Abstract: Next Point-of-Interest (POI) recommendation aims to predict a user’s next location based on historical check-in data. However, real-world check-in records often contain uncertain check-ins, in which ambiguous spatial, temporal, or behavioral information obscures true mobility patterns and degrades prediction accuracy. To mitigate this issue, this study first learns user preferences from historical trajectories and adjusts transition importance based on temporal and spatial proximity, before modeling transition relationships using three complementary features: category, spatial area, and routine/non-routine behavior patterns. Based on transition probability analysis, feature-level dependencies in user mobility are systematically examined. The results indicate that these transition features contribute unequally to prediction performance, with area-based transitions being the most effective when considered individually. Nevertheless, their integration consistently yields the highest accuracy, highlighting the importance of transition-aware modeling. Experiments on two real-world datasets demonstrate that the proposed framework outperforms state-of-the-art methods in terms of Recall and NDCG, confirming the effectiveness of the proposed approach.

Abstract: In this paper, we present a deep neural network–based approach for computing radar cross section (RCS) over a wide frequency band and a broad range of incident angles.The proposed network, termed WBRCS-Net, is designed to converge to the solution of the method of moments (MoM) formulation by minimizing a mean-squared residual loss without explicitly solving the MoM linear system, thereby avoiding the numerical instabilities commonly encountered in conventional iterative solvers. Moreover, by using only the frequency and incident angle as inputs, WBRCS-Net enables wideband RCS prediction over a broad range of incident angles while substantially simplifying the network architecture. The performance of WBRCS-Net is evaluated on perfectly electrically conducting (PEC) spheres and cubes and compared with the Maehly approximation based on Chebyshev polynomials. Experimental results show that, once trained, WBRCS-Net provides accurate and stable wideband RCS computations over a wide range of incident angles with instantaneous inference speed, highlighting a key advantage of the neural network–based approach.

Abstract: Coronary artery anomalies (CAAs) represent a rare but clinically significant group of congenital abnormalities, often implicated in sudden cardiac death among young individuals. Despite increasing recognition, standardized diagnostic and management pathways remain underdeveloped, particularly regarding surgical intervention. This narrative review aims to synthesize current evidence on the classification, pathophysiology, and contemporary treatment options for CAAs, with a specific focus on the surgical indications and guidelines. A systematic literature search was conducted through February 2025 across PubMed, Embase, Google Scholar, and UpToDate, using relevant MeSH terms and keywords related to coronary anomalies, sudden cardiac death, imaging modalities, and treatment strategies. Priority was given to high-level evidence including guidelines, systematic reviews, and large observational studies, in English and French. CAAs encompass a wide anatomical and clinical spectrum. Among these, anomalies of origin and course, such as anomalous aortic origin of a coronary artery (AAOCA), are now better defined and integrated into international guidelines, with surgical repair (e.g., unroofing, reimplantation, bypass) increasingly recommended in symptomatic patients or those with high-risk anatomical features. In contrast, anomalies like myocardial bridges, coronary artery fistulas, or ectasia remain controversial in both diagnosis and management, with inconsistent thresholds for medical versus surgical treatment. The variability in care stems from the lack of a unified classification system, limited prospective data, and underutilization or misinterpretation of imaging modalities such as coronary CT angiography and intravascular imaging. The management of coronary artery anomalies is evolving, particularly in the domain of surgical indications. Clear consensus exists only for select anomalies, leaving others subject to individualized, often non-standardized decisions. There is an urgent need for a harmonized diagnostic framework and outcome-based criteria to guide surgical and non-surgical interventions. A multidisciplinary, evidence-informed approach is essential to optimize outcomes and reduce the risk of sudden cardiac events.

