Search Results (4,784)

Recent advancements in large language models (LLMs) have significantly enhanced the effectiveness of Retrieval-Augmented Generation (RAG) systems. This study focuses on the development and evaluation of a domain-specific AI chatbot designed to support international student admissions by leveraging LLM-based RAG pipelines. We implement and compare multiple pipeline configurations, combining retrieval methods (e.g., Dense, MMR, Hybrid), chunking strategies (e.g., Semantic, Recursive), and both open-source and commercial LLMs. Dual evaluation datasets of LLM-generated and human-tagged QA sets are used to measure answer relevancy, faithfulness, context precision, and recall, alongside heuristic NLP metrics. Furthermore, latency analysis across different RAG stages is conducted to assess deployment feasibility in real-world educational environments. Results show that well-optimized open-source RAG pipelines can offer comparable performance to GPT-4o while maintaining scalability and cost-efficiency. These findings suggest that the proposed chatbot system can provide a practical and technically sound solution for international student services in resource-constrained academic institutions. Full article

This paper presents the design, implementation, and experimental validation of a Condition Monitoring (CM) circuit for SiC-based Power Electronics Converters (PECs). The paper leverages in situ drain–source resistance (R_ds,on) measurements, interfaced with cloud connectivity for data processing and lifetime assessment, addressing key limitations in current state-of-the-art (SOTA) methods. Traditional approaches rely on expensive data acquisition systems under controlled laboratory conditions, making them unsuitable for real-world applications due to component variability, time delay, and noise sensitivity. Furthermore, these methods lack cloud interfacing for real-time data analysis and fail to provide comprehensive reliability metrics such as Remaining Useful Life (RUL). Additionally, the proposed CM method benefits from noise mitigation during switching transitions by utilizing delay circuits to ensure stable and accurate data capture. Moreover, collected data are transmitted to the cloud for long-term health assessment and damage evaluation. In this paper, experimental validation follows a structured design involving signal acquisition, filtering, cloud transmission, and temperature and thermal degradation tracking. Experimental testing has been conducted at different temperatures and operating conditions, considering coolant temperature variations (40 °C to 80 °C), and an output power of 7 kW. Results have demonstrated a clear correlation between temperature rise and R_ds,on variations, validating the ability of the proposed method to predict device degradation. Finally, by leveraging cloud computing, this work provides a practical solution for real-world Wide Band Gap (WBG)-based PEC reliability and lifetime assessment. Full article

Demand Side Management (DSM) plays a crucial role in modern energy systems, enabling more efficient use of energy resources and contributing to the sustainability of the power grid. This study examines DSM strategies within a multi-environment context encompassing residential, commercial, and industrial sectors, with a focus on diverse appliance types that exhibit distinct operational characteristics and user preferences. Initially, a single-objective optimization approach using Genetic Algorithms (GAs) is employed to minimize the total energy cost under a real Time-of-Use (ToU) pricing scheme. This heuristic method allows for the effective scheduling of appliance operations while factoring in their unique characteristics such as power consumption, usage duration, and user-defined operational flexibility. This study extends the optimization problem to a multi-objective framework that incorporates the minimization of CO₂ emissions under a real annual energy mix while also accounting for user discomfort. The Non-dominated Sorting Genetic Algorithm II (NSGA-II) is utilized for this purpose, providing a Pareto-optimal set of solutions that balances these competing objectives. The inclusion of multiple objectives ensures a comprehensive assessment of DSM strategies, aiming to reduce environmental impact and enhance user satisfaction. Additionally, this study monitors the Peak-to-Average Ratio (PAR) to evaluate the impact of DSM strategies on load balancing and grid stability. It also analyzes the impact of considering different periods of the year with the associated ToU hourly schedule and CO₂ emissions hourly profile. A key innovation of this research is the integration of detailed, category-specific metrics that enable the disaggregation of costs, emissions, and user discomfort across residential, commercial, and industrial appliances. This granularity enables stakeholders to implement tailored strategies that align with specific operational goals and regulatory compliance. Also, the emphasis on a user discomfort indicator allows us to explore the flexibility available in such DSM mechanisms. The results demonstrate the effectiveness of the proposed multi-objective optimization approach in achieving significant cost savings that may reach 20% for industrial applications, while the order of magnitude of the trade-offs involved in terms of emissions reduction, improvement in discomfort, and PAR reduction is quantified for different frameworks. The outcomes not only underscore the efficacy of applying advanced optimization frameworks to real-world problems but also point to pathways for future research in smart energy management. This comprehensive analysis highlights the potential of advanced DSM techniques to enhance the sustainability and resilience of energy systems while also offering valuable policy implications. Full article

