Search Results (701)

This study examines how emotional and reflexive processes jointly relate to transfer of learning in workplace training contexts. Drawing on organizational learning theory, it introduces Reflexivity on Emotions (RoE) as a metacognitive capability through which individuals become aware of, critically examine, and respond to their emotional experiences. Integrating RoE, reflexivity on practice, positive affect, and negative affect within a person-centered framework, the study applies Latent Profile Analysis (LPA) to data collected from 609 correctional officer cadets enrolled in a six-month training program. The analysis identified four emotional–reflexive profiles (Generative–Reflexive, Balanced–Reflexive, Detached–Unreflexive, and Inhibited–Unreflexive), which showed different levels of transfer of learning. Notably, the Generative–Reflexive profile, characterized by elevated negative affect alongside strong reflexive resources, was associated with the highest levels of transfer, suggesting that negative emotions are not uniformly associated with poorer learning outcomes. More broadly, the findings indicate that transfer of learning is better understood through emotional–reflexive configurations rather than through isolated factors. The study contributes to organizational learning research by extending reflexivity into the emotional domain and by demonstrating the value of person-centered approaches for understanding individual differences in workplace learning. Practical implications for training design and the development of emotionally reflective learning environments are discussed. Full article

Background/Objectives: Wheelchair users with multiple sclerosis (MS) often face barriers that restrict participation in physical activity and exercise training. This manuscript reports on participant feedback to guide evaluating and refining a novel exercise training program, Seated Physical activity INtervention (SPIN). SPIN was adapted from the Guidelines for Exercise in MS (GEMS) approach using a three-step community-engaged research framework based on meeting the needs of wheelchair users with MS. Methods: Semi-structured interviews were conducted with 9 participants who completed the 16-week SPIN intervention. The key SPIN intervention components were the exercise prescription, exercise equipment, and behavioral coaching grounded in Social Cognitive Theory. Formative interview domains included overall experience, enjoyable and missing components, delivery modifications, barriers, lessons learned, and additional research topics of interest. Data were analyzed and reported using a rapid qualitative analysis approach. Results: Interviews averaged 16 ± 10 min. Participants reported enjoying SPIN, noting program strengths as being flexible and appropriate for individuals with MS, receiving coaching calls by knowledgeable staff that offered support and accountability, and receiving exercise equipment and video demonstrations. Participants also identified strategies for enhancing the program such as including peer support, offering real-time feedback during exercise, and adding other wellness behavior topics (e.g., diet). Conclusions: The results offer helpful ideas to consider when developing exercise training programs for wheelchair users with MS and other disabilities that may improve health and well-being. Full article

This study examines complex thinking as a higher-order cognitive competence in international graduate education. Drawing on Edgar Morin’s theoretical perspective, it analyzes how master’s students perceive this competence through four interrelated dimensions: systemic, scientific, critical, and innovative thinking. A total of 491 international students from a graduate university in Japan participated in the study. Using a quantitative, cross-sectional design, data were collected with the validated eComplexity instrument and analyzed through PERMANOVA with 999 permutations. The analysis examined differences in self-perceived complex thinking by sex, academic field, nationality, and academic semester. Results showed moderately high levels of self-perceived complex thinking across the sample, with systemic and critical thinking emerging as the strongest dimensions. Significant differences were found by nationality and academic semester, while no significant differences were observed by sex or academic field. These findings suggest that students’ perceptions of complex thinking are associated with cultural and academic trajectories, although the cross-sectional and self-report design requires cautious interpretation. The study contributes to competence-based graduate education by showing that complex thinking can be examined as a multidimensional and context-sensitive form of perceived cognitive development. Educational implications are discussed in relation to curriculum design, intercultural learning, global citizenship, and inclusion in international master’s programs. Full article

Prior research suggests that programming is a fundamental competency for all students. Due to its importance, programming education is integrated across many disciplines beyond computer science (e.g., humanities, social sciences, and engineering). Also, many existing courses report increasing enrollment trends. However, these changes have also introduced instructional challenges, particularly in supporting students with diverse backgrounds at scale. In this context, many studies have explored the use of AI code assistants as tools that may support instruction and learning. In these studies, while examining the use of AI code assistants, researchers have reported variation in educational contexts, implementation approaches, and outcomes. With this paper, we argue that synthesized information of such variations could help in understanding the effective use of such tools in programming education. To create a synthesized resource on AI code assistants, in this paper, we present a narrative review that synthesizes existing research. We reviewed 29 peer-reviewed studies identified through searches across three databases. The studies were analyzed to identify reported patterns of use, student and instructor perceptions, limitations in existing research, and suggested directions for future research. Across the reviewed studies, AI code assistants were commonly discussed for tasks such as code generation, debugging support, and real-time feedback, with ChatGPT reported most often (16 mentions), followed by GitHub Copilot (6 mentions). Disciplinary information was available in 24 studies, which helped identify the academic settings where AI code assistants were reported. Students generally describe these tools as useful, while also expressing concerns related to over-reliance and accuracy. Student perceptions were reported in 10 studies, while instructor perceptions were reported in 4 studies. Common reported limitations include small sample sizes, short intervention durations, reliance on self-reported data, and limited examination of long-term learning outcomes. Overall, this review consolidates current evidence on how AI code assistants are used and perceived in programming education and identifies areas where more research is needed. Full article