Abstract: A Ti6Al4V alloy fabrication via laser powder bed fusion (L-PBF) leads to the formation of coarse columnar β grains that give rise to anisotropic mechanical properties and inadequate strength. Incorporating the rare earth oxide, yttrium oxide (Y₂O₃), has proven an effective strategy in enhancing the mechanical performance of Ti6Al4V al-loys. This study systematically investigates the effects of various Y₂O₃ contents on the microstructure and mechanical properties of Ti6Al4V alloys fabricated via L-PBF. The results demonstrate that a Y₂O₃ addition of 0.2 wt.% produces β grains and α phases with average sizes of 61.6 and 7.6 μm, respectively. Transmission electron microscopy observations reveal that Y₂O₃ nanoparticles, together with elemental Y nanoparticles formed by reduction, are distributed both within the α-Ti matrix and along phase boundaries. This distribution effectively reinforces grain boundaries and promotes heterogeneous nucleation, thereby refining the microstructure. Mechanical property tests indicate that the alloy strength significantly improves as the Y₂O₃ content in-creases. Specifically, the alloy with 0.2 wt.%Y₂O₃ exhibits a tensile strength of 1106 MPa, a yield strength of 1074 MPa, and an elongation of 10.0%. This study proposes an in-novative rare earth strengthening method for refining the microstructure of L-PBF-fabricated titanium alloys and comprehensively enhancing their mechanical properties.

Abstract: The scarcity of high-quality, labeled audio data for legal proceedings remains a significant barrier to developing robust speech-to-text and speaker diarization systems for the judiciary. This paper in- troduces Deepcounsel, a high-fidelity synthetic speech dataset simulating courtroom environments. Utilizing a multi-agent system powered by the Gemini 2.5 Pro model, we orchestrated complex interactions between eleven distinct roles, including judges, attor- neys, witnesses, and court staff. By leveraging native multimodal generation, Deepcounsel provides a diverse range of legal termi- nology, emotional prosody, and multi-speaker overlaps. Our results demonstrate that synthetic datasets generated via multi-agent Large Language Models (LLMs) can serve as a viable proxy for training specialized legal AI models where real-world data is restricted by privacy laws.

Abstract: Cosmic radiation represents a critical barrier to long-term human presence beyond Earth’s magnetosphere, particularly in lunar and Martian environments [1]. Traditional shielding materials—such as regolith, water, and metallic alloys— face significant logistical, economic, and structural limitations [2]. This study investigates the potential of fungal melanin, a biological pigment known for its radiation-shielding properties in extreme environments (e.g., Chernobyl and spaceflight), as a lightweight and sustainable alternative for space architecture [3,4,5]. We propose an architectural framework for integrating fungal melanin into bio-inspired coatings, analyzing species-specific variations and production feasibility [6]. Comparative assessments indicate that melanin offers superior mass efficiency and architectural flexibility over conventional materials [7]. The research concludes with a roadmap for hybrid material integration and experimental validation, establishing a biologically-driven paradigm for resilient extraterrestrial habitats [8,9].

Abstract: Chronic obstructive pulmonary disease (COPD) and bronchial asthma (BA) are very common pathologies, both of them being characterized by a chronic bronchopulmonary inflammation. This paper aims to present the mechanisms of the two pathologies in comparison, starting from the classical approach, entering the cellular level (effector cells), then the molecular one (lipid mediators, cytokines, chemokines, reactive oxygen species, proteases, ATP, cellular senescence markers), and finally addressing the mechanisms at the quantum level. It will be explained that electron transfer through the interfacial water is essential for all cellular energy metabolism associated events. It will also be presented how biochemical reactions do not occur instantaneously and randomly, but depend on exceeding a threshold of free energy of activation and satisfying steric requirements. Another important topic addressed will be the electron-accepting property of ionic reactive oxygen species (ROS) and how cellular metabolism is regulated by the formation and decomposition of collective ROS states (this is a quantum regulated phenomenon involving a large number of entangled ROS molecules simultaneously). Finally, it will be presented how these mechanisms are altered in COPD and BA as well as the consequences of pulmonary fibrosis at the quantum level. We believe it is important for physicians to understand how the principles of quantum physics applied to experimental biology deepen the understanding of the normality and disease origin at the level of subatomic particles, molecules, and their associated electromagnetic fields.

