Neural Network Technology Research Articles

Neural network technologies in mining data on particle size distribution of muck pile rocks

Neural network technologies in mining data on particle size distribution of muck pile rocks

Enhancing fraud detection efficiency in mobile transactions through the integration of bidirectional 3d Quasi-Recurrent Neural network and blockchain technologies

Cases of financial fraud are increasing despite recent technical breakthroughs. It is tough to find authentic financial transaction data due to privacy issues and a lack of inter-organization synergy. However, for technologies based on data, such as machine learning to function accurately inside practical systems, they require actual data. This manuscript proposes a Fraud Detection (FD) Efficiency in Mobile Transactions through the Integration of Bidirectional 3D Quasi-Recurrent Neural Networks and Proof of Voting consensus Blockchain Technologies (Bi-3DQRNN-PoV-FD-MT). A dataset of Bitcoin transactions is employed in the suggested model. The banking sector’s Bitcoin transactions serve as the foundation for this dataset. Then the data balancing is performed using a Self-Adaptive Synthetic Over-Sampling Technique (SASOS). Then, the proposed framework utilizes the Bi-3DQRNN to determine if the information is fraudulent or not. Moreover, an intelligent Enhanced Artificial Gorilla Troops (EAGTO) Optimization algorithm is introduced to tune the weights of the model parameters. A blockchain-based proof of Voting (PoV) consensus algorithm is integrated with the proposed model for forecasting upcoming transactions. The developed scheme is instigated in Python and the performance provides 12.09%, 8.91%, and 6.92% higher accuracy compared with existing techniques like K-nearest neighbor- Distributed Blockchain Consortium (KNN-DBC), Decision tree-Ethereum blockchain-enabled smart contract (DT-EBSC) and Heterogeneous Graph Transformer Networks-Ethereum smart contract (HGTN-ESC).

Phosphoenolpyruvate carboxykinase-2 (PCK2) is a therapeutic target in triple-negative breast cancer.

Metabolic rewiring in malignant transformation is often accompanied by altered expression of metabolic isozymes. Phosphoenolpyruvate carboxykinase-2 (PCK2) catalyzes the rate-limiting step of gluconeogenesis and is the dominant isoform in many cancers including triple-negative breast cancer (TNBC). Our goal was to identify small molecule inhibitors of PCK2 enzyme activity. We assessed the impact of PCK2 down regulation with shRNA on TNBC cell growth in vitro and used AtomNet® deep convolutional neural network software to identify potential small molecule inhibitors of PCK2-based structure. We iteratively tested candidate compounds in an in vitro PCK-2 enzyme assay. The impact of the top hit on metabolic flux and cell viability was also assessed. PCK2 downregulation decreased growth of BT-549 and MDA-MB-231 cells and reduced metabolic flux through pyruvate carboxylase. The first AtomNet® in silico structural screen of 7 million compounds yielded 86 structures that were tested in PCK2 enzyme assay in vitro. The top hit (IC50 = 2.4µM) was used to refine a second round of in silico screen that yielded 82 candidates to be tested in vitro, which resulted in 45 molecules with inhibition > 20%. In the second in vitro screen we also included 3-(3,4-dihydroxyphenyl)-2-hydroxypropanoate, previously suggested to be PCK2 inhibitor based on structure, which emerged as the top hit. The specificity of this compound was tested in PCK1 and PCK2 enzymatic assays and showed IC50 of 500nM and 3.5-27nM for PCK1 and PCK2, respectively. 3-(3,4-dihydroxyphenyl)-2-hydroxypropanoate is a high affinity PCK2 enzyme inhibitor that also has significant growth inhibitory activity in breast cell lines in vitro and represents a potential therapeutic lead compound.

Риски и вызовы при внедрении искусственного интеллекта в систему высшего образования

<p style="text-align: justify;">The purpose of this article is to study possible risks and challenges in the higher education system that arise in connection with the growing speed of implementation of technologies, where artificial intelligence (AI) and neural network technologies are already part of its structure. The analysis is carried out with the aim of predicting the future nature of higher education, as well as exploring changes in the entire teaching paradigm under the influence of AI. The article identifies a number of expected problems when teaching students, in the pedagogical process, organizing their independent work, as well as in the management and administration of higher educational institutions when introducing AI capabilities into the process of higher education. The question is raised about the impact of restrictions on the development of Russian science, namely the disastrous consequences for its development due to the monopolization of AI algorithms and the subjective approach to managing a neural network by an organized technocratic group. In the course of analyzing the intersections of opposing issues of problematic issues and the positive dynamics of the influence of AI on the educational system of higher education, a vector has been set for further research on this topic.</p>