Humanoid robots have attracted considerable attention for their anthropomorphic structure, extended workspace, and versatile capabilities. This paper presents a novel humanoid upper-body robotic system comprising a pair of 8-degree-of-freedom (DOF) arms, a 3-DOF head, and a 3-DOF torso—yielding a 22-DOF architecture inspired by human biomechanics and implemented via standardized hollow joint modules. To overcome the critical reliance of zeroing neural network (ZNN)-based trajectory tracking on the Jacobian matrix derivative, we propose an integration-enhanced matrix derivative observer (IEMDO) that incorporates nonlinear feedback and integral correction. The observer is theoretically proven to ensure asymptotic convergence and enables accurate, real-time estimation of matrix derivatives, addressing a fundamental limitation in conventional ZNN solvers. Workspace analysis reveals that the proposed design achieves an 87.7% larger total workspace and a remarkable 3.683-fold expansion in common workspace compared to conventional dual-arm baselines. Furthermore, the observer demonstrates high estimation accuracy for high-dimensional matrices and strong robustness to noise. When integrated into the ZNN controller, the IEMDO achieves high-precision trajectory tracking in both simulation and real-world experiments. The proposed framework provides a practical and theoretically grounded approach for redundant humanoid arm control. Full article

This paper presents an annual, real-world evaluation of the performance and economics of a residential photovoltaic (PV) system coupled with a battery energy storage system (BESS) in southern Poland. The system, monitored with 5 min resolution, operated under time-of-use (TOU) electricity tariffs. Seasonal variation was significant; self-sufficiency exceeded 90% in summer, while winter conditions increased grid dependency. The hybrid system reduced electricity costs by over EUR 1400 annually, with battery operation optimized for high-tariff periods. Comparative analysis of three configurations—grid-only, PV-only, and PV + BESS—demonstrated the economic advantage of the integrated solution, with the shortest payback period (9.0 years) achieved with financial support. However, grid voltage instability during high PV production led to inverter shutdowns, highlighting limitations in the infrastructure. This study emphasizes the importance of tariff strategies, environmental conditions, and voltage control when designing residential PV-BESS systems. Full article

Background/objectives: This was a post hoc analysis of safety data across the bivalent respiratory syncytial virus prefusion F (RSVpreF) vaccine clinical trial development program. Methods: Data from eight clinical trials in 46,913 immunocompetent adults who received RSVpreF or placebo were analyzed. Local reactions and systemic events were assessed among non-pregnant ≥18-year-olds (n = 9517); adverse events (AEs) among pregnant and non-pregnant 18–59-year-olds (n = 9238); and vaccine-related AEs among non-pregnant ≥18-year-olds (n = 39,314). Post-marketing data in non-pregnant adults were considered. Results: Local reactions and systemic events were reported more frequently in RSVpreF versus placebo recipients; injection site pain was the most common local reaction (RSVpreF, 18.9%; placebo, 7.4%), and fatigue (23.5%; 18.4%) and headache (19.5%; 15.0%) were the most common systemic events. Percentages of AEs within 1 month after vaccination were similar across groups (RSVpreF, 12.8%; placebo, 13.1%); severe AEs were reported in ≤1.5% of participants. Differences in percentages of individuals reporting vaccine-related AEs between the RSVpreF and placebo groups were <0.2% for all related AEs. Serious AEs throughout the study were reported in ≤14.0% (RSVpreF, 12.6%; placebo, 14.0%). No atrial fibrillation, Guillain-Barré syndrome, or acute polyneuropathy cases were reported. The AE data from post-marketing data sources were consistent with the safety profile from the clinical trial program, with no new safety concerns. Conclusions: Integrated data demonstrated that RSVpreF was well tolerated with a favorable safety profile in non-pregnant and pregnant adults. Ongoing surveillance through real-world use and clinical trial experience continue to support the safety profile of RSVpreF. ClinicalTrials.gov: CT03529773/NCT04071158/NCT04785612/NCT05035212/NCT05096208/NCT05842967/NCT04032093/NCT04424316. Full article

