Search Results (27,297)

Large Language Models (LLMs) like BERT have gained significant prominence due to their remarkable performance in various natural language processing tasks. However, they come with substantial computational and memory costs. Additionally, they are essentially black-box models, being challenging to explain and interpret. In this article, Persistent BERT Compression and Explainability (PBCE) is proposed, a Green AI methodology to prune BERT models using persistent homology, aiming to measure the importance of each neuron by studying the topological characteristics of their outputs. As a result, PBCE can compress BERT significantly by reducing the number of parameters (47% of the original parameters for BERT Base, 42% for BERT Large). The proposed methodology has been evaluated on the standard GLUE Benchmark, comparing the results with state-of-the-art techniques achieving outstanding results. Consequently, PBCE can simplify the BERT model by providing explainability to its neurons and reducing the model’s size, making it more suitable for deployment on resource-constrained devices. Full article

With the advancement of federated learning (FL), there is a growing demand for schemes that support multi-task learning on multi-modal data while ensuring robust privacy protection, especially in applications like intelligent connected vehicles. Traditional FL schemes often struggle with the complexities introduced by multi-modal data and diverse task requirements, such as increased communication overhead and computational burdens. In this paper, we propose a novel privacy-preserving scheme for multi-task federated split learning across multi-modal data (MTFSLaMM). Our approach leverages the principles of split learning to partition models between clients and servers, employing a modular design that reduces computational demands on resource-constrained clients. To ensure data privacy, we integrate differential privacy to protect intermediate data and employ homomorphic encryption to safeguard client models. Additionally, our scheme employs an optimized attention mechanism guided by mutual information to achieve efficient multi-modal data fusion, maximizing information integration while minimizing computational overhead and preventing overfitting. Experimental results demonstrate the effectiveness of the proposed scheme in addressing the challenges of multi-modal data and multi-task learning while offering robust privacy protection, with MTFSLaMM achieving a 15.3% improvement in BLEU-4 and an 11.8% improvement in CIDEr scores compared with the baseline. Full article

The primary objective of this study is to validate the Sustainability Consciousness Questionnaire (SCQ) for the Portuguese population, ensuring its reliability and applicability across the dimensions of knowledge, attitudes, and behaviours related to sustainability. This validation is crucial for ensuring the SCQ captures local cultural nuances and provides reliable data to inform educational and policy strategies for promoting sustainability. To achieve this goal, a quantitative methodology was adopted, involving the translation and cultural adaptation of the SCQ into Portuguese. Data were collected from a convenience sample of 630 participants, aged 17 to 83, using an online platform. Ethical procedures were rigorously followed, including obtaining informed consent from all participants and ensuring data confidentiality. The factor structure of the SCQ was analysed using structural equation modelling (SEM). The analysis confirmed a three-dimensional factor structure aligned with the environmental, social, and economic pillars of sustainability, as well as significant correlations between these dimensions and real-world sustainable practices such as recycling and energy conservation. The results confirmed the construct validity of the SCQ, demonstrating robust reliability indicators across its scales and acceptable model fit indices (CFI = 0.860; TLI = 0.851; RMSEA = 0.045). These findings highlight the questionnaire’s utility as a measurement tool for sustainable consciousness in the Portuguese context. The SCQ provides a valuable resource for educators, policymakers, and researchers. For instance, educators can use the SCQ to identify gaps in students’ sustainability knowledge, policymakers can prioritise areas for intervention based on public attitudes, and researchers can explore relationships between awareness and sustainable behaviours to design effective programs. Furthermore, this study contributes to Sustainable Development Goal 4 (Quality Education) by enabling data-driven strategies to integrate sustainability education into curricula, fostering a deeper understanding of sustainable practices and behaviours essential for achieving global education goals. Full article

Building employee loyalty is a prerequisite for a company to achieve a competitive advantage, high organizational performance, and sustainability. The lack of voluntary leaves does not result in recruitment costs or reduced efficiency during the adaptation period of a new employee. It helps retain knowledge and experience within the organization. The article aims to explore employees’ loyalty in terms of voluntary employment continuity during the pandemic slowdown of COVID-19, when employee loyalty was put to an exceptional test, and identify the factors that have had the most significant impact. This empirical study was carried out for Germany, mainly due to the strength and position of the German economy in Europe and the availability of a large, detailed micro dataset necessary for in-depth econometric analyses. The dataset used in the survey is the fifth wave of the German Linked Personnel Panel—LPP in 2020/21 (N = 7397). A multinomial logit model was used as a research tool. Loyalty appears as an explained variable in four ordered logit models that differ in the set of explanatory variables. The explanatory variables include demographics, job title, working conditions, compensation and rewards, job content, training and career development, teamwork, and relationships with colleagues and superiors. The results confirm the influence of extra-organizational factors, such as age and living in a four- or five-person household, on employee loyalty. However, age seems to be a factor of decreasing importance. Too much complexity of work, manifested by great task variety, working in multiple teams, and the requirement to perform work remotely, harmed employee loyalty during the pandemic. Findings justify building loyalty based on sustainable human resource policies to increase income satisfaction, reasonable workload, competence development, and greater autonomy at work. It is also clear that leadership issues (fairness in contact with superiors and recognition for work) mattered during this challenging time and have a high potential to improve employee loyalty in the future. Full article

