Network Update Research Articles

AoI Optimization in Multi-Source Update Network Systems Under Stochastic Energy Harvesting Model

AoI Optimization in Multi-Source Update Network Systems Under Stochastic Energy Harvesting Model

Finite-time robust control of morphing aircraft based on time-varying observer

Morphable vehicles have strong nonlinear, strong coupling and strong time-varying characteristics in the process of variation, which brings great challenges to the design of their flight control systems. To address this problem, a time-varying gain extended state observer is proposed to accurately approximate the system coupling terms. Combining adaptive backstepping control technology and finite-time control theory, a finite-time bounded robust adaptive controller is designed. By designing the barrier Lyapunov function, the actual control law and the adaptive network update law are obtained step by step, ensuring that the system command tracking error can converge to a predetermined range within a finite time. Finally, the effectiveness of the designed control system is verified by attitude control simulation. The simulation results show that the morphable vehicle can track the command signal well in the process of variation and is basically not affected by the variation rate.

CLlight: Enhancing representation of multi-agent reinforcement learning with contrastive learning for cooperative traffic signal control

Multi-agent reinforcement learning has shown great potential for coordinating multi-intersection traffic signals due to its powerful adaptive capabilities, treating each intersection as an agent. However, in the real world, different intersections possess differentiating characteristics such as unique vehicle distributions and traffic patterns. Most existing methods directly add neighboring intersection states to local intersections and optimize the cooperative policy network based on synthesized global features. This indirect optimization approach makes it difficult to thoroughly explore the mutual interactions among different intersection agents, preventing agents from truly learning features with cooperative awareness. To resolve these challenges, we introduce contrastive learning as representation task to the multi-intersection traffic signal control approach named CLlight for two-stage policy network updating. In the first stage, we utilize policy-based or actor-critic-based reinforcement learning methods such as A2C, SAC, and PPO to train policy networks with certain representational capabilities. In the second stage, by extracting pre- and post-masked features and reconstructing the post-masked features, the agents are encouraged to learn the similarities and differences between different intersection policies, which in turn enhances the cooperative and individual representation capabilities of the policy network. To the best of our knowledge, this is the first application of contrastive learning in the field of traffic signal control. Experimental results demonstrate, compared to other state-of-the-art traffic signal control methods, superior average travel time and average waiting time performance under various scenarios, tested on synthetic and real-world datasets.

AOCBLS: A novel active and online learning system for ECG arrhythmia classification with less labeled samples

Electrocardiogram (ECG) is a pivotal determinant of cardiac arrhythmia. In practice, ECG data are often acquired as continuous unlabeled chunks with high labeling costs and severe class imbalances. This poses formidable challenges to existing offline learning methods. To address these issues, we propose a novel model called active online class-specific broad learning system (AOCBLS) to learn streaming imbalanced ECG and classify arrhythmias. First, a two-stage active learning strategy combining diversity and uncertainty queries was designed to select the most valuable beats from the current chunk for further labeling and learning, thereby reducing labeling costs. Subsequently, a stable training weight update rule was developed to address the dynamic imbalances in the streaming data. Finally, an incremental learning algorithm was proposed to update network weights online using selected beats, eliminating the need for retraining from scratch. Thus, the AOCBLS can effectively learn online from imbalanced data streams with less label information. Intra-patient experiments on the benchmark MIT-BIH arrhythmia database showed our model achieved an overall accuracy of 99.10% and a G-mean of 96.35% in identifying the four AAMI-recommended beat types, with only 10% additional annotation, outperforming advanced methods at lower labeling cost. Additionally, patient-specific experimental results on the benchmark database and two other databases, INCARTDB and SVDB, suggest the potential of AOCBLS for personalized ECG monitoring on portable devices.

