Search Results (817)

Accurate and interpretable prediction of student academic performance is a cornerstone of data-driven educational support systems, enabling timely interventions, personalized learning pathways, and equitable resource allocation. While ensemble machine learning models such as Random Forest, Extra Trees, and CatBoost have shown promise in educational data mining, their predictive power and generalizability are often limited by suboptimal weighting schemes and sensitivity to hyperparameter configurations. To address this, we propose a Grey Wolf Optimizer (GWO)-guided ensemble framework that dynamically optimizes each base regressor’s contribution to minimize prediction error while preserving model transparency. Evaluated on a real-world student performance dataset, the proposed approach achieves a coefficient of determination (R²) of 0.93, significantly outperforming individual and conventional ensemble baselines. Furthermore, we integrate SHAP (SHapley Additive exPlanations) to provide educator-friendly interpretability, revealing that daily study hours, study effectiveness, lifestyle score, and screen time are the most influential predictors of exam outcomes. By bridging an optimized machine learning model with educational analytics, this work delivers a robust, transparent, and high-performing AI solution tailored for intelligent tutoring systems, early-warning platforms, and adaptive learning environments. The methodology exemplifies how nature-inspired optimization can enhance not only accuracy but also actionable insight for stakeholders in smart education ecosystems. Full article

This study explores the dynamic relationship between industrial robots, education, and environmental sustainability in the United States, emphasizing their role in reducing CO₂ emissions. The research aims to quantify how automation, human capital, and the energy transition contribute to carbon mitigation within a data-driven, AI-oriented policy framework. Quarterly data spanning 2011Q1–2024Q4 were analyzed using the advanced Quantile-on-Quantile Autoregressive Distributed Lag (QQARDL) model, which captures heterogeneous long- and short-run effects across emission distributions. Results reveal that industrial robot adoption, education, and renewable energy transition significantly reduce emissions, with the strongest effects occurring at both high- and low-emission quantiles. Economic growth and financial development also support decarbonization when complemented by green finance and innovation, while urbanization increases emissions unless aligned with compact urban design and clean energy systems. The findings imply that AI-driven industrial robotics and education jointly foster sustainability through efficiency, innovation, and awareness. Policymakers are encouraged to integrate automation strategies, renewable energy incentives, and sustainability education into climate policy. This study provides empirical evidence supporting the Resource-Based View, highlighting human capital and intelligent automation as strategic assets for achieving long-term carbon neutrality. Full article

The integration of digital technologies in mathematics is becoming increasingly significant, particularly in promoting conceptual understanding and student engagement. This study systematically reviews the literature on applications of Computer Algebra Systems, Artificial Intelligence, Visualisation Tools, augmented-reality technologies, Statistical Software, game-based learning and cloud-based learning in higher education mathematics. This meta-analysis synthesises findings from 88 empirical studies conducted between 1990 and 2025 to evaluate the impact of these technologies. The included studies encompass diverse geographical regions, providing a comprehensive global perspective on the integration of digital technologies in higher mathematics education. Using the PRISMA framework and quantitative effect size calculations, the results indicate that all interventions had a statistically significant impact on student performance. Among them, Visualisation Tools demonstrated the highest average percentage improvement in academic performance (39%), whereas cloud-based learning and game-based approaches, while beneficial, showed comparatively modest gains. The findings highlight the effectiveness of an interactive environment in fostering a deeper understanding of mathematical concepts. This study provides insights for educators and policymakers seeking to improve the quality and equity of mathematics education in the digital era. Full article

This study examined stakeholders’ satisfaction with current student classification and progress monitoring systems and explored their perceptions of the potential role of artificial intelligence (AI) in enhancing these processes. A cross-sectional survey was administered to 313 stakeholders, including teachers, administrators, decision-makers, and educational service providers. Descriptive statistics, multiple regression analysis, and group comparisons were employed to examine satisfaction levels, predictors of satisfaction, and expectations regarding AI integration. Despite high satisfaction with the current systems (85%), nearly 80% of stakeholders rated AI integration as essential. The most frequently expected functions of an AI-enabled system were predicting student challenges (33.2%), generating detailed analyses and reports (32.9%), customizing individual learning plans (22.7%), and providing immediate feedback (11.2%). Anticipated challenges focused on acceptance and adaptation by teachers and students (40.9%) and concerns about privacy and system integration. Regression analysis revealed that perceptions of classification practices (β = 0.473, p < 0.001) were a stronger predictor of satisfaction than perceptions of progress monitoring practices (β = 0.315, p < 0.001). Demographic analyses revealed greater dissatisfaction among non-teaching staff, females, and mid-career professionals. The findings show that stakeholders are broadly satisfied with existing systems while simultaneously demanding AI-driven innovation, suggesting satisfaction reflects acceptance rather than alignment with stakeholders’ needs and expectations. Full article

