Search Results (731)

Federated learning (FL) has emerged as a promising paradigm for privacy-preserving edge intelligence because it enables geographically distributed devices to collaboratively train a shared model without transferring raw data to a central cloud. This capability is particularly valuable for 5G and emerging 6G networks, where edge-native services are required to satisfy stringent latency, bandwidth, and privacy constraints while operating on highly heterogeneous devices and time-varying wireless channels. In practice, however, synchronous FL is often constrained by straggling clients with limited computation capability or unfavorable communication conditions, which increases round latency and reduces overall resource efficiency. To address this challenge, this study develops a rigorously structured framework for dynamic client selection and radio resource allocation in heterogeneous wireless edge environments. Each FL round is formulated as a latency-aware scheduling problem that jointly captures local computation time, uplink transmission time, minimum participation constraints, and resource block assignment. On this basis, we propose a Dynamic Client Selection and Resource Allocation (DCS-RA) method that integrates computation-aware, channel-aware, and fairness-aware scoring with greedy resource block allocation guided by marginal completion time reduction. The study further provides a clear methodological structure, workflow visualization, literature-grounded justification, dataset documentation, and uncertainty-aware result reporting. Under the reported simulation setting with 100 clients and 20 resource blocks, DCS-RA reduces the average round completion time from 1.92 s to 1.55 s on MNIST and from 2.02 s to 1.57 s on CIFAR-10, corresponding to improvements of 19.39% and 22.47%, respectively. Standard deviation reductions of 70.59% and 80.77% further indicate improved round-to-round stability and more reliable training behavior. These results support the central conclusion that lightweight joint scheduling can materially improve wall-clock FL efficiency in heterogeneous 5G/6G edge networks. Full article

The rapid growth of heterogeneous IoT devices in 5G environments has created stringent requirements for low-latency edge-based protocol processing. Existing static parsing frameworks lack adaptability to dynamic multi-protocol traffic, resulting in increased processing delays and quality-of-service (QoS) violations under bursty workloads. This paper presents HeteroEdge, a latency-aware adaptive protocol parsing framework for 5G Multi-access Edge Computing (MEC) environments. HeteroEdge integrates four tightly coupled components: (i) a lightweight machine-learning-based Heterogeneous Protocol Parsing Layer (HPPL) built on gradient-boosted decision trees (XGBoost); (ii) a Network Digital Twin (NDT) that maintains a compressed and continuously updated representation of IoT endpoint states; (iii) a Real-Time Inference Engine (RTIE) that dynamically reallocates parsing resources at 50 ms intervals; and (iv) a What-If Simulation (WIS) module that proactively evaluates resource-allocation strategies under hypothetical traffic scenarios. Experimental evaluation on a physical 5G MEC testbed comprising four Intel Xeon Silver 4316 edge nodes and 2000 emulated IoT endpoints spanning twelve protocol classes demonstrates the effectiveness of the proposed framework. HeteroEdge reduces median edge parsing latency (including parsing, classification, and queuing delays, but excluding the 5G radio component) by up to 44.7% compared with static MEC baselines, achieves a macro-averaged protocol classification accuracy of 97.8%, and sustains sub-7 ms edge parsing latency at a line-rate NIC injection throughput of 18 Gbps. Furthermore, latency spikes under bursty traffic are reduced by 39% at the 95th percentile, while SLA violation rates decrease by a factor of 3.9 relative to static resource allocation. These results demonstrate that HeteroEdge provides an effective and scalable solution for latency-critical IoT applications, including smart manufacturing, connected vehicles, and urban sensing. Full article