Effective earthquake early warning (EEW) is essential for disaster prevention in the built environment, enabling a rapid structural response, system shutdown, and occupant evacuation to mitigate damage and casualties. However, most current EEW systems lack rigorous reliability analyses of their predictive outcomes, limiting their effectiveness in real-world scenarios—especially for on-site warnings, where data are limited and time is critical. To address these challenges, we propose a Bayesian neural network (BNN) framework based on Stein variational gradient descent (SVGD). By performing Bayesian inference, we estimate the posterior distribution of the parameters, thus outputting a reliability analysis of the prediction results. In addition, we incorporate a continual learning mechanism based on elastic weight consolidation, allowing the system to adapt quickly without full retraining. Our experiments demonstrate that our SVGD-BNN model significantly outperforms traditional peak displacement (Pd)-based approaches. In a 3 s time window, the Pearson correlation coefficient R increases by 9.2% and the residual standard deviation $SD$ decreases by 24.4% compared to a variational inference (VI)-based BNN. Furthermore, the prediction variance generated by the model can effectively reflect the uncertainty of the prediction results. The continual learning strategy reduces the training time by 133–194 s, enhancing the system’s responsiveness. These features make the proposed framework a promising tool for real-time, reliable, and adaptive EEW—supporting disaster-resilient building design and operation. Full article

Electroless nickel deposition (ELD) is an autocatalytic technique extensively used to impart conductive, protective, and mechanical functionalities to inherently non-conductive synthetic substrates. This review systematically explores the fundamental mechanisms of electroless nickel deposition, emphasising recent advancements in surface activation methods, solvent systems, and microstructural control. Critical analysis reveals that bio-inspired activation methods, such as polydopamine (PDA) and tannic acid (TA), significantly enhance coating adhesion and durability compared to traditional chemical etching and plasma treatments. Additionally, solvent engineering, particularly using polar aprotic solvents like dimethyl sulfoxide (DMSO) and ethanol-based systems, emerges as a key strategy for achieving uniform, dense, and flexible coatings, overcoming limitations associated with traditional aqueous baths. The review also highlights that microstructural tailoring, specifically the development of amorphous-nanocrystalline hybrid nickel coatings, effectively balances mechanical robustness (hardness exceeding 800 HV), flexibility, and corrosion resistance, making these coatings particularly suitable for wearable electronic textiles and smart materials. Furthermore, commercial examples demonstrate the real-world applicability and market readiness of nickel-coated synthetic fibres. Despite significant progress, persistent challenges remain, including reliable long-term adhesion, internal stress management, and environmental sustainability. Future research should prioritise environmentally benign plating baths, standardised surface activation protocols, and scalable deposition processes to fully realise the industrial potential of electroless nickel coatings. Full article

