Sovereign large language models (LLMs), emerging as strategic assets in global information ecosystems, represent advanced AI system developed under distinct national governance regimes. This study examines how model origin and governance context influence AI-generated narratives on international territorial disputes. The study compares outputs from three prominent sovereign LLMs - OpenAI’s GPT-4o (United States), DeepSeek-R1 (China), and Mistral (European Union), across 12 high-profile territorial conflicts. Statistically significant differences in each model's sentiment distribution and geopolitical framing are identified using a mixed-methods approach that combines sentiment analysis with statistical evaluation (chi-square tests and analysis of variance, ANOVA) on responses to 300 standardized prompts.

The findings indicate model provenance substantially shapes the tone and stance of outputs, with each LLM reflecting distinct biases aligned with its national context. These disparities carry important policy and societal implications: reliance on a single sovereign model could inadvertently bias public discourse and decision-making toward that model's native perspective. The study highlights ethical considerations such as transparency and fairness and calls for robust governance frameworks. It underscores the need for careful oversight and international cooperation to ensure that sovereign LLMs are deployed in a manner that supports informed and balanced geopolitical dialogue.

Each employee has a number of workloads in the form of tasks and responsibilities that must be completed within a certain period of time. There are several aspects that are assessed in performance evaluations, such as the number of working hours per week, the number of projects handled, the number of overtime hours to complete work, the number of sick days, the number of team members, the number of hours to develop self-skills, job promotion offers, etc. All of these aspects certainly impact performance scores, employee satisfaction scores, and the ability to survive. Excessive workload will have a negative impact on physical and mental health, performance, and employee satisfaction levels. This study aims to analyze the results of employee performance evaluations based on workload factors using machine learning approaches such as linear regression and random forest. The computational results will be used to compare the effectiveness of machine learning models and analyze the accuracy of the assessment results. The significance of this study lies in its potential to enhance employee performance management systems by providing accurate, data-driven insights for decision-making processes such as promotions, compensation, and workforce planning. Practical and fair employee performance assessments will enable decision-makers to make informed choices regarding job promotions, salary increases, annual bonuses, and employee career development.

Soiling, defined as the accumulation of dirt, dust, and other particles on the surface of photovoltaic (PV) panels, is a significant issue that substantially impacts solar panel efficiency and performance. This accumulation leads to energy losses and decreased electricity output. Numerous research papers have proposed various systems to address this issue. This paper provides a comprehensive review of recent publications on soiling detection in solar panels. The review methodology includes literature retrieval, screening, content analysis, and bibliometric analysis, utilizing the Scopus database to compile a final selection of 75 papers. This review identifies gaps in previous research, such as the need for more robust and cost-effective detection systems and the integration of emerging technologies like artificial intelligence and remote sensing. Key findings highlight that deep learning models and advanced sensor technologies show promising results in improving soiling detection accuracy. The review also suggests potential areas for future work, emphasizing the development of innovative inspection tools, models, and cleaning systems that can enhance efficiency and reduce operational costs.

The Musi River in Palembang City plays an important role in meeting human needs, such as food, drink, and clean water. The high level of activity of the community around the river has led to an increase in macroplastic (MP) waste pollution. This study aims to (1) determine the condition of macroplastic waste pollution in the Musi River; (2) determine sustainable macroplastic waste management strategies; and (3) sustainable plastic waste reduction participation. The method uses a qualitative description of semi-structured interviews. Data collection used documentation and interviews. Data analysis used descriptive to describe the current condition and analysed using SWOT. The research results showed that the source of macroplastic waste pollution came from community activities such as domestic activities, industry, markets and tourism. The strategy for capturing macroplastic waste involved placing waste containment devices (trash booms and waste containment fences) in the streams of densely populated tributaries. To reduce the amount of macroplastic waste entering the environment, all parties, including the government, the community, and non-governmental organisations, must participate and work together. The implication of this research is that preventing macroplastic waste from entering the sea and effective management can reduce the threat to shrinking endemic river populations.