Frailty is a multidimensional geriatric syndrome that lacks a consistent definition, complicating its clinical management. Epigenetic data suggest that frailty involves altered CpG sites, potentially driven by environmental epigenetic factors (the exposome) that influence aging. Systematically reviewing studies from 2009 to 2025, we quantified frailty prevalence, pooled weighted methylation beta values for associated CpG sites, performed enrichment analysis, and conducted structural network analysis to evaluate chemical interactions, following the PRISMA 2020 guidelines and with the study prospectively registered in PROSPERO (ID 1159037). Results showed a pooled frailty prevalence of 17.4% with extreme heterogeneity (I² = 98.88%), and a combined methylated beta effect of −0.1378 (CI: −0.4156, 0.1400) with high heterogeneity (I² = 100%), highlighting sources of variability. Interestingly, we found a CpG site (cg04772644) shared between Chinese and German cohorts, and, upon mapping, four frailty-related genes (CDC42BPB, SLC1A5, RXRB, and SLC22A18AS) were shared across cohorts. Indeed, these genes are significantly enriched in pathways including thrombin signaling, G protein-coupled receptor signaling, and immune cell differentiation signaling. Finally, our system toxicology analysis demonstrated that arsenite, bisphenol A, benzamide, dorsomorphin, and trichostatin A directly interact with the four shared genes, suggesting that the chemical exposome contributes to the observed epigenetic heterogeneity of frailty and the concomitant clinical manifestations. Full article

Background: Microcalcifications (MCs) are among the earliest mammographic signs of breast cancer, yet their detection remains challenging due to small size, low contrast, and dense breast tissue. This scoping review synthesizes AI-driven methods for MC detection in digital mammography, focusing on three dimensions: comparative performance of deep learning (DL) versus traditional methods, the clinical impact of explainable artificial intelligence (XAI), and the role of synthetic data in addressing dataset limitations. Methods: Following PRISMA-ScR guidelines, we systematically searched seven databases for studies published between January 2000 and January 2026. Of 366 initial records, 72 peer-reviewed studies were included in the final synthesis. Results: DL architectures, particularly convolutional neural networks (CNNs), have generally reported higher diagnostic performance (accuracy up to 99.71% and Area Under the Curve (AUC) up to 0.998) than traditional machine learning methods, although direct comparisons are hindered by heterogeneous datasets and evaluation protocols. XAI techniques have yet to undergo rigorous validation in real-world clinical settings, with very low certainty of evidence regarding their impact on radiologists’ trust or workflow integration. Synthetic data generation mitigates some data scarcity and privacy constraints but introduces artifacts (e.g., checkerboard patterns in 39–46% of cases) that limit clinical realism. Conclusions: DL offers substantial promise for MC detection, but translation to clinical practice requires robust XAI validation, higher-quality synthetic data, and prospective studies on diverse, longitudinal datasets. Full article

The rapid proliferation of Internet of Things (IoT) devices within 5G-enabled smart city environments has introduced unprecedented challenges in communication protocol management across heterogeneous network architectures. With connected IoT devices projected to reach 21.1 billion by the end of 2025 and approximately 39 billion by 2030, existing static protocol selection mechanisms are unable to accommodate the dynamic Quality of Service (QoS) requirements of different smart city applications, such as enhanced Mobile Broadband (eMBB), Ultra-Reliable Low-Latency Communication (URLLC), and massive Machine-Type Communication (mMTC). This paper presents APO-DRL (Adaptive Protocol Optimization using Deep Reinforcement Learning), a framework that utilizes a Dueling Double Deep Q-Network (D3QN) combined with a Prioritized Experience Replay mechanism for intelligent, real-time communication protocol selection and parameter optimization in heterogeneous IoT networks. The proposed framework formulates the protocol optimization problem as a Markov Decision Process (MDP), wherein the DRL agent dynamically selects the optimal communication protocol (NB-IoT, LTE-M, LTE Cat-1, or 5G NR) and adaptively tunes transmission parameters based on real-time network conditions. Experimental evaluation in a 3GPP TR 38.901 Urban Macro simulation environment with N = 30 devices demonstrates that APO-DRL achieves a 138.9% improvement in average throughput compared to Static Allocation (60.00 vs. 25.12 Mbps), while simultaneously achieving the highest QoS satisfaction (83.38%) across all methods, albeit with higher energy consumption and packet loss than Static Allocation. Relative to D3QN+PER, APO-DRL exhibits substantially lower cross-seed throughput variance (±0.88 vs. ±11.03 Mbps), confirming that QA-PER produces a more stable and reproducible learned policy. Full article

