Search Results (10,930)

As the scale of graph-structured data continues to grow, graph summarization has become an important technique for storage efficiency and high-level visualization. This study investigates a Node Importance (NI) approach to graph summarization that prioritizes structural integrity over simple size reduction. The NI approach selects super nodes by ranking vertices through centrality and propagation metrics. Experimental results demonstrate that the proposed NI method achieves compression rates comparable to or slightly lower than traditional Minimum Description Length (MDL) methods across various datasets while maintaining structural integrity. However, today, the high dimensionality and complexity of modern graph data are making deep learning techniques more popular. Great progress in deep learning summarization techniques is achieved with Graph Neural Networks (GNNs). This study investigates the structure and suitability of different GNN architectures for graph summarization using the NI approach. Graph Attention Networks (GATs) and their variants are discussed as a flexible, learned notion of node importance via attention. We present an examination of GATs, covering both diverse approaches and improvements. This study also discusses extensions that enhance the concept of node importance established by the GAT model, GAT variants for node importance estimation, and application-specific GAT research. Full article

Decarbonizing downstream steel logistics remains underexplored in sustainable supply chain management. This study proposes a machine learning-enhanced tri-objective optimization framework for the ship stowage planning problem (SSPP). The framework handles heterogeneous finished steel products, including coils, plates, ingots, tubes, and sections. The model simultaneously maximizes deadweight utilization and minimizes a novel Adaptive Weighted Moment Balance (AWMB) index. It also minimizes voyage carbon emissions through a trim-and-heel resistance penalty. A spatial-to-sequential discretization strategy transforms the NP-hard placement problem into a tractable permutation optimization. A deep neural network (DNN) surrogate achieves a 3.57-fold speedup with only 1.52% hypervolume degradation. An improved NSGA-III algorithm with adaptive operators ensures Pareto front exploration. Embedded step-wise moment verification guarantees dynamic stability throughout loading and unloading. Validated on real data from a Chinese steel enterprise, the framework achieves 99.88% deadweight utilization, reduces transverse and longitudinal imbalance by 48.27% and 90.54%, and cuts CO₂ emissions by 95.5% per voyage. SOLAS constraints, load line limits, and CII/FuelEU targets are addressed through embedded stability and capacity constraints. Multi-route and weather-dependent validation remains necessary before fleet-scale deployment. Full article

Accurate crop yield prediction is fundamental to sustainable agricultural management, enabling optimized resource allocation and informed decision-making. However, a critical gap exists in current prediction models: existing approaches overlook the temporal alignment between meteorological conditions and production management phases—defined as the intervals between consecutive agronomic operations (e.g., sowing, fertilization, thinning). This oversight results in suboptimal predictive performance, as conventional whole-season weather aggregation fails to capture phase-sensitive crop–weather interactions. While machine learning (e.g., XGBoost) and deep learning approaches (e.g., CNN, LSTM) have been applied to yield prediction, these models typically treat weather variables as temporally homogeneous inputs, inadequately modeling the correlation between historical yields and phase-specific meteorological patterns. To address this gap, this study proposes CNN-LSTM-AM, an innovative hybrid deep learning model that integrates convolutional neural networks (CNNs), long short-term memory (LSTM), and attention mechanisms (AMs), utilizing weather data explicitly aligned with production management phases as input. The CNN component extracts cross-phase weather patterns, the LSTM captures sequential dependencies across growth stages, and the attention mechanism dynamically weights phase importance based on meteorological conditions. The proposed model is validated using a real-world case study of Bok choy production from an agricultural cooperative in Yunlin County, Taiwan, encompassing 1714 production cycles over eight years (2011–2019). Experimental results demonstrate that CNN-LSTM-AM achieves an RMSE of 1448.24 kg/ha, MAPE of 3.60%, and R² of 0.98, outperforming five baseline models—CNN (RMSE = 2919.18), LSTM (RMSE = 2529.74), CNN-LSTM (RMSE = 1516.44), LSTM-AM (RMSE = 2284.64), and XGBoost (RMSE = 3452.47)—representing a notable reduction in prediction error (58% lower RMSE) compared to XGBoost. Furthermore, prediction accuracy improves progressively as harvest time approaches, and phase-specific weather encoding enhances accuracy by 16.5% compared to whole-season averaging. These findings underscore the critical importance of integrating agronomic domain knowledge into data-driven prediction frameworks. Full article

