Search Results (3,123)

Accurate protein function prediction (PFP) is essential for understanding biological systems. However, structure-based graph neural networks often rely on fixed-distance contact maps, which may inadequately capture continuous, multi-scale spatial topologies, while the long-tail distribution of Gene Ontology (GO) labels may bias prediction toward frequent functions. We propose DHST, a deep hybrid structure–topology framework that integrates sequence semantics from a pretrained protein language model with local structural information learned by a residual graph convolutional network. DHST further introduces site-specific persistent homology to encode multi-scale topological invariants and a topology-guided residue-wise gated fusion module to modulate structure–semantics representations using local topological embeddings. The fused residue features are aggregated through dual-path pooling, and a weighted binary cross-entropy loss is used to mitigate the adverse effects of label imbalance. On the PDB dataset, DHST achieved area under the precision–recall curve (AUPR) scores of 0.779, 0.481, and 0.557 for molecular function (MF), biological process (BP), and cellular component (CC), respectively; on the AF2 dataset, the corresponding scores were 0.729, 0.390, and 0.459. The model also demonstrated robust generalization to low-homology proteins and maintained strong predictive performance across GO terms with different levels of functional specificity. Ablation results supported the contributions of the main components. Full article

Airside vehicles operate close to aircraft, service equipment and safety-critical surface markings, yet driver-assistance functions developed for road traffic do not directly transfer to airport aprons, taxiways and service roads. This article presents a vision-based driver-assistance and warning prototype for recognizing airport surface types and horizontal markings in video recorded at Poznan Airport. Two manually annotated segmentation datasets were prepared from GoPro HERO8 video acquired from a vehicle perspective: a four-class surface dataset covering asphalt, concrete, paving blocks and grass, and a three-class marking dataset covering red, white and yellow lines. The study compares You Only Look Once (YOLO) variants YOLOv8 and YOLOv11 with U-Net, DeepLabV3 and SegFormer under a common 512-by-512 input resolution and evaluates both model-level quality and complete video-application behavior. For semantic segmentation, SegFormer achieved the highest validation results, with Intersection over Union (IoU)/Dice of 0.7657/0.8624 for surfaces and 0.8852/0.9380 for markings. Among YOLO models, YOLOv8m obtained the highest surface mean average precision at an IoU threshold of 0.50 (mAP@50) of 0.7847, whereas YOLOv8s obtained the highest marking mAP@50 of 0.8449. On video recordings, paired YOLO configurations processed approximately 15–16 frames per second (FPS) on a personal computer (PC), while U-Net, DeepLabV3 and SegFormer processed approximately 10–11 FPS. A YOLOv8n pair compiled for Raspberry Pi 5 with Raspberry Pi AI HAT+ Hailo-8 reached 10.05 detection FPS and 18.15 processing FPS without GUI rendering. Under the adopted evaluation protocol, SegFormer achieved the highest mask quality in the conducted comparison, while the paired YOLOv8n configuration demonstrated embedded throughput feasibility. Full article

Knowledge tracing (KT) is a core technology in intelligent tutoring systems that predicts students’ future responses by analyzing their historical interaction sequences. Although existing deep learning-based KT models achieve high predictive accuracy, their “black-box” nature severely hinders practical deployment in educational scenarios. To balance predictive performance and interpretability, this paper proposes AMDKT, an interpretable dual-stream Transformer model grounded in the Student Proficiency–Item Competency Matching (SPIM) mechanism. The model employs two parallel Transformer branches to separately model the temporal evolution of student proficiency and the competency demands of each item and defines the discrepancy between their outputs as “proficiency surplus.” A non-negative regularization loss is further introduced to constrain the proficiency surplus to be non-negative for correctly answered samples, thereby embedding pedagogical rules into the optimization objective and ensuring that the model outputs conform to educational cognitive priors. Experiments on five public datasets show that AMDKT achieves AUC performance comparable to the state-of-the-art AKT model, and obtains statistically superior results against DKT, DKVMN, DEEP-IRT, and DIMKT on most datasets, with relatively weaker performance observed on the KDD Cup 2010 dataset. Ablation studies verify the effectiveness of each component, and visualization results demonstrate that AMDKT produces smooth, coherent, and interpretable student proficiency trajectories, providing a fine-grained tool for quantifying individual learning progress. Therefore, AMDKT offers a feasible solution for applications such as weak knowledge point localization, adaptive exercise recommendation, and learning risk warning. Full article