Path loss modeling is essential for the design, analysis, and applications (e.g., localization) of low-power wide-area networks (LPWANs). Conventional models typically rely on coarse regional land cover classifications (e.g., urban or suburban), which fail to capture the direction-dependent path loss variations of long-range LPWAN links that traverse heterogeneous environments. Although per-link modeling and geographical clustering have individually shown promise in addressing these limitations, their combined potential remains unexplored. This paper presents GeoSeg, a path loss modeling approach that integrates per-link modeling with geographical clustering. GeoSeg represents the propagation environment between each transmitter-receiver pair as a variable-length sequence that encodes both land cover types and their spatial arrangement and employs a hidden Markov model (HMM)-based clustering method to group these sequences into subregions. A per-subregion path loss exponent is then estimated for each identified subregion, enabling spatially adaptive path loss estimation. Evaluated using an open-access LoRaWAN dataset, the preliminary results demonstrate median MAE reductions of up to 96% across the evaluated clusters compared with the standard log-distance path loss model. These results suggest that integrating per-link environmental characterization with geographical clustering can potentially improve path loss estimation accuracy in heterogeneous LPWAN deployments. Full article

5G New Radio Vehicle-to-Everything (NR V2X) networks must support heterogeneous traffic with strict and diverse latency requirements. Conventional proportional-fair (PF) scheduling does not explicitly account for packet deadlines, which can lead to deadline violations for critical vehicular services under congestion. This paper studies deadline-aware MAC scheduler-weight adaptation for 5G NR V2X using probabilistic prediction and reinforcement learning. We implement a closed-loop ns-3/5G-LENA framework in which network telemetry is exchanged with a Python control agent through ns3-ai shared memory. Gaussian Mixture Model (GMM), Hidden Markov Model (HMM), and Bayesian Logistic Regression (BLR) classifiers are used to predict imminent deadline violations. Their outputs are either mapped directly to scheduler weights or provided as additional state information to a Proximal Policy Optimization (PPO) agent. We evaluate ten scheduling strategies: PF, a non-learning Slack-Based Deadline-Aware Scheduler (SB-DAS), three classifier-only controllers, three classifier-assisted PPO variants, PPO-only, and PPO-only with safety shielding. Experiments are conducted across three vehicle densities and three random seeds per density, using the Deadline-Constrained Packet Reception Ratio (DC-PRR) as the main metric. The PF baseline achieves 61.55% mean DC-PRR and degrades from 75.2% at 30 vehicles to 44.1% at 60 vehicles. In contrast, all adaptive strategies exceed 95% mean DC-PRR and recover 34–38 percentage points over PF in every paired density/seed comparison. The main result is therefore the robust gap between PF and deadline-aware adaptation. Differences among the adaptive controllers are much smaller and fall within the observed seed-to-seed variability. In particular, SB-DAS, which uses no classifier, neural network, or training, achieves DC-PRR statistically indistinguishable from the learned and probabilistic controllers. This indicates that, in the evaluated scenarios, most of the gain comes from deadline awareness itself rather than from learning. We also find that adding classifier-derived violation probabilities to PPO does not consistently improve performance over PPO using raw telemetry alone. To support reproducibility and deployment assessment, the paper includes detailed parameter tables, reward-coefficient and sensitivity analysis, scheduler-weight sensitivity, and per-controller inference-latency and complexity measurements. Full article