This article introduces an intelligent framework using deep learning to recognize and classify different faults through the real-time detection of multiple faults in power distribution systems. A collection of data representing normal operating conditions, alongside various fault scenarios including line-to-ground (LG), line-to-line (LL), double line-to-ground (LLG), and three-phase line (LLL) faults, was created using three phase current signals obtained from the Real-Time Digital Simulator (RTDS) microgrid test system. To properly model the system dynamics, a feature extraction method that integrates phase currents, differential currents, summation currents and magnitude results was developed. The temporal features of the fault signals were identified by using a sliding window approach to fit the data. A one-dimensional convolutional neural network (CNN) was developed to identify different types of faults. This model performed well, obtaining nearly 96.15% accuracy while testing. In order to evaluate the feasibility of the approach, the trained model was loaded on Raspberry Pi 5, NodeMCU, ESP32 and existing sensing devices. The fault classification performed in real-time was time-sensitive. The proposed intelligent framework is applicable to low-scale operation for smart grid fault monitoring and protection and it is an economically viable solution. Full article

Deep learning models have achieved high accuracy in forecasting problems, but at the cost of large computational energy demand. Brain-inspired spiking neural networks (SNNs) offer a promising, low-power alternative, yet their adoption for time-series forecasting has been limited by information loss from binary spikes and degraded performance in deeper networks. This paper proposes a fully spiking framework that bridges this gap by improving both the encoding and propagation of information in SNNs. The framework introduces a hybrid Delta-Rate encoding mechanism that captures both abrupt changes and gradual trends in time-series data, and a Mem-Spike mechanism that transmits analog membrane potential values to preserve fine-grained information between spiking layers. We further employ residual membrane connections to maintain signal flow in deep spiking networks. Using two public energy load datasets, our enhanced SNNs consistently outperform conventional spiking models, improving prediction accuracy by up to 61.6% and mitigating degradation in multi-layer networks. Notably, it narrows the gap to the selected deep learning baseline (LSTM), achieving comparable accuracy in some settings while requiring only about 10% of the estimated inference energy of that baseline under a common operation-level model. These results show that, within the empirical scope considered here, enhanced conventional SNNs can improve time-series forecasting accuracy while retaining favorable estimated efficiency. Full article

Climate change disproportionately threatens low- and middle-income countries, yet integrated assessments combining socio-economic fragility with physical hazards remain limited. This study quantifies multi-dimensional climate vulnerability and derives optimized adaptation policies for six representative nations (Bangladesh, Colombia, Kenya, Morocco, Pakistan, Vietnam) by fusing socio-economic indicators with climate risk data (2000–2024). A computational framework integrating unsupervised learning, dimensionality reduction, and predictive modeling was employed. Principal Component Analysis synthesized eight indicators into a Compound Vulnerability Score (CVS), while K-Means and DBSCAN identified distinct vulnerability regimes. XGBoost quantified driver importance, and Graph Neural Networks captured systemic interconnections. XGBoost identified projected drought risk (31.2%), precipitation change (18.1%), and poverty headcount (14.3%) as primary drivers. Graph networks demonstrated significant risk amplification in African nations (Morocco SRS: 0.728–0.874; Kenya SRS: 0.504–0.641) versus damping in Asian countries. A Reinforcement Learning (RL) agent was trained using Deep Q-Networks with experience replay to optimize intervention portfolios under budget constraints. The RL policy achieved a 23% reduction in systemic risk compared to uniform allocation baselines, generating context-specific priorities: drought management for Morocco (score 50) and Pakistan (40); poverty alleviation for Kenya (40); coastal protection for Bangladesh (40); agricultural resilience for Vietnam (35); and institutional capacity building for Colombia (50). In conclusion, socio-economic fragility non-linearly amplifies climate hazards, with poverty and drought risk constituting critical vulnerability multipliers. The AI-driven framework demonstrates that targeted interventions in high-sensitivity systems maximize systemic risk reduction. This integrated approach provides a replicable, evidence-based foundation for strategic adaptation finance allocation in an increasingly uncertain climate future. Full article