DOUNet: Dynamic Optimization and Update Network for Oriented Object Detection

DOUNet: Dynamic Optimization and Update Network for Oriented Object Detection

Upgoing and downgoing wavefield separation in VSP data using CGAN based on asymmetric convolution blocks

Abstract Accurately upgoing and downgoing wavefield separation is a critical step in vertical seismic profile (VSP) data processing, as its accuracy is directly related to the imaging quality of VSP data. Traditional methods are based mainly on transforms, and their windows are manually set in the transformed domain to obtain the target wavefield. The manual operations often cause errors and affect the accuracy of wavefield separation. In contrast, deep learning algorithms are more automatic, and have achieved a lot in seismic data processing. We propose to employ a conditional generative adversarial network (CGAN) for wavefield separation in VSP data. A CGAN consists of two main components: a generator, which generates new data samples, and a discriminator, which evaluates the generated samples against real data, with both components trained simultaneously in an adversarial manner to improve the quality of generated samples. The full wavefield serves as a constraint to link the generator and discriminator, ensuring that the separated up- and downgoing wavefields align better with the full wavefield. An asymmetric convolution block is introduced to more effectively capture the directional features of the VSP wavefield. To mitigate the influence of amplitude differences between the waves on the network update, the relative downgoing wavefield (obtained by subtracting the predicted upgoing wavefield from the full wavefield) is included in the loss function. Numerical experiments demonstrate that the trained network can effectively learn the characteristics of the up- and downgoing waves, especially their propagation directions, achieving high-precision wavefield separation.

Comparative efficacy of different single drugs to prevent necrotizing enterocolitis in preterm infants: an update systematic review and network meta-analysis.

To investigate an optimal regimen of six drugs, including lactoferrin, probiotics, prebiotics, glutamine, arginine and erythropoietin (EPO), for the prevention of necrotizing enterocolitis (NEC) in preterm infants. PubMed, Embase, Ovid, The Cochrane Library, and Web of Science databases were searched for randomized controlled trials (RCTs) investigating the efficacy of lactoferrin, probiotics, prebiotics, glutamine, arginine, and EPO in preventing NEC in preterm infants, with a cutoff date of June 20, 2024. Two authors independently screened studies and extracted all the data. Network meta-analysis (NMA) was conducted to compare the outcomes of different interventions, and group rankings were determined using the surface under the cumulative ranking curve (SUCRA). A total of 89 RCTs with 26,861 preterm infants were included. Arginine demonstrated the highest clinical efficacy in reducing the incidence of NEC, with probiotics being the next most effective and the placebo being the least effective. Lactoferrin was identified as the most effective intervention for reducing the incidence of NEC-associated sepsis. Prebiotics showed the highest effect on overall mortality, reducing the beginning of enteral feeding, and were associated with the shortest hospital stay. Glutamine significantly decreased the time to full enteral feeding. Existing literature highlights arginine as the most efficacious pharmacological agent in preventing NEC in preterm infants. It has been shown to effectively lower the rates of NEC, septicemia, and mortality, warranting its recommendation as the first-line clinical intervention. Following this, probiotics are recommended as a second option.

Linear Model Predictive Control and Back-Propagation Controller for Single-Point Magnetic Levitation with Different Gap Levitation and Back-Propagation Offline Iteration

Magnetic suspension balance systems (MSBSs) need to allow vehicle models to levitate stably in different attitudes, so it is difficult to ensure the stable performance of the system under different levitation gaps using a controller designed with single balance point linearization. In this paper, a levitation controller based on linear model predictive control and a back-propagation neural network (LMPC-BP) is proposed and simulated for single-point magnetic levitation. The deviation of the BP network is observed and compensated by an expansion state observer (ESO). The iterative BP neural network model is further updated using current data and feedback data from the ESO, and then the performance of the LMPC-BP controller is evaluated before and after the update. The simulation results show that the LMPC-BP controller can achieve stable levitation at different gaps of the single-point magnetic levitation system. With further updating and iteration of the BP network, the controller anti-jamming performance is improved.