Background: Bimekizumab, secukinumab, and ixekizumab are IL-17-targeting biologics approved for the treatment of moderate-to-severe plaque psoriasis. While secukinumab and ixekizumab selectively inhibit IL-17A, bimekizumab targets both IL-17A and IL-17F, potentially providing greater anti-inflammatory efficacy. This study aimed to compare the real-world effectiveness, safety, and tolerability of these agents in a Polish dermatology center between 2019 and 2024. Methods: We conducted a retrospective analysis of 98 patients meeting at least one of the following criteria: PASI ≥ 10, BSA ≥ 10, DLQI ≥ 10, or involvement of special areas with inadequate response or contraindications to ≥2 systemic therapies. Patients with prior exposure only to IL-17 inhibitors were excluded. PASI, BSA, and DLQI scores were recorded at baseline, week 4, and week 12. Due to differences in dosing schedules, outcomes were aligned using standardized timepoints and exponential modeling of continuous response trajectories. Mixed-effects ANOVA was used to assess the influence of baseline factors (age, BMI, PsA status) on treatment outcomes. Adverse events were documented at each monthly follow-up visit. Results: Bimekizumab showed the greatest effect size for PASI reduction (Hedges’ g = 3.662), followed by secukinumab (2.813) and ixekizumab (1.986). Exponential modeling revealed a steeper response trajectory with bimekizumab (intercept = 0.289), suggesting a more rapid PASI improvement. The efficacy of bimekizumab was particularly notable in patients who were previously treated with IL-23 inhibitors. All three agents demonstrated favorable safety profiles, with no serious adverse events or discontinuations. The most frequent adverse events were mild and included upper respiratory tract infections and oral candidiasis. Conclusions: This real-world analysis confirmed that IL-17 inhibitors effectively improved PASI, BSA, and DLQI scores in moderate-to-severe psoriasis. Bimekizumab demonstrated the most rapid early improvements and a higher modeled likelihood of complete clearance, without significant differences at week 12. All agents were well tolerated, underscoring the need for further individualized, large-scale studies. Full article

This paper addresses the problem of optimizing obesity, which has been a challenging issue in the last decade based on recent data revealed in 2024 by the World Health Organization (WHO). The current work introduces a new mathematical model of the dynamics of weight over time with embedded control parameters to optimize the number of obese, overweight, and comorbidity populations. The mathematical formulation of the model is developed under certain sufficient conditions that guarantee the positivity and boundedness of solutions over time. The model structure exhibits inherent symmetry in population group transitions, particularly around the equilibrium state, which allows the application of analytical tools such as the Routh–Hurwitz and Metzler criteria. Then, the analysis of local and global stability of the obesity-free equilibrium state is discussed based on these criteria. Based on the Pontryagin maximum principle (PMP), the deviation from the obesity-free equilibrium state is controlled. The model’s effectiveness is demonstrated through simulation using the Forward–Backward Sweeping algorithm with parameters derived from recent research in human health. Incorporating symmetry considerations in the model enhances the understanding of system behavior and supports balanced intervention strategies. Results suggest that the model can effectively inform strategies to mitigate obesity prevalence and associated health risks. Full article

This study evaluates the impact of digital systems and IT audit frameworks on tax performance and integrity within tax administrations. Using international data from organizations like the World Bank, OECD (Organisation for Economic Co-operation and Development), and IMF (International Monetary Fund), the research examines the relationship between tax revenue as a percentage of GDP, digital infrastructure, corruption perception, e-government development, and cybersecurity readiness. Quantitative analysis, including correlation, regression, and clustering methods, reveals a strong positive relationship between digital maturity, e-governance, and tax performance. Countries with advanced digital governance systems and robust IT audit frameworks, such as COBIT, tend to show higher tax revenues and lower corruption levels. The study finds that e-government development and anti-corruption measures explain over 40% of the variance in tax performance. Cluster analysis distinguishes between digitally advanced, high-compliance countries and those lagging in IT adoption. The findings suggest that digital transformation strengthens fiscal integrity by automating compliance and reducing human contact, which in turn mitigates bribery risks and enhances fraud detection. The study highlights the need for adopting international best practices to guide the digitalization of tax administrations, improving efficiency, transparency, and trust in public finance. Full article

Obstructive sleep apnea (OSA) is a common condition associated with intermittent hypoxia, systemic inflammation, and vascular dysfunction; mechanisms implicated in retinal disease pathogenesis. This real-world retrospective cohort study used data from the TriNetX Research Network to assess whether continuous positive airway pressure (CPAP) therapy reduces retinal disease incidence among adults with OSA and BMI between 25.0 and 30.0 kg/m². After 1:1 propensity score matching, 101,754 patients were included in the analysis. Retinal outcomes included diabetic retinopathy (DR), age-related macular degeneration (AMD), retinal vein occlusion (RVO), and central serous chorioretinopathy (CSC). CPAP use was associated with a modest but statistically significant reduction in DR (3.2% vs. 3.4%, RR: 0.922, p = 0.016) and AMD (2.1% vs. 2.3%, RR: 0.906, p = 0.018), while no significant differences were found for RVO or CSC. These findings support prior evidence linking CPAP to improved retinal microvascular health and suggest a protective effect against specific retinal complications. Limitations include a lack of data on CPAP adherence, OSA severity, and imaging confirmation. Still, this study highlights the importance of interdisciplinary care between sleep and eye health, and the need for further prospective studies to validate CPAP’s role in preventing retinal disease progression in OSA patients. Full article

