Search Results (905)

Expansive clays damage the foundations, slabs, and utilities of low- and mid-rise buildings, threatening daily operations and incurring billions of dollars in costs globally. This study pioneers a domain-informed machine learning framework, coupled with a collinearity-aware feature selection strategy, to predict soil swell potential solely from routine index properties. Following hard-limit filtering and Unified Soil Classification System (USCS) screening, 291 valid samples were extracted from a public dataset of 395 cases. A random forest benchmark model was developed using five correlated features, and a multicollinearity analysis, as indicated by the variance inflation factor, revealed exact linear dependence among the Atterberg limits. A parsimonious two-variable model, based solely on plasticity index (PI) and clay fraction (C), was retained. On an 80:20 stratified hold-out set, this simplified model reduced root mean square error (RMSE) from 9.0% to 6.8% and maximum residuals from 42% to 16%. Bootstrap analysis confirmed a median RMSE of 7.5% with stable 95% prediction intervals. Shapley Additive Explanations (SHAP) analysis revealed that PI accounted for approximately 75% of the model’s influence, highlighting the critical swell surge beyond PI ≈ 55%. This work introduces a rule-based cleaning pipeline and collinearity-aware feature selection to derive a robust, two-variable model balancing accuracy and interpretability, a lightweight, interpretable tool for foundation design, GIS zoning, and BIM workflows. Full article

Coronavirus disease 2019 (COVID-19) produced devastating health and economic impacts worldwide. While progress has been made in vaccine development, effective antiviral treatments remain limited, particularly those targeting the papain-like protease (PLpro) of SARS-CoV-2. PLpro plays a key role in viral replication and immune evasion, making it an attractive yet underexplored target for drug repurposing. In this study, we combined machine learning, molecular dynamics, and molecular docking to identify potential PLpro inhibitors in existing drugs. We performed long-timescale molecular dynamics simulations on PLpro–ligand complexes at two known binding sites, followed by structural clustering to capture representative structures. These were used for molecular docking, including a training set of 127 compounds and a library of 1107 FDA-approved drugs. A random forest model, trained on the docking scores of the representative conformations, yielded 76.4% accuracy via leave-one-out cross-validation. Applying the model to the drug library and filtering results based on prediction confidence and the applicability domain, we identified five drugs as promising candidates for repurposing for COVID-19 treatment. Our findings demonstrate the power of integrating computational modeling with machine learning to accelerate drug repurposing against emerging viral targets. Full article

Solar power generation is rapidly emerging within renewable energy due to its cost-effectiveness and ease of deployment. However, improper inspection and maintenance lead to significant damage from unnoticed solar hotspots. Even with inspections, factors like shadows, dust, and shading cause localized heat, mimicking hotspot behavior. This study emphasizes interpretability and efficiency, identifying key predictive features through feature-level and What-if Analysis. It evaluates model training and inference times to assess effectiveness in resource-limited environments, aiming to balance accuracy, generalization, and efficiency. Using Unmanned Aerial Vehicle (UAV)-acquired thermal images from five datasets, the study compares five Machine Learning (ML) models and five Deep Learning (DL) models. Explainable AI (XAI) techniques guide the analysis, with a particular focus on MPEG (Moving Picture Experts Group)-7 features for hotspot discrimination, supported by statistical validation. Medium Gaussian SVM achieved the best trade-off, with 99.3% accuracy and 18 s inference time. Feature analysis revealed blue chrominance as a strong early indicator of hotspot detection. Statistical validation across datasets confirmed the discriminative strength of MPEG-7 features. This study revisits the assumption that DL models are inherently superior, presenting an interpretable alternative for hotspot detection; highlighting the potential impact of domain mismatch. Model-level insight shows that both absolute and relative temperature variations are important in solar panel inspections. The relative decrease in “blueness” provides a crucial early indication of faults, especially in low-contrast thermal images where distinguishing normal warm areas from actual hotspot is difficult. Feature-level insight highlights how subtle changes in color composition, particularly reductions in blue components, serve as early indicators of developing anomalies. Full article

