Search Results (479)

Stereo vision is essential for passive three-dimensional perception in resource-constrained applications that require low power consumption, predictable latency, and explainable geometry. Although deep learning architectures dominate recent benchmarks, the classical block-matching pipeline remains a foundational approach. Optimizing this pipeline involves navigating complex trade-offs among matching robustness, map density, and computational efficiency. This study systematically surveys and physically validates the classical stereo framework. After revisiting geometric first principles, three matching costs (SAD, NCC, ZNCC) are benchmarked alongside Sobel preprocessing and structural refinements, with subsequent validation using a calibrated consumer webcam rig. Middlebury benchmarks (2001–2021) indicate that while SAD fails under complex radiometric distortion, NCC consistently achieves superior quantitative metrics, incurring only a 1.2-fold computational overhead. Extending the disparity search range improves foreground localization, while block size imposes a trade-off between resolving the aperture problem and preserving fine geometric detail. To bridge theoretical analysis and practical deployment, the pipeline is validated using a custom-calibrated consumer stereo rig. The optimized Sobel-NCC architecture is then evaluated for real-time edge deployment on constrained hardware (NVIDIA Jetson Nano) and narrow-baseline sensors (OAK-D SR) in the context of agricultural robotic manipulation. By prioritizing metric precision over dense prediction, the classical pipeline reconstructs target surfaces with approximately 1 cm depth accuracy at 21 frames per second. These results demonstrate that optimized local algorithms offer deterministic and reliable geometric foundations for real-time edge-computed robotics. Although neural networks are essential for dense reconstructions in ill-posed regions, the foundational principles established here remain indispensable for advanced stereo vision system deployment. Full article

Optimal sizing of hybrid renewable energy systems for desert camps is a multi-objective problem that must account for cost, reliability, component degradation, and uncertainty. This paper introduces an improved multi-objective neural architecture search (IMONAS) framework for hybrid nanogrid sizing in the desert environment of Hafr Al-Batin, Saudi Arabia. The framework combines neural optimization, stochastic uncertainty modeling, and explicit battery degradation modeling, a combination not addressed in the reviewed studies for this application. Six test cases are examined by varying uncertainty assumptions, battery degradation, and the annual duration of uncertain operation. For each case, IMONAS provides Pareto-front solutions that specify the photovoltaic, diesel generator, battery autonomy, and inverter choices while minimizing the cost of energy (COE) and the loss of power supply probability (LPSP). IMONAS is compared with the original MONAS and five other multi-objective optimization methods. In addition to visual Pareto-front comparisons, the assessment uses Pareto-dominance indicators, namely the C-metric and an aggregated score derived from pairwise C-metric comparisons across the algorithms and cases. The results provide a validated sizing framework for remote arid-region nanogrids under uncertainty and battery degradation. Full article

Price spikes are rare but economically significant events observed across electricity, financial, commodity, and cryptocurrency markets. Their abrupt magnitude, heavy-tailed distributions, and severe class imbalance make them difficult to forecast using conventional time-series methods. This systematic literature review, conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework, synthesises recent deep learning approaches to forward-looking price-spike prediction and classification. Searches of Scopus, Web of Science, and IEEE Xplore identified studies published between 2020 and 2026. Following screening and full-text eligibility assessment of approximately 300 studies, only 20 met the inclusion criteria and were included in the final synthesis, comprising 19 peer-reviewed papers and one doctoral thesis. The review develops a structured taxonomy spanning spike definitions, task formulations, model architectures, input design, and evaluation practices. A central finding is that predictive performance is driven more by problem formulation, label construction, and evaluation design than by model architecture. While architectures have diversified to include recurrent networks, transformers, graph neural networks, and hybrid frameworks, improvements are often attributable to differences in how the prediction problem is defined rather than the models themselves. Key limitations stem from inconsistent spike definitions and insufficient treatment of class imbalance, leading to a misalignment between modelling objectives and evaluation practices, further exacerbated by the absence of standardised benchmarks. These issues hinder comparability and can lead to overstated model performance by masking poor detection of rare but economically critical spike events. The review therefore identifies clear directions for future research, including standardised spike labelling, adoption of rare-event-appropriate evaluation frameworks, and problem formulations that explicitly target extreme-event prediction. Full article