Simulation of synaptic properties of ferroelectric memory capacitors and neural network applications

In this work, the electrical properties and synaptic characteristics of hafnium oxide-based ferroelectric memory capacitor with metal - ferroelectric layer - metal (MFM) structure were simulated using TCAD (technology computer aided design) software. Based on the synaptic potentiation/depression characteristics of the simulated memory capacitor, a multilayer perceptron (MLP) network was constructed, and the recognition accuracy and convergence speed of the MLP network in the MNIST recognition task were simulated, and the feasibility of the ferroelectric memory capacitor synaptic device for real neural network operation was analyzed. The results show that the recognition accuracy of the MLP network reaches 93% and stabilizes after 50 iterations of training, and the recognition accuracy of the MLP network is already at a high usable level after a smaller number of training times of 20, which suggests that the synaptic plasticity of the ferroelectric memory capacitor has a good potential for the practical application of the weight updating of the MLP network.

Template‐Refine Network for Siamese Object Tracking

Various mainstream target tracking algorithms based on Siamese networks are gradually becoming a trend in the field of deep learning tracking due to their concurrent advantages of accuracy and speed. Most Siamese network‐based trackers describe the tracking of a target object as a similar matching problem, and these trackers have achieved more advanced performance in several public tests. Most trackers often suffer from tracking drift or performance degradation owing to the non‐updating of the template in the first frame and the target appearance encounters disturbing environments such as occlusion and drastic deformation. Therefore, to address this problem, this paper introduces a template updating mechanism and proposes a refine structure network based on the template updating of Siamese networks as well as the greater similarity of target features in two adjacent frames, which improves the tracking accuracy while limiting the amount of computation using an anchor‐free method in order not to lose the tracking speed, and only needs to be trained by selecting the most suitable pre‐training network, thus greatly reducing the amount of network computation. Meanwhile, in the application of the refine structure, with the aim of making the weight design of the target localisation module more reasonable, we propose a new Refine Head section and analyze and design the update threshold to optimize the overall network. This method is practiced in SiamFC++ algorithm, which firstly designs the template refine module, inputs the image that needs to be improved, and then outputs it to the Refine Head to complete the template update and applies it to the tracking of the subsequent frames, thereby constituting the SiamTRN (Template‐Refine Network). According to the experiments, the improved structure of the method can effectively implement the refine module function and enhance the performance of the tracker on public datasets, such as OTB100, VOT2016, UAV123 and GOT‐10 k. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

MCAN: Multimodal Causal Adversarial Networks for Dynamic Effective Connectivity Learning From fMRI and EEG Data.

Dynamic effective connectivity (DEC) is the accumulation of effective connectivity in the time dimension, which can describe the continuous neural activities in the brain. Recently, learning DEC from functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) data has attracted the attention of neuroinformatics researchers. However, the current methods fail to consider the gap between the fMRI and EEG modality, which can not precisely learn the DEC network from multimodal data. In this paper, we propose a multimodal causal adversarial network for DEC learning, named MCAN. The MCAN contains two modules: multimodal causal generator and multimodal causal discriminator. First, MCAN employs a multimodal causal generator with an attention-guided layer to produce a posterior signal and output a set of DEC networks. Then, the proposed method uses a multimodal causal discriminator to unsupervised calculate the joint gradient, which directs the update of the whole network. The experimental results on simulated data sets show that MCAN is superior to other state-of-the-art methods in learning the network structure of DEC and can effectively estimate the brain states. The experimental results on real data sets show that MCAN can better reveal abnormal patterns of brain activity and has good application potential in brain network analysis.

Immunomodulatory signalling networks in glioblastoma multiforme: a comprehensive review of therapeutic approaches.

Glioblastoma multiforme is a very aggressive type of cancer with high mortality and poor prognosis worldwide. Advanced treatment options with an understanding of the molecules and signalling mechanisms involved in this type of cancer have the potential to increase targeted therapy and decrease off-target effects, resistance, and recurrence. Glioblastoma multiforme (GBM) presents a complex tumour microenvironment with numerous cellular components and an extracellular matrix comprising multiple components. A deeper understanding of these components and corresponding signalling pathways can increase the success of immune checkpoint inhibitors in the treatment of GBM. The discovery of specific molecular changes and biomarkers has led to the investigation of tailored treatments for individual patients. Combination therapies targeting multiple pathways or utilizing different modalities are emerging as a promising strategy albeit with challenges in drug delivery to the brain. The review presents a comprehensive update of the various immunomodulatory signalling networks in GBM and highlights the corresponding therapeutic approaches by targeting them.

