Search Results (20,941)

Domain-Specific Languages (DSLs) are essential in software engineering for safely expressing complex domain logic. However, Large Language Models (LLMs) struggle to generate syntactically and semantically correct code for niche DSLs due to sparse representation in pre-training corpora. While Grammar-Constrained Decoding (GCD) resolves syntactic hallucinations by masking logits through a formal Context-Free Grammar (CFG), empirical evidence shows that strict GCD disrupts the autoregressive Chain-of-Thought (CoT) reasoning of modern models, frequently forcing them into irreversible semantic dead-ends. To overcome the friction between internal neural reasoning and external symbolic constraints, we propose a Dual-Phase Cascaded Neurosymbolic framework. In the first phase, the model is provided with dynamically injected grammar rules and is permitted to reason unconstrained, producing an optimistic code draft. If the draft fails native compiler checks, the system enters a second phase: it preserves the successful semantic reasoning from Phase 1 but re-generates the code under strict GCD enforcement. This cascaded architecture utilizes the formal FSM not as an adversarial constraint, but as a localized syntax repair engine guided by the model’s own prior reasoning. We construct a comprehensive benchmark of 100 MiniZinc constraint programming tasks and evaluate our approach using a $pass@k$ metric with a strict semantic LLM judge. Our findings demonstrate that this dual-phase “think-then-constrain” approach significantly outperforms zero-shot, pure few-shot, and pure GCD baselines, achieving highly reliable, training-free code generation for unseen DSLs. Full article

Based on survey data from 579 farmer households in major non-timber forest product (NTFP) regions of Zhejiang Province, this study comprehensively employs binary logit/ordered probit and machine learning methods for cross-validation. It systematically examines the effects of multiple factors, including perceived property rights security, technical training, village rules and regulations, ecological awareness, and economic incentives, on forest farmers’ adoption of green production technologies. The cross-validation between econometric and machine learning approaches enhances the reliability of the findings. Results show that perceived property rights security is robustly and positively associated with green production behavior, while village rules and regulations and ecological awareness emerge as the two most critical driving factors. These associations exhibit significant NTFP-type heterogeneity: the impacts of technical training and forestry subsidies vary in direction depending on the crop cultivated, rendering traditional one-size-fits-all policies ineffective. This study highlights the crucial role of informal institutions and environmental awareness in the green transition, offering empirical evidence for designing differentiated training programs, optimizing penalty gradients, and implementing targeted subsidy policies. Full article

Background: Digital behavioral phenotyping of autism spectrum disorder (ASD) offers a promising approach for developing more scalable diagnostic frameworks across diverse global contexts. Machine learning (ML) models show promise for ASD diagnosis using behavioral videos, but critical questions remain regarding whether models trained on data from one country work in another, and how the background of the raters affects the accuracy. Our work addresses these questions by testing whether ML models can accurately diagnose ASD across different populations and rater groups. Methods: This work evaluates the performance of a supervised ML framework for binary classification of ASD versus non-ASD [speech, language and communication disorders (SLC) + neurotypical (NT)] in a cohort of 227 children in Bangladesh. We first assessed the cross-domain model transferability of a clinical-instrument-trained logistic regression model (LR-9) on behavioral ratings that were based on videos of Bangladeshi children interacting with caregivers and toys at two major child development centers in Dhaka, Bangladesh. We then trained five diverse classifiers (Logistic Regression, Random Forest, XGBoost, SVM, and RuleFit) on the full annotated Bangladeshi dataset. Using SHAP-based consensus elbow feature selection, we identified a compact set of features that maintained the performance. Finally, we developed ensemble models to improve predictive stability. Results: The LR-9 model, originally trained on U.S. clinical instrument data, was evaluated on video-based behavioral ratings from 214 Bangladeshi children. When tested on Bangladeshi clinician ratings, the LR-9 model achieved a sensitivity of 86.1% (95% CI: [0.78–0.93]) and AUC of 0.79 (95% CI: [0.73–0.86]). The distinction across rater groups was between trained raters (clinicians and students) and crowd workers, who showed lower sensitivity 28.5% (95% CI: [0.21, 0.39]). When tested on the aggregated ratings from all groups, the model achieved an AUC of 0.78 (95% CI: [0.72–0.84]). Inter-rater reliability followed the same pattern: individual agreement was fair (Krippendorff’s α = 0.26), but the multi-rater consensus was reliable (ICC(1,k) = 0.84), with Bangladeshi clinicians showing the highest agreement (α = 0.34) and crowd workers the lowest (α = 0.20). We then trained new models directly on the Bangladeshi ratings. All model types achieved similar AUC values (0.86–0.89), with overlapping confidence intervals. Using just 8–11 key behaviors kept the similar performance while cutting the features by 66–75%. Combining ensembles gave similar results (e.g., Bayesian averaging: AUC 0.88 [0.78, 0.95]) but with more stable predictions. Conclusion: This study provides evidence that mobile video-based ASD diagnosis can achieve comparable performance (AUC: 0.89 [0.76, 0.96]) to models trained on clinical instrument data. This work contributes to the development of broader adaptable autism detection tools, bypassing the dependence on traditional clinical instrument data. Full article

