Search Results (5,216)

Large Language Models (LLMs) exhibit remarkable generative capabilities but remain vulnerable to hallucinations—outputs that are fluent yet inaccurate, ungrounded, or inconsistent with source material. To address the lack of methodologically grounded surveys, this paper introduces a novel method-oriented taxonomy of hallucination mitigation strategies in text-based LLMs. The taxonomy organizes over 300 studies into six principled categories: Training and Learning Approaches, Architectural Modifications, Input/Prompt Optimization, Post-Generation Quality Control, Interpretability and Diagnostic Methods, and Agent-Based Orchestration. Beyond mapping the field, we identify persistent challenges such as the absence of standardized evaluation benchmarks, attribution difficulties in multi-method systems, and the fragility of retrieval-based methods when sources are noisy or outdated. We also highlight emerging directions, including knowledge-grounded fine-tuning and hybrid retrieval–generation pipelines integrated with self-reflective reasoning agents. This taxonomy provides a methodological framework for advancing reliable, context-sensitive LLM deployment in high-stakes domains such as healthcare, law, and defense. Full article

Cognitive biases continue to pose significant challenges in executive decision-making, often leading to strategic inefficiencies, misallocation of resources, and flawed risk assessments. While traditional decision-making relies on intuition and experience, these methods are increasingly proving inadequate in addressing the complexity of modern business environments. Despite the growing integration of big data analytics into executive workflows, existing research lacks a comprehensive examination of how AI-driven methodologies can systematically mitigate biases while maintaining transparency and trust. This paper addresses these gaps by analyzing how big data analytics, artificial intelligence (AI), machine learning (ML), and explainable AI (XAI) contribute to reducing heuristic-driven errors in executive reasoning. Specifically, it explores the role of predictive modeling, real-time analytics, and decision intelligence systems in enhancing objectivity and decision accuracy. Furthermore, this study identifies key organizational and technical barriers—such as biases embedded in training data, model opacity, and resistance to AI adoption—that hinder the effectiveness of data-driven decision-making. By reviewing empirical findings from A/B testing, simulation experiments, and behavioral assessments, this research examines the applicability of AI-powered decision support systems in strategic management. The contributions of this paper include a detailed analysis of bias mitigation mechanisms, an evaluation of current limitations in AI-driven decision intelligence, and practical recommendations for fostering a more data-driven decision culture. By addressing these research gaps, this study advances the discourse on responsible AI adoption and provides actionable insights for organizations seeking to enhance executive decision-making through big data analytics. Full article

The integration of emerging technologies is reshaping the teaching skills required in the 21st century, yet little evidence exists on how contextualized training supports rural teachers in adopting active methodologies and critically incorporating AI into entrepreneurship education. This study evaluated the impact of a 40-h professional development program implemented in Educational District 15D01 in the Ecuadorian Amazon. Thirty-nine secondary school teachers participated (mean age = 43.1 years); 36% lacked prior entrepreneurship training, and 44% had not recently mentored student projects. A sequential explanatory mixed-methods design was employed. The quantitative phase employed a 22-item questionnaire that addressed four dimensions: entrepreneurial knowledge, competencies, methodological strategies, and AI integration. Significant pre–post improvements were found (p < 0.001), with large effects for knowledge (d = 1.43), methodologies (d = 1.39), and AI integration (d = 1.30), and a moderate effect for competences (d = 0.66). The qualitative phase analyzed 312 open-ended responses, highlighting greater openness to innovation, enhanced teacher agency, and favorable perceptions of AI as a resource for ideation, prototyping, and evaluation. Overall, the findings suggest that situated, contextually aligned training can strengthen digital equity policies, foster pedagogical innovation, and empower educators in underserved rural communities, contributing to sustainable pathways for teacher professional development. Full article