Regenerative aluminum melting serves as a core process in recycled aluminum production. In the regenerative aluminum melting process, the furnace temperature is a key variable which affects product performance and energy costs. The extreme in-furnace temperature necessitates sensors equipped with protective jackets, which increases measurement costs and severely compromises real-time monitoring capability. Accordingly, accurate furnace temperature prediction is highly valuable for regenerative aluminum melting. In regenerative aluminum melting furnaces, periodic burner nozzle commutation and frequent material charging and discharging lead to complex and time-varying operating conditions, posing considerable challenges to high-precision furnace temperature prediction. To address these issues, a condition-adaptive multi-scale convolution-enhanced Transformer (CA-MC-Transformer) model is proposed for furnace temperature prediction. Firstly, an agglomerative hierarchical clustering algorithm based on the weighted dynamic time warping (WDTW) distance is designed to perform unsupervised clustering on historical process data, thereby extracting physically interpretable prior labels for macroscopic operating conditions. Secondly, multi-scale dilated causal convolutions are utilized to capture local dynamic features at diverse temporal resolutions. A soft attention mechanism is further introduced to dynamically assign fusion weights to condition embeddings and local features, enabling condition-adaptive feature reconstruction. Finally, the fused adaptive features are fed into an encoder-only Transformer network to capture the global long-range temporal dependencies and achieve accurate furnace temperature prediction. Comparative experiments conducted on real operational datasets from an aluminum plant verify that the proposed method effectively eliminates the inherent tracking lag of conventional deep learning models, and substantially improves prediction accuracy and anti-noise robustness under complex and variable operating conditions. Full article

Accurate estimation of subsurface acoustic impedance is fundamental to quantitative reservoir characterization in seismic exploration. Nevertheless, a single network architecture cannot adequately represent both the local details and the global trends of seismic records within a unified framework, while the severe scarcity of annotated well-log data substantially constrains the generalization capability and predictive accuracy of deep-learning-based inversion approaches. To overcome these limitations, a semi-supervised acoustic impedance inversion framework based on a hybrid deep learning architecture is proposed. The framework employs a cascaded architecture consisting of a multi-scale depthwise separable convolution with channel attention (MSDSE) module and a convolution-augmented Transformer encoder. Seismic data are first processed by the MSDSE module to extract local multi-scale temporal features, and are subsequently passed to the convolution-augmented Transformer encoder, which captures global long-range sequence dependencies while retaining complementary local temporal information. The two modules progress hierarchically and jointly achieve a feature representation that spans from local details to global trends, and the initial low-frequency model is fused with the network output via channel-wise concatenation. Meanwhile, an initial-model constraint together with a physical-consistency constraint are simultaneously imposed within the loss function, thereby improving training stability while fully leveraging the physical information embedded in unlabeled traces. Experiments on both synthetic and field data confirm the effectiveness of the proposed method. The results show that, even with a small number of labels, the method produces stable impedance estimates and outperforms conventional deep learning methods in both generalization and prediction accuracy. Full article