Introduction: Hypocalcemia represents a major cause of seizures in children in the absence of fever or infections. Hypovitaminosis D, usually associated with a lack of proper prophylactic regimens, can trigger those events. This case series aims to highlight two cases of hypocalcemic seizures in infants, attributed to severe vitamin D deficiency, with the aim of raising awareness regarding the importance of vitamin D prophylaxis. A narrative review of the literature is also provided, which highlights similar reported cases. Methods: We hereby report two cases of male infants (aged 4 months and 5 months) who presented to the emergency department of a tertiary pediatric center with seizures in the absence of fever or infections. Diagnostic work-ups revealed low levels of total serum calcium and ionic calcium deficiency. The cause of the hypocalcemia turned out to be severe vitamin D deficiency in both cases, caused by complete absence of vitamin D supplements since birth and during maternal pregnancy. In both cases, combined oral calcium and vitamin D supplementation led to complete resolution of symptoms and restoration of normal calcium levels. A comprehensive review of the literature is also provided, which focuses on pediatric studies and case reports that analyzed the prevalence, particularities, and outcomes of hypocalcemic seizures related to vitamin D deficiency. Articles including subjects with congenital or endocrine disorders that could have caused hypocalcemia were ruled out. Results: After accessing the full-length form of each article and applying the inclusion and exclusion criteria, 27 articles were included, namely 14 studies and 13 case reports. Hypocalcemic seizures caused by vitamin D deficiency are more common in infants, are usually linked to maternal hypovitaminosis D, and have a higher prevalence in developing countries. Most of the case reports depicting hypocalcemic seizures were distinguished through very low vitamin D levels. Conclusions: These case series emphasize the importance of vitamin D supplementation for the prevention of hypocalcemia, which constitutes a major metabolic cause of seizures in infancy. Nevertheless, maternal vitamin D supplementation during pregnancy is the only factor that can ensure the presence of satisfactory deposits in the newborn and prevent hypovitaminosis D-related complications. Full article

The electrification of residential demand through electric vehicles (EVs), heat pumps (HPs), photovoltaic (PV) systems, and battery energy storage systems (BESSs) creates new congestion challenges in low-voltage (LV) grids. This study evaluates a transparent, deterministic, and real-time-capable rule-based energy management system (EMS) for curative thermal congestion management within a §14a EnWG-oriented setting. The EMS is implemented in MATLAB/Simulink and tested on a representative four-feeder LV network supplying 56 households. Congestion is detected from maximum phase root-mean-square currents using conservative transformer and feeder thresholds. After a threshold is reached, the EMS first activates available BESS support and then applies simultaneous feeder-wide EV limitation, batched round-robin curtailment, or staged feeder-wide reduction toward $4.2$ kW. In the uncontrolled case, the Feeder 3 and transformer overload areas are $62.84$ Ah and $48.50$ Ah, respectively. All controlled scenarios remove at least 98.70% of the feeder overload and eliminate the transformer overload within the reported numerical precision. The batched strategy requires $328.54$ Ah of cumulative feeder-current reduction, compared with $977.34$ Ah for simultaneous control and $816.00$ Ah for staged control, and achieves the highest feeder-relief efficiency. It therefore provides a balanced trade-off between congestion relief and intervention intensity for the investigated deterministic case. Full article