This paper presents results from the project schAUT, a participatory research project initiated by Humboldt University Berlin, Goethe University Frankfurt a.M. and White Unicorn e.V., funded by the German Federal Ministry of Education and Research (BMBF; FKZ: 01NV2104). It aimed to identify and reduce barriers to learning and participation in mainstream schools, with a particular focus on autistic students. This paper introduces a questionnaire and a program to support School Organizational Development (SOD), aiming to provide equitable and accessible learning environments grounded in international frameworks on inclusive education. This study combines qualitative and quantitative approaches to examine the subjective experiences of barriers. We present data obtained through a multi-phase development and validation phase. The results show that neurodivergent participants generally experienced higher subjective barriers, although we observed that barriers affect neurotypical students as well, highlighting a subjective nature. We argue that these findings support neurodiversity as a relevant concept, especially in educational contexts. This supports Larrauri et al.’s Big-Tent approach to neurodiversity, emphasizing individual variability while acknowledging structural biases towards neurotypical norms in educational environments. The study highlights the value of multiperspective approaches in (participatory) research and SOD, to develop strategies for an inclusive educational environment through neurodiversity-informed decision processes and enable equitable learning environments for all students. Full article

The urban tree canopy is an important resource that spans public and private property and whose form and quantity change over short distances. Although remote sensing and deep learning approaches have been used to map urban tree canopy, the high cost of commercial imagery and the technical complexity of model development have limited their adoption by urban forestry practitioners. We developed a structured and reproducible deep learning workflow optimized for freely available USDA National Agriculture Imagery Program (NAIP) imagery. The workflow incorporates a reproducible U-Net segmentation model for canopy delineation and a YOLOv9e object detection model for individual tree identification, enabling complementary estimation of the canopy extent and individual tree locations. Across two neighborhoods in Los Angeles, the optimized U-Net achieved a Dice coefficient of 0.824 for canopy segmentation, while YOLOv9e reached an F1-score of 0.687 for individual tree detection on a held-out test set with 17,466 annotated trees. A data sufficiency experiment showed that model performance stabilizes when approximately 130 trees are annotated per 320 × 320 pixel (px) tile, corresponding to about 25,379 training and 2641 validation labels, providing a practical target for annotation effort. Additional experiments demonstrate a structured workflow for spatial sampling, training data requirements, and the use of model inferences to estimate tree canopy extent and individual tree locations. The workflow also shows encouraging evidence of transferability to previously unseen urban areas without retraining. By relying solely on NAIP-optimized approaches, this new workflow bridges the gap between complex deep learning techniques and the practical needs of urban foresters; empowers local stakeholders to create accurate, affordable, and timely urban tree inventories; and fosters data-driven decision-making for the sustainable management of urban green infrastructure. Full article