The automatic inspection of airport infrastructure, specifically horizontal surface markings and Airfield Ground Lighting (AGL), is a critical task for maintaining aviation safety and operational efficiency. As the aviation industry shifts from manual surveys to automated visual inspections utilizing unmanned aerial vehicles (UAVs) and smart service vehicles, the demand for robust, real-time computer vision systems has surged. However, deploying computationally intensive deep learning models in the field introduces severe Size, Weight, and Power (SWaP) constraints. This paper presents a comprehensive framework for the semantic segmentation of runway/taxiway markings and the point-localization of AGL lamps, specifically focusing on the deployment paradigm shift from the expensive, GPU-accelerated heterogeneous system on chip (SoC) to highly efficient, dedicated Neural Processing Units (NPUs). We evaluate the performance of U-Net, LinkNet, U-Net-Point and HRNet-Lite-Point architectures trained on a custom dataset from the Poznań-Ławica Airport. Crucially, this study conducts a rigorous comparative hardware analysis between the flagship NVIDIA Jetson AGX Orin and a highly cost-effective heterogeneous setup comprising a Raspberry Pi 5 augmented with an NPU Hailo-8 AI accelerator. Experimental results demonstrate that while both platforms achieve real-time inference, the Hailo-8 integration fundamentally disrupts the traditional cost-to-performance ratio. Furthermore, the Hailo-8 configuration consumed less electrical power and memory footprint required by the Jetson, proving that dedicated NPUs are vastly superior for continuous, battery-operated edge deployment in autonomous airport maintenance systems. Specifically, the Raspberry Pi setup with the Hailo-8 accelerator demonstrated superior energy efficiency, requiring a significantly lower energy consumption per processed video frame compared to the Jetson platform. Full article

Quality control in green coffee bean production is critical for food safety and economic sustainability. A persistent gap in existing automated inspection systems is their inability to capture both sides of each bean, which leads to systematic under-detection of surface defects. This work presents three contributions: (i) a novel mechatronic apparatus that mechanically guarantees dual-sided imaging of every bean, (ii) a public 12-class dataset of green coffee bean defects, and (iii) an embedded, real-time inspection pipeline validated on low-cost hardware. The apparatus sequentially presents each bean, from a standard 350 g sample, to two 16-megapixel cameras under controlled LED illumination. A dataset of 9600 images spanning 12 classes (11 defects and 1 normal) was generated from expert-classified samples and enriched through data augmentation. Four convolutional neural network (CNN) architectures, VGG-16, VGG-19, ResNet-50 and YOLOv8, were trained and benchmarked using precision, recall, F1-score and mean average precision. YOLOv8 achieved the best overall performance, with a precision of 97.4%, a recall of 99.6%, an F1-score of 0.930 and a mean average precision of 96.5%, outperforming VGG-16 (accuracy 86.07%), VGG-19 (accuracy 67.03%) and ResNet-50 (accuracy 87.76%). Dual-sided acquisition raised mean per-class detection accuracy from 0.727 to 0.908, a relative gain of 25.7% over an equivalent single-sided configuration. Deployed in real-time “track” mode on a Raspberry Pi 4, the system simultaneously classifies defects and counts beans by category, processing a 350 g sample in approximately 38 min. Combining mechanical innovation with lightweight deep learning enables practical, scalable, and cost-effective quality control for laboratories specialized in coffee analysis. Full article

High-bandwidth memory (HBM) faces critical power integrity challenges in high-stack configurations due to elongated power delivery paths and increased parasitic inductance. This paper proposes a dual-interposer HBM architecture with an interposer–HBM stack–interposer configuration, integrating an additional top interposer embedded with chip capacitors. This topology redesigns the HBM’s power distribution network, reducing PDN impedance, and this technology enables bidirectional vertical power supply to DRAM chips during moments when they require current. The PDN impedance is systematically compared with a conventional trench-capacitance-enhanced structure (Structure A) and a deep-trench-capacitance-enhanced structure (Structure B). Results show that at 0.1–11.2 GHz, the proposed structure reduces peak PDN impedance by 66.41% and 65.7% versus Structures A and B, respectively, and decreases the loop inductance of the top-layer DRAM chip by 66.71%. The top interposer’s redistribution layer forms a parallel-plate capacitor complementing the embedded chip capacitors, achieving wideband impedance suppression. Without modifying existing protocols, this architecture provides a system-level PDN optimization strategy for high-stack HBM, offering quantitative insights for capacitor selection and layout design. Full article