Abstract: Eating disorders (EDs) are complex conditions that can significantly affect health and productivity, yet their assessment in occupational settings remains underexplored. This study aimed to evaluate the psychometric properties of the Italian version of the Eating Disorder Examination Questionnaire–Short Form (EDE-QS) among 1,912 workers undergoing health surveillance. Using an Item Response Theory framework, we tested dimensionality, reliability, and measurement invariance across gender, applying a graded response model to assess item discrimination and threshold parameters. Results supported a unidimensional structure with excellent internal consistency (ω ≈ .95) and strong indices of factor score determinacy and construct replicability. Measurement invariance analyses indicated configural and metric invariance but not full scalar invariance, due to differential item functioning in a subset of items. Latent mean differences were small, with women scoring slightly higher than men, and associations with psychological, occupational, and health-related variables did not differ by gender. These findings indicate that the Italian EDE-QS is a reliable and valid instrument for rapid screening of ED symptoms in workplace contexts. However, gender-related item bias warrants cautious interpretation of specific behaviors, suggesting the need for tailored assessments to enhance diagnostic accuracy and inform preventive interventions.

Abstract: This study examines how AI-assisted artistic practices reshape authorship, cultural ownership, and museum governance through the lens of cultural sustainability. Drawing on qualitative methods including literature analysis, expert interviews, and exhibition case studies, it explores emerging ethical challenges related to data provenance, creative agency, and institutional responsibility. The findings reveal hybrid forms of authorship that disrupt conventional intellectual property frameworks and highlight museums’ growing role as mediators between technological innovation and cultural preservation. While AI-driven exhibitions expand accessibility and engagement, they also risk cultural homogenization. The study offers strategic insights for policymakers and cultural institutions on fostering ethical, inclusive, and sustainable AI integration in artistic practice.

Abstract: Calcium imaging with miniscopes allows researchers to record the activity of many neurons over long periods in freely moving animals. While data collection has become easier, analysis have not. Typical calcium imaging analysis requires many processing steps, uses multiple software tools, and depends heavily on parameter choices. In practice, these details are often poorly recorded, rely on proprietary software, or are lost when large intermediate files are deleted to save disk space. As a result, analyses are hard to reproduce, compare, or rerun reliably. This paper describes an open, Python-based framework designed to make calcium imaging analysis clearer, more reproducible, and easier to manage. The framework treats each analysis step as an explicit, recorded operation with defined inputs, outputs, parameters, and software providers. All steps are logged in lightweight trace files saved to disk, allowing analyses to be resumed, audited, or exactly reproduced later, even if large intermediate data have been removed. Algorithm-specific code is isolated behind standardized wrappers, so users can switch between proprietary and open-source tools without changing the overall workflow.The framework also supports branching to compare different methods, batch processing across multiple animals or sessions, controlled cleanup to reduce disk usage, and a modular design. The result is a practical system that makes calcium imaging analyses easier to follow, repeat, and reuse.

Abstract: Polycystic ovary syndrome (PCOS) is a heterogeneous endocrine–metabolic disorder as-sociated with insulin resistance (IR), visceral adiposity, and increased cardiometabolic risk. The visceral adiposity index (VAI) is a validated surrogate marker of adipose tissue dysfunction, but its relationship with circulating neurotrophins and adipokine balance in PCOS remains incompletely understood. In this study, 100 women with PCOS were strati-fied into lower- (n = 50) and higher-risk (n = 50) groups according to VAI. Anthropometric measures, fasting glucose and insulin concentrations, lipid profile, and serum levels of brain-derived neurotrophic factor (BDNF), nerve growth factor-β (NGFβ), leptin, adi-ponectin, and resistin were assessed. HOMA-IR, adipokine ratios and atherogenic indices were calculated. Multivariate regression revealed that BDNF was independently associ-ated with VAI and non-HDL-cholesterol, whereas NGFβ was independently associated with HDL-cholesterol and estradiol, indicating neurotrophin associations with metabolic and endocrine parameters independent of general adiposity. Correlation heatmap and network analyses demonstrated interconnected clusters linking visceral adiposity, IR, dyslipidemia, adipokine imbalance, and neurotrophins, with the leptin/adiponectin ratio emerging as a central integrative marker. These findings indicate that VAI-defined car-diometabolic risk in PCOS is accompanied by distinct and opposing neurotrophin–adipokine signatures, highlighting neurotrophin–adipokine networks underlying visceral adiposity-driven cardiometabolic and endocrine risk.