Background: Osteosarcoma remains a biologically complex and clinically challenging malignancy, with survival gains plateauing despite decades of multimodal therapy incorporating surgery and cytotoxic chemotherapy. Unlike cancers in which mutation-centric precision oncology has yielded transformative advances, osteosarcoma is characterized by profound structural variation, copy number alteration dominance, and dynamic clonal evolution, limiting the effectiveness of single-target approaches. These realities motivate alternative strategy-level frameworks that better align treatment selection with evolving disease behavior. Methods: This narrative educational review synthesizes contemporary evidence from osteosarcoma biology, radiobiology, and translational oncology to propose a state-based framework for integrating radiotherapy—particularly stereotactic body radiotherapy (SBRT/SABR) and spatially fractionated radiotherapy (SFRT)—into osteosarcoma management and clinical trial design. Rather than relying solely on static anatomic stage, this framework emphasizes clinically actionable, time-varying state variables, including disease burden patterns (localized, oligometastatic, polymetastatic), tempo of progression, prior systemic response, and feasibility of complete local control. Results: Within this context, radiotherapy is presented not only as a local control modality but also as a hypothesis-generating biologic intervention, capable of perturbing tumor vasculature, inflammatory signaling, innate DNA-sensing pathways, and immune/myeloid programs in a dose-, fractionation-, and spatial-distribution-dependent manner. The review critically examines both the potential opportunities (e.g., local eradication, immune modulation) and limitations (e.g., rarity of abscopal responses, risk of unintended systemic signaling) of radiobiotherapy combinations, emphasizing the need for cautious interpretation and prospective validation. Conclusions: Finally, the article outlines practical implications for state-stratified, biomarker-embedded clinical trials, highlighting endpoints beyond conventional response criteria, including circulating tumor DNA dynamics, immune and myeloid signatures, and long-term patterns of disease progression. Overall, this review frames radiobiotherapy as an educational and investigational paradigm intended to support rational hypothesis generation, multidisciplinary decision-making, and learning-oriented trial designs in osteosarcoma, rather than as definitive clinical guidance. Full article

Breast cancer is a heterogeneous disease characterized by substantial molecular diversity and variable treatment outcomes across patients. Despite advances in targeted and systemic therapies, anticipating individual benefit remains a major clinical challenge. In this context, Artificial Intelligence (AI) can support precision oncology by integrating high-dimensional molecular profiles with clinical and pharmacological information. Here, we present B.R.E.A.S.T. (Breast canceR Enhanced AI-Supported Therapy), an interpretable machine learning framework designed to predict therapy outcome from tumor proteomic profiles integrated with clinical and treatment annotations. Proteomic data from The Cancer Genome Atlas (TCGA) and The Cancer Proteome Atlas (TCPA) were harmonized with outcome and therapy information, and thirteen supervised classifiers were systematically evaluated using stratified 5-fold cross-validation. Therapeutic outcome labels were operationally defined by integrating available treatment response annotations with complementary clinical outcome information. Across both cohorts, ensemble-based models consistently achieved the most stable and highest discriminative performance, supported by learning-curve analyses and consistent behavior across independent datasets. To enhance interpretability, we implemented a two-step feature selection strategy combining model-specific importance measures with a global consensus ranking, enabling the identification of a compact set of robust proteomic biomarkers associated with therapeutic outcome. Top-ranked features mapped to molecular programs relevant to breast cancer progression and treatment sensitivity, including regulators of cell survival, DNA damage response, PI3K/AKT/mTOR signaling, and invasion-related processes. Re-evaluation using only the top 30 globally ranked features preserved high predictive performance across both independent breast cancer cohorts, indicating that a parsimonious proteomic signature captures core molecular determinants of outcome. Overall, B.R.E.A.S.T. provides a robust and generalizable proteomics-driven framework for modeling outcome-associated therapeutic response patterns and supporting biologically informed biomarker discovery in breast cancer. Full article

Background/Objectives: Children’s mental health and positive development are shaped by family, environmental, and individual factors. Although neurodevelopmental disabilities (NDDs) are well-established correlates of poorer mental health outcomes, few national-scale studies have simultaneously modeled positive (flourishing) and negative (anxiety, depression) outcomes within a unified ecological framework. This study examined how parent mental health, peer victimization, neighborhood and school context, and four NDD diagnoses (autism spectrum disorder [ASD], attention-deficit/hyperactivity disorder [ADHD], developmental delay, and learning disability) are associated with flourishing, current anxiety, and current depression in a national sample of U.S. children aged 6–17 years. Methods: Cross-sectional data from the 2023–2024 National Survey of Children’s Health (NSCH; N = 71,172) restricted to ages 6–17 with complete data (unweighted n = 64,263; weighted population estimate ≈ 44.6 million children) were analyzed using Complex Sample logistic regression (SPSS 30), accounting for stratified design (state × stratum), household clustering, and sampling weights. Three hierarchical models were estimated for each outcome. NDD-stratified subgroup analyses (n = 13,971; weighted ≈ 8.6 million) triangulated moderation findings. Multiple imputation (m = 5) sensitivity analyses confirmed robustness. Results: Weighted prevalence was 60.7% for flourishing, 13.2% for current anxiety, and 5.1% for current depression. In Block 2 models, poorer parent mental health and more frequent bullying victimization were robustly associated with all outcomes (flourishing OR = 0.62 and 0.65; anxiety OR = 1.64 and 1.63; depression OR = 1.95 and 1.75; all p < 0.001). Supportive neighborhood (flourishing OR = 1.40, depression OR = 0.80) and safe school (flourishing OR = 1.20, anxiety OR = 0.87) were protective. ADHD was the strongest disability-specific correlate (flourishing OR = 0.29; anxiety OR = 4.69; depression OR = 4.27). Three of the twelve interaction terms were significant, all involving ADHD. Relative to children without any NDD, subgroup analyses suggested attenuated associations of parent mental health and bullying with anxiety and depression among children with any NDD (e.g., bullying on anxiety: no-NDD aOR = 1.73 vs. Any-NDD 1.52); however, formal interaction tests identified ADHD as the only significant moderator of these associations. On the absolute-risk scale, however, the increase in internalizing problems with more frequent bullying was larger in children with ADHD. Conclusions: Family mental health support and bullying prevention are universally relevant levers for improving children’s mental health and flourishing. Although attenuation of the odds-ratio associations was observed primarily in ADHD-related analyses, specifically for the internalizing outcomes (anxiety and depression), universal anti-bullying and parent mental health interventions remain relevant for children with NDDs, supporting integration into pediatric clinical and public-health programs alongside disability-specific support pathways. Full article