The increasing penetration of renewable energy sources and the consequent rise in forecast uncertainty have underscored the need for robust operational strategies in transmission power systems. This paper introduces a risk-averse, data-driven distributionally robust optimization framework that integrates unit commitment and power flow constraints to enhance both reliability and operational security. Leveraging advanced forecasting techniques implemented via gradient boosting and enriched with cyclical and lag-based time features, the proposed methodology forecasts renewable generation and demand profiles. Uncertainty is quantified through a quantile-based analysis of forecasting residuals, which forms the basis for constructing data-driven ambiguity sets using Wasserstein balls. The framework incorporates comprehensive network constraints, power flow equations, unit commitment dynamics, and battery storage operational constraints, thereby capturing the intricacies of modern transmission systems. A worst-case net demand and renewable generation scenario is computed to further bolster the system’s risk-averse characteristics. The proposed method demonstrates the integration of data preprocessing, forecasting model training, uncertainty quantification, and robust optimization in a unified environment. Simulation results on a representative IEEE 24-bus network reveal that the proposed method effectively balances economic efficiency with risk mitigation, ensuring reliable operation under adverse conditions. This work contributes a novel, integrated approach to enhance the reliability of transmission power systems in the face of increasing uncertainty. Full article

The methylation of CpG sites with 5mC mark is a dynamic epigenetic modification. However, the relationship between the methylation and the surrounding genomic sequence context remains poorly explored. Investigation of the allele methylation provides an opportunity to decipher the interplay between differences in the primary DNA sequence and epigenetic variation. Here, we performed high-coverage long-read whole-genome direct DNA sequencing of one individual using Oxford Nanopore technology. We also used Illumina whole-genome sequencing of the parental genomes in order to identify allele-specific methylation sites with a trio-binning approach. We have compared the results of the haplotype-specific methylation detection and revealed that trio binning outperformed other approaches that do not take into account parental information. Also, we analysed the cis-regulatory effects of the genomic variations for influence on CpG methylation. To this end, we have used available Deep Learning models trained on the primary DNA sequence to score the cis-regulatory potential of the genomic loci. We evaluated the functional role of the allele-specific epigenetic changes with respect to gene expression using long-read Nanopore RNA sequencing. Our analysis revealed that the frequency of SNVs near allele-specific methylation positions is approximately four times higher compared to the biallelic methylation positions. In addition, we identified that allele-specific methylation sites are more conserved and enriched at the chromatin states corresponding to bivalent promoters and enhancers. Together, these findings suggest that significant impact on methylation can be encoded in the DNA sequence context. In order to elucidate the effect of the SNVs around sites of allele-specific methylation, we applied the Deep Learning model for detection of the cis-regulatory modules and estimated the impact that a genomic variant brings with respect to changes to the regulatory activity of a DNA loci. We revealed higher cis-regulatory impact variants near differentially methylated sites that we further coupled with transcriptomic long-read sequencing results. Our investigation also highlights technical aspects of allele methylation analysis and the impact of sequencing coverage on the accuracy of genomic phasing. In particular, increasing coverage above 30X does not lead to a significant improvement in allele-specific methylation discovery, and only the addition of trio binning information significantly improves phasing. We investigated genomic variation in a single human individual and coupled computational discovery of cis-regulatory modules with allele-specific methylation (ASM) profiling. In this proof-of-concept analysis, we observed that SNPs located near methylated CpG sites on the same haplotype were enriched for sequence features suggestive of high-impact regulatory potential. This finding—derived from one deeply sequenced genome—illustrates how phased genetic and epigenetic data analyses can jointly put forward a hypotheses about the involvement of regulatory protein machinery in shaping allele-specific epigenetic states. Our investigation provides a methodological framework and candidate loci for future studies of genomic imprinting and cis-mediated epigenetic regulation in humans. Full article