Personality is a fundamental psychological trait that exerts a long-term influence on human behavior patterns and social interactions. Automatic personality recognition (APR) has exhibited increasing importance across various domains, including Human–Robot Interaction (HRI), personalized services, and psychological assessments. In this study, we propose a multimodal personality recognition model that classifies the Big Five personality traits by extracting features from three heterogeneous sources: audio processed using Wav2Vec2, video represented as Skeleton Landmark time series, and text encoded through Bidirectional Encoder Representations from Transformers (BERT) and Doc2Vec embeddings. Each modality is handled through an independent Self-Attention block that highlights salient temporal information, and these representations are then summarized and integrated using a late fusion approach to effectively reflect both the inter-modal complementarity and cross-modal interactions. Compared to traditional recurrent neural network (RNN)-based multimodal models and unimodal classifiers, the proposed model achieves an improvement of up to 12 percent in the F1-score. It also maintains a high prediction accuracy and robustness under limited input conditions. Furthermore, a visualization based on t-distributed Stochastic Neighbor Embedding (t-SNE) demonstrates clear distributional separation across the personality classes, enhancing the interpretability of the model and providing insights into the structural characteristics of its latent representations. To support real-time deployment, a lightweight thread-based processing architecture is implemented, ensuring computational efficiency. By leveraging deep learning-based feature extraction and the Self-Attention mechanism, we present a novel personality recognition framework that balances performance with interpretability. The proposed approach establishes a strong foundation for practical applications in HRI, counseling, education, and other interactive systems that require personalized adaptation. Full article

Background: Native vertebral osteomyelitis (NVO) presents diagnostic and therapeutic challenges requiring adherence to complex clinical guidelines. The emergence of large language models (LLMs) offers new avenues for real-time clinical decision support, yet their utility in managing NVO has not been formally assessed. Methods: This study evaluated four LLMs—Consensus, Gemini, ChatGPT-4o Mini, and ChatGPT-4o—using 13 standardized questions derived from the 2015 IDSA guidelines. Each model generated 13 responses (n = 52), which were independently assessed by three orthopedic surgeons for accuracy (4-point scale) and comprehensiveness (five-point scale). Results: ChatGPT-4o produced the longest responses (428.0 ± 45.4 words), followed by ChatGPT-4o Mini (392.2 ± 97.4), Gemini (358.2 ± 60.5), and Consensus (213.2 ± 68.8). Accuracy ratings showed that ChatGPT-4o and Gemini achieved the highest proportion of “Excellent” responses (54% and 51%, respectively), while Consensus received only 20%. Comprehensiveness scores mirrored this trend, with ChatGPT-4o (3.95 ± 0.79) and Gemini (3.82 ± 0.68) significantly outperforming Consensus (2.87 ± 0.66). Domain-specific analysis revealed that ChatGPT-4o achieved a 100% “Excellent” accuracy rating in therapy-related questions. Statistical analysis confirmed significant inter-model differences (p < 0.001). Conclusions: Advanced LLMs—especially ChatGPT-4o and Gemini—demonstrated high accuracy and depth in interpreting clinical guidelines for NVO. These findings highlight their potential as effective tools in augmenting evidence-based decision-making and improving consistency in clinical care. Full article

Data security is of the utmost importance in the domain of real-time environmental monitoring systems, particularly when employing advanced context-aware intelligent visual analytics. This paper addresses a significant deficiency in the Global System for Mobile Communications (GSM), a widely employed wireless communication system for environmental monitoring. The A5/1 encryption technique, which is extensively employed, ensures the security of user data by utilizing a 64-bit session key that is divided into three linear feedback shift registers (LFSRs). Despite the shown efficacy, the development of a probabilistic model for assessing the vulnerability of breaking or intercepting the session key (Kc) has not yet been achieved. In order to bridge this existing knowledge gap, this study proposes a probabilistic model that aims to evaluate the security of encrypted data within the framework of the Global System for Mobile Communications (GSM). The proposed model implements alterations to the current GSM encryption process by the augmentation of the quantity of Linear Feedback Shift Registers (LFSRs), consequently resulting in an improved level of security. The methodology entails increasing the number of registers while preserving the session key’s length, ensuring that the key length specified by GSM standards remains unaltered. This is especially important for environmental monitoring systems that depend on real-time data analysis and decision-making. In order to elucidate the notion, this analysis considers three distinct scenarios: encryption utilizing a set of five, seven, and nine registers. The majority function is employed to determine the registers that will undergo perturbation, hence increasing the complexity of the bit arrangement and enhancing the security against prospective attackers. This paper provides actual evidence using simulations to illustrate that an increase in the number of registers leads to a decrease in the vulnerability of data interception, hence boosting data security in GSM communication. Simulation results demonstrate that our method substantially reduces the risk of data interception, thereby improving the integrity of context-aware intelligent visual analytics in real-time environmental monitoring systems. Full article