Unsupervised Social Bot Detection via Structural Information Theory

Research on social bot detection plays a crucial role in maintaining the order and reliability of information dissemination while increasing trust in social interactions. The current mainstream social bot detection models rely on black-box neural network technology, for example, Graph Neural Network, Transformer, and so on, which lacks interpretability. In this work, we present UnDBot, a novel unsupervised, interpretable, yet effective, and practical framework for detecting social bots. This framework is built upon structural information theory. We begin by designing three social relationship metrics that capture various aspects of social bot behaviors: posting type distribution , posting influence , and follow-to-follower ratio . Three new relationships are utilized to construct a new, unified, and weighted social multi-relational graph, aiming to model the relevance of social user behaviors and discover long-distance correlations between users. Second, we introduce a novel method for optimizing heterogeneous structural entropy. This method involves the personalized aggregation of edge information from the social multi-relational graph to generate a two-dimensional encoding tree. The heterogeneous structural entropy facilitates decoding of the substantial structure of the social bots network and enables hierarchical clustering of social bots. Third, a new community labeling method is presented to distinguish social bot communities by computing the user’s stationary distribution, measuring user contributions to network structure, and counting the intensity of user aggregation within the community. Compared with 10 representative social bot detection approaches, comprehensive experiments demonstrate the advantages of effectiveness and interpretability of UnDBot on 4 real social network datasets.

Artificial Intelligence Measurement of Preoperative Radiographs in Adolescent Idiopathic Scoliosis Based on Multiple-View Semantic Segmentation.

Cross-sectional study. Imaging classification of adolescent idiopathic scoliosis (AIS) is directly related to the surgical strategy, but the artificial classification is complex and depends on doctors' experience. This study investigated deep learning-based automated classification methods (DL group) for AIS and validated the consistency of machine classification and manual classification (M group). A total of 506 cases (81 males and 425 females) and 1812 AIS full spine images in the anteroposterior (AP), lateral (LAT), left bending (LB) and right bending (RB) positions were retrospectively used for training. The mean age was 13.6 ± 1.8. The mean maximum Cobb angle was 46.8 ± 12.0. U-Net semantic segmentation neural network technology and deep learning methods were used to automatically segment and establish the alignment relationship between multiple views of the spine, and to extract spinal features such as the Cobb angle. The type of each test case was automatically calculated according to Lenke's rule. An additional 107 cases of adolescent idiopathic scoliosis imaging were prospectively used for testing. The consistency of the DL group and M group was compared. Automatic vertebral body segmentation and recognition, multi-view alignment of the spine and automatic Cobb angle measurement were implemented. Compare to the M group, the consistency of the DL group was significantly higher in 3 aspects: type of lateral convexity (0.989 vs 0.566), lumbar curvature modifier (0.932 vs 0.738), and sagittal plane modifier (0.987 vs 0.522). Deep learning enables automated Cobb angle measurement and automated Lenke classification of idiopathic scoliosis whole spine radiographs with higher consistency than manual measurement classification.

Data-driven stock forecasting models based on neural networks: A review

As a core branch of financial forecasting, stock forecasting plays a crucial role for financial analysts, investors, and policymakers in managing risks and optimizing investment strategies, significantly enhancing the efficiency and effectiveness of economic decision-making. With the rapid development of information technology and computer science, data-driven neural network technologies have increasingly become the mainstream method for stock forecasting. Although recent review studies have provided a basic introduction to deep learning methods, they still lack detailed discussion on network architecture design and innovative details. Additionally, the latest research on emerging large language models and neural network structures has yet to be included in existing review literature. In light of this, this paper comprehensively reviews the literature on data-driven neural networks in the field of stock forecasting from 2015 to 2023, discussing various classic and innovative neural network structures, including Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs), Transformers, Graph Neural Networks (GNNs), Generative Adversarial Networks (GANs), and Large Language Models (LLMs). It analyzes the application and achievements of these models in stock market forecasting. Moreover, the article also outlines the commonly used datasets and various evaluation metrics in the field of stock forecasting, further exploring unresolved issues and potential future research directions, aiming to provide clear guidance and reference for researchers in stock forecasting.

Spatio-temporal analysis and prediction of land use landcover (LULC) change in Wular Lake, Jammu and Kashmir, India.