Arbitrary zones are commonly used to describe and monitor external load (EL) during training and competitions. However, in recent years, relative speed zones have gained interest as they allow a more detailed description of the demands of each individual player, with their benefits largely unknown. This study aimed to (i) identify differences in EL methodological approaches using arbitrary and relative running speed zones; (ii) examine the effect of the methodological approaches to identify fast and slow basketball players during competition and training; and (iii) determine the effect of the season stage on the methodological approaches. Twelve players from a Spanish fourth-division basketball team were observed for a full season of matches and training using inertial devices with ultra-wideband indoor tracking technology and micro-sensors. Relative velocity zones were based on the maximum velocity achieved during each match quarter and were retrospectively recalculated into four zones. A linear mixed model (LMM) compared fast and slow players based on speed profiles between arbitrary and relative thresholds and during each competition stage. All players surpassed peak speeds of 24 km·h⁻¹ during the season, exceeding typical values reported in elite basketball (20–24.5 km·h⁻¹). Arbitrary thresholds produced greater distances in high-speed running (Zones 3 and 4) and yielded lower values in low-speed activity (Zone 1), with differences of ~100 m and ~120–250 m, respectively (p < 0.001), particularly for fast-profile players. These discrepancies were consistent across most stages of the season, although relative zones better captured variations in Zone 1 across time. Training sessions also elicited +8.7% to +40.7% greater distances > 18 km·h⁻¹ compared to matches. The speed zone methodology substantially influenced EL estimates and affected how player EL was interpreted across time. Arbitrary and relative approaches offer unique applications, with coaches and sport scientists encouraged to be aware that using a one-size-fits-all approach may lead to misrepresentation of individual player demands, especially when tracking changes in performance or managing fatigue throughout a competitive season. Full article

Power outages pose considerable risks to the reliability of electric grids, affecting both consumers and utilities through service disruptions and potential economic losses. This study analyzes a historical outage dataset from a Regional Transmission Organization (RTO) to reveal key patterns and trends that suggest outage management strategies. By integrating exploratory data analysis, predictive modeling, and a Large Language Model (LLM)-based interface integration, as well as data visualization techniques, we identify and present critical drivers of outage duration and frequency. A random forest regressor trained on features including planned duration, facility name, outage owner, priority, season, and equipment type proved highly effective for predicting outage duration with high accuracy. This predictive framework underscores the practical value of incorporating planning information and seasonal context in anticipating outage timelines. The findings of this study not only deepen the understanding of temporal and spatial outage dynamics but also provide valuable insights for utility companies and researchers. Utility companies can use these results to better predict outage durations, allocate resources more effectively, and improve service restoration time. Researchers can leverage this analysis to enhance future models and methodologies for studying outage patterns, ensuring that artificial intelligence (AI)-driven methods can contribute to improving management strategies. The broader impact of this study is to ensure that the insights gained can be applied to strengthen the reliability and resilience of power grids or energy systems in general. Full article

This research presents a comprehensive analysis of green development policies in the building industry in New South Wales (NSW), Australia, examining their evolution and development over the past two decades. The research adopts a structured methodology comprising a policy review to identify relevant policy documents, content analysis to trace the policy framework, and SWOT (Strengths, Weaknesses, Opportunities, and Threats) analysis to evaluate the policy development, and then generate evidence-based recommendations. As the first comprehensive assessment of green development policy in the Australian building industry, the study proposes targeted policy recommendations based on analyzing the SWOT factors, including policy for the non-residential sectors and construction phase, education and training, financial support and incentives, and innovation and technology adoption. The insights offer guidance for policymakers to strengthen policy integration and accelerate the transition toward a low-carbon building industry. Full article

This paper offers content priorities, justifications, and pedagogical approaches for the integration of climate change education into the training of teachers, and thus into public schooling. To meet urgent imperatives presented by the polycrisis of the Anthropocene, climate change education must be inclusive, comprehensive, flexible, and regionally responsive. Climate change education can be achieved by adapting regional programs for teacher education to meet those requirements. An example is the Climate Education in Teacher Education (CETE) project in northern British Columbia, Canada. Using the Education Design-Based Research methodology, the project addresses critical questions for curricular and pedagogical development of teachers to address the following three questions: (a) what content and outcomes to prioritize, (b) why these elements matter, and (c) how to implement them effectively. Over two years, CETE engaged pre-service and in-service teachers through workshops, reflective practices, and consultations with Indigenous communities. Our tentative answers emphasize the importance of adapting curriculum and pedagogy to foster community resilience, address climate anxiety, and promote an ethical renewal toward sustainability. The iterative development of objectives as “High-Level Conjectures” provides flexibility and reflexivity in the design process in the face of rapid contextual change. CETE developed practical pedagogical tools and workshop strategies that align educational priorities with local and global needs. This study offers a replicable framework to empower educators and communities in diverse locations to navigate the complexities of the climate crisis in their quest for a more secure and sustainable future. Full article