The article reflects on the care for the work of creation within Benedictine spirituality and within the broader tradition of the Cistercian Order. It seeks to demonstrate that monastic spirituality, though developed in a pre-modern context, contains enduring theological and ethical intuitions that remain relevant for contemporary ecological reflection. The starting point of the analysis is the Rule of Saint Benedict written by Benedict of Nursia, which, although it does not employ the modern vocabulary of ecology or environmental ethics, nevertheless articulates a coherent vision of human life characterized by reverence, moderation, and responsibility toward the created world. Particular attention is given to key spiritual and anthropological dimensions of the Benedictine tradition, such as attentive listening (obsculta), humility, moderation, stability, and responsible work, expressed synthetically in the principle ora et labora. These elements are interpreted not only as monastic disciplines shaping individual and communal life, but also as forming a relational anthropology in which the human person is understood as fundamentally embedded within a network of relationships with God, other people, and creation. In this sense, Benedictine spirituality already implies a non-exploitative attitude toward the natural world. The article then examines how concern for creation has been deepened and concretized in the spirituality, institutional practice, and historical development of the Cistercian Order. Drawing on selected Cistercian documents as well as contemporary monastic reflections, the study shows that care for the natural environment is inseparably linked with spiritual attentiveness to God, interior silence, openness of heart, and readiness to serve. The Cistercian tradition, with its emphasis on simplicity of life, manual labor, and harmonious integration with the surrounding environment, develops these Benedictine intuitions in a particularly concrete and historically embodied form. From this perspective, a pro-ecological attitude is not interpreted merely as a reactive response to contemporary environmental crises, but rather as a structural consequence of a theologically grounded vision of reality. It emerges from a deep relationship with God as Creator and from a sense of entrusted responsibility for creation understood as a gift rather than as an object of domination. Ecological sensitivity is thus presented as an intrinsic dimension of monastic spirituality rather than an external adaptation to modern concerns. The main research question of the article is: to what extent and in what way do Benedictine and Cistercian spiritual traditions provide theological, anthropological, and ethical foundations for an ecological attitude toward creation within the framework of contemporary integral ecology. To address this problem, the study employs qualitative research methods grounded in theological inquiry. The primary method is textual analysis of the Rule of Saint Benedict and selected Cistercian documents, complemented by theological–hermeneutical interpretation aimed at identifying key spiritual principles and their implicit ecological implications. In addition, the study applies a historical–critical perspective in order to situate monastic sources within their original cultural and theological context, as well as a comparative theological approach, which enables a dialogue between classical monastic spirituality and contemporary ecological theology. These methods are used not only to reconstruct historical meanings, but also to interpret monastic texts in light of contemporary debates on integral ecology, sustainability, and the ethics of responsibility for our common home. In this way, the article highlights a significant continuity between monastic wisdom and modern ecological reflection, showing their mutual capacity for enrichment. Full article

This article presents an empirical comparison of four automatic face recognition methods: Principal Component Analysis (PCA), Local Binary Patterns (LBP), the deep neural network FaceNet, and an original hybrid approach proposed by the authors, referred to as FLLF (Feature-Level Late Fusion). The experiments were conducted on a subset of the VGGFace2 database, comprising approximately 480 classes in the training set and 60 classes in the validation set. For closed-set identification, a new test set (20%) was extracted from the training set. Classification accuracy, training and inference times, and prediction confidence distributions were evaluated for each method. The results show a clear advantage of the deep learning approaches: FaceNet achieved an accuracy of approximately 98% with only five training images per person, whereas the classical methods—PCA and LBP—reached only approximately 7% and 22%, respectively. The proposed FLLF method, which fuses FaceNet embeddings with PCA-whitened LBP descriptors at the feature level and classifies them with a calibrated linear SVM, further improved accuracy to approximately 98.5% and produced the highest prediction confidence values of all tested methods. However, calibration quality was not directly assessed using standard metrics such as expected calibration error or reliability diagrams, so this observation should be interpreted as a confidence-distribution shift rather than a formal calibration improvement. The article also discusses the theoretical foundations of each algorithm, their respective advantages and limitations, and the architecture of the software system implemented for this study. Full article