Landsat land use/land cover (LULC) data analysis to establish freshwater lakes' temporal and spatial distribution can provide a solid foundation for future ecological and environmental policy development to manage ecosystems better. Analysis of changes in LULC is a method that can be used to learn more about direct and indirect human interactions with the environment for sustainability. Neural network technology significantly facilitates mapping between asymmetric and high-dimensional data. This paper presents a methodological advancement that integrates the CA-ANN (cellular automata-artificial neural network) technique with the dynamic characteristics of the water body to forecast forthcoming water levels and their spatial distribution in "Wular Lake." We used remote sensing data from 2001 to 2021 with a 10-year interval to predict spatio-temporal change and LULC simulation. The validation of the calibration of predicted and accurate LULC maps for 2021 yielded a maximum kappa value of 0.86. Over the past three decades, the study region has seen an increase in a net change % in the impervious surface of 22.41% and in agricultural land by 52.02%, while water decreased by 14.12%, trees/forests decreased by 40.77%, shrubs decreased by 11.53%, and aquatic vegetation decreased by 4.14%. Multiple environmental challenges have arisen in the environmentally sustainable Wular Lake in the Kashmir Valley due to the vast land transformation, primarily due to human activities, and have been predominantly negative. The research acknowledges the importance of (LULC) analysis, recognizing it as a fundamental cornerstone for developing future ecological and environmental policy frameworks.

ВЛИЯНИЕ НЕЙРОСЕТЕЙ НА РАЗВИТИЕ ПРАКТИКИ УПРАВЛЕНИЯ В СОВРЕМЕННЫХ КОМПАНИЯХ

Relevance. In the modern world, neural networks are playing an increasingly significant role in various fields of activity, including management in companies. The use of neural networks in management practice can lead to significant changes in the methods of decision making, data analysis, forecasting and optimization of business processes. Therefore, it is important to explore exactly how neural networks influence the development of management practices and what new opportunities they provide for companies to improve efficiency. Goal. The purpose of this study is to study the influence of neural networks on the development of management practices in modern companies and to identify the main aspects that determine this process. Tasks: - Analysis of current trends and applications of neural networks in business management. - Study of the advantages and disadvantages of using neural networks. - Development of recommendations for the effective use of neural networks in management in modern companies. Research results: The specifics and main current trends in the development of neural networks are analyzed. Examples of their use in different companies are considered. Both the advantages and disadvantages of neural network technologies are identified. Recommendations have been developed for the effective use of neural networks in management in modern companies.

Regularity analysis of resilient modulus for hot-mix asphalt with large temperature fluctuations

To evaluate the regularity of resilient modulus for hot-mix asphalt (HMA) under large temperature fluctuations, back propagation (BP) neural network technology was used to analyze the continuous change of HMA resilient modulus. Firstly, based on the abundant data, the training model of HMA resilient modulus was established by using BP neural network technology. Subsequently, BP neural network prediction and regression analysis were performed, and the prediction model of HMA resilient modulus at different temperatures (−50 °C to 60 °C) was obtained, which fully considered multi-factor and nonlinearity. Finally, the fitted theoretical model can be used to evaluate the HMA performance under the condition of large temperature fluctuations, and the rationality of theoretical model was verified by taking Harbin region as an example. It was found that the relationship between HMA resilient modulus and temperatures can be described by inverse tangent function. And the key parameters of theoretical model can be used to evaluate the continuous change characteristics of HMA resilient modulus with large temperature fluctuations. The results can further improve the HMA performance evaluation system and have certain theoretical value.

Artificial neural network-based sparse channel estimation for V2V communication systems

Abstract Artificial neural networks (ANNs) have gained a lot of attention from researchers in the past few years and have been employed on a large scale. They have also been gaining momentum in wireless communication systems. For efficient vehicle-to-vehicle (V2V) channel communication, a sparse multipath channel issue must be studied. To minimize the multipath effect, a time reversal (TR) operation and time division synchronization orthogonal frequency division multiplexing (TDS-OFDM) have been appealing because of their fast synchronization and active spectral efficiency. To improve the transceiver's execution in a frequency-selective fading channel environment, an OFDM system is used to reduce inter- symbol interference (ISI). Simultaneous Orthogonal Matching Pursuit (SOMP) channel state estimator algorithm suffer from high computational cost and high computational complexity. The ANN algorithm has better performance than SOMP algorithm. The proposed neural network technologies have lower complexity than the SOMP algorithm. The application of ANN is capable of solving complex problems, such as those encountered in image, signal processing and have been implemented for channel estimation in OFDM. The proposed ANN outperformed the SOMP algorithm with regard to signal compensation. Overall, the ANN algorithm achieved the best performance. This study proposes an ANN-based sparse channel state estimator. Regarding the bit error rate (BER) metric, the proposed estimator outperforms the channel estimation approach based on the SOMP. The simulation results confirm the efficacy of the proposed approach.

