Search Results (85,526)

This paper presents an integrated mathematical model to improve the safety and operational efficiency of train traffic in centralised railway dispatching systems. The proposed approach combines the alternative graph model with a Mealy automaton to synchronously address route planning, delay minimisation, and strict compliance with safety requirements. Formal control theory based on finite-state automata is employed to describe routing logic and signal control through state transitions, while the alternative graph model represents scheduling constraints and resource conflicts. To enhance real-time adaptability, a tabu search algorithm is implemented for train schedule optimisation, enabling dynamic rescheduling under changing operational conditions. The mathematical formulation incorporates blocking time parameters, a system of discrete constraints, and automaton-based safety conditions governing train movements and route authorisation. The integrated model explicitly formalises the processes of block section occupation and release, ensuring consistency between control logic and scheduling decisions. Practical testing and computational experiments demonstrate that the proposed approach effectively reduces train delays, improves the reliability of dispatch control, and increases system resilience to dynamic disturbances. The results confirm that the developed model can be implemented within existing centralised dispatching infrastructures without requiring a complete system overhaul. Overall, the proposed framework expands the functional capabilities of centralised dispatch systems by enabling efficient schedule generation, minimising the propagation of delays, and ensuring reliable command exchange between central control posts and field-level railway infrastructure. Full article

Multi-view 3D reconstruction is essential for smart city, supporting applications such as smart city planning and autonomous navigation. While traditional reconstruction pipelines and recent neural implicit methods, such as NeRF, achieve high visual fidelity, they often struggle with geometric accuracy and sparse-view scenarios. To address this challenge, we present DiffLiGS, a novel multi-modal 3D reconstruction framework that integrates LiDAR point clouds and LiDAR-guided diffusion-based priors into the 3D Gaussian Splatting (3DGS) pipeline, enabling high-fidelity and geometrically accurate models. Our method first densifies sparse LiDAR depths using a diffusion model and refines them through multi-view geometric constraints, producing dense LiDAR depth maps that provide robust supervision for 3DGS optimization. Leveraging these dense depth maps, we guide a Stable Video Diffusion model to synthesize novel view images, which are incorporated into training to enhance reconstruction completeness and visual realism. By jointly fusing rich appearance cues from multi-view images with precise LiDAR-derived geometry and diffusion priors, DiffLiGS achieves unified, geometry-aware 3D scene representations. Our extensive experiments demonstrate that our approach significantly improves both geometric accuracy and rendering quality compared to existing 3D reconstruction methods, enabling real-time, high-precision modeling of complex urban environments. Full article

Reliable detection of previously unseen classes under distributional shift remains a central challenge in concept learning and explainable artificial intelligence. In particular, high-performance deep learning models often lack statistically grounded mechanisms to signal when an instance deviates from learned concepts. This paper addresses this limitation by investigating whether conformal prediction can be effectively combined with a YOLOv5 deep learning classifier to enable principled counterfactual detection without prior exposure to the counterfactual class. As a controlled testbed, we employ Kandinsky patterns, a structured benchmark widely used in explainable AI research due to its rule-based generative transparency and suitability for concept learning studies. The proposed framework first classifies valid and invalid patterns and subsequently applies inductive conformal prediction to obtain calibrated prediction sets at a user-defined significance level. Counterfactual instances are, at start, identified based solely on information from known true and false patterns, without explicit training examples of the counterfactual class. Experimental results demonstrate that the conformalized detector reliably identifies a substantial proportion of previously unseen counterfactual patterns while maintaining statistical validity. In addition, the method flags unlabeled (“empty”) instances, thereby providing a principled signal for the emergence of new concepts. By conformalizing YOLOv5 outputs, the approach establishes a statistically sound mechanism for uncertainty-aware detection of divergent classes, contributing to robust and explainable concept learning in structured visual pattern recognition. Full article