The SCARECROW (SCR) transcription factor governs cell-type patterning in plant roots and Kranz anatomy of leaves, serving as a master regulator of root and shoot morphogenesis. Foxtail millet (Setaria italica), characterized by a compact genome, self-pollination, and a short growth cycle, has emerged as a C₄ model plant. Here, we revealed two SCR paralogs in foxtail millet—SiSCR1 and SiSCR2—which exhibit high sequence conservation with ZmSCR1/1h (Zea mays), OsSCR1/2 (Oryza sativa), and AtSCR (Arabidopsis thaliana), particularly within the C-terminal GRAS domain. Both SiSCR genes exhibited nearly identical secondary structures and physicochemical profiles, with promoter analyses revealing five conserved cis-regulatory elements. Robust phylogenetic reconstruction resolved SCR orthologs into monocot- and dicot-specific clades, with SiSCR genes forming a sister branch to SvSCR from its progenitor species Setaria viridis. Spatiotemporal expression profiling demonstrated ubiquitous SiSCR gene transcription across developmental stages, with notable enrichment in germinated seeds, plants at the one-tip-two-leaf stage, leaf 1 (two days after heading), and roots during the seedling stage. Co-expression network analysis revealed that there is a correlation between SiSCR genes and other functional genes. Abscisic acid (ABA) treatment led to a significant downregulation of the expression level of SiSCR genes in Yugu1 roots, and the expression of the SiSCR genes in the roots of An04 is more sensitive to PEG6000 treatment. Drought treatment significantly upregulated SiSCR2 expression in leaves, demonstrating its pivotal role in plant adaptation to abiotic stress. Analysis of heterologous expression under the control of the 35S promoter revealed that SiSCR genes were expressed in root cortical/endodermal initial cells, endodermal cells, cortical cells, and leaf stomatal complexes. Strikingly, ectopic expression of SiSCR genes in Arabidopsis led to hypersensitivity to ABA, and ABA treatment resulted in a significant reduction in the length of the meristematic zone. These data delineate the functional divergence and evolutionary conservation of SiSCR genes, providing critical insights into their roles in root/shoot development and abiotic stress signaling in foxtail millet. Full article

Despite the rapid adoption of green buildings as a sustainable development strategy, robust, data-driven approaches for assessing and predicting project risks remain limited. This study proposes an innovative hybrid framework combining the fuzzy analytic hierarchy process (FAHP), multilayer perceptron neural networks (MLPNNs), and particle swarm optimization (PSO) to quantify and forecast the impact of critical risks on green buildings’ performance. Drawing on structured input from 30 domain experts in Shenzhen, China, ten risk categories were identified and prioritized, with economic, market, and functional risks emerging as the most influential. Using these expert-derived weights, an MLP was trained to predict the effects of the top five risks on four core performance metrics—cost, time, quality, and scope. PSO was applied to optimize the model’s architecture and hyperparameters, improving its predictive accuracy. The optimized framework achieved RMSE values ranging from 0.06 to 0.09 and R² values of up to 0.95 across all outputs, demonstrating strong predictive capability. These results substantiate the framework’s effectiveness in generating actionable, quantitative risk predictions under uncertainty. Full article

Background/Objectives: This study evaluates whether combining ⁶⁸Ga-PSMA-11-PET/CT derived imaging biomarkers with clinical risk factors improves the prediction of early biochemical recurrence (eBCR) or clinical progress in patients with high-risk prostate cancer (PCa) after primary treatment, using machine learning (ML) models. Methods: We analyzed data from 93 high-risk PCa patients who underwent ⁶⁸Ga-PSMA-11 PET/CT and received primary treatment at a single center. Two predictive models were developed: a logistic regression (LR) model and an ML derived probabilistic graphical model (PGM) based on a naïve Bayes framework. Both models were compared against each other and against the CAPRA risk score. The models’ input variables were selected based on statistical analysis and domain expertise including a literature review and expert input. A decision tree was derived from the PGM to translate its probabilistic reasoning into a transparent classifier. Results: The five key input variables were as follows: binarized CAPRA score, maximal intraprostatic PSMA uptake intensity (SUVmax), presence of bone metastases, nodal involvement at common iliac bifurcation, and seminal vesicle infiltration. The PGM achieved superior predictive performance with a balanced accuracy of 0.73, sensitivity of 0.60, and specificity of 0.86, substantially outperforming both the LR (balanced accuracy: 0.50, sensitivity: 0.00, specificity: 1.00) and CAPRA (balanced accuracy: 0.59, sensitivity: 0.20, specificity: 0.99). The decision tree provided an explainable classifier with CAPRA as a primary branch node, followed by SUVmax and specific PET-detected tumor sites. Conclusions: Integrating ⁶⁸Ga-PSMA-11 imaging biomarkers with clinical parameters, such as CAPRA, significantly improves models to predict progression in patients with high-risk PCa undergoing primary treatment. The PGM offers superior balanced accuracy and enables risk stratification that may guide personalized treatment decisions. Full article