Electric buses (EBs) are increasingly adopted for their environmental and operational benefits, yet their real-world efficiency is influenced by climate, route characteristics, and auxiliary energy demands. While most existing research identifies winter as the most energy-intensive season due to cabin heating and reduced battery performance, this study presents a contrasting perspective based on a three-year longitudinal analysis of the LYMMO fleet in Orlando, Florida—a subtropical U.S. region. The findings reveal that summer is the most energy-intensive season, primarily due to sustained HVAC usage driven by high ambient temperatures—a seasonal pattern rarely reported in the current literature and a key regional contribution. Additionally, idling time exceeds driving time across all seasons, with HVAC usage during idling emerging as the dominant contributor to total energy consumption. To mitigate these inefficiencies, a proxy-based HVAC energy estimation method and an optimization model were developed, incorporating ambient temperature and peak passenger load. This approach achieved up to 24% energy savings without compromising thermal comfort. Results validated through non-parametric statistical testing support operational strategies such as idling reduction, HVAC control, and seasonally adaptive scheduling, offering practical pathways to improve EB efficiency in warm-weather transit systems. Full article

This study presents a comprehensive analysis of CO₂ trapping mechanisms in subsurface reservoirs by integrating numerical reservoir simulations, geochemical modeling, and machine learning techniques to enhance the design and evaluation of carbon capture and storage (CCS) strategies. A two-dimensional reservoir model was developed to simulate CO₂ injection dynamics under realistic geomechanical and geochemical conditions, incorporating four primary trapping mechanisms: residual, solubility, mineralization, and structural trapping. To improve computational efficiency without compromising accuracy, advanced machine learning models, including random forest, gradient boosting, and decision trees, were deployed as smart proxy models for rapid prediction of trapping behavior across multiple scenarios. Simulation outcomes highlight the critical role of hysteresis, aquifer dynamics, and producer well placement in enhancing CO₂ trapping efficiency and maintaining long-term storage stability. To support the credibility of the model, a qualitative validation framework was implemented by comparing simulation results with benchmarked field studies and peer-reviewed numerical models. These comparisons confirm that the modeled mechanisms and trends align with established CCS behavior in real-world systems. Overall, the study demonstrates the value of combining traditional reservoir engineering with data-driven approaches to optimize CCS performance, offering scalable, reliable, and secure solutions for long-term carbon sequestration. Full article

In today’s fast-changing world, the higher education system must evolve to enhance the quality of learning and teaching. Fulfilling the role of a university is a major challenge. Universities must implement strategies that place the student at the center of their concerns; so, these strategies must be designed for and by the student. However, the high university dropout rate is one of the current problems faced by many universities. This suggests that there are some issues that hinder the learning process. Several studies have highlighted the advantage of artificial intelligence (AI) technologies in providing explorative and predictive analyses that explain why students are dropping out, with the aim of improving the quality of teaching and providing an integrated learning environment. This paper proposes a framework that predicts student dropout rates using machine learning techniques, based on data collected from various sources. Data collection was carried out between 2022 and 2024. We used a quantitative analysis method employed through a questionnaire distributed to 120 students (aged 18–26) from open access faculties of a Moroccan public university to identify the factors leading to an increase in university dropout rates. We discuss the impact of selected variables, and the findings show that several factors are related to university dropout rates, such as social background, psychological and health problems, insufficient motivation of professors, limited perspective on educational programs, changes in language and teaching methodologies, absenteeism, student attitude, and a lack of interaction between professors and students. Full article