Environmental science communication has become a core competence for addressing global challenges such as climate change, glacier recession, and hydrometeorological risks. Yet university curricula often prioritize technical knowledge over communicative skills, limiting students’ ability to engage with diverse audiences. This study proposes a structured three-level framework (i.e., micro-, meso-, and macro-communication) for teaching environmental science communication. The framework is explored across six applied case studies, including glaciological thematic trails, dual-training programs, a climate-education game, an international higher-education project, immersive 360° field experiences, and an AI-enhanced scientific exhibition. Drawing on qualitative and descriptive evidence, the cross-case analysis suggests that communication competencies may develop progressively from synthesis and clarity (micro-communication), to multimodal visualization and structured argumentation (meso-communication), to stakeholder-oriented and intercultural dialogue (macro-communication). The findings indicate that multimodal, immersive, and AI-supported approaches may support accessibility, engagement, and inclusivity, while authentic learning environments contribute to the development of transferable communication skills. This study provides an exploratory and practice-based framework that may inform curriculum design and pedagogical innovation, suggesting that communication could be more systematically embedded across environmental science programs in order to strengthen evidence-informed societal engagement and support sustainable environmental governance. Full article

Accurate maize yield prediction across diverse environments is pivotal for modern breeding programs. While machine learning (ML) excels at capturing non-linear environmental effects, Genomic Best Linear Unbiased Prediction (GBLUP) remains a benchmark for modeling polygenic small-effect contributions. However, principled integration of these paradigms—while explicitly accounting for genotype-by-environment interaction (G×E)—remains a formidable challenge. We propose a two-step framework evaluated on the Genomes to Fields (G2F) 2022 dataset. In Step 1, ML models are employed to fit environmental main effects; in Step 2, genomic residuals are modeled via additive-dominance relationship matrices, augmented by an explicit low-rank G×E matrix. Candidate interaction markers were screened through plasticity-based genome-wide association studies (GWAS) across six phenotypic stability metrics and used to construct a low-rank candidate G×E representation, with a cross-validation-selected scaling parameter applied to control the contribution of the predicted G×E component. TwoStep_G×E_alpha0.33, achieved a within–environment Pearson correlation coefficient (PCC) of 0.376, outperformed both GBLUP and the competition-winning model (PCC = 0.357) in within-environment ranking. Furthermore, environment-adaptive selection yielded a genetic gain of 0.454 Mg ha⁻¹, representing a 34.7% improvement over GBLUP. Overall, the proposed framework provides a practical approach for environment-specific yield prediction and adaptive selection in maize breeding. Full article

Efficient Distributed Computing is still a major challenge, especially in networks composed of very-low-resource embedded systems, e.g., tiny microcontrollers deployed in sensor networks. This work will, firstly, address the design and implementation of event-driven and real-time capable low-resource Virtual Machines (VMs) tightly coupled to communication-centric systems, and secondly, address messaging and routing in mesh-grid networks. The distributed VM network herein forms one big virtual computer executing typically the same program on each node, but processing different data with different control states. The VM provides an integrated program code compiler and an optimized Bytecode processor. The programming language of the VM supports channel-based communication, multi-tasking, and event-based (asynchronous) data processing following the CSP model. The VM fits in microcontrollers with only a few kB of RAM and ROM. A major part of this work is dedicated to network messaging (supported by the VM, too) and routing in two-dimensional mesh-grid networks with a varying degree k of communication ports per node (connectivity degree k), and especially considering the odd but technical relevant case, k = 3, which introduces challenges in message routing that are solved herein. This study demonstrates the performance and suitability of our VM approach for distributed sensor networks performing distributed Machine Learning and clustering by using local sensor data only. Full article

