Chúng tôi thu được các kết quả tiệm cận về số trung bình của các biểu diễn Goldbach của một số nguyên dưới dạng tổng của hai số nguyên tố trong cấp số cộng. Chúng tôi cũng chứng minh một kết quả Omega cho thấy rằng kết quả tiệm cận về cơ bản là kết quả tốt nhất có thể.

This paper presents a novel rate-splitting multiple access (RSMA) framework with partial reconfigurable intelligent surface (PRIS) selection over Nakagami-m fading channels, a scenario that has not been comprehensively analyzed in prior works. To this end, we conduct a detailed performance analysis of a wireless communication system that incorporates multiple reconfigurable intelligent surfaces (RISs) as cooperative relays. The proposed framework integrates RSMA with a PRIS selection strategy, enabling efficient data transmission from the base station (BS) to two users with distinct channel conditions. Closed-form expressions for the outage probability and achievable throughput are derived over independent and non-identically distributed (i.n.i.d.) Nakagami-m fading channels. The analysis highlights the role of PRIS in enhancing system efficiency and ensuring fair resource allocation between users. In addition, asymptotic evaluations offer deeper insights into the system’s behavior under varying channel dynamics. Simulation results are provided to validate the theoretical findings, demonstrating a close match with the analytical expressions and confirming the robustness of the proposed approach.

Enabling Power Beacon for Downlink NOMA Security Systems Employing RIS with Fountain Codes

Analysis of Security–Reliability Tradeoff of Two-Way Hybrid Satellite–Terrestrial Relay Schemes Using Fountain Codes, Successive Interference Cancelation, Digital Network Coding, Partial Relay Selection, and Cooperative Jamming

This article explores the integration of a relay station with two intelligent reflecting surfaces (IRSs) to enhance energy efficiency (EE) and system throughput (ST) in low-power wide-area networks (LPWAN) modelled for Internet of Things (IoT) applications. By leveraging IRS technology, the system improves communication between IoT sensors, the relay station, and the IoT gateway, mitigating signal degradation and optimizing energy consumption. To demonstrate these improvements, we derive theoretical expressions for ST and EE in an LPWAN environment, considering IRS-assisted relay transmission over Nakagami-m fading channels. Our analysis demonstrates that incorporating IRSs leads to significant performance enhancements compared to conventional non-IRS systems. More specifically, the proposed network achieves a power reduction of 10 dBm while achieving the same ST and EE targets, highlighting its energy efficiency advantages. Furthermore, our results show that an LPWAN employing a relay station and IRSs is capable of maintaining reliable operation in high-frequency regimes, such as 5 GHz, while maintaining stable communication over distances of up to 200 meters. We conduct an in-depth investigation into the impact of various factors, including the number of reflecting elements, IRS placement, data transmission rate, available bandwidth, and operating frequency, on overall system performance. These insights provide valuable guidelines for the future design and deployment of IRS-assisted IoT frameworks, ensuring efficient, high-performance communication in LPWAN environments.

This paper develops a Q-Learning framework for adaptive MMSE regularization in massive MIMO systems with 1-bit ADCs. The receiver adjusts its regularization parameter to instantaneous channel conditions via a tabular agent operating on a 6-dimensional state and 75 discrete actions. Evaluation on 20,000 independent test samples shows consistent gains: 1.27–1.70% over the best Bussgang-MMSE baseline and up to 47% over MRC across antenna settings N ∈ {16, 32, 64, 96, 128}. Training exhibits stable learning with the positive-reward rate rising from 16.1% to 49.8% over 20,000 episodes. This gain is achieved with negligible inference overhead, adding only an O(1) table lookup to the standard O(N3) MMSE complexity.

In the context of the growing prominence of artificial intelligence (AI) as a strategic tool in visual culture creation, the application of AI is based on the digitization of heritage platforms, specifically the exploitation of Bình Định pagoda patterns (prior to 01/07/2025) for book cover design. This process involves data digitization, as well as the application of deep learning and biogenerative models (GANs, Diffusion Models) to identify, reproduce, and transform traditional patterns. The results demonstrate that AI not only supports the preservation of heritage in the form of digital data but also expands creative space and contributes to the formation of a visual language that harmonizes tradition and modernity. The case of book cover design illustrates the feasibility of this approach, while simultaneously highlighting the important role of AI in preserving and promoting cultural innovation in the era of international integration.