Abstract: Background: Sepsis remains a leading cause of mortality in Intensive Care Units (ICUs) worldwide. Machine learning models for clinical prediction must be accurate, fair, transparent, and reliable to ensure that physicians feel confident in their decision-making process. Methods: We used the MIMIC-IV, version 3.1, database to evaluate several machine learning architectures, including Logistic Regression, XGBoost, LightGBM, LSTM (Long Short-Term Memory) networks and Transformer models. We predicted three main clinical targets: hospital mortality, length of stay, and septic shock onset. Model interpretability was assessed using Shapley Additive Explanations (SHAP). Results: The XGBoost model demonstrated superior performance in prediction tasks, particularly for hospital mortality (AUROC 0.874), outperforming traditional LSTM networks, transformers and linear baselines. Importance analysis of the variables confirmed the clinical relevance of the model. Conclusions: While XGBoost and ensemble algorithms demonstrate superior predictive power for sepsis prognosis, their clinical adoption necessitates robust explainability mechanisms to gain the doctors trust.

Abstract: Introduction: Delirium in older adults is strongly associated with and can predate dementia. While GLP-1 receptor agonists (GLP-1 RAs) may provide neuroprotective benefits, their role in reducing dementia risk among older patients with type 2 diabetes mellitus (T2DM) and prior delirium is unclear. This study compares the effectiveness of GLP-1 RAs and metformin in preventing dementia among patients with a history of delirium using an extensive healthcare database. Methods: A retrospective cohort study was conducted from 2005 to 2025 using data from the TriNetX Global Federated Research Network. We identified adults aged 65 and older with T2DM and delirium, who received either GLP-1 RAs (exposure) or metformin (control). Propensity score matching (PSM) was performed. We determined dementia outcomes and all-cause mortality using Kaplan–Meier survival curves, Cox proportional hazards models, and estimated odds ratio (OR). Subgroup analyses were conducted by age, sex, race, body mass index (BMI), and dementia types. Results: In a study involving 23,980 patients treated with GLP-1RAs and 23,980 matched patients treated with metformin, the mean age was 74 years, with 53% being female. The use of GLP-1RAs therapy was associated with a significantly lower risk of dementia compared to metformin among adults aged 65 years and older with T2DM with delirium with the adjusted hazard ratio (AHR) for dementia was 0.778 [95% confidence interval (CI): 0.736-0.822 p < 0.0001], and OR was 0.617 (95% CI: 0.582-0.655 p < 0.0001). The reduction in dementia risk varied by age, race, BMI , and dementia type. We observed a time-dependent decrease in mortality risk with GLP-1 RA users. Conclusion: In older adults with T2DM and delirium, GLP-1 RA therapy was associated with a reduced risk of dementia compared with metformin. Variations in subgroups suggest individualized treatment. Prospective randomized controlled trials are needed to confirm these findings and clarify differential effects across various subgroups.

Abstract: This work presents Recurrence with Correlation Network (RWCNet), a novel multi-scale recurrent neural network architecture for medical image registration that integrates core principles from optical flow, including correlation volume computation and inference-time instance optimization. In evaluations on the large-displacement National Lung Screening Test (NLST) dataset, which features large displacements, RWCNet exhibited superior performance (total registration error (TRE) of 2.11mm) to deep learning alternatives, and on par results with variational optimization techniques. In contrast, on the OASIS dataset characterized by smaller displacements, RWCNet's results (average dice similarity of 81.7\%) were superior to variational optimization techniques and showed a small improvement over other multi-scale deep learning models. Ablation experiments showed that multi-scale features consistently improved performance, where as the correlation volume, number of recurrent steps, and inference-time instance optimization only had large positive impacts on performance in the NLST dataset. The performance of RWCNet compared to approaches that use instance optimzation show that deep learning based methods can find local minima that escape instance optimization methods. The results highlight the need for algorithm hyperparameter selection that adjusts with the dataset characteristics. RWCNet's promising results may imporve registration performance and the speed of computation, allowing many potential applications including, treatment planning, intra-procedural guidance, and longitudinal monitoring.