Model Parameter Calibration for Vibration Fatigue Analysis by Means of Bayesian Updating and Artificial Neural Network Based Surrogate Models

Abstract In this paper, a methodology for model parameter calibration for vibration fatigue analysis is proposed. It combines Bayesian updating of uncertain model parameters and artificial neural networks (ANNs). The calibrated parameters are used to increase the accuracy of fatigue lifetime calculations for components submitted to vibrational loads. The Bayesian updating uses eigenfrequencies, mode shapes, total mass, and the frequency response functions (FRFs). These quantities are predicted by ANN-based surrogate models to accelerate the Bayesian updating process. A novel strategy for the prediction of the magnitude and phase of FRFs with ANNs is proposed. The frequency is used as an additional input variable, and a schematic selection of significant points of the FRF curves is presented. A high prediction accuracy of the surrogate models could be achieved. The procedure includes the analysis of the relevant frequency range and a sensitivity analysis based on the Morris method to identify appropriate modes and the most-influential parameters. The proposed framework is applied to a current vehicle component subjected to vibrational loads. An experimental modal analysis is used for the calibration and consideration of real parameter uncertainty. First, the accuracy of the surrogate models and Bayesian updating is verified by a nominal reference simulation and then validated with experimental data. The measurable control parameter thickness and component mass are used to examine the calibration accuracy. Finally, a decrease in the dispersion of the vibration fatigue distribution is obtained with the calibrated parameters.

Interactive dynamic trust network for consensus reaching in social network analysis based large-scale decision making

With the advances in social media and e-democracy technologies, large-scale decision making (LSDM) problems with trust relationships demands effective consensus-reaching processes. Existing literature has identified a need for improvement in the updating method of trust relationships and the recommendation-feedback mechanism, while the overall effect of trust relationships on consensus reaching remains unexplored and not quantitatively examined. This study proposes a novel interactive dynamic trust-lead consensus reaching process (IDT-CRP) based model to address these challenges. The model integrates a dynamic Trust network updating (DTNU) process and a trust-lead and efficiency-driven recommendation-feedback (TERF) process. In the DTNU process, secondary trust is defined to represent the trust derived from decision makers’ modification behaviors in each iteration. In the TERF process, a novel trust-lead and efficiency-driven recommendation-feedback mechanism is proposed, in which two different decision scenarios are provided. Furthermore, the role of trust relationships in enhancing group consensus level is analyzed, revealing its significant impact while identifying an upper limit. Novelties of this paper include 1) the improved trust updating algorithm which derived secondary trust from modification behaviors instead of opinion similarities; 2) the improved recommendation-feedback mechanism with different trust utilization scenarios while retaining minority opinions; 3) the novel IDT-CRP model based on the interactive mechanism between trust relationship and opinion evolution. A series of experiments are implemented, validating that the proposed model is of good feasibility, effectiveness and robustness, and can be applied in future LSDM research.

Research on multiscale OpenStreetMap in China: data quality assessment with EWM-TOPSIS and GDP modeling