Respiratory Syncytial Virus infection is a significant cause of morbidity and mortality in infants. Maternal vaccination with the bivalent vaccine Abrysvo^® in the third trimester (24–36 weeks) is an effective strategy to prevent severe respiratory illnesses in newborns. However, the introduction of this new technology faces structural obstacles that amplify inequalities. This rapid literature review sought to map and synthesize evidence on inequalities and inequities in adherence and accessibility to maternal vaccination among Black pregnant women. A rapid literature review was conducted using a mixed-methods approach (narrative synthesis and thematic analysis), following guidelines adapted from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses and the Cochrane Handbook. The research question was structured using the acronym Population/Problem, Exposure, Comparison, and Outcome, focusing on Black pregnant women, maternal vaccination, comparison with other groups, and barriers/determinants. The search was conducted in databases such as PubMed (via Medical Literature Analysis and Retrieval System Online), Scopus and Literatura Latino-Americana e do Caribe em Ciências da Saúde, covering studies published between 2022 and 2025 that presented disaggregated analysis by race. The analysis and interpretation of the findings were guided by Critical Race Theory. The analysis of the twelve included studies (mainly from the United States, the United Kingdom, and Brazil) revealed systematic and robust disparities. Black pregnant women had lower vaccination coverage and were less likely to receive timely recommendations compared to White pregnant women. The barriers identified include: institutional distrust (resulting from structural racism), poor access to prenatal care, inadequate communication, and socioeconomic factors. Inequities are structural and multifactorial phenomena. To ensure that the benefits of the vaccine are distributed equitably, strategies such as anti-racist training for healthcare teams, active vaccination outreach, and continuous monitoring of data disaggregated by race are essential. Full article

In practical customer-support dialogue systems, responses must simultaneously deliver factually grounded information and context-appropriate empathy, yet existing single-stage generation models often exhibit specialization bias, favoring one objective at the expense of the other. To address this limitation, we propose a dual-stage generation framework that explicitly decouples factual grounding from empathetic modulation. Our primary configuration follows a fact-to-empathy order, in which the system first generates a fact-centric draft via structured query interpretation and optional retrieval-augmented generation, then applies empathy-aware tuning conditioned on inferred emotion type, intensity, and empathy necessity. To enable deployment in resource-constrained environments, only the query interpretation module is explicitly trained using knowledge distillation, allowing the overall system to operate with compact 4B–8B backbone language models. Furthermore, we construct a customer-support dialogue dataset designed to reflect realistic interactions involving both informational and emotional demands. Extensive experiments with compact models show that the proposed approach generally improves key dimensions of empathetic response quality while maintaining overall factual performance, thereby helping mitigate the representational entanglement empirically observed in single-stage baselines. Both quantitative metrics and scenario-based analyses confirm that decoupled generation enables a more balanced integration of factuality and empathy than single-stage generation. These results suggest that dual-stage generation provides a practical and extensible foundation for deployable, real-world customer-support dialogue systems. Full article

Decision Transformers (DTs) reformulate reinforcement learning as a conditional sequence modeling problem and have demonstrated competitive performance in offline Reinforcement Learning (RL) scenarios. However, their behavior in card games, specifically partially observable imperfect-information, trick-taking games remains underexplored. In parallel, general-purpose card-game toolkits have shown the value of unified environments and standardized evaluation protocols for accelerating research in imperfect-information games. Motivated by the goal of creating a general card-game-playing framework, we present a unified RL pipeline for trick-taking card games using DTs. While classical learning methods have demonstrated strong performance in card games, transformer-based reinforcement learning remains comparatively underexplored in this domain. This paper studies the applicability of DTs to the core play-phase of trick-taking games and evaluates whether a single, reusable pipeline can be transferred across multiple games in this class with minimal game-specific engineering. We propose a unified framework integrating offline pretraining, online selective expert iteration, and inference-time legal-action filtering. Crucially, our proposed approach demonstrates two key advantages over standard implementations. First, the model successfully internalizes complex game rules (e.g., follow-suit constraints) implicitly from the empirical data distribution, completely eliminating the need for explicit action masking during training. Second, we introduce a selective expert iteration mechanism equipped with strict acceptance filtering, which effectively prevents distribution collapse and enables safe, monotonic offline-to-online policy refinement. Ultimately, we show that this single, reusable transformer-based pipeline achieves competitive performance across multiple trick-taking domains (Hearts, Whist, and Spades) with minimal game-specific engineering. Full article