The increasing worldwide energy requirements, combined with sustainable urban growth, drive the need for inventive building technologies. Building integrated photovoltaic and thermal systems (BIPV/T) generate both electricity and thermal energy while enabling nearly zero-energy buildings (NZEBs) to achieve their energy goals. Our analysis examines the technological, economic, and environmental aspects of BIPV/T systems and their application within building elements such as roofs, facades, and glazing areas. The review also examines supportive policy frameworks for BIPV/T implementation while pinpointing adoption barriers like steep initial investments and incomplete regional policies. The present review is based on a systematic literature review from January 2010 to March 2025 from the Scopus, Web of Science, and IEEE Xplore databases. A search string consisting of the combination of the keywords building integrated photovoltaic and thermal systems (BIPV/T), nearly zero-energy buildings (NZEBs), solar technologies, Aesthetic and Architectural Integration, Energy Efficiency, and Building codes. A total of 75 articles were selected after screening and eligibility assessment. This study seeks to provide guidance for researchers, architects, and policymakers to progress BIPV/T integration towards sustainable urban development.

The study investigates the concentration of chemical elements in biological tissues (placenta and blood) of women from the Akmola region, Kazakhstan to assess the impact of environmental pollution on maternal and newborn health. The research conducted from 2018 to 2021 involved 67 placental and umbilical cord blood samples collected from women in four Akmola districts. The study utilized instrumental neutron activation analysis and electronic microscopy to determine the concentration of 28 chemical elements. Statistical methods were applied to analyze the distribution, including the mean values, standard deviations, and frequency distribution curves. Significant variability in chemical element concentrations was observed across samples, with notable differences in rare earth elements and heavy metals. Elements such as sodium (Na), calcium (Ca), and chromium (Cr) displayed high variation. The study identified a strong environmental influence on the accumulation of toxic elements in the placenta and blood. The accumulation of chemical elements in biological tissues was heterogeneous, influenced by natural and anthropogenic factors. Blood was found to be more sensitive to environmental contamination compared to the placenta, indicating the need for enhanced environmental health monitoring in the region.

This study evaluates the operational cost efficiency and environmental implications of transitioning from diesel to electric buses, using the Trans Jogja public transport system in Indonesia as a case study. Employing a total cost of ownership (TCO) framework and emissions analysis, the study compares the financial performance and greenhouse gas (GHG) emissions between diesel and battery electric buses. Results show that electric buses incur approximately 50% higher operating costs, primarily due to elevated capital expenditures and depreciation. Moreover, under Indonesia's coal-dominated electricity grid, electric buses generate higher indirect CO₂ emissions than their diesel ones, highlighting a critical energy-emission paradox. However, electric buses eliminate tailpipe pollutants such as NO_x and PM2.5, offering considerable public health benefits. A systemic scenario analysis reveals that full fleet electrification without concurrent reform in the energy sector could raise annual emissions by over 2,200 tons. The study identifies key barriers—including high upfront costs, limited charging infrastructure, and regulatory misalignment—and proposes phased policy interventions. Recommendations include targeted subsidies, contract revisions, integration with renewable energy, and technical capacity-building. The findings offer valuable insights for Indonesian cities seeking to scale sustainable urban mobility through electric transportation.

The freight forwarding industry serves as a crucial bridge between importers, exporters, and shipping and transportation companies. By facilitating the smooth movement of goods across borders, freight forwarders play a vital role in global trade. However, this industry also significantly impacts environmental stability due to the emission of harmful gases, carbon footprints, waste generation, and improper disposal practices, such as dumping waste into the soil. These activities contribute to environmental degradation and pose serious threats to natural ecosystems. Therefore, it is essential for the freight forwarding industry to adopt green initiatives and sustainable practices to minimize its environmental impact and promote long-term ecological balance. This study attempts exploratory research on green logistics practices and the challenges of their implementation in the case of freight forwarding industry in Hyderabad, India. Using primary research with 150 employees in freight forwarding companies, the paper explores the levels of awareness and adoption, as well as challenges to green logistics management. The study tested the following five hypotheses: educational gaps, economic barriers, customer demand, industry structure, and heterogeneity. Using convenience random sampling and quantitative data analysis, the results show that employees have considerable gaps in education and awareness, as only 28.0% of employees are also very familiar with green logistics concepts. The major barriers inhibiting the widescale adoption included high upfront costs (74.7%), education and awareness challenges (65.3%), customer expectations for competitive pricing (62.7%), and extended installation time (60.0%) All five hypotheses were confirmed with chi-square statistics from 19.76 to 45.72 (p<0.05). We highlight that the diversity of company sizes within the freight forwarding industry results in a spectrum of behavior when it comes to adopting green practices. Micro-level enterprises are facing much more significant challenges (58.0%) than higher-level firms at these conditions, coupled with highly uneven resource distribution (60.7%) Despite these barriers, the majority of respondents acknowledge the significance of green logistics concerning his/her company for operational efficiency (93.3%) and competitive advantage (86.7%). The results highlight a vital relationship where comprehensive education programs, targeted financial support and collaborative efforts from stakeholders can help highlight the more sustainable environmental approach to this activity.