ABSTRACT OpenStreetMap (OSM) is a voluntary platform designed to provide free and up-to-date geographic data. Since OSM is based on multisource geographic data provided by the public, the quality of the data has become a concern of researchers. In this study, a unified measure for evaluating OSM data quality was constructed, and the entropy weight method (EWM) and Technique for Order Performance by Similarity to Ideal Solution (TOPSIS) model were used to evaluate the quality of multiscale OSM data in China from 2014 to 2020. In addition to evaluating the data quality, the use of OSM data quality index factors in economic modeling at different spatial scales in China was explored by using a geographic information system (GIS) analysis method and a geographically and temporally weighted regression (GTWR) model. Four machine learning models, SVM, RF, XGBoost and CatBoost, were used to simulate the grid-scale GDP, and the effectiveness of these simulations was discussed. The results showed that (1) the weights of OSM data quality indicator factors vary across different spatial scales. (2) From 2014 to 2020, the quality of national-scale OSM data first increased, then decreased and then gradually stabilized. In addition, the quality of OSM data at the provincial and municipal scales is significantly different, and the distribution is affected by the population and geographical environment. (3) Over time, the spatial clustering characteristics of OSM data quality at different spatial scales in China has continuously strengthened. In addition, the quality of Chinese OSM data displays obvious local spatial autocorrelation characteristics, which are dominated by H-H clustering and L-L clustering. (4) The GTWR model performs well in predicting and revealing the spatiotemporal correlation characteristics between GDP and OSM data quality indicators at the provincial and municipal scales in China. The correlations increase with decreasing spatial scale (provincial to municipal). Moreover, the GDP modeling ability is better in economically underdeveloped Northwest China and economically developed East China. (5) Four machine learning models coupled with road network length completeness, relative linear density, road name attribute completeness, POI name attribute completeness, road network accuracy, road network update frequency, POI update frequency, topological consistency and directional similarity yielded the best grid-scale GDP simulation values. Notably, the CatBoost model provided the best accuracy, and the results further verified that it is feasible to use the proposed OSM data quality index system to predict the regional economic development level in China.

Optimizing reverse osmosis desalination from brackish waters: Predictive approach employing response surface methodology and artificial neural network models

A mathematical model frame has been developed to optimize the performance of reverse osmosis (RO) desalination in brackish waters with medium to high salt concentrations using response surface methodology (RSM) and artificial neural network (ANN), specifically catered towards efficiently operating RO desalination systems. The RSM model aims at predicting RO performance with second-order polynomials based on 15 experimental datasets. Inputs for RSM model include feed TDS concentration, feed water flow rate, and concentration-to-desalination ratio, while the outputs consider rejection, membrane flux, performance index, and specific energy consumption. The Genetic Algorithm optimized back-propagation Neural Network was employed to update network weights and biases, leading to an optimal ANN network structure of 3–10-4. A ratio of 70 % training, 15 % validation, and 15 % test data was chosen to establish an ANN model that predicts the performance of RO desalination. The ANN model demonstrated significantly higher prediction accuracy than the RSM model, with impressive correlation coefficients of 0.9936 for the performance index and 0.9932 for SEC prediction, and low RMSE of 0.85 for rejection. To maximize the performance index and minimize specific energy consumption, a two-step optimization method was further set up by applying the RSM model to initially determine optimal process conditions followed by the ANN model to enhance the accuracy of performance forecasts. Optimal process conditions were proposed for specific feed TDS scenarios, which involved setting a feed TDS concentration limit to meet drinking water standards.

Data-driven ensemble optimal compensation control for partially known delayed and persistently disturbed nonlinear systems

In almost all industrial processes, the development of an accurate mathematical model that comprehensively characterizes the regulated system poses a substantial obstacle. Diverse factors, such as unidentified system components and time-varying external disturbances, contribute to this phenomenon. This can result in a disparity between the model and the plant and thereby compromising the performance of the designed controller. In this study, a novel data-driven ensemble optimal compensation control framework is proposed for partially known nonlinear systems with state delays and structural disturbances. Initially, the estimated system model is leveraged to construct a robust fuzzy model predictive controller (RFMPC) based on the Lyapunov–Krasovskii functional (LKF), ensuring the input-to-state stability (ISS) of the plant. Then, a reinforcement learning (RL)-based compensator is assembled to gradually minimize the model-plant mismatch across the entire state space. To streamline the training process of the compensator and facilitate its stability, a dynamic network update scheme is devised, incorporating a moving average gradient calculation method and a learning rate tuning strategy. Furthermore, comparative discussions with some prevalent methods are demonstrated. With the established control framework, the performance of the RFMPC can be securely enhanced, and convergence is ensured within a narrow proximity to equilibrium, despite time-varying perturbation dynamics. By integrating this approach with deep learning (DL) methods, the efficacy of numerous industrial processes could be significantly improved. A numerical simulation and a discretized continuous-time stirring tank reactor (CSTR) example validate the effectiveness and benefits of the proposed control strategy.