In edge-intelligent systems, efficient resource management and task scheduling are critical but challenging due to the dynamic and heterogeneous nature of edge nodes (e.g., IoT devices, drones). We model this dynamic resource allocation challenge as an online sleeping Restless Multi-Armed Bandits (RMAB) problem, where each edge node (arm) operates as a Markov decision process. Unlike prior RMAB frameworks assuming perpetual availability, our setting captures the stochastic availability of edge nodes across rounds. The system controller (learner) is unaware of the transition functions, reward distributions, and node availability a priori. The goal is to maximize expected cumulative rewards through adaptive node selection. To explore this target problem, we first derive an asymptotically optimal sleeping-index policy (SIP) as the oracle based on the fluid process transformation. Then we propose OSILA (Online Sleeping Index-aware Learning Algorithm), featuring a Minimum Exploration Guarantee (MEG) mechanism for efficient exploration. This is coupled with a modified Linear Programming-based exploitation mechanism to construct an online sleeping index, effectively handling dynamic node availability. To the best of our knowledge, this work is the first to provide the theoretical analysis (which achieves $\tilde{\mathcal{O}}\left(K{T}^{2/3}\sqrt{logT}\right)$ regret where K is the number of arms and T is the time horizon) to the online sleeping RMAB problem. Empirical results validate both theoretical guarantees and practical effectiveness in dynamic edge computing environments. Full article

This article reports an exploratory, small-cohort mixed-methods case study of how an e-learning platform and a virtual reality (VR) workshop simulator can be integrated into a traditional craft apprenticeship without displacing workshop-based learning. Drawing on the Craeft glassblowing Pilot 1 at the European Centre for Research and Training in Glassmaking (CERFAV), it reports a two-phase mixed-methods study contrasting a Traditional Augmented (TA) group, which used a Craeft e-learning platform and a VR glassblowing simulator, with a Traditional (T) control group following the standard Certificate of Professional Competence (CPC) program. Quantitative data from formative assessments and CPC examination results are combined with qualitative feedback, satisfaction surveys, self-assessment questionnaires, and interviews with apprentices and trainers. In Phase 1, where digital tools were deployed in a separate mode alongside existing instruction, the e-learning platform was perceived as pedagogically valuable, but descriptive differences in assessment outcomes were limited and uneven, with greater score dispersion in the TA group. In Phase 2, redesigned hybrid usage scenarios assigned distinct and complementary roles to the e-learning platform, VR, and workshop practice within an iterative learning cycle, and the descriptive results suggest more consistent patterns of higher scores for the TA group in cross-cutting theoretical subjects, with less variance in their scores. Qualitative analyses show that apprentices adopt a pragmatic stance towards digital tools, using the e-learning platform primarily for revision and exam preparation and VR for workshop discovery and tool recognition, while maintaining a strong attachment to material practice. The study suggests that, in small, high-stakes craft VET program, the perceived value of virtual learning environments depends less on their intrinsic properties than on their orchestration within coherent hybrid designs and on trainers’ capacity to align them with authentic tasks and assessment regimes. All findings should be interpreted as exploratory given the small sample size (n < 20), non-random group assignment, and potential self-selection biases. Full article

Safety assurance is uncompromisable in safety-critical environments, especially when drastic model uncertainties (e.g., distributional shift) exist, especially with humans in the loop. However, incorporating uncertainty in safe learning will naturally lead to a bi-level problem, where at the lower level, the worst-case safety constraint is evaluated within the uncertainty ambiguity set. In this paper, we present a tractable distributionally safe reinforcement learning framework that enforces safety under a distributional shift, as measured by a Wasserstein metric. To improve the tractability, we first use duality theory to transform the lower-level optimization from the infinite-dimensional probability space where distributional shift is measured, to a finite-dimensional parametric space. Moreover, by differentiable convex programming, the bi-level safe learning problem is further reduced to a single-level one with two sequential computationally efficient modules: a convex quadratic program to guarantee safety, followed by a projected gradient ascent to find the worst-case uncertainty simultaneously. This end-to-end differentiable framework with safety constraints offers a tractable single-level approach to addressing distributional safety. We test our approach on first- and second-order systems with varying complexities, including hardware demonstration on a 6-DOF drone. Compared with both uncertainty-agnostic policies and robust policies, our approach demonstrates a significant improvement in safety guarantees. Full article