The fusion of information from multiple imaging modalities plays a very important role in medical diagnosis. Wavelet-based transformations have been identified as powerful methods for this purpose due to their multiresolution nature, which enables the simultaneous preservation of both structural and fine-detail information across different frequency bands. In this paper we present experimental results obtained from the wavelet-based decomposition and fusion of medical images using Python and PyWavelets. Seven wavelets from four wavelet families—Daubechies (‘db1’, ‘db10’), biorthogonal (‘bior1.3’, ‘bior4.4’), coiflets (‘coif1’, ‘coif10’), and discrete Meyer (‘dmey’)—were systematically evaluated across three decomposition levels. An emphasis was put on the preservation of approximation and detail sub-images. Results outline that simpler wavelets used for the wavelet-based decomposition of grayscale medical images produce more details when compared with the colored medical images. From the tested fusion rules, and for the specific image pairs used in the analyses, we conclude that the average fusion rule gives the best information, without a lack of or excess of information regarding the visual quality of the fused image. Considering entropy as a quality metric and according to its higher values at all levels, the ‘bior4.4’ wavelet emerges as the best for wavelet-based image fusion. These findings could provide practical guidance for wavelet selection in multimodal medical image fusion pipelines. Full article

This paper addresses human error in network orchestration systems and the high cost and resource requirements of integrating artificial intelligence (AI) for network orchestration. It proposes a framework for implementing an AI decision-maker and automation. Data are fed into the AI decision-maker to trigger designated automation robots’ tasks or notify IT specialists to gradually implement automated robots, ensuring efficient resource use, reducing costs, and enhancing productivity. We evaluated the proposed method in a simulation with genuinely uncertain outcomes, across 20 independent runs: the AI decision-maker reached 78.3% accuracy against an estimated 79.1% achievable ceiling, and the proposed framework reduced operational cost by 61.4 ± 0.7% relative to fully manual operation—the best of six operating policies in the training environment—while an explicit sensitivity guard, rather than the learned model, accounts for the absence of security incidents; under distribution shift, the framework retains 43.2 ± 0.8% savings, second only to a hand-tuned rule-based router that requires environment-specific threshold calibration. However, the proposed method requires an IT specialist to implement it properly, and the AI model’s accuracy depends on the amount of input data. In the end, we recommend that future work conduct a study focused on AI decision-makers, test the proposed method on real-world companies, and implement AI decision-makers across various departments to cover a broader range of the company’s systems. Full article

Federated learning enables collaborative model training without centralizing private data, yet its deployment on edge devices remains constrained by communication overhead, heterogeneous data distributions, and privacy leakage from shared model updates. Spiking neural networks offer an energy-efficient alternative to conventional artificial neural networks by transmitting sparse binary events rather than dense activations, but existing federated spiking learning methods still suffer from inefficient gradient exchange and insufficient privacy protection under non-independent and identically distributed data. This paper proposes PriSparK, a privacy-preserving and communication-efficient federated spiking neural learning framework that jointly exploits temporal event sparsity, adaptive client-side spike-gradient compression, and privacy-calibrated aggregation. The core of PriSparK is a novel Event-Sparse Differentially Private Federated Spiking Optimization algorithm, which converts local surrogate gradients into spike-saliency-aware sparse updates, dynamically allocates communication budgets across layers and time steps, and injects calibrated Gaussian noise after clipping in a low-dimensional event subspace. To mitigate accuracy degradation caused by aggressive compression and privacy perturbation, PriSparK further introduces a membrane-aware error-feedback mechanism and a heterogeneity-adaptive server aggregation rule that weights client updates according to spike activity stability and local distribution drift. Experiments on neuromorphic and vision benchmarks, including N-MNIST, DVS128 Gesture, CIFAR-10, and Fashion-MNIST, show that PriSparK achieves competitive or superior accuracy compared with federated artificial neural and spiking baselines while substantially reducing uplink communication. Under strong privacy constraints, PriSparK maintains stable convergence and improves the accuracy–communication–privacy trade-off, demonstrating its potential for privacy-sensitive edge intelligence with event-driven neural computation. Full article