The continued scaling of semiconductor devices at advanced technology nodes introduces significant challenges in maintaining performance, reliability, and design efficiency. This work presents a data-driven framework for the modeling and optimization of nanosheet (NS) and forksheet (FS) transistors using deep learning and Bayesian optimization. An extensive dataset is generated through LTSpice-based circuit simulations, enabling efficient exploration of the design space while incorporating key device parameters, including channel length, channel width, supply voltage, temperature, and threshold voltage, together with variability and noise effects. A deep neural network (DNN) is developed as a surrogate model to learn the nonlinear relationship between input parameters and transistor switching behavior, achieving strong predictive performance with a coefficient of determination ($R^2\approx 0.91$), mean absolute error (MAE $\approx 0.024$), and root mean square error (RMSE $\approx 0.031$) on unseen test data. To improve physical consistency, a bounded-output formulation is introduced to guarantee physically admissible voltage predictions, while device-level benchmarking is performed to assess agreement with expected transistor characteristics. The results demonstrate accurate modeling of transient behavior across the sampled operating conditions. Comparative analysis shows that NS devices achieve faster switching and lower propagation delay, whereas FS devices exhibit improved stability under certain conditions. Bayesian optimization is employed to efficiently explore the design space and identify high-performing transistor configurations without exhaustive simulation-based searches. The proposed framework provides a scalable and computationally efficient methodology for surrogate modeling, design-space exploration, and early-stage assessment of advanced transistor architectures. Full article

Gliomas constitute the majority of primary brain tumors, and accurate diagnosis through MRI is essential for patient management. Existing computer-aided diagnosis approaches frequently rely on tumor segmentation frameworks. In this study, a segmentation-independent framework for volumetric low-grade versus high-grade glioma (LGG/HGG) classification is proposed using a Convolutional Neural Network (CNN) designed through task-oriented Neural Architecture Search (NAS). The proposed method was evaluated on a multi-center dataset comprising 1194 patients with pre-operative MRI scans, including T1-CE and FLAIR sequences from four publicly available cohorts. NAS was conducted within a controlled search space to optimize a 3D U-Net–based backbone using Tree-structured Parzen Estimator (TPE) combined with Hyperband pruning. The optimized backbone was enhanced with residual connections and Squeeze-and-Excitation (SE) attention mechanisms to improve feature representation and training stability. Internal validation employed repeated 5-fold cross-validation across all four multi-center datasets. An external experiment used REMBRANDT as a test cohort (49 LGG, 19 HGG). The proposed model achieved 88.25% internal accuracy and 75.51% external accuracy (macro-F1: 87.37% internal, 73.77% external), outperforming benchmark 3D CNNs. Explainable Artificial Intelligence (XAI) analysis based on Grad-CAM revealed robust tumor localization without segmentation supervision, validated against available ground-truth masks. Additional experiments demonstrated the model’s generalization capacity, achieving 89.51% accuracy for IDH mutation prediction and 78.74% for multi-grade classification. Full article

With the increasing penetration of renewable energy systems, multilevel inverters have been widely adopted to meet the growing demand for high-power and high-quality energy conversion. Among various multilevel topologies, cascaded H-bridge multilevel inverters (CHMIs) are particularly attractive due to their modular structure and improved output voltage quality. However, the increased number of power semiconductor devices and switching states significantly complicates fault diagnosis under practical operating conditions. Currently, most existing neural networks for fault diagnosis are manually designed based on domain expertise. This may limit their adaptability to task-specific fault patterns as well as edge-side inference performance. To reduce the dependence on manually designed diagnostic networks, an edge-oriented fault diagnosis framework based on differentiable architecture search (DARTS) is proposed to automatically design task-specific diagnostic networks. A simplified special cell search strategy is adopted to improve search efficiency and facilitate practical deployment. The searched architectures are lightweight and suitable for deployment on edge platforms. The experiments show that the proposed method achieves an average diagnostic accuracy of 99.44% on the test set under the RL load of $(7\mathrm{\Omega },6\mathrm{mH})$. Furthermore, the searched model contains only 0.2417 M trainable parameters, and edge deployment experiments on the Jetson Orin Nano platform show low-latency inference capability. Full article

Neural architecture search (NAS) is a core technology in AutoML, but it faces challenges such as high computational costs and inefficient evaluation. Traditional NAS methods require fully training each candidate architecture, leading to substantial resource consumption and long evaluation times. This paper introduces a four-stage progressive multi-task learning framework that shifts from brute-force search to performance prediction. By progressively training from simple synthetic data to complex real data, the framework enables efficient architecture performance prediction. The main contributions are a unified progressive predictor paradigm, a deployment-aware multi-task prediction mechanism with dual lightweighting, and a benchmark-aware data-and-transfer framework based on NATS-Bench reconstruction and progressive knowledge distillation. Experiments on 15,000 NATS-Bench architectures with a fixed train–validation–test split (8:1:1) and consistent hyperparameters show a 95.56% correlation-based prediction score, computed as Pearson correlation expressed as a percentage (Pearson correlation 0.9556, R-squared 0.9134), 32-fold training efficiency improvement (37.87 s ± 2.1 s vs. 1247.6 s), and 95.7% convergence stability across five random seeds. Ablation studies, including a Direct Stage-3-Only comparison, quantify component contributions, and benchmarks compare against random forest, XGBoost, and MLP under identical data splits and feature spaces. Full article