This study comprehensively analyses how AI tools scaffold and share metacognitive processes, thereby facilitating students’ learning in STEM classrooms through a mixed-method research synthesis combining bibliometric analysis and systematic review. Using a convergent parallel mixed-methods design, the study draws on 135 peer-reviewed articles published between 2005 and 2025 to map publication trends, author and journal productivity, keyword patterns, and theoretical frameworks. Data were retrieved from Scopus and Web of Science using structured Boolean searches and analysed using Biblioshiny and VOSviewer. Guided by PRISMA 2020 protocols, 24 studies were selected for in-depth qualitative review. Findings show that while most research remains grounded in human-centred conceptualisations of metacognition, there are emerging indications of posthumanist framings, where AI systems are positioned as co-regulators of learning. Tools like learning analytics, intelligent tutoring systems, and generative AI platforms have shifted the discourse from individual reflection to system-level regulation and distributed cognition. The study is anchored in Flavell’s theory of metacognition, General Systems Theory, and posthumanist perspectives to interpret this evolution. Educational implications highlight the need to reconceptualise pedagogical roles, integrate AI literacy in teacher preparation, and prioritise ethical, reflective AI design. The review provides a structured synthesis of theoretical, empirical, and conceptual trends, offering insights into how human–machine collaboration is reshaping learning by scaffolding and co-regulating students’ metacognitive development in STEM education. Full article

Student classroom behavior recognition is a core research direction in intelligent education systems. Real-time analysis of students’ learning states and behavioral features through classroom monitoring provides quantitative support for teaching evaluation, classroom management, and personalized instruction, offering significant value for data-driven educational decision-making. To address the issues of low detection accuracy and severe occlusion in classroom behavior detection, this article proposes an improved YOLOv11n-based algorithm named YOLOv11-GLIDE. The model introduces a Channel Prior Convolutional Attention (CPCA) mechanism to integrate global and local feature information, enhancing feature extraction and detection performance. A scale-based dynamic loss (SD Loss) is designed to adaptively adjust the loss weights according to object scale, improving regression stability and detection accuracy. In addition, Sparse Depthwise Convolution (SPD-Conv) replaces traditional down-sampling to reduce fine-grained feature loss and computational cost. Experimental results on the SCB-Dataset3 demonstrate that YOLOv11-GLIDE achieves an excellent balance between accuracy and lightweight design. Compared with the baseline YOLOv11n, mAP@0.5 and mAP@0.5-0.95 increase by 2.5% and 7.6%, while Parameters and GFLOPS are reduced by 9.4% and 11.1%, respectively. The detection speed reaches 127.9 FPS, meeting the practical requirements of embedded classroom monitoring systems for accurate and efficient student behavior recognition. Full article

This paper presents an Explainable Artificial Intelligence (XAI) framework for the automated generation of lesson plans aligned with Bloom’s Taxonomy. The proposed system addresses the dual challenges of accurate cognitive classification and pedagogical transparency by integrating a multi-task transformer-based classifier with a taxonomy-conditioned content generation module. Drawing from a locally curated dataset of 3000 annotated lesson objectives, the model predicts both cognitive process levels and knowledge dimensions using attention-enhanced representations, while offering token-level explanations via SHAP to support interpretability. A GPT-based generator leverages these predictions to produce instructional activities and assessments tailored to the taxonomy level, enabling educators to scaffold learning effectively. Empirical evaluations demonstrate strong classification performance (F1-score of 91.8%), high pedagogical alignment in generated content (mean expert rating: 4.43/5), and robust user trust in the system’s explanatory outputs. The framework is designed with a feedback loop for continuous fine-tuning and incorporates an educator-facing interface conceptually developed for practical deployment. This study advances the integration of trustworthy AI into curriculum design by promoting instructional quality and human-in-the-loop explainability within a theoretically grounded implementation. Full article

We introduce a novel, privacy-preserving AI authoring copilot designed for educational content creation, which uniquely combines a Markov Decision Process (MDP) as a reinforcement learning orchestrator with a locally deployed Qwen3-1.7B-ONNX large language model to iteratively refine text for clarity, unity, and engagement—all running on a modest CPU-only system (Intel i7, 16 GB RAM). Unlike cloud-dependent models, our agent treats writing as a sequential decision problem, selecting refinement actions (e.g., simplification, elaboration) based on real-time LLM and sentiment feedback, ensuring pedagogically sound outputs without internet dependency. Evaluated across five diverse topics, our MDP-orchestrated agent achieved an overall average quality score of 4.23 (on a 0–5 scale), statistically equivalent to leading cloud-based LLMs like ChatGPT and DeepSeek. This performance was validated through blind evaluations by four independent LLMs and human raters, supported by statistical consistency analysis. Our work demonstrates that lightweight local LLMs, when guided by principled MDP policies, can deliver high-quality, context-aware educational content, bridging the gap between powerful AI generation and ethical, on-device deployment. This advancement empowers educators, researchers, and curriculum designers with a trustworthy, accessible tool for intelligent content augmentation aligning with the Quality Education Sustainable Development Goal through innovations in educational technology, inclusive education, equity in education, and lifelong learning. Full article