The growing complexity of electric vehicle charging station (EVCS) operations—driven by grid constraints, renewable integration, user variability, and dynamic pricing—has positioned reinforcement learning (RL) as a promising approach for intelligent, scalable, and adaptive control. After outlining the core theoretical foundations, including RL algorithms, agent architectures, and EVCS classifications, this review presents a structured survey of influential research, highlighting how RL has been applied across various charging contexts and control scenarios. This paper categorizes RL methodologies from value-based to actor–critic and hybrid frameworks, and explores their integration with optimization techniques, forecasting models, and multi-agent coordination strategies. By examining key design aspects—including agent structures, training schemes, coordination mechanisms, reward formulation, data usage, and evaluation protocols—this review identifies broader trends across central control dimensions such as scalability, uncertainty management, interpretability, and adaptability. In addition, the review assesses common baselines, performance metrics, and validation settings used in the literature, linking algorithmic developments with real-world deployment needs. By bridging theoretical principles with practical insights, this work provides comprehensive directions for future RL applications in EVCS control, while identifying methodological gaps and opportunities for safer, more efficient, and sustainable operation. Full article

Land Surface Temperature (LST) is a critical parameter for analyzing urban heat islands, surface–atmosphere interactions, and environmental management. This study enhances the temporal resolution of LST data by leveraging machine learning and deep learning models. A novel methodology was developed using Landsat 8 thermal data and Sentinel-2 multispectral imagery to predict LST at finer temporal intervals in an urban setting. Although Sentinel-2 lacks a thermal band, its high-resolution multispectral data, when integrated with Landsat 8 thermal observations, provide valuable complementary information for LST estimation. Several models were employed for LST prediction, including Random Forest Regression (RFR), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) network, and Gated Recurrent Unit (GRU). Model performance was assessed using the coefficient of determination (R²) and Mean Absolute Error (MAE). The CNN model demonstrated the highest predictive capability, achieving an R² of 74.81% and an MAE of 1.588 °C. Feature importance analysis highlighted the role of spectral bands, spectral indices, topographic parameters, and land cover data in capturing the dynamic complexity of LST variations and directional patterns. A refined CNN model, trained with the features exhibiting the highest correlation with the reference LST, achieved an improved R² of 84.48% and an MAE of 1.19 °C. These results underscore the importance of a comprehensive analysis of the factors influencing LST, as well as the need to consider the specific characteristics of the study area. Additionally, a modified TsHARP approach was applied to enhance spatial resolution, though its accuracy remained lower than that of the CNN model. The study was conducted in Tehran, a rapidly urbanizing metropolis facing rising temperatures, heavy traffic congestion, rapid horizontal expansion, and low energy efficiency. The findings contribute to urban environmental management by providing high-temporal-resolution LST data, essential for mitigating urban heat islands and improving climate resilience. Full article

Language learners increasingly rely on intelligent digital tools to supplement their learning experiences, yet existing chatbots often provide limited support, lacking adaptability, personalization, or domain-specific intelligence. This study introduces a novel AI-powered multi-agent chatbot architecture designed to support English–Arabic translation and language learning. Developed through a three-phase methodology, offline preparation, real-time deployment, and evaluation, the system employs both retrieval-based and generative AI models, with specialized agents managing tasks such as translation, example retrieval, user translation review, and learning feedback. The chatbot was developed using a hybrid architecture incorporating fine-tuned Generative Pre-trained Transformer (GPT) model, sentence embedding techniques, and similarity evaluation metrics. A user study involving 40 undergraduate students and 4 faculty members evaluated the system across usability, effectiveness, and pedagogical value. Results show that the multi-agent chatbot significantly enhanced learner engagement, provided accurate and contextually appropriate language support, and was positively received by both students and instructors. These findings demonstrate the value of multi-agent design in language learning applications and highlight the potential of AI-driven chatbots as intelligent educational assistants. Full article

