Network Method Research Articles

Mapping Anti-Refugee Attitude Networks in the Classroom: Examining the Relationship Between Migrant Children’s Attitude Alignment and National Identification

ABSTRACT Irish schools have been rapidly diversifying since the early 2000s. As increased diversity often leads to increased prejudicial attitudes, children with migration backgrounds may struggle to gain acceptance and a sense of belonging as Irish. There are several strategies migrant children may take in response to prejudicial attitudes they perceive in their classroom, including adopting prejudicial attitudes toward other migrant groups – such as refugees – to align with the majority, or distancing themselves from the negative attitudes of the majority at the expense of national identification. We examine these strategies using a novel network method which identifies the positions of each child in relation to others in their classroom based on their attitude similarity. The study uses data from 54 primary school classes (N = 969 children). We find that migrant children were not identifiable by distinct patterns of anti-refugee attitudes compared to those of Irish national children. National identification was not related to the alignment of attitudes to Irish peers. Alignment of attitudes toward one’s classmates, however, was associated with higher identification with the class, for both migrant and non-migrant children. These findings highlight the class context as a relevant environment in which attitudes and identity are jointly navigated.

Identification of Cactus Species Using the MobileNet Convolutional Neural Network Model

The objective of this design is to create a system capable of accurately and precisely classifying different types of succulent cacti. The system design aims to create a succulent cactus classification application using a Convolutional Neural Network (CNN) with the MobileNet architecture. The process involves collecting cactus images, dividing them into training and test datasets, and developing a CNN model to recognize patterns in succulent cactus species. The training data is used to train the CNN model, and the test data evaluates the model's accuracy. The trained model, stored in Tf.lite format, successfully classifies 15 cactus types, achieving high accuracy by employing preprocessing steps like resizing, normalization, and background removal. Over 1,200 cactus images were taken with a smartphone, categorized into 15 classes, and prepared to ensure optimal lighting, angle, background, and resolution (224x224 pixels). The MobileNet model was chosen for its high accuracy and efficiency. Hardware used includes a Samsung A54 smartphone and an Intel i7 laptop, with software such as Python, Kotlin, and Android Studio facilitating development. This design ensures the application’s accessibility, making it a valuable tool for cactus enthusiasts and the general public to easily identify different succulent cactus types. Testing of the cactus species classification program using the Convolutional Neural Network (CNN) method with the MobileNetV2 architecture yielded strong results, achieving over 90% accuracy in classifying 15 cactus species. The highest training accuracy of 0.9837 was achieved at 150 epochs without early stopping, outperforming other epoch configurations. The model successfully classified species across five main genera—Kalanchoe, Crassula, Echeveria, Haworthia, and Euphorbia. This high accuracy highlights the model's effectiveness, making it a useful tool for cactus enthusiasts and the public to accurately identify and distinguish cactus species.

Optimizing intelligent penetration path planning using reinforcement learning: A focus on valid action masking and sample enhancement

Abstract. As penetration testing is a crucial technique for the rapid evolution of cybersecurity to identify system vulnerabilities, traditional penetration testing methods face the difficulty of large amounts of manual labor and professional knowledge, which makes them difficult to scale. This paper proposes a highly automated and efficient penetration testing approach based on the MASK-SALT-DQN algorithm. MASK-SALT-DQN is a reinforcement learning model that effectively optimizes penetration path planning by valid action masking and sample enhancement. Through valid action masking, the MASK-SALT-DQN model filters out redundant and invalid actions in the solution space, reducing the complexity of the action space and improving the convergence speed and efficiency. In addition, the sample enhancement method increases the frequency of critical exploitation actions and positive rewards in the sparse-reward environment, speeding up the learning process of the agent. Experiments have been conducted on the NASim attack simulation network to demonstrate the advantages of the proposed model over the baseline method. The experimental results show that our model has a significant performance advantage over baseline methods in large networks, and it can be applied to the field of scalable and efficient automated penetration testing.

Injection locking in DC-driven spintronic vortex oscillators via surface acoustic wave modulation

Control of the microwave signal generated by spin-transfer torque oscillators (STOs) is crucial for their applications in spin wave generation and neuromorphic computing. This study investigates injection locking of a DC-driven vortex STO using surface acoustic waves (SAWs) to enhance the STO’s signal and allow for its synchronization with external inputs. We employ a simplified model based on Thiele’s formalism and highlight the role of vortex deformations in achieving injection locking. Micromagnetic simulations are conducted to validate our theoretical predictions, revealing how the locking bandwidth depends on SAW amplitude, as well as on the amplitude and direction of an applied external field. Our findings are pivotal for advancing experimental research and developing efficient low-power synchronization methods for large-scale STO networks.