To address the challenge of physically realizing fractional-order electrical networks, this study proposes an implementation method for a mechatronic inerter–spring–damper (ISD) suspension based on a fractional-order biquadratic transfer function. Building upon a previously established model of a mechatronic ISD suspension, the influence of parameter perturbations on the suspension’s dynamic performance characteristics was systematically investigated. Positive real synthesis was employed to determine the optimal five-element passive network structure for the fractional-order biquadratic electrical network. Subsequently, the Oustaloup filter approximation algorithm was utilized to realize the integer-order equivalents of the fractional-order electrical components, and the approximation effectiveness was analyzed through frequency-domain and time-domain simulations. Bench testing validated the effectiveness of the proposed method: under random road excitation at 20 m/s, the root mean square (RMS) values of the vehicle body acceleration, suspension working space, and dynamic tire load were reduced by 7.86%, 17.45%, and 2.26%, respectively, in comparison with those of the traditional passive suspension. This research provides both theoretical foundations and practical engineering solutions for implementing fractional-order transfer functions in vehicle suspensions, establishing a novel technical pathway for comprehensively enhancing suspension performance. Full article

Dicke superradiance, a single-mode burst of radiation emitted by an ensemble of two-level atoms, has garnered tremendous attention within the physics community. Its extension to multi-level systems introduces additional degrees of freedom, such as mode-selective control over well-known Dicke superradiant behaviors. However, previous work on the extension to two-mode superradiance in three-level atoms has been largely overlooked for over five decades. In this study, we revisit the two-mode superradiance model for a Λ-type three-level system, where two modes couple to a common excited state and two separate lower levels, offering new insights. For the first time, we obtain exact numerical solutions of the two-mode rate equations for this model. We analyze the temporal evolution of two-mode intensities, superradiance time delays, and quantum noise in the time domain as the number of atoms varies. We believe this work will enable external mode-selective control over superradiance processes—a capability unattainable in the single-mode case. Full article

Various areas of natural language processing (NLP) have greatly benefited from the development of large language models in recent years. This research addresses the challenge of developing efficient tokenizers for transformer-based domain-specific language models. Tokenization efficiency within transformer-based models is directly related to model efficiency, which motivated the research we present in this paper. Our goal in this research was to demonstrate that the appropriate selection of data used for tokenizer training has a significant impact on tokenizer performance. Subsequently, we will demonstrate that efficient tokenizers and models can be developed even if language resources are limited. To do so, we will present a domain-adapted large language model tokenizer developed for masked language modeling of the Serbian legal domain. In this paper, we will present a comparison of the tokenization performance for a domain-adapted tokenizer in version 2 of the SrBERTa language model we developed, against the performances of five other tokenizers belonging to state-of-the-art multilingual, Slavic or Serbian-specific models—XLM-RoBERTa (base-sized), BERTić, Jerteh-81, SrBERTa v1, NER4Legal_SRB. The comparison is performed using a test dataset consisting of 275,660 samples of legal texts written in the Cyrillic alphabet gathered from the Official Gazette of the Republic of Serbia. This dataset contains 197,134 distinct words, while the overall word count is 5,265,352. We will show that our tokenizer, trained upon a domain-adapted dataset, outperforms presented tokenizers by at least 4.5% ranging to 54.62%, regarding the number of tokens generated for the whole test dataset. In terms of tokenizer fertility, we will show that our tokenizer outperforms compared tokenizers by at least 6.39% ranging to 56.8%. Full article

