The rapid digitalisation of the construction and built environment sectors has increased demand for innovative teaching methods in higher education. Among emerging digital tools, Metaverse applications offer immersive and interactive learning environments with the potential to transform traditional pedagogical practices. This study evaluates the adoption of the metaverse application (EON-XR) in built environment education by examining educator training outcomes using the Kirkpatrick model. The evaluation focused on a structured training workshop designed to equip built environment educators with skills to integrate metaverse-based applications into first-year teaching modules. A post-training evaluation employed a Google Forms survey instrument explicitly aligned with Kirkpatrick’s four evaluation levels: reaction, learning, behaviour, and results. The instrument captured educators’ perceptions of usability, relevance, pedagogical value, behavioural intention to apply acquired skills, and the broader institutional and disciplinary implications of metaverse adoption. Both qualitative and quantitative data were analysed to assess the training programme’s effectiveness across these four dimensions. Findings indicate a generally positive reaction to the metaverse training, with educators recognising the relevance and user-friendliness of the EON-XR application. Evidence of learning was demonstrated by increased confidence and perceived competence in applying metaverse tools for teaching. At the behavioural level, participants expressed strong intentions to integrate metaverse applications into their teaching practice. At the results level, the training was perceived as enhancing teaching engagement, stimulating interest in construction digitalisation, and contributing to the institutional standing of higher education providers. The study concludes that metaverse-based educator training can play a critical role in advancing digital pedagogy within built environment education. By applying the Kirkpatrick model, this research provides a structured, replicable framework for evaluating the adoption of immersive technology in higher education contexts.

This chapter presents a Retrieval-Augmented Generation (RAG) framework to evaluate the originality of student projects in a Project-Based Learning (PBL) context within an advanced digital systems course of a Master’s degree in Electronic Engineering at the University of Seville . While PBL fosters creativity, technical competence, and active learning, assessing originality becomes increasingly complex as repositories of past projects grow and distinctions between inspiration, reuse, and plagiarism blur. Traditional plagiarism-detection tools focus on textual similarity and fail to capture conceptual novelty, especially in engineering contexts where code reuse is common.

The proposed system combines structured summarization using an open-weight large language model (DeepSeek-R1) with semantic retrieval over an indexed knowledge base of previous projects. Each historical project is automatically preprocessed and summarized into a structured format, including objectives, modules, hardware, and contextual keywords. These summaries are embedded and indexed using an RAG pipeline, enabling semantic comparison between new proposals and prior work. The system generates originality scores, identifies similar projects, and provides explanatory feedback.

Validation was conducted using 91 historical projects and 10 controlled test cases spanning the full originality spectrum. The system’s scores showed a strong correlation with expert evaluations (Pearson r = 0.87, p = 0.006), though it tended to be slightly more conservative. Qualitative feedback from instructors and students highlighted its usefulness as a cognitive support tool that enhances transparency, reduces memory bias, and promotes reflective refinement of project ideas. Rather than replacing human judgment, the framework functions as an explainable co-evaluation assistant, offering a scalable, locally deployable, and ethically grounded approach to originality assessment in engineering education.

Artificial Intelligence is increasingly embedded in contemporary military operations, yet its effective and ethical integration depends largely on how defense institutions educate and prepare personnel. While technological advancements receive significant attention, the pedagogical systems that enable responsible adoption remain underexamined. This study examines how AI can be systematically incorporated into professional military education and training for irregular warfare. Using qualitative doctrinal analysis and thematic synthesis of United States and NATO policy documents, scholarly literature, and recent operational practices, the research develops a five layer instructional framework linking strategic alignment, doctrinal integration, operational learning, data driven intelligence education, and human AI governance. The framework demonstrates how AI competencies can be embedded within curricula, training programs, simulations, and institutional learning systems in a structured and scalable manner. Findings indicate that successful AI adoption requires coordinated reforms in doctrine, instructional design, leadership development, and ethical oversight rather than isolated technical training initiatives. The study further emphasizes the importance of experiential learning environments, interdisciplinary collaboration, scenario based exercises, and continuous professional development in cultivating AI literacy and adaptive decision making among military personnel. By reframing AI integration as a comprehensive pedagogical transformation rather than a purely technical upgrade, this research contributes to education sciences by illustrating how emerging technologies reshape professional learning ecosystems. The proposed model offers practical guidance for educators, policymakers, and training institutions seeking to enhance operational readiness while maintaining transparency, accountability, and principled human judgment.