Transmission tower climbing is a critical specialized operation in ultra-high-voltage power maintenance and communication infrastructure servicing. However, existing lower-limb exoskeletons used for tower climbing still suffer from insufficient motion intention recognition accuracy under complex operational environments. To address this issue, this study proposes an inertial measurement unit (IMU)-based bidirectional temporal deep learning method for motion intention recognition. First, a one-dimensional convolutional neural network (1D-CNN) is employed to extract local temporal features from multi-channel IMU signals. Subsequently, a bidirectional long short-term memory network (Bi-LSTM) is introduced to model the forward and backward temporal dependencies of motion sequences. Furthermore, a temporal attention mechanism is incorporated to emphasize discriminative features at critical movement phases, enabling the precise recognition of short-duration and transitional motions. Experimental results demonstrate that the proposed method outperforms traditional machine learning approaches and unidirectional temporal models in terms of accuracy, F1-score, and other evaluation metrics. In particular, this method demonstrates significant advantages in identifying the flexion/extension phases and transitional states. This study provides an offline method for analyzing movement intentions in lower-limb exoskeleton control for power transmission tower climbing scenarios and offers a reference for developing assistive control strategies for assisted climbing tasks in this specific context. Full article

In underground coal mining, the gear of a shearer’s rocker arm endures extreme stress and environmental fluctuations. Failures in this vital component can pose serious safety hazards, cause prolonged operational downtime, and result in significant financial losses. Therefore, accurate gear fault diagnosis is crucial. However, conventional diagnostic methods often struggle with limited feature extraction and poor performance when dealing with non-stationary, noisy signals typical of this environment. To address these challenges, a hybrid model consisting of Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) network, and Markov Transition Model (MTM) is proposed. In this framework, the CNN is used to extract both global and local features related to gear fault. A time-distributed feature extractor is then integrated with the LSTM to capture the temporal progression of these features, aiding in effective modeling of fault evolution over time. Finally, the MTM further refines classification by incorporating probabilistic state transition between fault conditions, thereby improving diagnostic stability and robustness under noise. Experimental validation was done using vibration data from the Taizhong Coal Machinery rocker arm test platform and gear data from Southeast University and achieved up to 99.79% accuracy. These results show this proposed method outperformed other advanced diagnostic methods, offering dependable fault diagnosis and strong noise resistance even under extreme noise conditions of −5 dB SNR. Full article

Cloud cover significantly reduces the electrical power output of photovoltaic systems, making accurate short-term cloud movement predictions essential for reliable solar energy production planning. This article presents a deep learning framework that directly estimates cloud movement from ground-based all-sky camera images, rather than predicting future production from past power data. The system is based on a three-step process: First, a lightweight Convolutional Neural Network segments cloud regions and produces probabilistic masks that represent the spatial distribution of clouds in a compact and computationally efficient manner. This allows subsequent models to focus on the geometry of clouds rather than irrelevant visual features such as illumination changes. Second, a Vector Quantized Variational Autoencoder compresses these masks into discrete latent token sequences, reducing dimensionality while preserving fundamental cloud structure patterns. Third, a GPT-style autoregressive transformer learns temporal dependencies in this token space and predicts future sequences based on past observations, enabling iterative multi-step predictions, where each prediction serves as the input for subsequent time steps. Our evaluations show an average intersection-over-union ratio of 0.92 and a pixel accuracy of 0.96 for single-step (5 s ahead) predictions, while performance smoothly decreases to an intersection-over-union ratio of 0.65 and an accuracy of 0.80 in 10 min autoregressive propagation. The framework also provides prediction uncertainty estimates through token-level entropy measurement, which shows positive correlation with prediction error and serves as a confidence indicator for downstream decision-making in solar energy forecasting applications. Full article