In today’s data-driven business landscape, effective and transparent decision making becomes relevant to maintain a competitive advantage over the competition, especially in customer service in B2B environments. This study presents a methodological framework that integrates Explainable Artificial Intelligence (XAI), C-means clustering, and the Analytic Hierarchical Process (AHP) to improve strategic decision making in business environments. The framework addresses the need to obtain interpretable information from predictions based on machine learning processes and the prioritization of key factors for decision making. C-means clustering enables flexible customer segmentation based on interaction patterns, while XAI provides transparency into model outputs, allowing support teams to understand the factors influencing each recommendation. The AHP is then applied to prioritize criteria within each customer segment, aligning support actions with organizational goals. Tested with real customer interaction data, this integrated approach proved effective in accurately segmenting customers, predicting support needs, and optimizing resource allocation. The combined use of XAI and the AHP ensures that business decisions are data-driven, interpretable, and aligned with the company’s strategic objectives, making this framework relevant for companies seeking to improve their customer service in complex B2B contexts. Future research will explore the application of the proposed model in different business processes. Full article

The COVID-19 pandemic offers an opportunity for the revitalisation of empty Spain and the development of new sustainable business models in a healthier environment, taking the competitive advantages of digitalisation and the benefits of contact with nature. This study presents a positive analysis of the situation after three months of confinement with the research objective of evaluating the potential for development a sustainable social economy in empty Spain based on the hypotheses presented. In order to demonstrate the six hypotheses put forward in the research, a review of the existing literature was conducted, socio-economic and environmental indicators from official sources were consulted, and descriptive statistics methods have been applied. Digitalisation, the social economy, the bio-economy, and the revitalisation of heritage seem to be the drivers for achieving the challenges proposed. By perceiving reality through a lens that values nature and creative intelligence, a new avenue of opportunities may be opened up, leading to an improvement in quality of life and well-being, and potentially retaining the rural population. Following this study, which assesses the opportunities, risks, and challenges and establishes a plan of measures, players, and resources for future implementation in vacant Spain, new lines of work will become available. Full article

Integrating new generation and storage resources within power systems is challenging because of the stochastic nature of renewable generation, voltage regulation, and the use of microgrids. Classical optimization methods struggle with these nonlinear, multifaceted issues. This paper presents a novel optimization framework for integrating, sizing, and siting distributed renewable generation and energy storage systems in power distribution networks. To accurately reflect load variability, the framework considers four distinct load models—constant impedance, current, power, and ZIP (constant impedance, constant current, constant power). Our approach utilized three metaheuristic approaches to enhance the efficiency of power system management. The validation results on the IEEE 33 Bus System conclude that the Elephant Herding Optimization (EHO) emerged as the best performer regarding voltage stability and real power loss reduction with a voltage stability index of 0.0031346. Modified Ant Lion Optimization (ALO) achieved a best voltage stability index of 0.0024115 and power losses of 7.5092 MVA. The Red Colobus Monkey Optimization (RMO) algorithm realized a voltage stability index of 0.0052053 and real power losses of 20.7564 MVA. Overall, the results conclude that ALO is the most effective approach for optimizing distributed renewable energy systems under different climatic conditions. According to the analysis, the algorithm works best in ideal circumstances when the percentages of wind and irradiance are 60% or greater. Full article

To address the problems of virgin plastic production from fossil resources and the growing amount of plastic waste, a rapid transition to a circular economy is being pursued. The separation of mixed plastics into pure fractions is of paramount importance for promoting recycling and preventing downcycling. In this study, experimental parameters were determined for the selective bulk dissolution of polyamide 6 (PA 6) filaments (1.75 mm diameter, 1 cm length) in CaCl₂-EtOH-H₂O mixtures (CEW) at 75 °C. These parameters included the energy supply mode, dissolution time, CEW composition and CEW:PA mass ratio. Compared with energy supply by microwaves, energy supply by ultrasound improved the yield of dissolved and recovered PA 6 after 5 h from 31% to 52%. In total, the yield of PA 6 after 3 h of bulk dissolution increased from 18% to 69% when the energy supply mode was changed from microwave to ultrasound and the H₂O:EtOH molar ratio of CEW was increased from 0.40 to 1.33 while maintaining an optimal CEW:PA mass ratio of 8.5. Additionally, master plot analysis suggested that dissolution under microwave energy supply followed a contracting cylinder model, whereas dissolution under ultrasonic energy supply aligned with a 2D diffusion or third-order kinetic model. Microscopic observations suggested that, in the case of ultrasonic energy supply, oscillating bubbles on the particle surface enhanced the dissolution rate of PA 6 filaments in CEW. Full article