Ensuring reliable transmission and secure communication remains a critical challenge for next-generation wireless networks, particularly in the context of sixth-generation (6G) and Internet of Things (IoT) systems. This paper investigates the reliability and physical layer security of active reconfigurable intelligent surface (ARIS)-assisted dual-hop relaying in non-orthogonal multiple access (NOMA) networks over Nakagami-m fading channels. We develop a comprehensive system model where a base station communicates with two users via a relay and an ARIS, in the presence of a potential eavesdropper. Closed-form expressions for outage probability (OP) and intercept probability (IP) are derived for both near and far users, capturing the effects of ARIS configuration, power allocation, and channel fading. Extensive Monte Carlo simulations confirm the accuracy of the analytical results and reveal that increasing the number of ARIS elements, optimizing amplification factors, and carefully allocating power can significantly enhance both reliability and secrecy performance. The findings provide valuable insights for the design of robust and secure ARIS-assisted NOMA networks in practical wireless environments

This paper introduces MobiIris, a lightweight deep network for mobile iris recognition that enhances attention and specifically addresses the balance between accuracy and efficiency on devices with limited resources. The proposed model is based on the large version of MobileNetV3 and adds more spatial attention blocks and an embeddingbased head that was trained using margin-based triplet learning, enabling fine-grained modeling of iris textures in a compact representation. To further improve discriminability, we design a training pipeline that combines dynamicmargin triplet loss, a staged hard/semi-hard negative mining strategy, and feature-level knowledge distillation from a ResNet-50 teacher. Finally, we investigate the use of post-training float16 quantization to reduce memory footprint and latency for deployment on mobile hardware. Experiments on the challenging CASIA-IrisV4-Thousand dataset show that the full-precision MobiIris model requires only 12 MB of storage and 27 ms inference latency, while achieving an EER of 1.409%, VR@FAR = 1% of 98.184%, and CMC@1 of 94.785%, closely matching a ResNet-50 baseline that is more than 7× larger and slower. Under post-training quantization, the model shrinks to 5.94 MB with 13 ms latency and maintains a competitive balance between accuracy and efficiency compared to other optimized variants. These results demonstrate that a coherent combination of lightweight architecture design, attention mechanisms, metric-learning objectives, hard negative mining, and knowledge distillation yields a practical iris recognition solution suitable for secure, real-time authentication on mobile and embedded platforms.

The rise in heart-related diseases has driven the development of advanced techniques for identifying irregular heart conditions, and with the progress in artificial intelligence and signal processing, automated arrhythmia classification using machine learning and electrocardiograms (ECG) has become increasingly effective and widely utilized by healthcare professionals. This paper presents an efficient machine learning solution for robust arrhythmia classification using 12-lead electrocardiograms (ECGs) by leveraging blending ensemble learning. Our developed blending model utilizing six base models-Adaptive boosting, Extreme gradient boosting, Decision trees, K-nearest neighbors, Random forest, and Support vector machine-with a Logistic regression meta-model, demonstrates enhanced efficiency in classifying arrhythmias based on 12-lead ECGs. This approach not only exploits the unique strengths of each base model but also captures diverse predictive patterns, which are combined by the LR meta-model into a refined and cohesive output. The use of LR as the meta-model enhances interpretability and generalization, reducing the risk of overfitting and optimizing overall performance. Experimental results demonstrate that the proposed blending ensemble outperforms conventional notable works in terms of accuracy and offers a robust and effective solution for accurate arrhythmia classification, i.e. up to 97.2% accuracy, supporting clinical decision-making.

Despite many research efforts, the biological insight related to Amyotrophic Lateral Sclerosis (ALS), a rare disease resulting in the loss of motor neurons and causing mortality, remains elusive and leads to challenges to the diagnosis of the disease. Fortunately, gene expression data has recently appeared as a potential approach for the functionality analysis of genes related to orphan diseases and for providing more accurate diagnosis outcomes. Moreover, with the explosion of machine learning (ML), implementing ML in analyzing biomedical data has become a promising direction with a notable effect on our lives. Leveraging these advantages, in this paper, we investigate to shed light on the effects of gene markers on ALS diagnosis and propose a novel gene combination that is effective in ALS diagnosis. We retrieve the datasets and perform the cleaning and pre-processing methods to obtain robust data for analysis. Then, the Max-Min Parents and Children (MMPC) and Sequential Forward Feature Selection (SFFS) algorithms are applied to achieve the optimal gene subsets that are effective for the final intelligent diagnosis model. Notably, the coefficient of the Ridge Classifier is utilized as the crucial score for determining the gene importance ranking table based on the selected gene signatures. All the possible gene combinations are evaluated and optimized in a set of robust machine learning algorithms. Consequently, a set of 20 genes identified through the Support Vector Machine (SVM) algorithm is selected as the optimal for the ALS diagnosis with an accuracy of 88.30% and an AUC score of 91.11%, which is dominant in comparison with notable traditional methods under the same datasets.

