complex environments.

Estimating high spatio-temporal resolution XCO2 using spatial features deep fusion model

The high temporal-spatial resolution estimation of XCO2 data is foundational for precision quantification of carbon dioxide sources and sinks at a regional scale. This study proposed an advanced XCO2 data estimation method, using spatial features deep fusion. Leveraging convolutional neural network (CNN) principles, SpatialFusionNet—a module was designed to amalgamate geographical features within a defined spatial range. This module captures and integrates the spatial characteristics of meteorological and surface environmental factors, enhancing its application to the XCO2 estimation model. Building upon machine learning methods including Extreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), and Deep Neural Network (DNN), combined with the SpatialFusionNet module, the spatial features deep fusion models were constructed utilizing relationships among OCO-2 satellite XCO2 trajectory monitoring data in China, Copernicus Atmospheric Monitoring Service (CAMS) XCO2 reanalysis data, and meteorological factors, surface vegetation, and meteorological factors. Model performance improvements were significant, with SVM, DNN, and XGBoost showing respective RMSE reductions of 1.297 ppm, 0.480 ppm, and 0.200 ppm in ten-fold cross-validation based on OCO-2 trajectory samples. In data validation with TCCON Hefei station, the correlation between inversion data and ground-based data reached 0.85, affirming the method's high accuracy. Employing the spatial feature extraction module combined with DNN, the 2015 XCO2 annual spatial distribution of China, analyzing temporal-spatial distribution characteristics in China was generated. The DNN model, combining the SpatialFusionNet module, significantly contributes to the estimation of high temporal-spatial resolution XCO2 datasets, facilitating fine-scale quantification of regional carbon cycling.

Energy balance and synchronization of the cross-ring photosensitive neural network

After detecting different external light stimuli, photosensitive neurons encode these stimuli and trigger different discharge patterns and membrane potentials, thereby transmitting signals in the neural network. The cross-ring structure can simulate complex dynamic behaviors in the neuron system, which helps to achieve synchronous behavior between neurons. This article, respectively, uses linear resistors and induction coils to connect four photosensitive neurons and constructs a cross-ring photosensitive neural network. We examine the energy balance and synchronization stability of the system by adjusting the coupling channels and external stimuli while also estimating the impact of noise. A system composed of identical neurons achieves energy balance and complete synchronization, whereas a system composed of diverse neurons is able to achieve phase synchronization by adjusting the coupling gain. However, under magnetic field coupling, a system composed of the same chaotic neurons can only achieve energy balance and complete synchronization with a larger coupling gain. For both types of coupled systems, noise may affect the synchronization of neuron systems in chaotic states, but it has no effect on the synchronization of spike-shaped neuron systems. Resistance coupling can synchronize more quickly than magnetic field coupling, but magnetic field coupling is more susceptible to variations in the coupling gain ratio. The study will give the domains of neuroscience and artificial intelligence a theoretical groundwork as well as a deeper knowledge of the workings and governing principles of neural networks.

Machine learning based screening of organic frameworks for separation of CF4/N2, C2F6/N2, and SF6/N2

Through molecular simulation, feature analysis and extraction, as well as the modeling and optimization of machine learning algorithms, we have developed exceptional regression prediction models for the high-throughput screening of organic framework materials in CF4/N2, C2F6/N2, and SF6/N2 separation. The Grand Canonical Ensemble Monte Carlo method was employed to simulate the adsorption behavior of these three gas mixtures in 603 organic framework materials at room temperature 298 K and different pressures, constructing a dataset suitable for subsequent machine learning studies. We analyzed the impact of six common structural features of adsorbents (PLD, LCD, Density, ASA, AVF, and AV) on separation performance, determining the optimal ranges of each structural feature for adsorption separation in different gas mixtures. Additionally, we introduced two custom descriptors (AVG_SIG, AVG_SQRT_EPS) to describe adsorbent force field parameters, revealing their significant correlation with adsorption separation performance. Using pressure, adsorbent structural features, and custom descriptors as features, and TSQ value representing adsorption separation performance as the target, we applied eight machine learning models based on linear regression (MLR, RR), decision tree (DT, RF, GBDT, XGBoost), and neural network (MLP, GN) principles to model and predict on three datasets. Results indicated that simple models struggle to reliably predict adsorption separation performance, while structurally complex machine models demonstrate significant potential. We utilized the Harris Hawks Optimization (HHO) algorithm to perform hyperparameter optimization on multiple machine learning models and introduced improvements to the GN network. The optimized models, especially XGBoost and GN, exhibited outstanding performance, significantly enhancing the accuracy of predicting adsorption separation.

Big data processing and analysis platform based on deep neural network model

Users are increasingly turning to big data processing systems to extract valuable information from massive datasets as the field of big data grows. Data analytics platforms are used by e-commerce enterprises to improve product suggestions and model business processes. In order to meet the needs of large-scale data center operation and maintenance management, Internet companies often use Flink to process log data. This paper takes the big data processing and analysis platforms built by Internet financial companies and large banks as examples, and implants a stock prediction model based on Deep Neural Network (DNN). In this context, this paper completes the following work: 1) The research status of big data processing and analysis platforms at home and abroad is introduced. 2) Drawing on the modular design idea, the commercial bank big data platform is designed and the functions of each sub-module are introduced. Then the basic principle and structure of Convolutional Neural Networks (CNN) are expounded. 3) The optimal parameters of Convolutional Neural Networks are selected through experiments, and then the trained model is used for experiments. It can be seen that the stock prediction model proposed in this article has a higher prediction accuracy compared to existing models, which also verifies the validity of the proposed model. Input the data and compare the obtained results with the actual results, and finally show that the model in this paper has a good performance on stock prediction.