A 40-year-old man presented with thunderclap headache (TCH) and vomiting post-masturbatory ejaculation. He reported a similar episode with posterior neck pain five days earlier after ejaculation. Unlike his migraine history, these headaches were unusually severe. Upon arrival, physical and neurological examinations were unremarkable except for an elevated blood pressure (154/110 mmHg). Brain MRI/MRA systematically ruled out intracerebral/subarachnoid hemorrhages, aneurysms, cerebral venous thrombosis, and cervical artery dissection, with preserved flow voids on B-PASS, T2, and T2-FLAIR. Notably, MRA demonstrated multiple segmental vasoconstrictions compared to a baseline MRA from five years prior. With an RCVS₂ score of 8, he was diagnosed with reversible cerebral vasoconstriction syndrome (RCVS). Symptoms resolved within 10 days of starting oral verapamil (120 mg/day), and follow-up MRA at one month showed partial improvement. This case highlights RCVS manifesting as a secondary cause sharing features with the proposed headaches associated with sexual activity (HSA) spectrum. RCVS is the predominant secondary cause of HSA (67–90%), often indistinguishable from primary HSA at onset due to the shared presentation of sudden TCH. Its pathophysiology likely involves an orgasm-induced sympathetic surge causing transient dysregulation of cerebral vascular tone. This case emphasizes that thorough neurovascular imaging is important in HSA to differentiate RCVS from primary headaches, ensuring the prompt initiation of calcium channel blockers and the avoidance of triptans. Full article

The paper proposes the Multi-ctx2 method, which enables image compression in lossless and near-lossless modes. It employs a generalized multi-context division method in the prediction stage, which is more efficient than other fast prediction methods. Five simple rules for computing the context number have been developed, which serve not only to identify an individualized linear predictor but also to correct the cumulative prediction error. In subsequent stages of the encoder, prediction errors are encoded in a two-stage process: first using an adaptive Golomb code, then a binary adaptive arithmetic encoder. The proposed method is characterized by short compression and decompression times while offering a good compromise between compression efficiency and encoding/decoding time. The proposed prediction method can be easily implemented in hardware due to its use of fixed-point arithmetic. Unlike many other solutions, there is no need to access the entire image data during encoding to tune the encoder parameters to a specific image. The paper demonstrates the efficiency of the proposed solution compared to competing solutions, showing improvements of 6.85% and 7.03% over JPEG-LS in lossless mode (and 10.2% in near-lossless mode) across two test sets. Full article

The sorting of Persian lime Citrus × latifolia (Yu.Tanaka) Tanaka intended for export is still performed manually in many production units, introducing variability and low repeatability. Deep learning-based vision systems offer high accuracy but at a high cost, and with decisions that are difficult to trace against a quality standard. A system was developed that integrates classical computer vision (grayscale conversion, Gaussian filtering, thresholding and edge detection) with three supervised symbolic classifiers (PRISM, ID3 and Naive Bayes) under a hierarchical decision scheme, validated against the criteria of the Mexican Standard NMX-FF-077-1996-SCFI. The models were evaluated using a set of 7017 images, with a 265-image test subset, and the physical prototype was validated with 200 fruits. The system reached an accuracy of 95.1% on the test set and 95.5% during physical operation, with an F1 score of 0.97 for the export-grade class; only 2 of 265 and 1 of 200 non-conforming fruits were wrongly admitted. The cost of the deployed prototype remained at 407 USD. Integrating classical vision with interpretable symbolic rules constitutes an accessible and auditable solution for Persian lime quality control in accordance with the standard, with reproducible performance between algorithmic evaluation and physical operation. Full article

With increasing renewable energy penetration, hydropower units face more frequent load rejection transients, which impose severe hydraulic excitation on Francis turbine runners. Although extensive studies have investigated flow dynamics and stress concentrations during transients, quantitative fatigue damage assessments for runners with pre-existing cracks remain scarce. To fill this gap, this study conducts CFD simulations coupled with one-way FSI to analyze a Francis turbine runner during load rejection, comparing uncracked and cracked configurations. Fatigue damage is evaluated using rain-flow counting, a modified S-N curve with Goodman mean stress correction, and the Palmgren–Miner linear damage rule. Results show that stress concentrations shift from the band-side to the crown-side T-junction during load rejection, with 4.5 times higher fatigue damage at the crown (D = 1.62 × 10⁻⁴) than at the band (D = 3.57 × 10⁻⁵). Pre-existing cracks increase local stress and reduce the allowable number of load rejection events from 6173 to 514 cycles. Reducing residual stress from 200 MPa to 100 MPa lowers fatigue damage by approximately 42%. This study provides a quantitative framework for transient fatigue assessments. Full article