Undergraduate life science education faces high attrition rates, especially among students from underrepresented groups. These disparities are often linked to differences in prior knowledge, self-efficacy, and interest, which are rarely addressed in traditional lecture-based instruction. This work explores the use of machine learning-based Intelligent Tutoring Systems (ITSs) to support personalized instruction in biology education by examining stochasticity in molecular systems. Accordingly, we developed and validated a Random Forest classification model and used it to assign instructional materials based on students’ prior knowledge and interests. We then applied the model in an introductory biology classroom and individually estimated the most promising instructional format. Results show that the most effective instruction can be reliably predicted from student performance and interest profiles, and model-based assignments may help reduce pre-existing opportunity gaps. Thus, machine-learning-driven instruction holds promise for enhancing equity in life science education by aligning materials with students’ needs, potentially reducing differences in achievement, self-efficacy, and cognitive load, which might be relevant to promoting underrepresented students. To facilitate a straightforward implementation for educators facing similar challenges associated with teaching molecular stochasticity, we developed an open-access ITS tool and provided a scalable approach for developing similar personalized learning tools. Full article

Generative models can generate art within a single modality with high fidelity. However, translating a work of art from one domain to another (e.g., painting to music or poem to painting) in a meaningful way remains a longstanding, interdisciplinary challenge. We propose a novel approach combining a multi-agent system (MAS) architecture with an ontology-guided semantic representation to achieve cross-domain art translation while preserving the original artwork’s meaning and emotional impact. In our concept, specialized agents decompose the task: a Perception Agent extracts symbolic descriptors from the source artwork, a Translation Agent maps these descriptors using shared knowledge base, a Generator Agent creates the target-modality artwork, and a Curator Agent evaluates and refines the output for coherence and style alignment. This modular design, inspired by human creative workflows, allows complex artistic concepts (themes, moods, motifs) to carry over across modalities in a consistent and interpretable way. We implemented a prototype supporting translations between painting and poetry, leveraging state-of-the-art generative models. Preliminary results indicate that our ontology-driven MAS produces cross-domain translations that preserve key semantic elements and affective tone of the input, offering a new path toward explainable and controllable creative AI. Finally, we discuss a case study and potential applications from educational tools to synesthetic VR experiences and outline future research directions for enhancing the realm of intelligent agents. Full article

Indoor Air Quality (IAQ) in educational environments is a critical determinant of students’ health, well-being, and learning performance, with inadequate ventilation and pollutant accumulation consistently associated with respiratory symptoms, fatigue, and impaired cognitive outcomes. Conventional monitoring approaches—based on periodic inspections or subjective perception—provide only fragmented insights and often underestimate exposure risks. Artificial intelligence (AI) offers a transformative framework to overcome these limitations through sensor calibration, anomaly detection, pollutant forecasting, and the adaptive control of ventilation systems. This review critically synthesizes the state of AI applications for IAQ management in educational environments, drawing on twenty real-world case studies from North America, Europe, Asia, and Oceania. The evidence highlights methodological innovations ranging from decision tree models integrated into large-scale sensor networks in Boston to hybrid deep learning architectures in New Zealand, and regression-based calibration techniques applied in Greece. Collectively, these studies demonstrate that AI can substantially improve predictive accuracy, reduce pollutant exposure, and enable proactive, data-driven ventilation management. At the same time, cross-case comparisons reveal systemic challenges—including sensor reliability and calibration drift, high installation and maintenance costs, limited interoperability with legacy building management systems, and enduring concerns over privacy and trust. Addressing these barriers will be essential for moving beyond localized pilots. The review concludes that AI holds transformative potential to shift school IAQ management from reactive practices toward continuous, adaptive, and health-oriented strategies. Realizing this potential will require transparent, equitable, and cost-effective deployment, positioning AI not only as a technological solution but also as a public health and educational priority. Full article