Hyperspectral anomaly detection (HAD) identifies targets by spectral differences. However, large deep learning models overfit under the small-sample conditions typical of remote sensing, where anomalies are sparse and annotations costly. We propose a lightweight network on the GT-HAD transformer backbone for hyperspectral imagery. The design includes: (1) capacity reduction via multi-layer perceptron (MLP) ratio and embedding dimension optimization; (2) a Decoupled Projection Gating Fusion (DPGF) module enforcing spatial–spectral decoupling via dual-path projection and complementary regularization; (3) analysis of capacity–generalization tradeoffs across six airborne hyperspectral datasets. Experiments reveal an inverted-U relationship between capacity and generalization. The Mini configuration (121 K parameters) achieves a 53% parameter reduction. It yields AUC improvements on three datasets, maintains negligible performance loss (<0.1%) on two datasets, and exhibits a measurable decline on only one dataset, compared with the 255 K baseline. Notably, the baseline attains higher accuracy with only 25% training data on Pavia (+3.69% area under the receiver operating characteristic curve, AUC). This dataset-specific observation provides empirical evidence that capacity–data mismatch can induce severe overfitting in deep HAD models. Under a 25% training ratio, DPGF shows observable performance trends on spectrally homogeneous scenes, and zero initialization ensures identical performance to that of the baseline at the initial training stage, eliminating insertion risk during module deployment. However, the limited diversity of sensors and scene types constrains the generalizability of the observed trends. These results demonstrate that spatial–spectral decoupling with lightweight design suppresses overfitting in few-shot HAD, guiding compact models for onboard satellite and unmanned aerial vehicle (UAV) applications. Full article

Maritime routes ensure connectivity between nations, carrying a vast flow of goods across borders, while ports serve as the critical junctions within this network, managing a wide spectrum of commodities from raw materials to finished goods. Ports also generate employment across numerous sectors and underpin a broad range of allied industries. A seaport is a maritime facility equipped with docks, cranes, and storage infrastructure for international trade, where ships load and unload cargo, containers, and passengers. Key functions of seaports include customs processing, warehousing, and vessel services, with major global hubs such as Shanghai, PSA Singapore, DP World, and Rotterdam handling immense volumes of cargo each year. In contrast, Bangladesh’s ports, Chittagong, Mongla, and Payra, play a vital role in sustaining regional commerce. Today, ports are widely recognized as essential capital infrastructure and prime movers of economic activity. Smart ports are automated facilities that leverage advanced digital technologies, including sensors, big data analytics, artificial intelligence (AI), machine learning (ML), deep learning (DL), augmented reality (AR), digital twins, the Internet of Things (IoT), and various automation systems, to optimize overall operational efficiency. These tools streamline cargo movement while embedding sustainable practices to protect the environment. Beyond operational gains, smart ports deliver faster, more advanced services to all stakeholders involved in port operations, including shipping companies, customs agencies, local communities, and other relevant parties. Renewable energy sources, electric vehicle charging stations, onshore power supply, and smart logistics infrastructure are among the defining sustainability features of smart ports in the present-day context. This study examines the current status and future development trajectory of Bangladesh’s sea ports in relation to the broader global imperative toward smart port transformation. Full article

Accurate land cover classification is critical for geographic information science. However, the fusion of hyperspectral and LiDAR data remains constrained by insufficient spectral-geometric coupling and limited scale representation. To address these challenges, we propose H-FANet, a hierarchical fusion attention network with a three-branch backbone (spectral, spatial, and elevation). Multi-scale enhancement modules are embedded in the spatial and elevation branches to capture scale-invariant features through hierarchical aggregation with convolutional splitting. For cross-modal interaction, H-FANet adopts a two-level fusion strategy: shape-level shallow cross-attention for geometric alignment and spectral-level deep residual fusion for semantic integration. Experiments on three benchmark datasets showed that H-FANet achieved overall accuracies of 99.13 ± 0.06% on Trento, 97.43 ± 0.06% on Houston 2013, and 92.11 ± 0.05% on the Muufl Gulfport datasets. The network outperformed five comparison methods by approximately 1.7–2.2% in overall accuracy. Ablation studies confirm the contributions of the multi-scale enhancement and hierarchical fusion modules. This GeoAI-driven framework improved land cover identification accuracy and could be applied to fields such as environmental monitoring, urban land use analysis, and ecological protection. Full article