Temperature plays a dominant environmental role in determining the efficiency of protein function. Accurately predicting protein thermal stability is crucial for fundamental biology, drug discovery, and protein engineering. Here, we introduce ThermoFormer, a transformer-based protein language model that learns both temperature-aware representation and sequence patterns. Specifically, we first built a large-scale dataset comprising more than 96 million protein sequences annotated with their optimal growth temperature (OGT). ThermoFormer is pre-trained with a supervised OGT prediction task and an unsupervised masked language modeling (MLM) task on the dataset. We evaluated ThermoFormer’s pre-training performance and its transferability to other temperature-prediction datasets, including two melting temperature (TM) datasets, an optimal catalytic temperature (OCT) dataset, and a thermophilic protein classification task. The results show that ThermoFormer achieves state-of-the-art performance across all evaluated tasks, outperforming prior unsupervised pre-trained models. In addition, we have also shown that ThermoFormer enables zero-shot temperature prediction, i.e., even without further fine-tuning, ThermoFormer can still achieve comparable performance. Our model can serve as a foundation for encoding protein sequences with temperature-aware representations, improving transferability to temperature-related downstream tasks. Full article

Metabolic dysfunction-associated steatotic liver disease (MASLD, introduced in 2023), formerly termed non-alcoholic fatty liver disease (NAFLD), is a leading cause of chronic liver disease worldwide and is closely linked to obesity, insulin resistance, and cardiometabolic dysfunction. Exercise is widely recommended as a cornerstone of MASLD management; however, the magnitude of its hepatic and metabolic benefits and the underlying molecular mechanisms remain incompletely defined. We aim to synthesize evidence from randomized controlled trials assessing how structured exercise interventions influence hepatic steatosis and metabolic dysfunction in adults with MASLD. A targeted search of PubMed from database inception to February 2025 identified eligible trials, of which eleven were included in the qualitative synthesis. Across studies, aerobic and resistance exercise interventions were consistently associated with reductions in hepatic fat content, improvements in plasma lipid profiles and liver enzyme concentrations, and enhanced indices of insulin sensitivity, frequently occurring independently of substantial weight loss. Mechanistically, exercise-induced activation of pathways related to mitochondrial function, lipid oxidation, inflammation modulation, and insulin signaling likely contributes to these benefits. Despite heterogeneity in intervention design, duration, and outcome assessment, the collective evidence supports structured exercise as an effective non-pharmacological strategy for improving hepatic steatosis and metabolic dysfunction in MASLD. Future studies integrating molecular biomarkers with clinical endpoints are warranted to refine exercise prescriptions and elucidate mechanisms of therapeutic response. Full article

Ginger supplementation is proposed as a natural ergogenic aid due to its anti-inflammatory and antioxidant properties, but its effects on endurance performance remain unclear. Methods: In this randomized, double-blind, placebo-controlled crossover trial, 30 trained cyclists (27 male, 3 females, aged 36 ± 11 yr) completed three visits: a baseline 75 km time trial, a post-supplementation time trial, and a second post-supplementation trial under the alternate condition. Participants consumed either ginger or placebo for four weeks. Statistics: Performance outcomes (i.e., 75 km time, VO₂, power output, heart rate, and RPE) were analyzed using repeated-measures ANOVA, with repeated-measures ANCOVA to assess dietary and age influences (p < 0.05). Results: Energy and carbohydrate intake were consistent across trials and unrelated to performance. Protein intake per kg body mass predicted performance time in the placebo trial and average VO₂ in the ginger trial; other macronutrients were not associated with outcomes. No significant differences were observed between ginger and placebo conditions for time to completion, VO₂, power output, heart rate, or perceived muscle soreness. Conclusions: Four weeks of ginger supplementation does not improve prolonged cycling performance, recovery, or muscle soreness in trained cyclists when dietary intake is controlled. Future research should explore cellular mechanisms to determine whether ginger supplementation could influence performance or recovery in endurance athletes. Full article