Artificial Intelligence (AI) has brought a revolution in many areas, including the education sector. It has the potential to improve learning practices, innovate teaching, and accelerate the path towards personalized learning. This work introduces Reinforcement Learning (RL) methods to develop a personalized examination scheduling system at a university level. We use two widely established RL algorithms, Q-Learning and Proximal Policy Optimization (PPO), for the task of personalized exam scheduling. We consider several key points, including learning efficiency, the quality of the personalized educational path, adaptability to changes in student performance, scalability with increasing numbers of students and courses, and implementation complexity. Experimental results, based on case studies conducted within a single degree program at a university, demonstrate that, while Q-Learning offers simplicity and greater interpretability, PPO offers superior performance in handling the complex and stochastic nature of students’ learning trajectories. Experimental results, conducted on a dataset of 391 students and 5700 exam records from a single degree program, demonstrate that PPO achieved a 42.0% success rate in improving student scheduling compared to Q-Learning’s 26.3%, with particularly strong performance on problematic students (41.3% vs 18.0% improvement rate). The average delay reduction was 5.5 months per student with PPO versus 3.0 months with Q-Learning, highlighting the critical role of algorithmic design in shaping educational outcomes. This work contributes to the growing field of AI-based instructional support systems and offers practical guidance for the implementation of intelligent tutoring systems. Full article

Over the last two years, with the rapid development of artificial intelligence, Large Language Models (LLMs) have obtained significant attention from the academic sector, making their application in higher education attractive for students, managers, faculty, and stakeholders. We conducted a Systematic Literature Review on the adoption of LLMs in the higher education system to address persistent issues and promote critical thinking, teamwork, and problem-solving skills. Following the PRISMA 2020 protocol, a systematic search was conducted in the Web of Science Core Collection for studies published between 2023 and 2024. After a systematic search and filtering of 203 studies, we included 22 articles for further analysis. The findings show that LLMs can transform traditional teaching through active learning, align curricula with real-world demands, provide personalized feedback in large classes, and enhance assessment practices focused on applied problem-solving. Their effects are transversal, influencing multiple dimensions of higher education systems. Consequently, LLMs have the potential to improve educational equity, strengthen workforce readiness, and foster innovation across disciplines and institutions. This systematic review is registered in PROSPERO (2025 CRD420251165731). Full article

Computational thinking (CT) is recognized as a core competency for the 21st century, and its development is shaped by multiple factors, including students’ individual characteristics and their use of information and communication technology (ICT). Drawing on large-scale international data from the 2023 cycle of the International Computer and Information Literacy Study (ICILS), this study analyzes a sample of 81,871 Grade 8 students from 23 countries and one regional education system who completed the CT assessment. This study is the first to apply a predictive modeling framework that integrates two machine learning techniques to systematically identify and explain the key variables that predict CT and their nonlinear effects. The results reveal that various student-level predictors—such as educational expectations and the number of books at home—as well as ICT usage across different contexts, demonstrate significant nonlinear patterns in the model, including U-shaped, inverted U-shaped, and monotonic trends. Compared with traditional linear models, the SHapley Additive exPlanations (SHAP)-based approach facilitates the interpretation of the complex nonlinear effects that shape CT development. Methodologically, this study expands the integration of educational data mining and explainable artificial intelligence (XAI). Practically, it provides actionable insights for ICT-integrated instructional design and targeted educational interventions. Future research can incorporate longitudinal data to explore the developmental trajectories and causal mechanisms of students’ CT over time. Full article

Conventional innovation management methodologies (IMMs) often struggle to respond to the complexity, uncertainty, and cognitive diversity that characterise contemporary innovation projects. This study introduces Innovation Flow (IF), a human-centred and adaptive framework grounded in Flow Theory and enhanced by Generative Artificial Intelligence (GenAI). At its core, IF operationalises Personalised Innovation Techniques (PInnTs)—adaptive variations of established methods tailored to project genetics and team profiles, generated dynamically through a GenAI-based system. Unlike traditional IMMs that rely on static toolkits and expert facilitation, Innovation Flow (IF) introduces a dynamic, GenAI-enhanced system capable of tailoring techniques in real time to each project’s characteristics and team profile. This adaptive model achieved a 60% reduction in ideation and prototyping time while maintaining high creative performance and autonomy. IF thus bridges the gap between human-centred design and AI augmentation, providing a scalable, personalised, and more inclusive pathway for managing innovation. Using a mixed-methods design that combines grounded theory with quasi-experimental validation, the framework was tested in 28 innovation projects across healthcare, manufacturing, and education. Findings show that personalisation improves application fidelity, engagement, and resilience, with 87% of cases achieving high efficacy. GenAI integration accelerated ideation and prototyping by more than 60%, reduced dependence on expert facilitators, and broadened participation by lowering the expertise barrier. Qualitative analyses emphasised the continuing centrality of human agency, as the most effective teams critically adapted rather than passively adopted AI outputs. The research establishes IF as a scalable methodology that augments, rather than replaces, human creativity, accelerating innovation cycles while reinforcing motivation and autonomy. Full article