In Space–Air–Ground Integrated Networks (SAGINs), the inherent high bit error rate (BER) and prolonged propagation latency of satellite links, compounded by the highly dynamic topologies and multi-hop nature of Unmanned Aerial Vehicle (UAV) networks, present severe bottlenecks to end-to-end transport performance. To mitigate performance degradation within these heterogeneously converged SAGIN-UAV architectures, this paper proposes a SAGIN-enabled Adaptive End-to-End Congestion Control scheme. By exploiting the distinct transmission characteristics of long-delay, high-BER satellite links alongside terrestrial mobile multi-hop UAV networks, the Proposed Scheme optimizes data injection during the slow-start phase and introduces a high-precision loss differentiation mechanism during the congestion avoidance phase. This framework accurately distinguishes non-congestive losses (e.g., channel errors or topology switching induced by UAV mobility) from genuine buffer overflows. The simulation results demonstrate that the proposed adaptive scheme significantly reduces queuing delays at UAV nodes, accelerates transmission efficiency across multi-hop terminals, and enhances data throughput in high-latency environments. Ultimately, this scheme offers a resilient solution for optimizing end-to-end transport control and maximizing the overall transmission capability of SAGIN-enabled UAV networks. Full article

In modern cloud-native systems, a single root cause can trigger cascading anomalies across multiple entities (e.g., microservices, databases, and hosts), generating alert storms with hundreds or thousands of heterogeneous alerts. Alert convergence (automatically grouping these alerts into actionable incident tickets) is critical for reducing operator burden and recovery time. Existing graph-based methods construct a topological graph from known entity dependencies and then leverage Graph Neural Networks (GNNs) for information propagation, but they rely on static physical topologies that fail to capture implicit fault propagation paths. Large Language Model (LLM)-based methods focus on reasoning about the textual information of alerts, yet they do not incorporate global topological structure and struggle with consistency at scale. Motivated by these limitations, we propose LLM-Guided Graph Structure Learning (LLM-GSL), a novel framework that combines the semantic reasoning ability of LLMs with the structural modeling power of GNNs for alert convergence. Specifically, LLM-GSL first leverages an LLM to evaluate pairwise entity relationships and discover implicit fault propagation paths that are absent from static topologies, thereby enhancing the physical-topology graph into a more complete structure. A Graph Attention Network (GAT) then refines alert representations over this enhanced graph via graph message passing, guided by a self-supervised graph affinity loss with continuous multi-modal supervision targets that fuse adjacency structure, textual affinity, and temporal affinity. Finally, density-based clustering groups the learned representations into incident tickets. Experiments on five public datasets, including four LogHub-derived datasets and one RCAEval microservice fault-injection subset, demonstrate that LLM-GSL achieves an average F1-score of 96.2%, outperforming six baselines including both traditional clustering and LLM-based methods by at least 14.0 percentage points. Full article

Background: Multimodal breast ultrasound, including B-mode imaging, color Doppler flow imaging, and elastography, provides complementary information for lesion characterization. However, effectively integrating heterogeneous modalities remains challenging due to inconsistent feature distributions, limited cross-modal interaction, computational cost in existing methods, and sensitivity to noise and missing data. Methods: We presented an efficient Cross-Modal Interaction and Dynamic Fusion Network (CIDFNet) for multimodal breast ultrasound analysis. The framework integrates a multi-scale feature enhancement module to improve modality-specific representations, a cross-modal interaction module to enable early-stage feature exchange across modalities, and a dynamic fusion strategy to adaptively combine modality information based on feature reliability estimation. In addition, an invertible neural network is incorporated to reconstruct missing modality features during training. Results: Experiments on an internal dataset of 248 patients with 1532 images show that CIDFNet obtains an AUC of 85.69%, accuracy of 75.51%, recall of 50.00%, F1-score of 62.50%, and precision of 83.33%, while requiring 49.51 M parameters and 79.79 G FLOPs, respectively. Under a simplified Gaussian noise perturbation setting, performance degradation is observed. Conclusions: CIDFNet presents a framework for multimodal breast ultrasound analysis that reflects a trade-off between performance and computational efficiency. Full article