With the fast-growing development of artificial intelligence (AI), large language models (LLMs) have become increasingly vital in technology-enhanced education, particularly in language learning and translation training. Tools such as ChatGPT-4o and DeepSeek V2 are increasingly used by students and educators to support bilingual communication tasks. As these systems become part of everyday academic practice, questions arise about how reliably they support learning and how differences between models may shape students’ understanding of translated content.
This study presents a comparative evaluation of ChatGPT-4o and DeepSeek V2 in English-Chinese news translation. News texts are selected because they involve linguistic complexity, contextual nuance, and sensitive discourse features that challenge automated systems.
Guided by Skopos Theory, Critical Discourse Analysis, Translation Quality Assessment, and Media Framing Theory, this study adopts a mixed-methods approach combining quantitative evaluation metrics with qualitative methods. A selected corpus of recent English-language news texts across political, economic, technological, and cultural domains is translated by both models under standardized prompts. Quantitative metrics assess surface-level correspondence, while qualitative analysis examines accuracy, fluency, terminology use, tone, and handling of sensitive content.
By comparing the outputs of ChatGPT-4o and DeepSeek V2, this study seeks to clarify how different LLMs function as learning tools in AI-supported educational contexts. Differences in wording, sentence structure, and tone may influence how students interpret translated materials and complete bilingual tasks. Rather than treating AI output as automatically reliable, this study highlights the importance of critical engagement when LLMs are incorporated into translation practice and language learning activities. By identifying where model outputs are helpful and where human guidance remains necessary, the research provides practical insights for educators who integrate AI tools into coursework and classroom tasks.

Introduction
Mathematics is essential for understanding mechanical and engineering phenomena, yet its practical applications are often underrepresented in secondary education. This study presents a teaching proposal that connects mathematical concepts—such as circular and linear motion—to real-world mechanisms. By exploring how curved paths can produce straight-line movements and vice versa, students gain insight into the mathematical principles behind everyday machines, including gears, pistons, and wheels, while developing technological competencies.

Methods
The project combined a literature review on mathematics in mechanics with hands-on, project-based learning activities. Students analyzed the mathematical relationships involved in transforming circular motion into linear trajectories. As a key component, learners designed and produced a 3D-printed model to observe and experiment with these movements directly. This approach integrates STEM/STEAM principles, linking mathematics, technology, engineering, and creativity. By working with 3D printing software and hardware, students also practiced digital design, problem-solving, and technical skills alongside mathematical reasoning.

Results
The 3D-printed model allowed students to visualize abstract concepts and test their understanding in a tangible way. Participants could connect mathematical transformations to real mechanical applications and explain how rotations translate into linear motion. Working with 3D printing and digital modeling increased engagement, promoted technological literacy, and strengthened problem-solving abilities. The combination of theory, design, and experimentation fostered both conceptual understanding and practical competence.

Conclusions
Integrating mathematics with hands-on 3D modeling and printing provides meaningful, technology-rich learning experiences. Project-based, STEM-oriented activities help students appreciate the relevance of mathematics in engineering, develop digital and technical skills, and deepen their understanding of both abstract and applied concepts.

Automatic Short-Answer Grading (ASAG) has become an increasingly relevant research area within technology-enhanced STEM education, where short open-ended responses are frequently used to evaluate students’ conceptual understanding. However, authentic classroom datasets often present low-resource characteristics, including small sample sizes, lexical sparsity, and class imbalance, which pose significant challenges for reliable model evaluation, didactic usability, and reproducibility.