Lead (Pb) is a highly toxic heavy metal that causes significant health hazards and environmental damage. Thus, the detection and removal of Pb²⁺ ions in freshwater sources are imperative for safeguarding public health and the environment. Moreover, the transformation of single resources into multiple high-value products is vital for achieving sustainable development goals (SDGs). In this regard, the present work focused on the preparation of two efficient materials, i.e., biochar (R-BC) and carbon dots (R-CDs) from a single resource (rice straw), via a novel approach by using extraction and hydrothermal process. The various microscopic and spectroscopy techniques confirmed the formation of porous structure and spherical morphology of R-BC and R-CDs, respectively. FTIR analysis confirmed the presence of hydroxyl (–OH), carboxyl (–COO) and amine (N–H) groups on the R-CDs’ surface. The obtained blue luminescent R-CDs were employed as chemosensors for the detection of Pb²⁺ ions. The sensor exhibited a strong linear correlation over a concentration range of 1 µM to 100 µM, with a limit of detection (LOD) of 0.11 µM. Furthermore, the BET analysis of R-BC indicated a surface area of 1.71 m²/g and a monolayer volume of 0.0081 cm³/g, supporting its adsorption potential for Pb²⁺. The R-BC showed excellent removal efficiency of 77.61%. The adsorption process followed the Langmuir isotherm model and second-order kinetics. Therefore, the dual use of rice straw-derived provides a cost-effective, environmentally friendly solution for Pb²⁺ detection and remediation to accomplish the SDGs. Full article

Due to climate, resource endowment, planting habits, policy publicity, subsidies, and constraints, there have been many problems in the utilization of straw resources in the cold, main grain-producing areas in northern China. Based on the theory of value perception, an analytical framework was constructed, and the ordered logistic model was used to form an empirical analysis of the questionnaire data of more than 60 townships in 7 cities of Heilongjiang Province, trying to analyze the problems existing in the utilization of straw resources. The results show that the external factors include policy subsidy, policy punishment, and transportation convenience. Among the internal reasons, farmers’ personal characteristics, production habits, and perception of technical effectiveness and convenience have a significant impact on the application of straw resource utilization technology. Therefore, improving the intensity and precision of subsidies, strengthening the intensity of punishment, improving the popularization of technology, strengthening the ideological education of farmers to clarify the ecological value of straw resource utilization, and strengthening the construction of infrastructure to improve the convenience of transportation are effective means to promote straw resource utilization technology and promote the green transformation of agriculture. Full article

Fairness in competitive games such as the Ultimatum Game is often defined theoretically. According to some of the literature, in which fairness is determined only based on resource allocation, a proposal splitting resources evenly (i.e., 5:5) is generally assumed as fair, and minimal deviation (i.e., 4:6) is considered enough to classify the proposal as unfair. Relying on multinomial processing tree models (MPTs), we investigated where the boundaries of fairness are located in the eye of responders, and pit fairness against relative and absolute gain maximization principles. The MPT models we developed and validated allowed us to separate three individual processes driving responses in the standard and Third-Party Ultimatum Game. The results show that, from the responder’s perspective, the boundaries of fairness encompass proposals splitting resources in a perfectly even way and include uneven proposals with minimal deviance (4:6 and 6:4). Moreover, the results show that, in the context of Third-Party Ultimatum Games, the responder must not be indifferent between favoring the proposer and the receiver, demonstrating a boundary condition of the developed model. If the responder is perfectly indifferent, absolute and relative gain maximization are theoretically unidentifiable. This theoretical and practical constraint limits the scope of our theory, which does not apply in the case of a perfectly indifferent decision-maker. Full article