This study assessed the effectiveness and sustainability of using barge versus feeder vessels to transport containerized cargo to Bangkok Port, Thailand. A survey of 387 stakeholders in marine logistics was conducted from October to December 2024. Multiple regression analysis (MRA) showed that cost-effectiveness, environmental impact, and operational flexibility primarily influenced transport mode choice, explaining 56.2% of the variance. Cost-effectiveness emerged as the key factor, while environmental impact was the strongest predictor of perceived sustainability. While operators favored feeders due to cost and time efficiency, barges scored higher due to environmental friendliness and operational flexibility. Notably, 68% of respondents preferred barges for short routes under 100 km due to their role in reducing road congestion and pollution. Furthermore, 73% expected greater barge use over the next five years, driven by technology and environmental policies. Improved waterway infrastructure would lead 82% to use barges more frequently, and 76% believed better intermodal integration would enhance logistics efficiency. This study is limited to the context of Thailand’s domestic maritime logistics and stakeholder perceptions, which may not be fully generalizable to other ASEAN or global port systems. Future research should explore multi-country comparative studies and assess longitudinal trends as green port policies evolve across Southeast Asia.

In this work, we introduce a hybrid method that combines Long Short-Term Memory (LSTM) neural networks with Taylor Series Expansion (TSE) to solve high-dimensional Fredholm Integral Equations of the second kind (SFIEs). Specifically, we focus on systems with up to 10000 dimensions, which are common in fields like fluid dynamics, electromagnetics, and quantum mechanics. Traditional methods for solving these equations, such as discretization, collocation, and iterative solvers, face significant challenges in high-dimensional spaces due to their computational cost and slow convergence. LSTM networks approximate the solution functions, and Taylor Series Expansion refines the approximation, ensuring higher accuracy and computational efficiency. Numerical experiments demonstrate that the hybrid method significantly outperforms traditional approaches in both accuracy and stability. This method provides a promising approach to solving complex high-dimensional integral equations efficiently in scientific and engineering applications.

Wireless propagation is a crucial technology in modern advancements, requiring highly accurate prediction. Path loss propagation is influenced by various parameters that must be accounted for to predict the signal route over the entire distance and refine breakpoint models with precise interference calculations. The breakpoint distance is defined as the point separating two distinct trends of path loss, each following a different path loss exponent. This paper reviews the Fresnel, Perera, and True breakpoints in a dual-slope model reference at 2 GHz, using a fixed exponent of n₁ = 2 before the breakpoint and n₂ = 4 after. It then proposes a distance-adaptive exponent model that considers a steady path by incorporating a flexible exponent based on environmental factors, mitigating the abrupt change in path loss exponent at breakpoints observed in the dual-slope model, which leads to discontinuities. The comparison results under similar conditions demonstrate that both models perform similarly over short distances of up to 100 meters, while the dual-slope model is more suitable for distances of up to 1 km. However, due to its stability and consistency, the distance-adaptive exponent model is more appropriate for longer distances. Validation using RMSE, followed by comparative analysis, confirms that our model offers higher stability in interference scenarios. These findings will assist researchers and wireless designers in predicting and selecting the most accurate and effective propagation model.