Piecewise Neural Network Method for Solving Large Interval Solutions to Initial Value Problems of Ordinary Differential Equations

Piecewise Neural Network Method for Solving Large Interval Solutions to Initial Value Problems of Ordinary Differential Equations

A Lightweight and Efficient Multimodal Feature Fusion Network for Bearing Fault Diagnosis in Industrial Applications

To address the issues of single-structured feature input channels, insufficient feature learning capabilities in noisy environments, and large model parameter sizes in intelligent diagnostic models for mechanical equipment, a lightweight and efficient multimodal feature fusion convolutional neural network (LEMFN) method is proposed. Compared with existing models, LEMFN captures rich fault features at multiple scales by combining time-domain and frequency-domain signals, thereby enhancing the model’s robustness to noise and improving data adaptability under varying operating conditions. Additionally, the convolutional block attention module (CBAM) and random overlapping sampling technology (ROST) are introduced, and through a feature fusion strategy, the accurate diagnosis of mechanical equipment faults is achieved. Experimental results demonstrate that the proposed method not only possesses high diagnostic accuracy and rapid convergence but also exhibits strong robustness in noisy environments. Finally, a graphical user interface (GUI)-based mechanical equipment fault detection system was developed to promote the practical application of intelligent fault diagnosis in mechanical equipment.

A study on Modeling and Simulation of Sports Training Injury Optimization from a Biomechanical Perspective

With the development of artificial intelligence technology today, sports training is one of the important ways to keep people healthy. Reasonable sports training can keep people happy, and can also improve the physical quality of athletes. However, because the intensity of sports training is greater than the physical endurance, it may lead to physical injury. The sports training injury model can collect the physical data of athletes, and simulate whether the actions that athletes want to practice are standard and correct through the collected data. If there are errors in the simulated exercise actions, the wrong exercise actions can be corrected in time, so as to reduce the incidence of sports injury accidents. The construction of the experimental model and the calculation and analysis process of the construction data are very complex, and the existing simulation models are difficult to achieve this, which has a significant impact on the subsequent development of the simulation model. In view of this phenomenon, this paper, on the basis of the sports training injury model, combined with the neural network method, conducted an effective research on the construction of the sports training injury model, and inspected the performance of the model by testing the accuracy of the evaluation of the injury risk level, the degree of sports training injury, and the accuracy of the evaluation of the prediction results of the sports training injury model. The experimental data shows that the maximum accuracy of the sports training injury model in the horizontal and vertical directions was 94.43% and 95.26% respectively, and the maximum accuracy of the single injury degree and the composite injury degree was 48.68% and 55.01% respectively, which had high accuracy and operational efficiency, and can effectively avoid the occurrence of training injury.

Task-oriented EEG denoising generative adversarial network for enhancing SSVEP-BCI performance

Objective.The quality of electroencephalogram (EEG) signals directly impacts the performance of brain-computer interface (BCI) tasks. Many methods have been proposed to eliminate noise from EEG signals, but most of these methods focus solely on signal denoising itself, disregarding the impact on subsequent tasks, which deviates from the original intention of EEG denoising. The main objective of this study is to optimize EEG denoising models with a purpose of improving the performance of BCI tasks.Approach.To this end, we proposed an innovative task-oriented EEG denoising generative adversarial network (TOED-GAN) method. This network utilizes the generator of GAN to decompose and reconstruct clean signals from the raw EEG signals, and the discriminator to learn to distinguish the generated signals from the true clean signals, resulting in a remarkable increase of the signal-to-noise ratio by simultaneously enhancing task-related components and removing task-irrelevant noise from the original contaminated signals.Main results.We evaluated the performance of the model on a public dataset and a self-collected dataset respectively, with canonical correlation analysis classification tasks of the steady-state visual evoked potential (SSVEP) based BCI. Experimental results demonstrate that TOED-GAN exhibits excellent performance in removing EEG noise and improving performance for SSVEP-BCI, with accuracy improvement rates reaching 18.47% and 21.33% in contrast to the baseline methods of convolutional neural networks, respectively.Significance.This work proves that the proposed TOED-GAN, as an EEG denoising method tailored for SSVEP tasks, contributes to enhancing the performance of BCIs in practical application scenarios.