The rapid development of renewable energy vehicles (REVs) has deepened the coupling between transportation and power systems, leading to the formation of the vehicle–station–traffic–grid (VSTG) coupled system. This paper provides a systematic review of multi-agent intelligent interaction technologies for REVs under the VSTG framework, covering the evolutionary process of VSTG systems, the composition and coupling mechanisms of vehicle–station–traffic–grid subsystems, the objectives and constraints of heterogeneous agents, representative V2X interaction modes, deployment-related standards, and collaborative optimization methods. First, the development trajectory of VSTG systems is traced, from independent planning and uncoordinated charging to V2G integration and V2X multi-network interaction. Second, a multi-agent interaction framework is established to characterize vehicle agents, charging station agents, grid agents, traffic management agents, user/operator agents, aggregator/platform agents, and roadside infrastructure agents. In addition, representative vehicle-to-everything (V2X) modes, including V2L, V2H, V2B, V2mG, and V2G, are compared in terms of their operating principles, application scenarios, and technical characteristics. Moreover, various optimization methods for the coupled system are reviewed. Finally, key challenges, including cross-domain coupling complexity, operational uncertainty, interoperability, battery degradation, and engineering deployment, are discussed, and future research directions are proposed. This review provides a structured reference for the modeling, optimization, and practical deployment of intelligent VSTG systems. Full article

Private Networks or Standalone Non-Public Networks (SNPNs) are essential for Industry 4.0 and enterprise connectivity. However, most existing studies rely on simulations, evaluate only a single radio access technology, or report raw key performance indicators (KPIs) without an interpretable quality assessment framework. In practical deployment, operators require measurement-driven evidence to assess the performance and feasibility of 4G LTE and 5G SNPN solutions. This study presents a controlled experimental comparison of software-defined radio (SDR)-based 4G LTE and 5G SNPNs using the same Universal Software Radio Peripheral (USRP) platform, Open5GS, srsRAN, and commercial off-the-shelf user equipment (COTS-UE). The evaluation was conducted in an indoor environment under line-of-sight (LOS) and non-line-of-sight (NLOS) conditions. Experimental iPerf3 results show that the SDR-based 5G SNPN achieves higher downlink and uplink throughput than the SDR-based 4G LTE SNPN across all tested scenarios. The 5G deployment reaches up to 55 Mbps downlink and 40.5 Mbps uplink under LOS conditions, while maintaining 42 Mbps downlink and 28 Mbps uplink under NLOS conditions. Furthermore, 5G achieves lower latency than 4G LTE, with average values ranging from 21 ms to 31 ms. To provide interpretable network quality assessment, a Mamdani fuzzy logic-based Network Quality Index (NQI) with 81 inference rules is proposed to map signal-to-interference-plus-noise ratio (SINR), throughput, latency, and jitter into linguistic quality levels. The proposed approach enables nonlinear integration of heterogeneous KPIs and provides a technology-agnostic framework for practical SNPN deployment. Full article

Synthetic aperture radar (SAR) ship detection faces three persistent challenges: coherent speckle noise that obscures target boundaries, heterogeneous background clutter in coastal and harbor scenes, and ship targets whose spatial extent varies by more than an order of magnitude within the same image. To address these issues jointly, this paper proposes MSA-DET, an improved SAR ship detection network built upon YOLOv11. In the backbone, a Multi-Scale Cross-axis Attention module (MSCAttention) runs horizontal and vertical axial attention branches in parallel across multiple receptive-field scales, sharpening feature representations for ship targets that vary widely in size and orientation. In the neck, the standard C3k2 block is redesigned as C3k2_SSA by embedding sparse self-attention, which selectively focuses on the most discriminative spatial tokens while suppressing speckle interference and reducing computational overhead. An Adaptive Spatial Feature Fusion detection head (ASFF) replaces fixed pyramid-level aggregation with learned per-pixel blending weights, resolving gradient conflicts across scales and improving localization consistency for both small and large ships. On the HRSID dataset, MSA-DET achieves an mAP@0.5:0.95 of 63.6% and mAP@0.5 of 88.1%, representing gains of 4.0% and 1.6% over the YOLOv11n baseline; on SSDD, it reaches 69.6% and 97.7%, surpassing the baseline by 7.2% and 2.1%, respectively. These results demonstrate that coordinated multi-stage redesign—rather than isolated module substitution—is an effective strategy for SAR-oriented ship detection. The accuracy gains are accompanied by a moderate increase in model size (8.9 M parameters versus 2.6 M for YOLOv11n) and computational cost (9.6 G FLOPs versus 6.3 G), a trade-off that is justified by the substantial improvement in detection quality. Full article