The increasing prevalence of heart diseases has driven the development of automated arrhythmia classification systems using machine learning and electrocardiograms (ECG). This paper presents a novel ensemble learning method for classifying multiple arrhythmia types using 12 lead ECG signals through a blending technique. The framework employs a predetermined meta model from foundation models, while the remaining models serve as potential base estimators, ranked by accuracy. Using sequential forward selection and meta-feature augmentation, the system determines an optimal base estimator set and creates a meta-dataset for the meta-model, which is optimized through grid search with k-fold cross-validation. Experiments conducted with seven diverse machine learning algorithms (Adaptive boosting, Extreme gradient boosting, Decision trees, 1 2 3 4 5 6 7 8 9 K-nearest neighbors, Logistic regression, Random Forest, and Support vector machine) demonstrate that the proposed blending solution, utilizing an LR meta-model with three optimal base models, 11 achieves superior classification accuracy of 96.48%, offering an effective tool for clinical decision support.

Recently, environment reconstruction (ER) in integrated sensing and communication (ISAC) systems has emerged as a promising approach for achieving high-resolution environmental perception. However, the initial results obtained from ISAC systems are coarse and often unsatisfactory due to the high sparsity of the point clouds and significant noise variance. To address this problem, we propose a noise–sparsity-aware diffusion model (NSADM) post-processing framework. Leveraging the powerful data recovery capabilities of diffusion models, the proposed scheme exploits spatial features and the additive nature of noise to enhance point cloud density and denoise the initial input. Simulation results demonstrate that the proposed method significantly outperforms existing model-based and deep learning-based approaches in terms of Chamfer distance and root mean square error.

This research aims to bridge this gap by employing the Expectation-Confirmation Theory to elucidate the underpinnings of customer loyalty in mobile payment. This study examines mobile payment customer loyalty using the Expectation- Confirmation theory and second-order constructs of perceived usefulness and perceived risk. The model was tested on 213 mobile payment users using a partial least square structural equation modeling-artificial neural network (PLS-SEM-ANN), capturing linear and nonlinear relationships. Mobile payment system adoption is more likely in people who can understand and analyze complex concepts like risk perception and perceived usefulness. ANN analyses show that confirmation and satisfaction are the most critical factors affecting satisfaction and continuance intention in mobile payment usage. An integrated model that accounts for 40.5% of customer loyalty variance examines feature-level perceived utility and risk to understand mobile payment better. The study ranks factors that sustain mobile payment usage, guiding executive decisions. Perceived usefulness, a second-order construct, positively affects continuance intention in mobile payment loyalty. Notably, the study reveals loyalty drivers essential for developing nations' post-pandemic economic recovery. These findings help developing countries' mobile payment service providers and management increase usage and customer loyalty.

Images have become a strategic digital asset that powers creative industries, e–commerce, and data–driven services. However, modern editing tools and large–scale sharing platforms have made copyright infringement, unauthorized redistribution, and covert manipulation easier to perpetrate and harder to detect. These risks lead to financial losses and weaken trust in digital ecosystems, creating an urgent need for technical protections that complement legal remedies. This paper presents a comprehensive survey of technologies and approaches for image copyright protection, with a particular emphasis on digital watermarking, deep learning-based methods, and blockchain-enabled frameworks. We systematically examine the principles, mechanisms, and applications of these techniques, evaluating their strengths, limitations, and potential synergies. In addition, we explore how these technologies can be effectively integrated into practical systems for secure, reliable, and scalable copyright protection of images. Finally, we identify existing challenges and propose promising future research directions to advance the state of the art in image copyright protection.

The proposed single-phase flexible AC/AC converter in an AC microgrid controlled by the PWM method is presented and tested with a small capacity. This converter uses a simple and small number of semiconductor switches and passive elements to limit power loss and increase efficiency. It has higher reliability, safety, and continuity of power supply in operation than traditional AC/AC converters due to the power circuit structure of the converter. It has the function of increasing or decreasing voltage when connecting to two microgrids and can be directly connected to distributed energy sources in microgrid systems with distributed power sources and loads. Besides, the AC/AC converter can be connected to the storage system to improve continuity and voltage stability for the grid. The performance of the proposed converter is compared with existing similar converters. The paper presents the analysis of simulation results by OrCAD with power values from 0.1-5 kW and experimental power with typical values in the range of 0.5-3.5 kW at different scenarios of the converter.