Rapid three-dimensional electromagnetic field simulation in stratified marine environments is essential for underwater target sensing system design. This paper presents a Physics-Informed Deep Operator Network (PI-DeepONet) that integrates analytical Sommerfeld integral solutions with seafloor experimental measurements to establish a validated, mesh-free forward modeling framework for extremely low-frequency (ELF) electromagnetic propagation. The architecture uses Fourier feature encoding to resolve multiscale dipole fields and incorporates Maxwell’s divergence constraint through automatic differentiation. The model is evaluated using a tiered validation strategy that combines analytical benchmarks, controlled seafloor experiments, and comparison with a purely data-driven DeepONet. The results show close agreement across stratified marine scenarios, improved accuracy and physical consistency from the embedded constraint, and substantially faster pointwise inference than conventional finite element solvers. Analysis of near-field discrepancies further identifies seabed anisotropy and environmental uncertainty as important sources of model–experiment mismatch, thereby clarifying the framework’s applicability and limitations for marine sensing-system design. Full article

Boundary-layer atmospheric stability, wind-shear variability, and thermodynamic forcing are all important factors for the initiation, intensification, and transport of dust storms. But there is limited knowledge of the quantitative evaluation of bulk-layer atmospheric stability and the relation between wind-driven dust dynamics and atmospheric stability in arid and semi-arid regions. In this research, a hybrid AI–meteorology framework, HyMet-Fusion, is presented that combines visual information derived from satellite observations with physics-based indicators of atmospheric stability to evaluate atmospheric stability during dust storm events over Iraq. The proposed framework is based on the use of deep features extracted from the satellite imagery through a frozen EfficientNetB0 backbone, combined with indicators derived from the ERA5 pressure level data for the atmosphere, such as the Bulk Richardson Number (Bulk Ri), the Wind Shear (WS) and the Dry Air Index (DAI). The two branches were merged using a late fusion (0.75 physics/0.25 image) and each hour was classified into three atmospheric stability conditions: Relatively Stable, Moderately Unstable and Unstable. The overall hourly accuracy using a Leave-One-Event-Out (LOEO) cross-validation scheme, where each dust event was used for independent testing and no dust event was used for training, was 72.4%, with 81.2% accuracy for the dominant stability state and 92.2% correct assessment of the unstable condition time for the severe dust events. Inaccuracies were mainly (66%) in the conservative direction (more instability). Unstable atmospheric conditions were also found to be associated with all severe dust storms and coincided with higher wind shear, lower Bulk Ri values and higher thermodynamic variability. Moderate and light dust events were primarily associated with transitional and relatively stable atmospheric conditions, and differed between the various regions, primarily in Kirkuk and Nasiriyah. Correlation analysis showed that wind shear had the highest correlation with atmospheric instability (r = 0.92), followed by DAI (r = 0.90) and Bulk Ri (r = −0.75). In addition, the wind shear also increased significantly from light to severe dust events at all stations investigated, showing that wind shear is a critical factor for turbulent mixing, vertical momentum exchange and dust uplift processes. The results suggest wind shear is the leading dynamics mechanism for bulk-layer instability in Iraqi dust storms. The findings highlight the complementary benefit of using physics-based atmospheric indicators embedded with deep learning satellite image analysis. The HyMet-Fusion system can be used as a transferable method for observing wind-driven instability of the atmosphere and related dust hazards, which could be employed in boundary-layer meteorology, air-quality forecasting, aviation safety and environmental risk assessment in arid and semi-arid areas. Full article