The oncocytic variant of adrenocortical carcinoma (OACC) represents an exceptional type of adrenal malignancy, with heterogenous presentation. Currently, the genetic and molecular spectrum remains an open matter. A 20-year-old adult was accidentally found with a 7.2 cm adrenal tumor and underwent an open right adrenalectomy with OACC confirmation. Post-adrenalectomy positron emission tomography/computed tomography was negative. Immunohistochemistry was positive for calretin, inhibin, steroidogenic factor 1; Ki67 of 20%. Microsatellite instability was 7.61. Lin–Weiss–Bisceglia score showed 2 major criteria [mitoses 6/50 HPF + positive atypical mitoses], the reticuline algorithm (disrupted reticuline network + mitoses 6/50 HPF) was consistent for a malignant behavior, the Helsinki score was of 48. Next generation sequencing identified a likely pathogenic variant of CEL gene (heterozygote, c.539-2A>G) in peripheral blood and two pathogenic variants in the tumor: exon 48, NF1 gene [c.7159_7164del p.(N2387_F2388del)] and exon 6, TP53 gene [c.596delG p.(G199Efs*48)]. Polycystic ovary syndrome type A has been diagnosed as teenager with no phenotype change before the tumor detection. After surgery, oocyte retrieval and cryopreservation upon ovarian stimulation protocol (OSP) was performed before starting mitotane therapy. To the best of our knowledge, this is a novel genetic configuration in OACC with an impact on prognosis to be determined. Hyperandrogenemia stands on a dual source (potential CEL-driven insulin resistance for the ovary and OACC-originating for the adrenal glands). Also, this is the first case to receive OSP in OACC, noting that a tailored multidisciplinary management is mandatory. Full article

Intrusion detection serves as a core technology for securing heterogeneous edge networks, including IoT, industrial edges, and 5G networks. However, existing federated learning-based intrusion detection systems suffer from environmental heterogeneity, limited sample availability, and severe class imbalance—issues that result in inefficient resource allocation and compromised detection performance against rare attacks. In this paper, we propose a novel lightweight intrusion detection model for heterogeneous edge networks, named FedNIDS-CNN, which is based on dynamic distillation-aided federated learning with a CNN backbone. In the data preprocessing phase, a two-level class balancing strategy integrating nearest-neighbor interpolation augmentation and adaptive synthetic sampling is employed to ensure distortion-free sample synthesis. For feature and model optimization, principal component analysis (PCA) is used to reduce the dimensionality of traffic features, while a lightweight 1D-CNN is adopted as the base model to alleviate computational overhead on edge devices. During federated training and knowledge aggregation, a dynamic weight distillation loss mechanism is designed to enhance the model’s ability to recognize minority-class attacks. Meanwhile, the federated framework supports client-side local training and server-side weighted soft-label aggregation, enabling effective knowledge fusion across heterogeneous models. Experimental results on the CICIDS2017 dataset demonstrate that the proposed method achieves an accuracy of 98.55% and an F1-score of 98.40%. Benefiting from the soft-label transmission and parameter-free aggregation design, the framework gets rid of the constraint of homogeneous model architecture and natively supports heterogeneous network models and edge devices with different computing capabilities. It also significantly reduces communication traffic and per-round training latency, confirming its excellent real-time performance and applicability in resource-constrained edge environments. Full article