In this article, we explore the possibility of using reinforcement learning to create convective flow approximation schemes that maintain accuracy and stability at high Courant-Friedrichs-Lewy (CFL) numbers in the finite-volume discretization of advection equations. Unlike most existing data-driven discretization methods, which primarily concentrate on spatial grid refinement, this work emphasizes increasing the allowable time step without compromising solution accuracy. This approach reduces the total number of time integration steps, thereby enabling faster computation. A neural network is used as a surrogate model for reconstructing the convective flow, which takes as input local information about the flow, scalars, and geometry and predicts scalar values at node points. Reinforcement learning is used for training and is formulated as a policy optimization problem, where the long-term reward is defined as the difference between the numerical and reference solutions over the entire simulation period. Both the genetic algorithm and the Deep Deterministic Policy Gradient (DDPG) method are investigated. The effectiveness of the approach is evaluated using a one-dimensional nonlinear advection problem with a constant velocity field. Despite the simplicity of the test case, the results demonstrate that the trained convective flux approximation scheme achieves accuracy comparable to or better than the classical second-order linear upwind (LUD) scheme, while operating at CFL numbers 2–50 times higher than the optimal CFL for LUD, thereby reducing the simulation time by the same factor. This allows for a wider range of stability and accuracy in the finite-volume method and the use of larger time steps without compromising the quality of the solution. The study is intentionally limited to a single spatial dimension and serves as a basic analysis of the method’s applicability. The results demonstrate that reinforcement learning can successfully find more convective flow approximation schemes that improve efficiency at high CFL numbers than conventional explicit second-order schemes, establishing a framework that is subsequently extended in our follow-up work to improve training methods and three-dimensional complex transport problems. The proposed method improves the spatial discretization of convective fluxes, which is independent of the choice of time integration scheme. Therefore, the neural reconstruction can in principle be used in both explicit and implicit finite-volume solvers. Full article

We investigate when LLM-based data augmentation can mitigate temporal collapse in web intrusion detection under extreme cross-temporal distribution shift. Under a strict hold-out protocol—training on CSIC-2010 and evaluating exclusively on the temporally separated SRBH-2020 golden set—the legacy-trained baseline exhibits near-collapse in balanced correlation (MCC ≈ 0) despite retaining high recall, revealing severe false-positive bias under drift. Rather than assuming uniform benefits from synthetic data, we analyze how augmentation effects vary with model scale. Using a fixed DeBERTa-v3-base backbone and five random seeds, we compare synthetic training corpora generated by multiple LLMs under identical schema-guided structured decoding and filtering constraints. The results reveal a scale-dependent threshold effect. Models below 12B parameters (i.e., the 4–8B settings in our experiments) frequently introduce structural artifacts that amplify false-positive bias and further destabilize cross-temporal performance. In contrast, models at or above 12B parameters consistently produce modest but statistically reliable recovery from correlation collapse (p < 0.001), with balanced metrics shifting toward the target-domain distribution. Although the absolute performance remains limited under extreme temporal separation, a confusion-matrix analysis shows that large-scale generation reduces false-positive skew and moves decision-boundary behavior closer to the modern-domain regime. These findings indicate that LLM-based augmentation is not inherently robustness-enhancing; rather, its effect depends critically on model scale and disciplined generation control. When properly constrained, ≥12B-scale models can partially stabilize cross-temporal behavior, whereas smaller-scale generation may exacerbate distributional fragility. Full article

Electroencephalogram (EEG)-based seizure classification remains challenging due to inter-subject variability and heterogeneous signal characteristics. Foundation models offer a promising alternative to dataset-specific training by leveraging pretrained priors. In this study, we evaluate a tabular foundation model, the Tabular Prior-Data Fitted Network (TabPFN), against classical ensemble baselines (gradient boosting, random forests, AdaBoost, and XGBoost) for EEG seizure segment classification. We use subject-independent GroupKFold cross-validation without out-of-fold evaluation to assess generalization to unseen individuals. Experiments on the Bangalore EEG Epilepsy Dataset (BEED) and the University of Bonn (Bonn) dataset show that TabPFN achieves higher accuracy than classical ensembles, reaching 99.7% on BEED and 99.6% on Bonn. These results suggest that pretrained tabular priors can be effective in feature-based EEG pipelines where subject-level generalization is required. Full article