Intelligent Pressure Monitoring Method of BP Neural Network Optimized by Genetic Algorithm: A Case Study of X Well Area in Yinggehai Basin

While drilling formation pressure monitoring is an important basis for ensuring drilling safety and oil and gas discovery, the calculation of existing pressure monitoring methods is complicated and the accuracy is difficult to improve. Taking the actual well data of well area X in Yinggehai Basin as the object, correlation analysis was first carried out to select and standardize the data features, and relevant effective parameters were extracted. Two kinds of neural networks, back-propagation network BP and back-propagation network GA-BP optimized by genetic algorithm, were used to establish artificial intelligence monitoring models of formation pressure based on 10 kinds of measuring and logging data, respectively. The application effect of the model was evaluated based on the results of monitoring the pressure while drilling. The results show that the monitoring accuracy of the BP neural network model is 91.25%, and that of the GA-BP neural network model is 92.89%. The latter has a better monitoring effect on formation pore pressure. In formation pressure monitoring in areas with a high degree of well control, the introduction of artificial intelligence technology has the advantages of simplicity, speed and high precision, and can provide a reference for other areas of pressure monitoring while drilling.

Short-term Load Forecasting using Genetic Algorithm based Artificial Neural Network

The electrical load Forecasting is considered one of the main and important points for the planning for the future in the electric power networks in both of transmission and distribution systems. In this paper, the genetic optimization algorithm (GA) with the artificial intelligent network (ANN) methods are proposed to forecast the electrical load of the building of Electrical Power Distribution office in the Iraq /Anbar Governorate. Moreover, the optimization method of GA is presented to train the ANN weights using Matlab software with simple programming code. Then, the proposed method is applied to the actual load and the results are compared with that obtained in traditional method such as ANN and regression trees (RG trees) methods. Hence, from the simulation results and the comparison, we can confirm that the proposed method is better than the traditional methods, and it achieved a good agreement with actual load and it provide a good total RMSE of 6% when compared with other methods

Bridging organ transcriptomics for advancing multiple organ toxicity assessment with a generative AI approach

Translational research in toxicology has significantly benefited from transcriptomic profiling, particularly in drug safety. However, its application has predominantly focused on limited organs, notably the liver, due to resource constraints. This paper presents TransTox, an innovative AI model using a generative adversarial network (GAN) method to facilitate the bidirectional translation of transcriptomic profiles between the liver and kidney under drug treatment. TransTox demonstrates robust performance, validated across independent datasets and laboratories. First, the concordance between real experimental data and synthetic data generated by TransTox was demonstrated in characterizing toxicity mechanisms compared to real experimental settings. Second, TransTox proved valuable in gene expression predictive models, where synthetic data could be used to develop gene expression predictive models or serve as “digital twins” for diagnostic applications. The TransTox approach holds the potential for multi-organ toxicity assessment with AI and advancing the field of precision toxicology.

A multi-task learning based line parameter identification method for medium-voltage distribution network

Accurate line parameters are critical for and dispatch in distribution systems. External operating condition variations affect line parameters, reducing the accuracy of state estimation and power flow calculations. While many methods have been proposed and obtained results rather acceptable, there is room for improvement as they don’t fully consider line connections in known topologies. Furthermore, inaccuracies in measurement devices and data acquisition systems can introduce noise and outliers, impacting the reliability of parameter identification. To address these challenges, we propose a line parameter identification method based on Graph Attention Networks and Multi-gate Mixture-of-Experts. The topological structure of the power grid and the capabilities of modern data acquisition equipment are utilized to capture. We also introduce a multi-task learning framework to enable joint training of parameter identification across different branches, thereby enhancing computational efficiency and accuracy. Experiments show that the GAT-MMoE model outperforms traditional methods, with notable improvements in both accuracy and robustness.