Robust identification of users by convolutional neural network in MATLAB and Raspberry Pi

The following article presents the development of an algorithm embedded in a Raspberry Pi 3B board, where a user identification was made, using the convolutional neural network (CNN) for 5 predefined users, with the option of loading remotely a new network for a new user. Comparatively, the same application was programmed in MATLAB programming software to evaluate the results and identify the advantages between them. Networks were trained for 5 different users, using the Caffe library on the Raspberry Pi, and the MATLAB neural network package on the computer. Where it was found that the training made by Caffe on an embedded system is much slower and less efficient than the ones performed in MATLAB, obtaining less than 55% accuracy with Caffe networks and more than 90% with MATLAB networks, training with the same number of samples, the same architecture, and the same database. Finally, the accuracy obtained through confusion matrix is over 88% in each case of users identification.

A bibliometric analysis of semantic segmentation in remote sensing images from 1981 to 2024

Semantic segmentation of remote sensing images plays a crucial role in various fields, such as disaster assessment, forestry measurement, and environmental monitoring. This study conducts a bibliometric analysis to investigate the research trends and development of semantic segmentation in remote sensing from 1981 to 2024. Utilizing the Web of Science Core Collection database, we retrieved representative literature data and analyzed the publication trends, main source journals, institutional cooperation, and keyword co-occurrence networks. The results show that the number of publications in this field has increased significantly since 2000, with an exponential growth pattern. Remote sensing is the most active research topic, followed by imaging science, photographic technology, and electrical and electronic engineering. The analysis of institutional cooperation reveals that Chinese research institutions, such as the Chinese Academy of Sciences and Wuhan University, play a leading role in this field and maintain close collaboration with research institutions worldwide. The keyword co-occurrence network analysis demonstrates the evolution of research focus from basic remote sensing and image processing techniques in the early stage to advanced image processing algorithms in the middle stage and deep learning and neural network technologies in the recent stage. The findings provide valuable insights into the current status and future trends of semantic segmentation in remote sensing, highlighting the importance of international cooperation and interdisciplinary exchange in driving technological innovation.

Inverse compensation mechanism-based adaptive fuzzy-neural fault-tolerant control for an uncertain quadrotor UAV

Actuator faults, system uncertainties, and external disturbances can affect tracking control performance of the quadrotor UAV, and even cause the instability of the system. To address this issue, an adaptive fuzzy-neural fault-tolerant control (AFNFTC) scheme is proposed for the quadrotor UAV integrated with inverse compensation strategy and fuzzy neural networks (FNNs) technology. Firstly, we design an inverse compensation mechanism to address actuator faults and utilize the approximation capabilities of the FNNs to compensate for system uncertainties. Then, based on the Lyapunov direct method, the AFNFTC scheme is designed to ensure the stability of the quadrotor UAV. Under the proposed scheme, the closed-loop system for the quadrotor UAV is proven to be uniformly bounded, and the tracking error of the quadrotor UAV is demonstrated to converge to a small compact set ultimately by appropriately selecting design parameters. Finally, the feasibility and effectiveness of the proposed scheme is validated through simulation results.

Потенциал использования технологии искусственного интеллекта при обучении физике в школе

This article examines the problem of using the potential of neural network technologies available in Russia for enhancing secondary school physics education. The relevance of the addressed issue is justified through an analysis of regulatory documents shaping Russia’s economic and educational strategies, scientific research in artificial intelligence (AI) technologies, and their application to improve educational processes. In addition, сurrent challenges in teaching the subject “Physics” are also considered. Results from a mini-study surveying physics teachers’ attitudes towards the adoption of new tools – AI technologies – are presented. The range of opportunities afforded by neural networks open up for both teachers and students is revealed. Сhallenges and risks that should be considered and necessary actions taken to minimize them are outlined. Fragments of an additional professional program (advanced training) “The use of Russian neural networks in the school educational process”, as well as fragments of practical work performed by physics teachers who attended the course, are presented. The conclusions of this article emphasize the methodological component of the application of AI technologies.

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.

BPP: a platform for automatic biochemical pathway prediction.