Smart Design methodologies represent a powerful approach for tackling optimization problems and exploring design spaces that would be unmanageable with traditional methods. By integrating computational approaches, optimization strategies and machine learning, it enables the systematic investigation of multiple configurations and the identification of optimal solutions with reduced computational effort. In the present work, Smart Design algorithms are implemented to investigate the influence of geometric parameters on a lattice–honeycomb–square structure. Results coming from finite element analysis and Life Cycle Assessment are exploited to train Random Forest and XGBoost machine learning models in order to find the lattice parameter set that ensures the optimal balance between mechanical performance and sustainability requirements. Full article

Grapes are high-value crops, but expanding cultivation has made manual harvesting inefficient and costly due to labor shortages and weather constraints. Automated harvesting requires accurate and lightweight image segmentation to ensure reliable visual perception. Improving segmentation precision, robustness, and model compactness is thus critical for intelligent grape harvesting. To enhance segmentation robustness in complex orchard environments, this study introduces a multimodal fusion and multi-scale enhancement strategy and develops a lightweight instance segmentation network. Using a multimodal grape dataset containing RGB, near-infrared (NIR), and depth information, a multi-resolution training scheme based on an image-pyramid framework was constructed. Among the three YOLOv11-based fusion strategies, early fusion achieved the best performance. Accordingly, the lightweight model YOLOv11-WFD was designed by integrating FasterNeXt, DySample, and WaveletPool to strengthen feature extraction, adaptive sampling, and small-object perception. The model delivers high segmentation accuracy and strong deployment suitability for intelligent harvesting applications. Experimental results show that YOLOv11-WFD achieves a mAP@50:95 of 79.3% on the validation set with only 2.25 M parameters, demonstrating outstanding performance in both precision and compactness. Compared with YOLOv3-tiny, YOLOv5n, YOLOv8n, YOLOv10n, YOLOv11n, and YOLOv12n, YOLOv11-WFD improves mAP@50:95 by 25.4, 3.0, 2.7, 2.8, 2.0, and 3.1 percentage points, respectively, while reducing parameters by 80.4%, 7.8%, 23.5%, 10.7%, 20.8%, and 18.8%. Overall, YOLOv11-WFD achieves an excellent balance among accuracy, speed, and complexity, verifying the effectiveness of the multimodal fusion and lightweight integration strategy. It shows strong potential for practical applications and large-scale deployment in complex agricultural environments such as intelligent grape harvesting. Full article

Background/Objectives: Simulation-based education often lacks personalization, focusing on technical competence rather than individual student profiles. This protocol describes a study designed to evaluate whether adapting gamified narratives to nursing students’ personality profiles has the potential to support academic performance in obstetrics. This study aims to validate the integration of psychometric profiling and AI as a sustainable strategy for personalized clinical training. Methods: A cluster-randomized controlled longitudinal pilot trial will be conducted at the University of Atlántico Medio. The protocol has been submitted for registration at ClinicalTrials.gov (Registration Pending). Thirty-eight second-year nursing students meeting inclusion criteria (excluding repeaters or those with prior specialized training) will be assigned by natural practice to either a control group (generic gamification) or an experimental group (gamification adapted according to Player Personality and Dynamics Scale profiles using AI-generated content). The intervention comprises four clinical simulation sessions focusing on pregnancy and childbirth, which are managed via the Wix platform. The primary outcome is academic performance, measured as “Learning Gain” (post-test scores minus pre-test scores). Secondary outcomes include student satisfaction measured via the Gameful Experience Scale. Data will be analyzed using Mann–Whitney U tests to compare overall efficacy and intragroup evolution. To minimize observer bias, knowledge assessments will utilize automated, objective scoring, and participants will be blinded to the study hypothesis. Expected Outcomes: The study aims to establish the technical and pedagogical feasibility of integrating AI-adapted narratives into nursing curricula. It is anticipated that the personalized approach will show positive trends in learning gains and engagement patterns, providing a baseline for larger multicenter trials. Conclusions: This protocol presents a framework for “Precision Education” in nursing, shifting from “one-size-fits-all” simulations to student-centered adaptive training. The use of Generative AI makes such personalization sustainable and cost-effective for health science faculties. Full article