Recent theoretical work has conceptualized computational thinking (CT) was presented as a dynamic psychological construct that evolves in parallel with Computer Science (CS). From this perspective, a 3-stage CT framework has been proposed to reflect how the construct has evolved and is being expanded due to recent advancements in the CS field. More specifically, this extended framework comprises the following stages: ‘rule-based’ CT (1.0), linked to procedural programming and deterministic logic; ‘data-driven’ CT (2.0), influenced by advancements in artificial intelligence (AI) and machine learning (ML) and incorporating inductive logic and probabilistic reasoning; and ‘quantum’ CT (3.0), the emerging stage, associated with quantum computing (QC) and embracing a holistic philosophical paradigm. Withing this boader perspective, the related psychological construct, namely computational talent, has been defined as the ability to progress from ‘visual block-based’ procedural programming environments to ‘text-based’ ones. Building on this framework, the present theoretical essay pursues two main objectives: (i) to delve deeper into the 3-stage CT framework, using a tripartite distinction between computational concepts, practices, and perspectives as its integrative analytical criterion; and (ii) to update the conception of computational talent by extending its definition from the ‘rule-based’ to the ‘data-driven’ and ‘quantum’ stages, distinguishing talent from mere proficiency at each stage through a triple criterion—quantitative, qualitative, and social. As a conceptual contribution, this essay aims to provoke reflection and discussion within the field, leaving its empirical validation for future research. Full article

Accessibility evaluation in built environments increasingly requires performance-based methods that capture dynamic user–environment interaction rather than static compliance with dimensional standards. Existing computational approaches, including BIM-based rule checking and path-finding analysis, are effective for geometric verification but cannot quantify the physical demands imposed on users during task execution. To date, no existing framework combines task-driven motion generation with quantitative biomechanical analysis specifically for accessibility evaluation in built environments. This paper presents a modular simulation framework that integrates procedural kinematic planning with inverse dynamics analysis to enable performance-based accessibility assessment. The system generates deterministic motion trajectories using a gait-based abstraction layer and inverse kinematics, then computes net joint forces and torques via a stabilized Newton–Euler recursive algorithm. External contact forces, including ground reactions during locomotion, are modeled through a dedicated force management subsystem. The framework is implemented in Unity and evaluated in three representative scenarios: level walking, stair ascent, and stair descent. Computed knee joint forces fall within magnitude and temporal ranges reported in the biomechanical literature (peak forces of 2.5–4.0 body weights for stairs and 1.5–2.0 body weights for level walking), indicating biomechanical plausibility rather than subject-specific experimental validation. The framework offers a reproducible, extensible foundation for evidence-based accessibility design, allowing designers to detect and quantify biomechanically demanding interactions in architectural environments prior to construction, and supporting a shift from prescriptive compliance toward dynamic, human-centered evaluation. Full article

The wide deployment of phasor measurement units has enabled data-driven transient stability prediction (TSP) of power systems. However, ensuring the reliability of TSP results is still a significant challenge that limits the practical application of data-driven methods. To this end, a convolutional neural network (CNN)-based deep ensemble model with optimized class-specific thresholds is proposed to achieve reliable TSP. Specifically, a CNN is utilized as the backbone predictor, where the time-series variables from multiple generators are transformed into image-like inputs, and a CNN-based deep ensemble model is developed to provide accurate confidence estimation for TSP. Subsequently, considering the asymmetric importance of different classes in TSP, a confidence-based class-specific thresholds rule is adopted, and a multi-objective optimization model for determining the class-specific thresholds is formulated. In this optimization model, the reliability requirement of TSP is imposed as a constraint, requiring that true unstable rate (TUR) equal to 100%, with the objectives of minimizing the rejection rate and maximizing the true stable rate (TSR). The Pareto front of the class-specific thresholds can be obtained by solving the optimization model. Test results on two benchmark power systems show that the proposed method achieves a TUR of 100% and a TSR of at least 99% with approximately 10% of the samples rejected, demonstrating its effectiveness and scalability. Full article