Purpose – The openness of the Android operating system offers users convenience but also exposes them to a multitude of malicious applications. Consequently, analyzing applications before installation has become a crucial research area in mobile security. Static analysis, known for its accuracy and low cost, is a prominent method within this field. This paper aims to propose an ML/DL-based approach to detect benign and malicious applications in APK format. Design/methodology/approach – The analysis method, detailed further in the paper, consists of five steps. Step 1, each APK file in the sample set undergoes decompilation to convert it into source code. Then, directed API call graph (DACG) generator is used to analyze the decompiled source code from Step 1 and extract API calls. After that, the authors apply the graph2vec method to convert the DACG data set into characteristic subgraphs. Next, saving the necessary features that each model needs to learn from the vector set. This helps reduce the vector dimensionality for each model type and reduces time and noise by eliminating unnecessary features. Finally, training and evaluating the ability to detect Android malware based on popular machine learning algorithms such as Random Forest, support vector machine, K-nearest neighbor, logistic regression and one of the most powerful machine learning algorithms currently available, gradient boosting regression. Findings – The authors come to the conclusion, feature graphs based on API call graphs are effective in detecting Android malware. Experimental results demonstrate the proposed method’s superiority over existing detection methods on a data set of 7,000 samples, achieving TPR > 97%, FPR < 1% and AUC∼0.98. Following these steps, a final classification will determine the safety of the tested application, aiding users in avoiding malware installation. Research limitations/implications – Although some limitations remain to be addressed, the DACG construction method holds significant potential for further exploration. Future research will focus on integrating dynamic analysis techniques to broaden the detectable and classifiable Android malware categories. In addition, the authors aim to adapt the methodology for broader applicability to other system types, including the widely used ELF systems in Linux. Originality/value – In the study, the authors addressed the issue of generating graph-based feature for Android malware detection in a meaningful, practical and efficient way. The results can be used as a pattern for similar scenarios and applications.

The adoption of digital technologies has emerged as a critical determinant of productivity growth and economic competitiveness in the era of the Fourth Industrial Revolution. For Đak Lak province—a developing and predominantly agricultural region in Vietnam—promoting digital transformation within the enterprise sector is both a strategic necessity and a key instrument for driving economic restructuring and improving labor productivity. This study aims to empirically assess the impact of digital technology adoption on the labor productivity of firms operating in Đắk Lắk. Using firm-level data from 1,315 enterprises, drawn from the 2024 Enterprise Survey conducted by the General Statistics Office of Vietnam, the research employs a linear regression model to estimate the effects of digital investment. The findings indicate that digital technology adoption contributes positively to enterprise performance, with software-related expenditures demonstrating a more substantial impact than hardware investments. However, the overall effect remains relatively modest, highlighting the need for more comprehensive digital adoption strategies. The paper concludes with a set of policy recommendations intended to foster a more enabling digital ecosystem and to enhance the effectiveness of technology adoption efforts in the province. The study provides timely empirical evidence to inform provincial digital transformation policies and support the productivity agenda of Vietnam’s emerging regions.

A 4 × 4 broadband patch antenna array has been developed for satellite communication downlink within the 3.7–4.2 GHz frequency range. The proposed design utilizes a hybrid feeding network combining both parallel and series configurations. In which, T-junction power splitters are employed to distribute equal power to each vertical sub-array, while quarter-wavelength transformers provide impedance matching between series elements. Fabricated on a 1.575 mm-thick Rogers 4350B substrate, the overall dimensions of the fabricated prototype are 2.11λ₀ × 2.35λ₀ × 0.018λ₀, where λ₀ is the free-space wavelength at the lowest frequency. Measurement results show that the array achieves a −10 dB impedance bandwidth of 16.3%, a peak realized gain of 14.5 dBi at 3.8 GHz, and an efficiency above 68% across the operating range of frequency. These features show the proposed antenna is a strong candidate for satellite communication systems for broadband applications

Psychological capital (PsyCap) plays an important role in increasing the learning capacity, professional skills, problem-solving, and innovation in the work of employees (Luthans et al., 2013). In developing countries, for example, Vietnam, research on PsyCap has not been mentioned much or fully. It extends beyond social capital and human capital to create profitability. The objective of this paper is to study the impact of PsyCap on the work engagement of sales staff in real estate enterprises in Vietnam. PsyCap includes the factors: 1) confidence, 2) optimism, 3) hope, and 4) resiliency. The research method we used is quantitative, with a sample size of 453 real estate sales staff in Vietnam. The research model is based on the theory of PsyCap and work engagement. The results of the study showed that resiliency had the strongest impact on dedication and passion at work, while confidence and optimism also had a positive effect, but to a lesser extent. However, factors such as hope did not have a significant impact. The study provides practical suggestions for real estate firms in developing training programs to enhance employees’ PsyCap to increase engagement and work performance.