Osteoarthritis (OA) is a complex degenerative joint disease for which early diagnosis and clear molecular characterization remain limited. DNA methylation has been increasingly recognized as an important regulatory factor in OA pathogenesis. In this study, we proposed an integrative computational framework combining statistical analysis, machine learning, deep learning, and functional genomics to identify and validate OA-associated genes and methylation biomarkers for diagnostic and biological interpretation. Candidate CpG sites were obtained using two complementary strategies: differential methylation analysis and selection of loci located near transcription start sites of previously reported OA-related genes. Key features were further refined using support vector machine recursive feature elimination and random forest algorithms. Based on the selected loci, we developed a feature-fusion diagnostic model that combines Transformer and convolutional neural networks with adaptive weighting to capture both global dependency structures and local methylation patterns. A panel of 220 methylation sites demonstrated stable and reproducible diagnostic performance in an independent cohort. Functional annotation and pathway analysis highlighted several established OA-associated genes, including TGFBR2, SMAD3, PPARG, and MAPK3, and suggested INHBB as a potential novel effector gene, with additional support for AMH and INHBE involvement. Overall, this study presents a robust methylation-based framework for identifying key OA-associated genes and provides new insights into the epigenetic mechanisms underlying OA. Full article

By 2024, global photovoltaic (PV) capacity exceeded 2000 GW, corresponding with a decline in levelized costs of approximately 90% since 2010. Artificial intelligence (AI) and machine learning (ML) are enabling novel approaches to solar energy system design and implementation. This survey offers a detailed evaluation of AI/ML methodologies utilized across the solar energy value chain, with a focus on solar irradiance forecasting, maximum power point tracking (MPPT), fault identification, and the expeditious discovery of system materials. The distinction between AI as the broader paradigm and ML as its data-driven subset is drawn and maintained throughout. The primary results cite forecasting improvements via deep learning architectures (LSTM, CNN, Transformer) of 10–40% over traditional methods, while hybrid numerical weather prediction and deep learning models achieve mean absolute error reductions of 15–25%. Reinforcement learning-based MPPT achieves tracking efficiencies in excess of 99% under partial shading, CNN-based fault classification reaches accuracies above 95%, and ML-based screening of materials accelerates perovskite optimization by a factor of 5–10×. Promising paradigms such as explainable AI, federated learning, digital twins, and physics-informed neural networks are evaluated alongside technical, economic, and regulatory constraints. This survey provides a consolidated reference and practical roadmap for the advancement of AI-driven solar energy technologies. Full article

Prolonged sitting with poor posture is associated with musculoskeletal disorders, reduced productivity, and long-term health risks. Many existing posture monitoring systems predominantly rely on single-modality sensing, such as pressure or vision-based approaches, limiting their ability to capture both static alignment and dynamic micro-movements. This study proposes a multimodal smart-skin system integrating pressure, temperature, and vibration sensors for sitting posture recognition. A total of 42 sensors distributed across 14 anatomical locations were deployed, generating 15,037 samples collected over three independent sessions to evaluate cross-session temporal generalization across nine posture classes under controlled experimental conditions. Two deep learning architectures—Temporal Convolutional Networks with Attention (TCN + Attn) and Convolutional Neural Network–Long Short-Term Memory (CNN − LSTM)—were compared under Leave-One-Session-Out (LOSO) cross-validation. TCN + Attn achieved 85.23% LOSO accuracy, outperforming CNN − LSTM by 2.56 percentage points while reducing training time by 36.7% and inference latency by 33.9%. Ablation analysis revealed that temperature sensing was the most discriminative unimodal modality (71.5% accuracy), and full multimodal fusion improved LOSO accuracy by 22.93% compared to pressure-only configurations. These results demonstrate the feasibility of multimodal smart-skin sensing combined with temporal convolutional modeling for cross-session posture recognition and indicate potential for efficient real-time, privacy-preserving ergonomic monitoring. This study should be interpreted as a controlled, single-subject proof-of-concept, and further validation in multi-subject and real-world environments is required to establish broader generalizability. Full article