Background/Objectives: In the past few years increasing attention has been given to the application of dry needling (DN) for spasticity in stroke survivors. Nevertheless, the underlying mechanisms of this technique have not yet been confirmed. The aim of this systematic review was to distinguish the effects of DN in post-stroke spasticity on both structural–mechanical muscle properties (SMMPs) and neurophysiological properties to address these mechanisms. Methods: A literature search was performed in Web of Science, PubMed, Scopus and Embase following PRISMA guidelines (PROSPERO ID: 1163064). Randomized controlled trials and case–control studies involving adults with post-stroke spasticity treated with DN were included. Outcomes were categorized as SMMPs (e.g., muscle architecture, passive stiffness, PROM) or neurophysiological measures (e.g., H-reflex, H/M ratio). Standardized effect sizes (Hedges’ g) were calculated when possible; however, heterogeneity in outcomes and incomplete variance reporting precluded meta-analysis. Results: Twelve studies met the inclusion criteria. Most of these studies assessed passive range of motion, reporting a significant increase following the intervention. Only two of the included studies examined structural characteristics, and five studies included neurophysiological outcomes. Correlations between mechanistic outcomes and clinical spasticity grading (MAS/MMAS) were weak. Emerging evidence suggests DN may additionally modulate local inflammatory mediators, indicating a potential neuroimmune contribution to its effects. Conclusions: DN appears to improve structural–mechanical muscle properties and produce moderate reductions in reflex excitability in individuals with post-stroke spasticity. Mechanical adaptations are more consistently demonstrated than neural changes, and neither domain is proportionally reflected in clinical spasticity scales. Evidence remains limited by small samples, methodological variability, and incomplete reporting. Further mechanistic research is needed to clarify how DN influences the complex pathophysiology of post-stroke spasticity. Full article

Target detection in complex marine environments serves as a core technology for marine environmental monitoring, underwater search and rescue, and vessel collision avoidance. However, traditional detection methods struggle to accurately identify low-pixel marine targets (e.g., small vessels, buoys) while balancing detection accuracy and computational efficiency in complex environments. To address the low accuracy and high computational costs of object detection in complex marine environments, this paper proposes YOLO-ELR, a lightweight model based on the YOLOv11 framework, designed to identify object categories and directional positions with enhanced precision while optimizing resource efficiency. The Efficient Multi-Branch Scale and Light Adaptive-weight downsampling (EMBSLaw) backbone network enhances multi-scale object detection in complex scenes by dynamically adjusting feature contributions through adaptive weight computation, while maintaining a lightweight architecture. To reduce computational parameters and complexity, a novel lightweight spatial multi-branch detector Lightweight Shared Convolutional Separamter BN Detection head (LSCSBD) is introduced. Furthermore, the RepGhostCSPELAN and Efficient Multi-Scale Conv (RGCEL_EMSC) module, integrating multi-neural networks, is proposed to improve detection accuracy and precision. Experimental results demonstrate that the YOLO-ELR model achieves an mAP@50 of 84.87%, surpassing the baseline YOLOv11 by 3.94% while reducing parameters by 35% and GFLOPs by 7.56%, which validates the effectiveness in balancing detection accuracy and computational efficiency. Full article