Abstract: Reliable visual characterization of food composition is a fundamental prerequisite for image-based dietary assessment and health-oriented food analysis. In fine-grained food recognition, models often suffer from large intra-class variation and small inter-class differences, where visually similar dishes exhibit subtle yet discriminative differences in ingredient compositions, spatial distribution, and structural organization, which are closely associated with different nutritional characteristics and health relevance. Capturing such composition-related visual structures in a non-invasive manner remains challenging. In this work, we propose a fine-grained food classification framework that enhances spatial relation modeling and key-region awareness to improve discriminative feature representation. The proposed approach strengthens sensitivity to composition-related visual cues while effectively suppressing background interference. A lightweight multi-branch fusion strategy is further introduced for stable integration of heterogeneous features. Moreover, to support reliable classification under large intra-class variation, a token-aware subcenter-based classification head is designed. The proposed framework is evaluated on the public FoodX-251 and UEC Food-256 datasets, achieving accuracies of 82.28% and 82.64%, respectively. Beyond benchmark performance, the framework is designed to support practical image-based dietary analysis under real-world dining conditions, where variations in appearance, viewpoint, and background are common. By enabling stable recognition of the same food category across diverse acquisition conditions and accurate discrimination among visually similar dishes with different ingredient compositions, the proposed approach provides reliable food characterization for dietary interpretation, thereby supporting practical dietary monitoring and health-oriented food analysis applications.

Abstract: Background/objectives: Emotional intelligence (EI) plays an important role in nursing education by supporting competencies such as communication, leadership, resilience, and clinical performance. In contemporary nursing education, students face increasing academic, clinical, and emotional demands, highlighting the need to identify modifiable factors that may be associated with EI and can inform student support strategies. Despite extensive EI research, evidence remains limited and inconsistent regarding how specific health-promoting lifestyle domains and sleep quality relate to EI among prelicensure nursing students. This study aimed to examine factors associated with EI and its relationship with health behaviors among prelicensure nursing students. Methods: A cross-sectional quantitative design was used. A convenience sample of 287 prelicensure nursing students from a local nursing school completed self-report questionnaires: the Schutte Self-report Emotional Intelligence Scale (SSEIS), the Health-Promoting Lifestyle Profile II (HPLP-II), and the Pittsburgh Sleep Quality Index (PSQI). Results: In structured multiphase regression, HPLP-II interpersonal relations (B = 4.42, 95% CI = 1.44 to 7.50, p = 0.004) and spiritual growth (B = 6.59, 95% CI = 3.81 to 9.37, p &lt; 0.001) were positively associated with EI. Poor sleep quality (PSQI &gt; 5) was negatively associated with EI (B = −1.95, 95% CI = −3.88 to −0.01, p = 0.049). Conclusion: Interpersonal relations, spiritual growth, and sleep quality were associated with EI among prelicensure nursing students. These factors may be relevant to consider when designing student support and EI-related educational initiatives; however, longitudinal and intervention studies are needed to clarify directionality and causality.

Abstract: This article presents the design and characterization process of a lightweight Vivaldi antenna for high-voltage ultra-wideband systems. The proposed antenna consists of two radiating arms with different exponential curves on their inner and outer edges fed with an insulated-coplanar-plates transmission line. The weight reduction is achieved by implementing the antenna with sheets composed of a polyester layer between two aluminum layers, with a polylactic acid insulator inserted between the arms. The reflection coefficient of the implemented antenna demonstrates an impedance bandwidth ranging from 0.3 GHz to 4.2 GHz. High voltage operation of up to 12.4 kV is also experimentally demonstrated. The transfer function between the voltage applied to the antenna, Vs, and the radiated electric field, Er, is measured. Using this transfer function, the radiated electric field is calculated for an input voltage pulse with a rise time of 110 ps to confirm the antenna capability of producing radiated pulses with low distortion. The calculated radiated electric field pulse closely matches the results obtained with full wave simulation. To assess the similarity between the radiated and applied pulses, the pulse width stretch ratio is calculated, yielding a variation of 3.86% for the direction of maximum gain and 9.36% for 30° in the H-plane of the antenna. This feature is desirable for EMC, EMI and sensing applications. The antenna is also characterized in the frequency domain, achieving a maximum gain of 13.9 dBi at 3 GHz and a 30° 3dB beamwidth for ultra-wideband pulses.