Inertial measurement units (IMUs) in low-cost navigation systems suffer from significant drift and noise errors due to sensor biases, scale factor instability, and nonlinear stochastic noise. This paper proposes a hybrid error compensation approach that combines Fast Orthogonal Search (FOS), a nonlinear system identification technique, with deep Long Short-Term Memory (LSTM) neural networks to improve IMU signal accuracy in GNSS-denied navigation. The FOS algorithm efficiently models deterministic error patterns (such as bias drift and scale factor errors) using a small training dataset, while the LSTM learns the IMU’s complex time-dependent error dynamics from much longer training data. In the proposed method, FOS is first used to predict the output of a high-end IMU based on that of a low-end IMU, and the trained FOS model is then used to extend the training data for an LSTM-based predictor. We demonstrate the efficacy of this FOS–LSTM hybrid on real vehicular IMU data by training with a limited segment of high-precision reference measurements and testing on extended operation periods. The hybrid model achieves high predictive accuracy for predicting the high-end signal based on the low-end signal, with a mean squared error below 0.1% and yields more stable velocity estimates than models using FOS or LSTM alone. Although long-term position drift is not fully eliminated, the proposed method significantly reduces short-term uncertainty in the inertial solution. These results highlight a promising synergy between model-based system identification and data-driven learning for sensor error calibration in navigation systems. Key contributions include FOS-based pseudo-label bootstrapping for data-efficient LSTM training and a navigation-level evaluation illustrating how signal correction impacts dead reckoning drift. Full article

With the rapid development of e-commerce and smart manufacturing, automated warehouse systems have become critical infrastructure for modern logistics. In China’s vast market, the dynamic scheduling of Rail-Guided Vehicles (RGVs) faces significant challenges due to complex task uncertainties, hierarchical supply chain structures, and real-time collision avoidance requirements. Traditional rule-based methods and static optimization models often fail to adapt to such dynamic environments. To address these issues, this paper proposes a novel hybrid deep reinforcement learning framework integrating a Dynamic Graph Neural Network (DGNN) and a Transformer model. The DGNN captures the spatiotemporal dependencies of the warehouse network topology, while the Transformer mechanism enhances long-range feature extraction for task prioritization. Furthermore, we design a centralized Deep Q-network (DQN) framework with parameterized action spaces to coordinate multiple RGVs collaboratively. While the system manages multiple physical vehicles, the learning architecture employs a single-agent global scheduler to avoid the non-stationarity issues inherent in multi-agent reinforcement learning. Experimental results based on real-world data from a large-scale electronics manufacturing warehouse demonstrate that our method reduces average task completion time by 18.5% and improves system throughput by 22.3% compared to state-of-the-art baselines. The proposed approach demonstrates potential for intelligent warehouse management in dynamic industrial scenarios. Full article

Background/Objectives: Peptide receptor radionuclide therapy (PRRT) is an established treatment for metastatic neuroendocrine tumors (NETs), yet long-term disease control occurs only in a subset of patients. Predicting progression-free survival (PFS) could support individualized treatment planning. This study evaluates laboratory, imaging, and multimodal deep learning models for PFS prediction in PRRT-treated patients. Methods: In this retrospective, single-center study 116 patients with metastatic NETs undergoing [177Lu]Lu-DOTATOC were included. Clinical characteristics, laboratory values, and pretherapeutic somatostatin receptor positron emission tomography/computed tomographies (SR-PET/CTs) were collected. Seven models were trained to classify low- vs. high-PFS groups, including unimodal (laboratory, SR-PET, or CT) and multimodal fusion approaches. Performance was assessed via repeated 3-fold cross-validation with area under the receiver operating characteristic curve (AUROC) and area under the precision–recall curve (AUPRC). Explainability was evaluated by feature importance analysis and gradient based saliency maps. Results: Forty-two patients (36%) displayed short PFS (≤1 year) and 74 patients displayed long PFS (>1 year). Groups were similar in most characteristics, except for higher baseline chromogranin A (p = 0.003), elevated $\gamma$-GT (p = 0.002), and fewer PRRT cycles (p < 0.001) in short-PFS patients. The Random Forest model trained only on laboratory biomarkers reached an AUROC of 0.59 ± 0.02. Unimodal three-dimensional convolutional neural networks using SR-PET or CT performed worse (AUROC 0.42 ± 0.03 and 0.54 ± 0.01, respectively). A multimodal fusion model integrating laboratory values, SR-PET, and CT—augmented with a pretrained CT branch—achieved the best results (AUROC 0.72 ± 0.01, AUPRC 0.80 ± 0.01). Explainability analyses provided insights into model predictions, with explainability patterns in the fusion model appearing physiologically plausible and predominantly tumor-focused. Conclusions: Multimodal deep learning combining SR-PET, CT, and laboratory biomarkers outperformed unimodal approaches for PFS prediction after PRRT. Upon external validation, such models may support risk-adapted follow-up strategies. Full article