Combining long short-term memory and shape superposition for estimating moving load-induced global responses of bridges

Recently, a method of estimating the global responses of bridges was introduced, by combining a deep neural network (DNN) and shape superposition method (SSM). This DNN-SSM model demonstrated superior performance to the traditional estimation method based on the least-squares method. However, the application and validation of this approach have been restricted to static problems. For bridges, the dynamic effect induced by the moving load requires consideration. Therefore, this study proposes a long short-term memory (LSTM)-SSM model to improve the estimation performance for dynamic global responses. The LSTM-SSM model consists of LSTM layers, followed by fully connected DNN layers and a post-process part based on SSM. Limited displacement, slope, and strain data were used as the inputs, and the global displacement, slope, and strain were estimated. To validate the effectiveness and improvement of the LSTM-SSM model, a numerical model of a cable-stayed bridge was used to compare the proposed model with the previously developed DNN-SSM model. The validation results indicated that the LSTM-SSM model can provide more accurate estimates of the dynamic global response than DNN-SSM models. Finally, a technique for structural safety monitoring based on the proposed method was introduced.

Pemanfaatan Metode Convolutional Neural Network (CNN) Dengan Arsitektur MobileNetV2 Untuk Penilaian Kelayakan Rumah

In the digitalization era, efforts to enhance efficiency and accuracy in assessing household eligibility for social assistance have become important. We plan to develop a household eligibility assessment system using convolutional neural network methods. Our system uses artificial intelligence to analyze uploaded house photos and provide eligibility assessments based on predefined criteria. The goal is to automate the assessment process, increase efficiency and accuracy, ensure social assistance is well-targeted, and reduce administrative workload. Through the implementation of this system, we hope to improve the effectiveness of social assistance and community development. The result of this research is a household eligibility assessment program that can be implemented in the future.

Use of ResNets for HLB Disease Detection on Orange Leaves Using Terrestrial Multispectral Images

Abstract. Huanglongbing (HLB) is a bacterial disease transmitted by different vectors of sap-sucking insects. It affects all crops of citrus trees, decreasing the values of those fruits in the market and eventually the decay of orchards. In Brazil, the world's leading orange producer, citriculture faces severe issues with HLB and substantial economic loss. Technical means of scanning the orchards with high-throughput becomes essential for the sustainability of this industry. In this study, we propose to investigate an operational strategy consisting of scanning large portions of foliage (the canopy of one tree or more) in which there can few early foliage symptoms. It is proposed to investigate deep learning tools to solve this complex binary classification problem. The study is based on a dataset comprising 1,297 terrestrial multispectral (14 channels) images captured at high spatial resolution in a commercial orange orchard in Brazil. It is proposed to adapt and retrain standard neural network architectures, namely ResNets18 and ResNets34, to process such images. Our analysis reveals promising results, with both models demonstrating convergence and achieving stable performance. Notably, ResNet18 outperformed ResNet34, achieving an accuracy of 76.45% compared to 66.79% from ResNet34. These findings suggest that deep neural network methods can effectively manage non-radiometrically calibrated data and accurately distinguish images with HLB symptoms from healthy plants . However, with reduced datasets and limited possibilities for transfer learning and fine-tuning, it seems that only reasonable sized networks can be trained. Thus, more advanced state-of-the-art tools of the are still challenging to deploy for agricultural multi- or hyperspectral data.

Renal injury in NSAIDs: a real-world analysis based on the FAERS database.

This study aims to evaluate the reporting risk of renal injury associated with non-steroidal anti-inflammatory drugs (NSAIDs), with a particular focus on the reporting risk levels and onset times of different NSAIDs. A pharmacovigilance study was conducted using data from the FAERS database from January 2004 to December 2023. Reports of renal injury were identified, and signal detection was performed using reporting odds ratio (ROR) and Bayesian confidence propagation neural network (BCPNN) methods. The study compared the incidence, mortality rates, and onset times of renal injury across five NSAIDs. Among the 7436 cases of NSAID-associated renal injury analyzed, elderly patients are at an increased risk of renal injury associated with NSAID usage. Ibuprofen had the highest number of reports (3475 cases, 46.7%), while celecoxib had the lowest (542 cases, 7.3%). Ibuprofen showed the highest signal with renal injury (ROR 3.3, IC025 1.7), whereas celecoxib exhibited the lowest (ROR 1.4, IC025 0.4). Aspirin had the highest mortality rate associated with renal injury (18.7%), while ibuprofen had the lowest (3.8%). The median onset time for renal injury was 6days, with 79.3% of adverse events occurring within the first 30days of use. The study indicates that ibuprofen presents the highest signal of renal injury, while celecoxib shows the lowest signal. The likelihood of NSAID-associated renal injury is heightened in elderly patients, and all five studied NSAIDs are linked to an increased likelihood of acute renal injury. NSAID-related renal damage tends to occur early in the treatment process, potentially leading to serious consequences. Due to the inherent limitations of pharmacovigilance studies, certain findings require additional validation like cohort studies. Nonetheless, the potential for an increased risk of renal injury must be taken into account in patient care.