This work presents a reproducible and didactically oriented machine learning pipeline designed to support the evaluation of short open-ended student responses under low-resource educational conditions. Rather than proposing novel algorithms, the study emphasizes methodological transparency by integrating established linear classifiers—Logistic Regression, Multinomial Naïve Bayes, and Linear Support Vector Machines—within a unified and interpretable evaluation framework. Textual responses are represented using TF–IDF features, while model performance is assessed through adaptive stratified cross-validation to ensure robust accuracy estimation and minimize information leakage.

The pipeline is evaluated across multiple concept-specific datasets derived from undergraduate teacher education contexts in mathematics and science. Results demonstrate stable performance across classifiers and conceptual domains, supporting the viability of interpretable linear models for small-scale classroom datasets. Additionally, the framework enables token-level inspection of discriminative lexical features, facilitating formative didactic feedback and supporting educators in monitoring students’ conceptual development.

By prioritizing reproducibility, interpretability, and didactic applicability, the proposed framework provides a transparent methodological reference for applied ASAG research. Furthermore, the pipeline establishes a foundation for future studies exploring automated feedback mechanisms and classroom-oriented learning analytics in STEM education.

Code similarity tools are widely used in programming courses to support academic integrity, yet fixed percentage cutoffs are often applied without considering how score distributions change over time, vary by assignment type, or differ across detection tools. This study examines similarity patterns over three academic years (2023/2024–2025/2026) using two programming assignments, with approximately 50 students per assignment each year, and multiple similarity detection tools. For each assignment–year–tool combination, we analyzed distributional indicators, including mean, median, quartiles, standard deviation, the 99th percentile, and maximum, to capture central tendency, spread, and upper-tail behavior.

The results show that similarity dynamics are not uniform. In one assignment pattern, central values rose consistently over time (e.g., increasing medians and lower quartiles), indicating that the shift was not driven only by a small number of extreme submissions. In the other pattern, similarity levels started higher and then showed mild increases or stabilization, depending on the tool. Across both patterns, extreme values and upper-tail indicators generally declined in later years, accompanied by lower variability, suggesting a transition toward higher baseline similarity with fewer outlier-heavy cases.

These findings support a more context-aware interpretation of similarity scores in programming education. Rather than relying on a single fixed threshold, educators may benefit from a distribution-based reading of results that accounts for assignment characteristics, year-to-year shifts, and tool-specific behavior. This perspective provides a more robust basis for course-level monitoring and more defensible decision-making in integrity and assessment workflows.

Wheel-spinning refers to a pattern of unproductive persistence in which students repeatedly practice a skill without achieving mastery, despite receiving feedback and multiple opportunities for correction. This phenomenon poses a significant challenge for intelligent tutoring systems (ITSs) and adaptive learning environments, as delayed identification of struggling learners can lead to frustration, disengagement, and attrition. Although prior studies have proposed various wheel-spinning detectors, many existing approaches rely on complex models, extensive response histories, or system-specific features, limiting their scalability and early-detection potential.

This study develops an early and simplified wheel-spinning detection model that relies exclusively on student response-sequence features. Using data from the Cognitive Model Discovery Experiment (Spring 2010) in the Carnegie Mellon University DataShop repository, the analysis included 123 students, 49 geometry skills, and over 45,000 response records. Wheel-spinning was operationalized as failure to achieve mastery—defined as three consecutive correct first attempts—within ten practice opportunities. Two models were compared: a baseline logistic regression model and an enhanced ensemble model using gradient boosted decision trees (GBDT).

Results indicate that both models achieved strong predictive performance at early practice stages. The logistic regression model demonstrated high accuracy and precision but moderate recall, identifying confirmed wheel-spinning cases while missing a subset of struggling students. In contrast, the GBDT model substantially improved recall while maintaining high accuracy, achieving over 70% accuracy and approximately 75% recall by the fifth practice opportunity. These findings suggest that ensemble learning methods are particularly effective for early identification of unproductive persistence.