Traumatic injuries often result in nerve tissue damage and functional deficits due to limited regeneration. Silk fibroin, a biopolymer with inherent biocompatibility and tunable properties, is a promising material for 3D-bioprinted neural tissue scaffolds. This review highlights recent advancements in silk-derived composite scaffolds, often incorporating additional materials like collagen or conductive polymers to enhance their performance. This review examines how material composition, scaffold architecture, and fabrication strategy influence biological response and functional recovery. This comprehensive review follows PRISMA guidelines and uses comprehensive searches of PubMed, MEDLINE, Embase, Web of Science, Cochrane Central, and ClinicalTrials.gov for studies published through 2025. Studies were screened for eligibility based on substance type, mechanical properties, production methods, and outcomes. Findings were synthesized qualitatively. Twelve studies were included, comprising rat (50%), canine (8.3%), and in vitro (41.7%) models. Analysis reveals that silk fibroin acts as a highly adaptable mechanical backbone. It can consistently integrate with bioactive additives (collagen, dECM) or conductive polymers (Polypyrrole, MXene) to meet specific therapeutic demands. For spinal cord injuries, composites reached a compressive modulus capable of resisting physiological pressures and preventing scaffold collapse. In soft tissue applications, silk–hydrogel blends provided localized release of exosomes and small molecules during the acute injury phase, reducing neuroinflammatory markers. Additionally, adding conductive materials allowed the scaffolds to bridge electrical gaps and promote Schwann cell proliferation and neuronal differentiation. Furthermore, 3D bioprinting enabled the creation of defined microchannels that replicate native fascicular architecture. In vivo outcomes consistently showed superior axonal regeneration, myelination, and synaptic reconnection compared to controls, correlating with significant improvements in electrophysiological and motor function. This review highlights the clinical potential of silk fibroin-based 3D-printed biomaterials for nerve regeneration, including neural repair and neural tissue engineering. More recent studies place greater emphasis on integrating mechanical, architectural, and biological considerations into scaffold design, resulting in increasingly multifunctional scaffold systems. Despite promising efficacy, the heterogeneity of fabrication methods and the predominance of rodent models highlight the need for standardized protocols and evaluations in relevant models to facilitate clinical translation. Full article

Securing IoT-enabled wireless sensor network links requires selecting a minimum-cost set of connected monitor nodes that observes every link while satisfying capacity constraints, a problem known as the minimum weighted connected capacitated vertex cover (MWCCVC). To the best of our knowledge, this work introduces the first learning-based framework for the MWCCVC through a three-stage pipeline that combines supervised graph neural networks, feasibility repair, and local search. We compare twelve graph neural network architectures, including graph convolutional network, graph attention network, GraphSAGE, Graph Isomorphism Network (GIN), and GraphTransformer, under unified features, loss functions, and hyperparameter tuning. Throughout the evaluation on 309 benchmark instances under a 5-fold cross-validation protocol, feasibility is guaranteed by the deterministic repair module instead of being learned by the network, resulting in 100% feasible covers across all evaluated instances. At the large scale, GIN, GraphSAGE, DeeperGIN, and EdgeAwareGIN reach parity with the state-of-the-art hybrid genetic algorithm (HGA), with GIN attaining a mean gap of $-0.37\%$ (a difference of less than one percentage point) while completing in seconds instead of HGA’s hours. Statistical tests across the full 309-instance benchmark confirm significant differences between the architectures, with Friedman ${\chi }^2=93.05$, $p<{10}^{-4}$. The best-performing architectures remain within about $2\%$ of HGA on small- and medium-scale instances, where HGA is near-optimal, and become the preferred choice at the large scale, mainly because their wall-clock time is much shorter than HGA’s at the same solution quality. Full article

As data volumes grow and business requirements become increasingly complex, Hybrid Transactional/Analytical Processing (HTAP) technologies, capable of handling both Online Transaction Processing (OLTP) and Online Analytical Processing (OLAP) workloads on a single platform, have gained prominence. HTAP databases typically maintain dual data storage formats and dual query engines: one row-oriented for OLTP, and another column-oriented for OLAP. Query plans, known as hybrid plans, can be segmented and pushed down to execute on these different formats. However, existing HTAP solutions still face challenges in optimizing these hybrid plans, struggling to explore the vast space of potential execution strategies effectively. To address these issues, this study introduces a learning-based query optimizer for row–column dual store HTAP database systems, which automatically generates multiple high-quality query optimizer hints (HINTs) to derive candidate plans. To balance plan generation efficiency with plan quality, a lightweight, learning-based algorithm using Monte Carlo Tree Search (MCTS) for generating hybrid access HINTs is proposed. Moreover, a Transformer-based neural network model coupled with a hybrid plan feature representation method is developed to select the candidate execution plan with the lowest predicted execution time. This work focuses on latency-oriented hybrid-plan selection for analytical queries in a row–column dual-store HTAP architecture; the current evaluation does not cover full mixed OLTP/OLAP workload scheduling, transactional interference, or concurrency control, which are left as future work. Experimental results on AlloyDB Omni, a recent row–column dual-store HTAP database, using the real-world IMDB dataset and JOB benchmark demonstrate that our system reduces execution time by 75.02% compared to the Cost-Based Optimizer (CBO) and by 62.23% compared to the state-of-the-art row-store-based learning query optimizer in this evaluated analytical-query setting. Full article