Background: High-performance tennis exposes players to repetitive high-load strokes and abrupt directional changes, which substantially increase musculoskeletal injury risk. This systematic review synthesized evidence on epidemiology, risk factors, and physiotherapy-led preventive strategies in elite adolescent and adult players. Methods: Following a PROSPERO-registered protocol, MEDLINE, Web of Science, and Scopus were searched (2011–2024) for observational studies reporting epidemiological outcomes in high-performance tennis. Methodological quality was appraised with NIH tools, and certainty of evidence was graded with GRADE. Results: Thirty-seven studies met inclusion criteria: 16 in adolescents, 18 in adults, and 3 mixed. Incidence ranged from 2.1 to 3.5 injuries/1000 h in juniors and 1.25 to 56.6/1000 h in adults. Seasonal prevalence was 46–54% in juniors and 30–54% in professionals. Lower-limb trauma (48–56%) predominated, followed by lumbar (12–39%) and shoulder overuse syndromes. Across age groups, abrupt increases in the acute-to-chronic workload ratio (≥1.3 in juniors; ≥1.5 in adults) were the strongest extrinsic predictor of injury. Intrinsic contributors included reduced glenohumeral internal rotation, scapular dyskinesis, and poor core stability. Three prevention clusters emerged: (1) External load control, four-week “ramp-up” strategies reduced injury incidence by up to 21%; (2) Kinetic-chain conditioning, core stability plus eccentric rotator-cuff training decreased overuse by 26% and preserved shoulder mobility; and (3) Technique/equipment adjustments, grip-size personalization halved lateral epicondylalgia, while serve-timing modifications reduced shoulder torque. Conclusions: Injury risk in high-performance tennis is quantifiable and preventable. Progressive load management targeted kinetic-chain conditioning, and tailored technique/equipment modifications represent the most effective evidence-based safeguards for adolescent and adult elite players. Full article

Organismal biosensing leverages the olfactory acuity of living systems to detect volatile organic compounds (VOCs) associated with cancer, offering a low-cost and non-invasive complement to conventional diagnostics. Early studies demonstrate its feasibility across diverse platforms. In C. elegans, chemotaxis assays on urine samples achieved sensitivities of 87–96% and specificities of 90–95% in case–control cohorts (n up to 242), while calcium imaging of AWC neurons distinguished breast cancer urine with ~97% accuracy in a small pilot cohort (n ≈ 40). Trained canines have identified prostate cancer from urine with sensitivities of ~71% and specificities of 70–76% (n ≈ 50), and AI-augmented canine breath platforms have reported accuracies of ~94–95% across ~1400 participants. Insects such as locusts and honeybees enable ultrafast neural decoding of VOCs, achieving 82–100% classification accuracy within 250 ms in pilot studies (n ≈ 20–30). Collectively, these platforms validate the principle that organismal behavior and neural activity encode cancer-related VOC signatures. However, limitations remain, including small cohorts, methodological heterogeneity, and reliance on binary outputs. This review proposes a Dual-Pathway Framework, where Pathway 1 leverages validated indices (e.g., the Chemotaxis Index) for high-throughput screening, and Pathway 2 applies machine learning to high-dimensional behavioral vectors for cancer subtyping, staging, and monitoring. By integrating these approaches, organismal biosensing could evolve from proof-of-concept assays into clinically scalable precision diagnostics. Full article

Dyadic co-construction, the mutual adaptation that occurs between dogs and their owners, is often discussed in terms of cooperation and participation, yet it remains poorly defined and under-conceptualized in the literature. This review proposed that self-determination theory (SDT), with its three core psychological needs—autonomy, competence, and relatedness (attachment)—offers a valuable framework for understanding this phenomenon within a dyadic context. The objectives of this review were twofold: (1) to conceptualize co-construction in owner–dog interactions through the lens of SDT, and (2) to propose methodological approaches for studying this process, while acknowledging their current limitations. Dyadic co-construction emerges as a dynamic, evolving process of mutual influence, shaped by biopsychosocial factors, individual and shared experiences, and the physical and social environments of both human and dog, as well as the dyad as a unit. Depending on the nature of the interaction, co-construction can be beneficial or detrimental. Positive training practices and secure attachment patterns in both humans and dogs tend to foster more harmonious co-construction, whereas aversive methods and insecure attachment may hinder it. Although existing methodologies offer promising insights into this process, they often lack standardization, statistical robustness, and true bidirectionality. This review underscores the need for more integrative, longitudinal, and empirically grounded approaches to fully capture the complexity and clinical relevance of owner–dog dyadic co-construction. Full article