This paper presents a specific analysis of an individual basic magnetically coupled direct current-to-direct current (DC–DC) converter specially designed for integration in a distributed architecture of renewable energy generators for smart grid applications. In such distributed architecture dedicated for renewable energy, parallel high-voltage DC presents many advantages over the classical centralized one. We show that in such setup, high voltage can be advantageously produced using a specific magnetically coupled boost converter, and we point out the influence of the coupling factor, generally considered equal to one, on the overall performance of the converter and on the global energy efficiency of the installation. In this study, the generalized concepts of system energy parameters of DC–DC converters are introduced and applied to the transient analysis. Consequently, the operation of a magnetic coupled DC–DC converter with a recovery stage is modeled. The simulation results are compared with those of the behavioral study, deduced from the model pointing out the large influence of the coupling factor value on the global behavior and mainly on the value of the recovery voltage, in all the various parts of the switching cycle. The renewable energy generator operating parameters, such as current and voltage values, can then be predicted in a more useful way to compute new similar DC–DC converter systems.

In this paper, we introduce the SDIR (Susceptible–Delayable–Infected–Recovered) model, an extension of the classical SIR epidemic framework, to provide a more explicit characterization of user behavior in online social networks. The newly merged state D (delayable) represents users who have received the information but delayed its spreading and may eventually choose not to share it at all. Based on the mean-field approximation method, we derive the dynamical equations of the model and investigate its convergence and stability conditions. Under these conditions, we further propose a greedy algorithm and a sandwich approximation algorithm for the edge-deletion problem, aiming to minimize the influence of information diffusion by identifying approximate solutions.

This paper proposes a novel approach for detecting vulnerabilities in source code written in C and C++, leveraging large language models (LLMs). Specifically, the study introduces a new model called RoS-Dex, based on ensemble learning techniques and comprising two main components: Code Understanding (CU) and Vulnerability Encoder (VE). Accordingly, the CU module is developed using code embedding techniques and a transformer-based architecture, enabling it to capture the semantic features of source code comprehensively, while the VE module focuses on encoding vulnerability-related features, thereby improving classification performance. In the experimental evaluation, the RoS-Dex model demonstrated effectiveness not only on a single dataset but also across four datasets with different structures and characteristics, including REVEAL, FFMQ+QEMU, BigVul, and RealVul. Furthermore, the RoS-Dex model also showcased its adaptability by successfully passing cross-data validation tests—one of the most rigorous evaluation methods that very few LLM-based models have managed to pass successfully. These results highlight the strong potential of the proposed model for real-world applications and pave the way for future research in C and C++ vulnerability detection.

The open and broadcast nature of wireless communication systems, while enabling ubiquitous connectivity, also exposes them to jamming attacks that may critically compromise network performance or disrupt service availability. The proliferation of Unmanned Aerial Vehicles (UAVs) introduces a new dimension to this threat, as UAVs can act as mobile, intelligent jammers capable of launching sophisticated attacks by leveraging Line-of-Sight (LoS) channels and adaptive strategies. This paper addresses a critical challenge of countering intelligent UAV jamming in the context of energy-constrained ambient backscatter communication systems. Traditional anti-jamming techniques often fall short against such dynamic threat sorare unsui table forlow-powerbacks catter devices. Hence, wepropose a novel anti-jamming framework based on Deep Reinforcement Learning (DRL) that empowers the transmitter to not only defend against but also strategically exploit the UAV’s jamming signals. In particular, our approach allows the transmitter to learn an optimal policy for switching between active transmission, energy harvesting from the jamming signal, and backscattering information using the jammer’s own emissions. We then formulate the problem as a Markov Decision Process (MDP) and employ a Deep Q-Network (DQN) to derive the optimal operational strategy. Simulation results demonstrate that our DQN-based method significantly outperforms conventional Q-learning in convergence speed and surpasses a greedy anti-jamming strategy in terms of average throughput, packet loss rate, and packet delivery ratio.