Abstract: This paper presents a deep learning-based adaptive sensor fusion framework for re-al-time control and fault-tolerant automation in Industrial IoT systems. The core of the framework is an attention-based CNN-Transformer model that dynamically fuses het-erogeneous sensor streams; its interpretable weighting signals are leveraged directly for fault detection and to inform a supervisory control policy. By dynamically weighting multiple heterogeneous sensor streams using an attention-based CNN-Transformer architecture, the proposed method reduces estimation error under noisy and fault-prone conditions, and seamlessly integrates with a closed-loop controller that adjusts to detected faults through a stability-aware supervisory policy. Experiments on synthetic IIoT data with injected transient faults demonstrate significant improvements in fusion accuracy (RMSE: 0.049 ± 0.003 vs 0.118 ± 0.008 for Kalman filter, p &lt; 0.001), faster fault detection (F1-score: 0.89 ± 0.02) and recovery (1.1 ± 0.2 seconds), and hard real-time performance suitable for edge deployment (99th percentile latency: 58ms). The results show that the proposed approach outperforms classical baselines in terms of RMSE, detection F1-score, recovery time, and latency trade-offs. This work contributes to more reliable, adaptive automation in industrial settings with minimal manual tuning and empirical stability validation.

Abstract: Human chromosome 2 (HSA2) originated from a telomere-to-telomere fusion event in thehuman lineage, supported by convergent cytogenetic and comparative-genomic signatures. The primaryunresolved questions are quantitative and empirical: how an (at least partially) underdominantrearrangement could establish under drift and realistic population structure, and whether fusion-proximalsequence behaved as a barrier to gene flow during later admixture with archaic hominins. Here, weintegrate (i) drift-aware Wright–Fisher simulations and a simple subdivided metapopulation model toquantify establishment probabilities under heterozygote fertility costs, including sensitivity to weaktransmission-ratio distortion (TRD; k>0.5); (ii) a tract-based assay of Neanderthal introgression at 2q13using a public IBDmix Vindija callset (hg19) benchmarked against a length-matched chromosome 2 null;and (iii) external evidence from recent T2T-CHM13 audits showing that reference completeness rescuessubstantial archaic sequence previously undetected in repeat-rich regions, constraining interpretations ofapparent 'introgression deserts' near pericentromeric sequence. Taken together, these results supportconservative, testable claims: establishment of an underdominant fusion is plausible under drift instructured populations and can be amplified by weak TRD, whereas introgression depletion at 2q13 inhg19-era callsets must be interpreted cautiously given callability vulnerabilities highlighted by T2T-basedremapping.

Abstract: A geomagnetic storm is a typical solar eruption activity. When high-speed solar wind or coronal mass ejections generated by solar eruptions impact the Earth, it can cause severe disturbances in the Earth's magnetic field within a short period of time, leading to changes in the ionosphere. For low-frequency time-code signals that rely on the ionosphere as a reflecting medium for long-distance propagation, the signal field strength and time deviation will also undergo corresponding changes, which will affect the normal propagation and reception of signal in space. This paper selects the geomagnetic storm phenomenon that occurred from November 10 to 15, 2025, and utilizes a low-frequency time-code signal monitoring system to conduct experimental research on the impact of geomagnetic storms on low-frequency time-code signals. The test group measured and analyzed the signal field strength data and time deviation data during this period, while combining parameters such as solar activity data and geomagnetic data, in an attempt to explore the causes and patterns of signal changes. The results indicate that a geomagnetic fluctuation occurred on November 12, resulting in a decrease in signal field strength over 2.3dBμV/m, time deviation data showed significant fluctuations, increasing by over 2.4ms, exhibiting a highly discrete state. The reason is that when a local geomagnetic storm occurs, the ionization level in the ionosphere increases. When low-frequency time code signals pass through the D layer, the equivalent emission height of the low ionosphere gradually decreases, resulting in signal attenuation and phase delay, which in turn leads to an increase in time deviation.