Internet Technology Letters最新文献

To address the problems of dynamic user experience optimization, adaptive resource allocation, and personalized service provisioning in modern wireless communication networks, this letter proposes a large model-based experiential landscape design method utilizing contrastive learning and pre-trained transformer architectures. Considering the high complexity and heterogeneous nature of wireless environments, along with the insufficient performance of traditional optimization methods for experiential landscape design tasks, wireless signal sequences are first transformed into token-like representations similar to those used in natural language processing. Subsequently, a pre-trained transformer model is employed to convert shallow representations into universal wireless experience representations suitable for various downstream landscape design tasks. By transforming the experiential landscape optimization problem into a similarity analysis problem, a diversity-sensitive transformer model architecture is designed based on contrastive learning principles, which enhances the model's sensitivity to experience differences through positive and negative sample pairs of wireless environments and proposes using information noise contrastive estimation as the loss function for fine-tuning downstream landscape design tasks. Experimental results demonstrate that the proposed method outperforms mainstream approaches in terms of user satisfaction accuracy, quality of experience precision, service continuity recall, and F1 score metrics, achieving 89.73% accuracy in wireless experiential landscape classification.

With the application of Internet of Things (IoT) technology, the field of online education has developed rapidly. Online users can flexibly obtain learning resources on the MOOC learning platform and conduct online course learning. The development of IoT enables various systems with sensing functions to continuously access the Internet. The interconnected data makes the development of education tend to be digital and intelligent. Semantic sensor associates semantic technology with a large number of sensors in IoT, providing effective technical means for data representation, management, and sharing. It can provide a theoretical basis for knowledge-based intelligent semantic sensor data processing in IoT. This paper analyzes conventional deep learning algorithms, that is, connectionist text proposal network (CTPN) and convolutional recurrent neural network (CRNN). Then we combine CTPN and CRNN to propose the improved algorithm. The improved algorithm can more accurately recognize text and be used in IoT to extract spatial features in semantics. It can solve contextual relationships between texts, capture temporal features, effectively handle the complexity and diversity of semantic sensor data in the MOOC system. Finally, compared with the traditional algorithms, the improved algorithm achieves higher accuracy and faster recognition speed.

The rapid advances of communication technologies have fueled significant IoT growth, with increasingly interconnected smart devices generating massive data. While various naming services and routing schemes have been proposed, existing solutions remain use-case specific, with Information-Centric Networking (ICN)/Named Data Networking (NDN) facing scalability challenges in edge networks and hybrid designs sacrificing naming flexibility, lacking a holistic architecture for future IoT networks. To address these challenges, we propose the Address Name Transfer Network (ANT-Net), a novel architecture that uniquely combines seamless integration of heterogeneous naming services with enhanced data retrieval efficiency. ANT-Net employs name-based routing at the edge network and IP-based routing within the core network, maintaining full TCP/IP compatibility for both naming and routing services. Furthermore, the distinct separation between edge and core networks in ANT-Net allows for the implementation of different routing strategies, which can significantly improve data sharing and overall communication efficiency.

With the rapid development of the global digital economy, digital transformation has become an essential path for enterprises to strengthen their market position and ensure sustainable growth. As an important force in China's economic development, private enterprises face both rare opportunities and unique challenges in promoting the digital transformation. This study focuses on industrial internet of things, a key digital technology, and conducts an in-depth analysis of case studies of private enterprises' transformation and upgrading based on industrial internet of things through a “technology–organization–environment” framework. It finds that industrial internet of things technology can effectively enhance the operational efficiency of private enterprises; however, successful digital transformation requires the coordinated advancement of technological support and organizational change. Therefore, this study proposes a step-by-step, modular implementation path tailored to the characteristics of private enterprises, providing theoretical guidance and development references for digital transformation based on industrial internet of things in private enterprises.

In response to the challenges of information fragmentation, ambiguous risk propagation mechanisms, and insufficient cross-stage collaboration in equipment full-lifecycle quality management, this study redefines the equipment quality chain in the context of Industrial Internet of Things (IIoT) technology enabling the interconnection of all equipment elements. By integrating real-time data collection technology from the IIoT, a quality risk propagation function is established to quantify single-factor sensitivity and multi-node coupling effects, revealing the laws of deviation accumulation and cascading amplification in risk propagation. An innovative fusion of the Multilayer Hypergraph Coupling Model (MHCM) and IoT technology is used to construct a risk propagation coupling network model, uncovering the general principles of risk propagation in the equipment quality chain. Monte Carlo simulation is used to identify key sensitive nodes and validate the rationality of the quality risk propagation function. Based on this, quality management strategies adapted to the resilient manufacturing requirements of Industry 5.0 are proposed, forming a full-chain control system and providing a quantifiable, traceable, and interceptable quality management paradigm for complex equipment systems, effectively enhancing equipment effectiveness and economy.