Theory-guided deep neural network for boiler 3-D NOx concentration distribution prediction

Timely and accurate three-dimensional (3-D) NOx concentration distribution prediction is essential for achieving low-emission and efficient operation in power plants. This study proposed a theory-guided data-driven prediction method for the 3-D NOx concentration distribution prediction. Firstly, the method created a foundational dataset by fusing numerical simulation data from the computational fluid dynamics (CFD) with operational data from the distributed control system (DCS). Then, the data was classified into three load condition categories, and the center operating conditions for each category were computed separately. Subsequently, the K-means algorithm was employed to extract representative data to address the computational challenges associated with big data. Finally, a Theory-Guided Deep Neural Network model (TG-DNN) was established leveraging the principle of carbon element mass conservation and deep neural network. Experimental results demonstrate that the method effectively monitors the 3-D NOx concentration distribution, potentially facilitating efficient production processes.

Neural Networks in Legal Theory

Abstract This article explores the domain of legal analysis and its methodologies, emphasising the significance of generalisation in legal systems. It discusses the process of generalisation in relation to legal concepts and the development of ideal concepts that form the foundation of law. The article examines the role of logical induction and its similarities with semantic generalisation, highlighting their importance in legal decision-making. It also critiques the formal-deductive approach in legal practice and advocates for more adaptable models, incorporating fuzzy logic, non-monotonic defeasible reasoning, and artificial intelligence. The potential application of neural networks, specifically deep learning algorithms, in legal theory is also discussed. The article discusses how neural networks encode legal knowledge in their synaptic connections, while the syllogistic model condenses legal information into axioms. The article also highlights how neural networks assimilate novel experiences and exhibit evolutionary progression, unlike the deductive model of law. Additionally, the article examines the historical and theoretical foundations of jurisprudence that align with the basic principles of neural networks. It delves into the statistical analysis of legal phenomena and theories that view legal development as an evolutionary process. The article then explores Friedrich Hayek’s theory of law as an autonomous self-organising system and its compatibility with neural network models. It concludes by discussing the implications of Hayek’s theory on the role of a lawyer and the precision of neural networks.

Application of Neural Network in Seismic Risk Survey

The accuracy of inversion data is determined by the reliability of linear inversion method. Therefore, the study of nonlinear seismic inversion method is a meaningful work. This paper mainly aims at the application of nonlinear neural network inversion method in Fujian Risk Survey, the following work has been accomplished: (1) The theory and algorithm of the conventional multilayer feedforward neural network and some problems in the design of the model are analyzed and summarized. Aiming at the gradient descent function search algorithm in the BP algorithm, a method for global optimization is proposed by combining simulated annealing (global search algorithm) with Conjugate gradient method (local search algorithm) , the principle and implementation of simulated annealing and conjugate gradient algorithm are analyzed in detail. (2) Based on the analysis of conventional multilayer feedforward neural networks, the model structure and basic principle of probabilistic neural networks are studied. The object function search algorithm and its realization are analyzed. The method is applied to the risk elimination prediction for the first time, and the results are satisfactory. (3) The method and principle of inversion of reservoir parameters by using various seismic attribute information and neural network technology are discussed. In order to make full use of the effective information contained in seismic signal, the method of seismic attribute extraction and optimization is studied, the principle and implementation method of attribute optimal ranking by stepwise recursive method and selecting the most important attribute combination by interactive verification method are presented.

A MOBILE APPLICATION FOR THE AUTOMATED DIAGNOSIS OF DISEASES OF AGRICULTURAL CROPS AND THE SELECTION OF RECOMMENDATIONS FOR THEIR TREATMENT

The aim of the study was a mobile development based on computer vision technologies and site parsing, which allows automating the process of diagnosing diseases of agricultural crops and issuing recommendations for treatment. The article discusses methods for recognizing plant diseases using computer vision, describes the principles of convolutional neural networks, selects the most appropriate machine learning model based on the accuracy, speed and efficiency of the model in conditions of limited resources of a mobile device, describes the tools: libraries and frameworks used for development. The detailed architecture of the application is presented, as well as the results of the developed software are demonstrated. A new contribution to the development of this topic is the experimental substantiation of the choice of a neural network model based on the analysis of its effectiveness on a prepared dataset, as well as the introduction of an automatic search for recommendations for a certain disease of agriculture. In the future, it is planned to introduce a voice assistant into this mobile application.

Quantum-Assisted Activation for Supervised Learning in Healthcare-Based Intrusion Detection Systems

Intrusion detection systems (IDS) are amongst the most important automated defense mechanisms in modern industry. It is guarding against many attack vectors, especially in healthcare, where sensitive information (patient’s medical history, prescriptions, electronic health records, medical bills/debts, and many other sensitive data points) is open to compromise from adversaries. In the big data era, classical machine learning has been applied to train IDS. However, classical IDS tend to be complex: either using several hidden layers susceptible to over-fitting on training data or using overly complex architectures such as convolutional neural networks (CNNs), long-short term memory systems (LSTMs), and recurrent neural networks (RNNs). This paper explored the combination of principles of quantum mechanics and neural networks to train IDS. A hybrid classical-quantum neural architecture is proposed with a quantum-assisted activation function that successfully captures patterns in the dataset while having less architectural memory footprint than classical solutions. The experimental results are demonstrated on the popular KDD99 dataset while comparing our solution to other classical models.