Overall, the study demonstrates that lightweight, response-based models can detect wheel-spinning early and reliably without reliance on tutor-specific parameters. The proposed approach offers a scalable and interpretable solution for adaptive learning systems, enabling timely instructional interventions and more responsive learner support.

The growing use of Artificial Intelligence (AI) in education is transforming teaching, learning, and assessment practices. While these AI-driven systems improve efficiency and personalisation, many existing approaches often prioritise technological performance over human-centred values. This raises many concerns about diminished human agency, trust in AI systems, and ethical governance in educational contexts. Thus, this paper aims to respond to these concerns by promoting a human-centred perspective on AI in education. This study adopts a conceptual and integrative approach, drawing on the interdisciplinary literature from fields such as education, AI–computer collaboration, and the intersection of ethics and AI, published in leading peer-reviewed journals and conference proceedings published in English between 2015 and 2025. Through theory development and critical synthesis, this paper develops a conceptual model that illustrates how human-centred AI principles can be integrated within technology-driven educational environments. The results present a human-centred AI framework for education around three interrelated dimensions: trust, human agency, and ethics. Transparency, explainability, and accountable governance are posited to enhance users’ trust, which in turn motivates learners and educators to exercise meaningful oversight and informed decision-making. Moreover, human agency acts as a mediating mechanism that ensures AI augments rather than replaces human judgement. Ethical governance is grounded in fairness, transparency, inclusivity, and accountability, under which trust and agency meaningfully inform responsible practice. Overall, this proposed integrated model highlights how the ethical interaction of these dimensions can foster ethically grounded human–AI collaboration while addressing ethical risks associated with over-automation and algorithmic bias. By re-centring key principles of trust, agency, and ethics, this study contributes to existing AI-enhanced educational research. This framework offers practical guidance for educators, policymakers, and practitioners by emphasising the need to implement AI systems that respect human values and enhance ethical responsibility, empowering educational practices. Future directions are also presented.

Introduction
Gamification improves memory by turning passive instruction into active learning. 3D simulations specifically replicate complex environments, building the spatial awareness and immersion needed for high-stakes training. This study evaluates the "Ready4Disasters" mobile application—an interactive platform developed within an Erasmus+ project—as a digital tool for simulating high-stress disaster scenarios.

Methods
By utilizing a modularized pedagogical approach, mobile applications facilitate the delivery of specialized disaster response training through several key instructional modalities: Scenario-Based Simulations, Gamified Engagement, and Just-in-Time Training. This framework promotes robust information encoding and mitigates the decay of technical competencies across critical hazard domains, specifically indoor fire, floods, and landslides.

This study employed a mixed-methods research design to establish a baseline for disaster response competencies. A questionnaire comprising 33 questions with a 5-point Likert scale was administered via MS Forms and Google Forms to a total sample of 159 experienced volunteers and trainers across Turkey (n=51), Italy (n=51), Greece (n=36), and Georgia (n=21). Analysis involved descriptive statistics to identify priority training areas and thematic coding of expert comments to drive iterative refinement of the 3D modules. Knowledge transfer was defined as the shift from theoretical instruction to applied procedural readiness, measured by analyzing the delta between initial competency gaps and final performance within the simulation.

Results
Pilot data prioritized Safety (66%) and Leadership (62%), revealing a 61% deficit in existing flood-response tools. Furthermore, 38% of participants (ages 18–25) rejected traditional manuals in favor of digital solutions. Final evaluations confirmed significant gains in knowledge transfer and high application usability (>80%), particularly regarding landslide and flood scenarios.

Conclusions
The Ready4Disasters framework validates 3D mobile gamification as a cost-effective, scalable training solution for multi-hazard preparedness. The results indicate that immersive environments effectively bridge the gap between theoretical instruction and applied response capabilities for decentralized volunteer networks.