Solving the university timetabling problem is crucial as it ensures efficient use of resources, minimises scheduling conflicts, and enhances overall productivity. This paper presents a comprehensive review of university timetabling problems using integer programming algorithms. This study explores various integer programming techniques and their effectiveness in optimising complex scheduling requirements in higher education institutions. We analysed 95 integer programming-based models developed for solving university timetabling problems, covering relevant research from 1990 to 2023. The goal is to provide insights into the evolution of these algorithms and their impact on improving university scheduling. We identify that the implementation rate of models using integer programming is 98%, which is much higher than 34% implementation rates using meta-heuristics algorithms from the existing review. The integer programming models are analysed by the problem types, solutions, tools, and datasets. For three types of timetabling problems including course timetabling, class timetabling, and exam timetabling, we dive deeper into the commercial solvers CPLEX (47), Gurobi (11), Lingo (5), Open Solver (4), C++ GLPK (4), AIMMS (2), GAMS (2), XPRESS (2), CELCAT (1), AMPL (1), and Google OR-Tools CP-SAT (1) and identify that CPLEX is the most frequently used integer programming solver. We explored the uses of machine learning algorithms and the hybrid solutions of combining the integer programming and machine learning algorithms in higher education timetabling solutions. We also identify areas for future work, which includes an emphasis on using integer programming algorithms in other industrial areas, and using machine learning models for university timetabling to allow data-driven solutions. Full article

The multi-armed bandit (MAB) problem is a foundational model for sequential decision-making under uncertainty. While MAB has proven valuable in applications such as clinical trials and online advertising, traditional formulations have limitations; specifically, they struggle to handle three key real-world scenarios: (1) when decisions must follow a hierarchical structure (as in autonomous systems where high-level strategy guides low-level actions); (2) when there are constraints at multiple levels of decision-making (such as both system-wide and component-level resource limits); and (3) when available actions depend on previous choices or context. To address these challenges, we introduce the hierarchical constrained bandits (HCB) framework, which extends contextual bandits to incorporate both hierarchical decisions and multilevel constraints. We propose the HC-UCB (hierarchical constrained upper confidence bound) algorithm to solve the HCB problem. The algorithm uses confidence bounds within a hierarchical setting to balance exploration and exploitation while respecting constraints at all levels. Our theoretical analysis establishes that HC-UCB achieves sublinear regret, guarantees constraint satisfaction at all hierarchical levels, and is near-optimal in terms of achievable performance. Simple experimental results demonstrate the algorithm’s effectiveness in balancing reward maximization with constraint satisfaction. Full article

Wind energy has emerged as a viable alternative to fossil fuels due to its clean and cost-effective nature. Pakistan, facing growing energy demands and the imperative to reduce carbon emissions, has invested significantly in wind power to supply electric power in rural and urban communities, particularly in the Thatta district of Sindh Province of Pakistan. However, the sustainability of wind energy generation is contingent upon consistent and sufficient wind resources. This study examines the wind potential of Thatta district from 2004 to 2023 to assess its suitability for large-scale wind power development. To evaluate the wind potential of Thatta district, seasonal wind speed and direction data were collected and analyzed. Wind shear at different heights was determined using the power law, and wind potential maps were generated using GIS interpolation techniques. Betz’s law was employed to assess wind turbine power density. Box–Jenkins ARIMA and SARIMA models were applied to predict future wind patterns. This study revealed that Thatta district experienced sufficient wind speeds during the study period, with averages of 9.7 m/s, 7.6 m/s, 7.4 m/s, and 4.8 m/s for summer, autumn, spring, and winter, respectively. However, a concerning trend of decreasing wind speeds has been observed since 2009. The most significant reductions occurred in summer, coinciding with Pakistan’s peak electricity demand. While Thatta district has historically demonstrated potential for wind energy, the declining wind speeds pose a challenge to the sustainability of wind power projects. Further research is necessary to identify the causes of this trend and to explore mitigation strategies. Full article

Fertilizer carbon emissions contribute the largest proportion to agricultural carbon emissions in China, while the extension of low-carbon fertilization technologies (LCFTs) is an effective measure to address this issue. Research suggests that the relational networks surrounding farmers significantly influence their carbon reduction behavior. This study conducted a field survey of 239 large-scale grain producers in August 2022 on China’s Poyang Lake Basin, which is the nation’s largest freshwater lake and a vital agricultural production area. Using cross-sectional data, probit and ordered probit models were employed to analyze the impacts of multidimensional relational networks (market, government, and social networks) on the adoption of LCFTs by large-scale grain producers. Additionally, a mediating-effect model was used to examine the pathways through which relational networks influence LCFT adoption. The findings indicated that relational networks not only increased the likelihood of large-scale grain producers adopting LCFTs but also enhanced the intensity of adoption. However, the effects of different relational networks on low-carbon behavior varied. The market network exerted the most prominent influence on LCFT adoption, followed by the social and government networks. A mediation analysis identified information sharing, demonstration effects, and resource guarantees as the mediating pathways between multidimensional relational networks and LCFT adoption by large-scale grain producers. Furthermore, a heterogeneity analysis revealed that the effects of multidimensional relational networks on LCFT adoption differed across generations and carbon intensity levels. The impact was greater among older grain producers than the younger generation, and those in the high-carbon-intensity group exhibited a stronger incentive compared to the medium- and low-carbon-intensity groups. Full article