Federated learning enables multiple clients to collaboratively train a global model without exchanging raw data, thereby alleviating data silos and privacy concerns. However, federated learning model design is still challenged by complex architecture and hyperparameter optimization, especially under heterogeneous data distributions. To address these issues, this paper proposes a multi-objective federated learning framework based on particle swarm optimization (PSO). Specifically, neural architecture search and training-parameter optimization are formulated as a multi-objective evolutionary optimization problem, where predictive performance and model complexity are optimized jointly. Furthermore, to improve robustness under non-independent and identically distributed (non-IID) data, a clustered multi-population extension is developed. In this framework, each cluster is associated with a dedicated particle population. An inter-population evolutionary knowledge transfer mechanism is then introduced to enable effective sharing of search experience across clusters, thereby improving the optimization efficiency and adaptability of federated learning under heterogeneous environments. Experiments on MNIST and CIFAR10 under both IID and non-IID settings demonstrate that the proposed methods achieve a superior trade-off between predictive accuracy and model complexity and outperform baseline approaches in robustness, search efficiency, and adaptability to imbalanced distributed environments. Full article

Dynamic gravity measurement, crucial for engineering application, suffers accuracy degradation due to errors induced by carrier maneuvers. To address the limitations of conventional swarm optimization algorithms in precision, stability, and generalizability, this study proposes a Swarm Cooperation Evolution Strategy (SCES) that efficiently integrates multiple algorithms. The proposed SCES is extensively evaluated on the CEC2022 benchmark suite in comparison with several cooperative fusion-related algorithms and representative single optimization algorithms. The experimental results demonstrate that SCES achieves an overall effectiveness score of 0.034 and an optimal accessibility rate exceeding 95%. Compared to the best-performing fusion-based algorithm, these metrics represent improvements of 54.67% and 31.11%, respectively. Moreover, relative to the best-performing single optimization algorithm, the improvements amount to 37.73% and 32.69%, respectively. These findings robustly validate the superior performance of the proposed algorithm. Moreover, an in-depth investigation based on SCES into dynamic error compensation methodologies is conducted. Firstly, a polynomial compensation model is established through error mechanism analysis, with parameters identified via SCES. Secondly, a data-driven compensation model employing a multi-layer long short-term memory (LSTM) network optimized via neural architecture search (NAS) guided by SCES is proposed, circumventing the performance limitations inherent in manually designed networks. Furthermore, an innovative two-stage hybrid strategy is introduced. Systematic trend errors are compensated using the polynomial model, followed by the NAS-LSTM model addressing complex residual nonlinear errors, effectively combining mechanism-based and data-driven approaches. Validation on three lines exhibiting varying maneuverability shows all methods significantly improve accuracy. The hybrid strategy delivers optimal performance, achieving 0.58 mGal internal coincidence accuracy on stable lines and up to 91.58% improvement in external coincidence accuracy under high maneuverability. This research provides an effective high-precision dynamic gravity measurement and compensation solution, advancing engineering applications. Full article

This paper investigates solar energy production forecasting at a monthly temporal resolution using a pooled neural network framework applied to the Chikalov photovoltaic systems in southwestern Bulgaria. The study considers several related PV installations with unequal time-series lengths and formulates the forecasting task as one-step-ahead prediction of the next monthly total energy yield, measured in kWh, in a global pooled setting. Two complementary neural architectures are compared: a multilayer perceptron (MLP), which serves as a nonlinear feed-forward benchmark based on lagged observations and seasonal descriptors, and a gated recurrent unit (GRU), which explicitly models sequential temporal dependence. In both cases, seasonality is represented through cyclical calendar encodings, while model selection is performed by chronological hyperparameter search using a separate validation block. Forecast accuracy is assessed by RMSE, MAE, coefficient of determination (R²), MAPE, and sMAPE, and uncertainty is quantified through validation residual prediction intervals. The results show that the MLP achieves stronger validation performance, whereas the GRU provides better final out-of-sample generalization after refitting on the combined training and validation data. For both architectures, the best configurations are obtained with a 12-month input horizon, indicating that one full annual cycle contains the most informative memory for forecasting monthly aggregated photovoltaic energy yield in the considered dataset. After refitting on the combined training and validation data, the GRU achieved the best final out-of-sample performance, with RMSE = 296.38 kWh, MAE = 213.16 kWh, R² = 0.9231, MAPE = 7.52%, and sMAPE = 7.49%. Overall, the findings demonstrate that pooled neural modeling is an effective framework for monthly PV production forecasting and can provide practically useful support for sustainable energy planning, monitoring, and optimization. Full article