Background: Large language models (LLMs) like ChatGPT are increasingly being explored for medical applications. However, their reliability in providing medication advice for patients with complex clinical situations, particularly those with multiple comorbidities, remains uncertain and under-investigated. This study aimed to systematically evaluate the performance, consistency, and safety of ChatGPT in generating medication recommendations for complex cardiovascular disease (CVD) scenarios. Methods: In this simulation-based study (21 January–1 February 2024), ChatGPT 3.5 and 4.0 were prompted 10 times for each of 25 scenarios, representing five common CVDs paired with five major comorbidities. A panel of five cardiologists independently classified each unique drug recommendation as “high priority” or “low priority”. Key metrics included physician approval rates, the proportion of high-priority recommendations, response consistency (Jaccard similarity index), and error pattern analysis. Statistical comparisons were made using Z-tests, chi-square tests, and Wilcoxon Signed-Rank tests. Results: The overall physician approval rate for GPT-4 (86.90%) was modestly but significantly higher than that for GPT-3.5 (85.06%; p = 0.0476) based on aggregated data. However, a more rigorous paired-scenario analysis of high-priority recommendations revealed no statistically significant difference between the models (p = 0.407), indicating the advantage is not systematic. A chi-square test confirmed significant differences in error patterns (p < 0.001); notably, GPT-4 more frequently recommended contraindicated drugs in high-risk scenarios. Inter-model consistency was low (mean Jaccard index = 0.42), showing the models often provide different advice. Conclusions: While demonstrating high overall physician approval rates, current LLMs exhibit inconsistent performance and pose significant safety risks when providing medication advice for complex CVD cases. Their reliability does not yet meet the standards for autonomous clinical application. Future work must focus on leveraging real-world data for validation and developing domain-specific, fine-tuned models to enhance safety and accuracy. Until then, vigilant professional oversight is indispensable. Full article

Preserving heritage sites is a complex challenge that requires multidisciplinary approaches, combining scientific accuracy with cultural and historical sensitivity. In alignment with UNESCO’s conservation guidelines, this study investigated the airflow dynamics and wind-induced structural effects within ancient architecture using advanced computational fluid dynamics (CFD). The study site was the Na Phra Meru Historical Temple in Ayutthaya, Thailand, where the shear stress transport $k-\omega$ turbulence model was applied to analyze distinctive airflow patterns. A high-precision 3D computational domain was developed using Faro focus laser scanning technology, with the CFD results being validated based on onsite experimental data. The findings provided critical insights into the temple’s ventilation behavior, revealing strong correlations between turbulence characteristics, wind speed, temperature, and relative humidity. Notably, the small slit windows generated complex flow mixing, producing a large internal recirculation zone spanning approximately 70% of the central interior space. In addition to airflow distribution, the study evaluated the aerodynamic forces and rotational moments acting on the structure based on five prevailing wind directions. Based on these results, winds from the east and northeast generated the highest aerodynamic loads and rotational stresses, particularly in the lateral and vertical directions. Overall, the findings highlighted the critical role of airflow and wind-induced forces in the deterioration and long-term stability of heritage buildings. The study demonstrated the value of integrating CFD, environmental data, and structural analysis to bridge the gap between conservation science and engineering practice. Future work will explore further the interactions between wall moisture and the multi-layered pigments in mural paintings to inform preservation practices. Full article

With the rapid advancement of digital technologies, smart construction has emerged as a transformative approach within the construction industry. Central to the success of human-machine collaboration is human-machine trust, which plays a critical role in safety, performance, and the adoption of intelligent systems. This study develops and empirically tests a comprehensive structural equation model to explore the formation mechanism of human-machine trust in smart construction. Drawing on the three-domain framework, five primary constructs—role cognition; controllability; technology attachment; equipment reliability; and autonomy—are identified across individual and system dimensions. The model also incorporates trust propensity and task complexity as contextual moderators. A questionnaire survey of 288 construction professionals in China was conducted, and partial least squares structural equation modelling (PLS-SEM) was employed to analyze the data. The results confirm that all five constructs significantly and positively influence human-machine trust, with role cognition and autonomy having the strongest effects. Furthermore, trust propensity positively moderates the impact of individual traits, while task complexity negatively moderates the effect of equipment characteristics on trust formation. These findings provide valuable theoretical insights and practical guidance for the design of trustworthy intelligent systems, which can foster safer and more effective human-machine collaboration in smart construction. Full article