Deep image prior inpainting of ancient frescoes in the Mediterranean Alpine arc

The unprecedented success of image reconstruction approaches based on deep neural networks has revolutionised both the processing and the analysis paradigms in several applied disciplines. In the field of digital humanities, the task of digital reconstruction of ancient frescoes is particularly challenging due to the scarce amount of available training data caused by ageing, wear, tear and retouching over time. To overcome these difficulties, we consider the Deep Image Prior (DIP) inpainting approach which computes appropriate reconstructions by relying on the progressive updating of an untrained convolutional neural network so as to match the reliable piece of information in the image at hand while promoting regularisation elsewhere. In comparison with state-of-the-art approaches (based on variational/PDEs and patch-based methods), DIP-based inpainting reduces artefacts and better adapts to contextual/non-local information, thus providing a valuable and effective tool for art historians. As a case study, we apply such approach to reconstruct missing image contents in a dataset of highly damaged digital images of medieval paintings located into several chapels in the Mediterranean Alpine Arc and provide a detailed description on how visible and invisible (e.g., infrared) information can be integrated for identifying and reconstructing damaged image regions.

Underlay Loosely Coupled Model for Public Safety Networks Based on Device-to-Device Communication

In several emergency situations, during natural or human-caused disasters, frontline responders need to be able to communicate and collaborate to properly carry out relief missions. Some countries build their national Public Safety Mobile Broadband based on cellular LTE technology to provide fast, safe, and secure emergency services. However, in several emergency situations, cellular antennas can be overloaded or partially damaged in a manner that affects group communication services. In the last few years, direct device-to-device (D2D) communications have been proposed by the 3GPP as an underlay of long-term evolution (LTE) networks based on proximity, reuse, and hop gains. This paper focuses on a loosely coupled model based on direct D2D communication in a public safety context. Many scenarios related to user membership and network management are detailed. Both the “less cost” and “optimized tree” approaches are proposed and implemented, and their performance is evaluated in terms of the network update number and the resulting average Channel Quality Indicator (CQI). Other optimization approaches, with different CQI thresholds and optimization interval parameters, are simulated to compare their performance with the “optimized tree” approach. By conducting simulations that combine a CQI threshold = 1 and optimization interval = 2 s, it becomes possible to keep an average CQI level close to the “optimized tree” approach, while the costs related to network updates significantly decrease by almost 35%. Other simulations are also carried out to measure the bandwidth required by the control messages between the server and active users. It was found that both inbound and outbound traffic on the server side can be well supported with LTE and 5G networks.

Hyperspectral image classification based on a novel Lush multi-layer feature fusion bias network

Convolutional neural networks (CNNs) exhibit excellent performance in hyperspectral image classification (HSIC) and have attracted significant interest. Nevertheless, the common CNN-based classification techniques still suffer from the following drawbacks. 1) Although deep CNNs can effectively extract features from hyperspectral images, a network that is too deep often leads to negative effects such as overfitting, vanishing gradients, and decreased accuracy. 2) Most of these models for HSIC do not fully consider the strong complementarity and correlation between features at different levels. To address these issues, firstly, a self-regularization parameter correction non-monotonic activation function, called Lush, was first proposed. Then, this study proposes a novel bias update method that can update network bias parameters more efficiently. On the basis of the above, a Lush multi-layer feature fusion polarization network (LMFFBNet) for hyperspectral image classification was proposed in this study. It designs a new after-melting technology to fuse deep and shallow features in multiple stages and layers, which can obtain more discriminative features, thereby improving the classification performance of hyperspectral images. Extensive experiments on five challenging datasets (Indian Pines, Pavia University, Kennedy Space Center, Salinas Valley, and Houston) have demonstrated that compared to some state-of-the-art methods, the proposed LMFFBNet method can provide better classification performance for hyperspectral images and has strong robustness. This fully proves the effectiveness of the LMFFBNet method.