A biochemical pathway consists of a series of interconnected biochemical reactions to accomplish specific life activities. The participating reactants and resultant products of a pathway, including gene fragments, proteins, and small molecules, coalesce to form a complex reaction network. Biochemical pathways play a critical role in the biochemical domain as they can reveal the flow of biochemical reactions in living organisms, making them essential for understanding life processes. Existing studies of biochemical pathway networks are mainly based on experimentation and pathway database analysis methods, which are plagued by substantial cost constraints. Inspired by the success of representation learning approaches in biomedicine, we develop the biochemical pathway prediction (BPP) platform, which is an automatic BPP platform to identify potential links or attributes within biochemical pathway networks. Our BPP platform incorporates a variety of representation learning models, including the latest hypergraph neural networks technology to model biochemical reactions in pathways. In particular, BPP contains the latest biochemical pathway-based datasets and enables the prediction of potential participants or products of biochemical reactions in biochemical pathways. Additionally, BPP is equipped with an SHAP explainer to explain the predicted results and to calculate the contributions of each participating element. We conduct extensive experiments on our collected biochemical pathway dataset to benchmark the effectiveness of all models available on BPP. Furthermore, our detailed case studies based on the chronological pattern of our dataset demonstrate the effectiveness of our platform. Our BPP web portal, source code and datasets are freely accessible at https://github.com/Glasgow-AI4BioMed/BPP.

Navigating the biomechanical landscape: Enhanced methods for drug analysis

In drug analysis and detection, chemiluminescence is a standard technical method based on the idea that the concentration of observed substances in the chemical detection system and the chemical luminescence intensity establish a linear quantitative relationship under certain conditions. Bio-mechanical landscape (BML), which uses bioengineered systems to simulate essential elements of the human tumour microenvironment, is already enhancing cancer biology and aiding clinical translation. The majority of illicit substances have dangerous cardiovascular side effects that might range from an irregular heartbeat to a heart attack. Cardiovascular disease, including blocked veins and bacterial infections of the blood arteries and heart valves, may also result from injecting illicit narcotics. The challenging characteristics of such a cardiovascular disease using drug analysis are heart failure, low blood pressure and headaches. Hence, in this research, drug analysis has improved wearable body sensor network-enabled artificial neural network (WBSN-ANN) technologies for cardiovascular disease. Medicine and drug development because they provide information on patient status in the clinical context and vital information on pharmacodynamics activity, effectiveness, and safety throughout the research process. The process results in a research-intensive organization finding a novel chemical or biological compound and developing it into a medicine that patients can legally utilize for drug discovery and development.

Fingerprint Identification Method Based on Convulsional Neural Networks

The article presents an advanced method of fingerprint identification based on convolutional neural network (CNN) technology. This work elaborately describes the development and implementation process of a specialized CNN architecture for detecting and verifying the authenticity of fingerprints. Utilizing the comprehensive Socofing dataset allowed for an in -depth analysis of the model’s ability to distinguish between genuine and fabricated fingerprints, where the model demonstrated impressive accuracy – up to 98.964%. Special attention is given to error analysis, including the false discovery and omission rates, pointing towards potential directions for further improvement. Besides highlighting the technical aspects and high identification accuracy, the article also addresses potential challenges and limitations that the method might encounter. This includes issues related to the imbalance and diversity of data in the Socofing set, as well as limitations associated with computational resources when training deep neural networks. Potential pathways for model optimization are discussed, particularly focusing on reducing the false omission rate, which could improve user experience in authentication. The concluding section of the article emphasizes the importance of the presented work for the security sector, where precise authentication of fingerprint images is critically needed. The obtained results can be considered a solid foundation for future scientific developments in this direction. Additionally, the need for systematic updates and modifications of the model is highlighted to adapt it to continually improved imitation techniques, ensuring its long-term relevance and effectiveness.

Neuroadaptive event-triggered tracking control for nonlinear systems with dynamic fault feedback and prescribed time convergence

This paper investigates the problem of event-triggered adaptive prescribed time tracking control for nonlinear systems with combining multiplicative and additive dynamic sensor fault. A simple state observer is designed by fusing fault output and reference signal to estimate unmeasurable states, where the unknown unmeasurable uncertain functions are approximated by neural network technology. Then, the prescribed time performance function is constructed based on the time scale function, and a suitable function transformation is introduced so that the tracking error can reach the desired residual set in the prescribed time, where the initial value of the error can be arbitrarily selected regardless of the prescribed time performance function. Moreover, the desired accuracy and the prescribed time can be specified on request. The designed adaptive event-triggered controller can effectively compensate for the sensor fault in the output feedback control and reduce the communication burden, and the feasibility of the proposed method is verified by the simulation result.