Exact results on the dynamics of the stochastic Floquet-East model

Abstract We introduce a stochastic generalisation of the classical deterministic Floquet-East model, a discrete circuit with the same kinetic constraint as the East model of glasses. We prove exactly that, in the limit of long time and large size, this model has a large deviation phase transition between active and inactive dynamical phases. We also compute the finite time and size scaling of general space-time fluctuations, which for the case of inactive regions gives rise to dynamical hydrophobicity. We also discuss how, through the Trotter limit, these exact results also hold for the continuous-time East model, thus proving long-standing observations in kinetically constrained models. Our results here illustrate the applicability of exact tensor network methods for solving problems in many-body stochastic systems.

GradToken: Decoupling tokens with class-aware gradient for visual explanation of Transformer network

Transformer networks have been widely used in the fields of computer vision, natural language processing, graph-structured data analysis, etc. Subsequently, explanations of Transformer play a key role in helping humans understand and analyze its decision-making and working mechanism, thereby improving the trustworthiness in its real-world applications. However, it is difficult to apply the existing explanation methods for convolutional neural networks to Transformer networks, due to the significant differences between their structures. How to design a specific and effective explanation method for Transformer poses a challenge in the explanation area. To address this challenge, we first analyze the semantic coupling problem of attention weight matrices in Transformer, which puts obstacles in providing distinctive explanations for different categories of targets. Then, we propose a gradient-decoupling-based token relevance method (i.e., GradToken) for the visual explanation of Transformer’s predictions. GradToken exploits the class-aware gradient to decouple the tangled semantics in the class token to the semantics corresponding to each category. GradToken further leverages the relations between the class token and spatial tokens to generate relevance maps. As a result, the visual explanation results generated by GradToken can effectively focus on the regions of selected targets. Extensive quantitative and qualitative experiments are conducted to verify the validity and reliability of the proposed method.

Data-driven prediction of higher-order structure in carbon nanotube films for high thermoelectric power factor

We optimized the higher-order structures and semiconducting purity of single-walled carbon nanotubes (SWCNTs) to enhance the thermoelectric power factor PF by combining the thermoelectric random stick network (TE-RSN) method and a genetic algorithm. The PF of the optimized films was increased approximately fivefold for initial random structures. In addition, while the random structures showed the maximum PF when the ratio of semiconducting to metallic SWCNTs R s exceeded 0.98, the optimized structures converged to an R s of approximately 0.9. The optimized structures exhibited an increased local density and the peak of alignment angle distribution, leading to an increase in both the electrical conductivity and the Seebeck coefficient. We interpreted the increase in the Seebeck coefficient using a serial model. The results indicated that the reduction in the number of contacts within the paths and the subsequent increase in temperature difference on semiconducting SWCNTs led to the increase in the Seebeck coefficient.

Analyzing tube arrangements of a finned-tube heat exchanger to optimize overall efficiency using artificial neural network and response surface methodology

Temperature is a critical factor in numerous equipment, industries, and daily life applications. Heat exchangers are essential devices that help regulate and optimize this factor. The finned-tube heat exchanger (FTHE) is widely favored due to its high efficiency in facilitating heat transfer between liquids and gases. Improving the performance of FTHE can significantly provide thermal requirements in industrial and engineering processes and reduce energy consumption. The present research includes a numerical study of an FTHE and the optimization of the performances based on the input variables by response surface methodology (RSM) and artificial neural network (ANN) methods. The tubes' transverse and longitudinal pitches and inlet Reynolds number were selected as input variables. Also, the examined responses were the Colburn and friction factors. Changing tubes' pitches makes it possible to significantly affect the thermo-hydraulic performance without incurring additional costs and processes. The acquired results showed that the responses predicted by the models are very close to the numerical results, which indicates the high accuracy and validity of these models. According to the results, the optimum heat exchanger's efficiency index was obtained at Pt=26.128mm, Pl=28mm, and Re=800. It was also observed that the overall performance of the optimal design is 185% higher than the weakest FTHE.