A key role for artificial intelligence (AI) is anticipated in the era of 5G networks. Effective radio resource management has become more and more important for network operators. But when new technologies, network topologies, and sophisticated equipment are integrated more quickly, it's getting harder to allocate enough radio resources for precise channel condition assessment in mobile networks. Predicting channel conditions automatically helps to make effective use of resources. Research presents the mutated gray wolf-driven spiking neural network (MGW-SNN) model, an ML-based method for signal-to-interference-and-noise ratio (SINR) estimation based on the cyber-physical system (CPS) location. An MGW-SNN model uses the present position of the CPS to forecast the SINR. Initially, obtain the data to validate suggested algorithms. Min–max normalization and noise reduction were used to preprocess after collecting data. The spiking neural network (SNN) structure's parameters are optimized with the usage of the gray wolf optimization (GWO) method to improve the network's performance. The research was implemented using MATLAB. The proposed strategy in terms of accuracy (93%), RMSE (1.35), $��R^2�$ (0.99), MAE (0.75), and average SINR (10 dB) is validated. It shows that the proposal is effective in predicting the SINR. The approach stabilizes the computational economy while simultaneously increasing accuracy, which makes it suitable for real-time applications in dynamic network environments.

Cardiovascular complications pertain to damages done to the myocardium and the arteries, as well as other parts of the circulatory system. Some important ischemic heart disease (IHD) risk factors include some health assessment metrics, age, and gender. In the prediction of cardiac diseases and in other tough decisions involving the analytics of large volumes of medical data, quantum computation techniques have become a staple in the healthcare field. Studies recently focus on the application of such procedures in quantum learning (QL) algorithms. For heart disease diagnosis, this work came up with a set of QL algorithms, including optimized quantum qubit vector learning (OQQVL) and quantum stacked convolutional neural networks (QSCNN), which we consider computationally light. By incorporating strong quantum learning based pre-processing and feature clustering techniques like quantum K means, the proposed models' prediction reliability along with accuracy is sharpened whilst ensuring robustness. A trade-off between the proposed and more sophisticated recently published models is done using available performance metrics. Thus, using the proposed techniques, the models achieved the highest accuracies of 99.1% and 99.5%, respectively. Moreover, the sophisticated structure provided reasonable execution times, making them suitable for real-time healthcare use.

Network traffic prediction stands as a cornerstone for the realization of intelligent network management. Traditional cloud-centric solutions encounter significant challenges, including latency and bandwidth limitations, when tasked with handling real-time predictive analytics. This paper introduces an innovative approach centered around an edge computing-based lightweight Graph Neural Network (GNN) model designed to facilitate AI-integrated traffic forecasting. The proposed methodology achieves model lightness through the employment of adaptive graph sampling techniques, while traffic predictions are executed by seamlessly integrating the formulated GNN model. Our comprehensive experimentation demonstrates that the proposed framework significantly outperforms existing state-of-the-art methodologies in multiple performance metrics. Specifically, our approach achieves a reduction in the mean absolute error by at least 7.9%, along with the lowest root mean squared error. Furthermore, the inference latency is reduced by approximately 10% or more, whereas the prediction accuracy reaches more than 93%, exceeding that of competing methods.

Photoplethysmography (PPG) signals are widely used in heart rate monitoring via wearable devices, but motion artifact interference and the need for lightweight methods remain the major obstacles to their practical implementation. To address the issues that existing deep learning methods have insufficient anti-artifact capabilities in strenuous exercise scenarios and excessively high model complexity, making them difficult to adapt to wearable devices, a lightweight Transformer method, called WearPPG-Former, is proposed for heart rate (HR) estimation. The proposed method incorporates a sparse attention mechanism into transformer modules, which selectively concentrates on heart rate-associated key feature regions. The design simultaneously reduces computational complexity while enhancing suppression capabilities against motion artifacts, including baseline wander and high-frequency noise. Experimental results demonstrate that WearPPG-Former outperforms the state-of-the-art methods in mean absolute error (MAE) on the PPG-DALIA dataset. Specifically, it achieves a reduced average MAE of 3.18 beats per minute (bpm) under intense exercise scenarios. When deployed on a resource-constrained wearable embedded platform, the model attains an inference speed of 410.4 ms, fulfilling real-time monitoring requirements for wearable devices. This delivers an efficient solution for PPG-based heart rate estimation in dynamic scenarios, thereby advancing the practical implementation of wearable health monitoring technology.

Recently, wireless communication networks have played a vital role in environmental monitoring and other data-driven applications. Even though these networks often struggle with limited energy and redundant data transmissions. Moreover, traditional routing protocols, such as the Cross-layer Opportunistic Routing Protocol (CORP), rely heavily on static routing decisions with fixed-cost functions, leading to a lack of adaptability. To address these issues, this study proposes a Mistral 7B-based Cross-layer Optimization (M7BCO), which integrates adaptive reasoning and prompt-based telemetry compression for energy-aware decisions. The proposed M7BCO model utilizes a Partially Informed Sparse Autoencoder (PISA) to select a minimal subset of informative nodes by learning spatial correlations while preserving data reconstructability. Then, the proposed M7BCO model generates a real-time decision for next-hop selection and transmits power adjustment as it replaces the static optimization with adaptive reasoning. Unlike pure sequential models, the proposed model introduced a lightweight training loop between PISA telemetry selection and Mistral 7B adaptive reasoning. From the results, the proposed M7BCO model achieved better results when compared to the existing CORP model in terms of Energy Efficiency (EE) of 22.5, 65.3, and 100.2 mJ for 150, 300, and 500 nodes respectively.