Search Results (165)

The rapid integration of large language models (LLMs) into education has shifted research focus from questions of capability, such as what LLMs can do and how accurately—to questions of deployability, including how they can be operated effectively for many learners at once. In school environments, system reliability, scalability, and real-time responsiveness are critical, as delays or interruptions can directly reduce learner engagement, particularly during synchronous activities. This study evaluates the performance of open-source LLaMA models ranging from 1 billion to 70 billion parameters across one-, dual-, triple-, and quad-GPU configurations suitable for educational settings. Performance is assessed using four key indicators: success rate (percentage of completed requests), generation speed (tokens per second), throughput (completed responses per second), and latency (time until full response generation). These metrics were measured under progressively increasing numbers of simultaneous users to identify system capacity limits and trade-offs between model size, responsiveness, and scalability. The results indicate that smaller models (1B–3B) deliver faster, more stable performance under concurrent use, while larger models (8B–70B) experience substantial slowdowns and reduced reliability, even on high-end GPU systems. These findings suggest that effective educational deployment should prioritize empirical performance and infrastructure compatibility over model size alone. The paper concludes by proposing a practical framework to guide educators, administrators, and developers in selecting and configuring locally hosted GPU systems that balance model capability, response speed, and resource efficiency for real-time applications such as AI tutors, classroom chatbots, and automated feedback tools. Full article

This article proposes an expanded framework of Professional Noticing (PN) for Sustainability in Higher Education by integrating intermedial semiotics and ecopoetry as pedagogical tools aligned with the United Nations Sustainable Development Goals (SDGs). Building on the PROMISE project, in which the author participated, the study conceptualises ‘noticing’ as an embodied, multimodal, and ethically inflected process of attending to human and more-than-human sign systems. The article introduces Intermedial Sustainability Noticing (ISN) as an extension of PN that foregrounds ecological awareness, intermedial perception, and cross-cultural interpretation. The study adopts a qualitative case study design based on the implementation of ISN within the Eurasia Foundation Cross-Cultural Partnerships hybrid course at Complutense University of Madrid. Participants included undergraduate students from diverse European and Asian institutions, who engaged in interdisciplinary and intercultural dialogues on sustainability through comparative literature and ecopoetry. In the course, students developed perceptual, interpretive, and ethical awareness of global challenges by emphasising ‘noticing’ and attentional depth while broadening understanding of ecological interdependence. Data were collected through reflective journals, written assignments, multimodal projects, and classroom discussions, and analysed using an interpretive, semiotically informed approach. Findings indicate that ISN fosters enhanced attentional depth, multimodal interpretive skills, and increased ecological awareness, particularly through structured engagement with ecopoetry. The work of Kathleen Jamie is presented here as exemplary of how literary texts can activate perceptual, interpretive, and responsive dimensions of noticing, enabling students to connect textual analysis with sustainability concerns. The article argues that ISN offers a transferable pedagogical model for embedding sustainability competencies within humanities curricula, contributing to Higher Education’s role in fostering ecological literacy, intercultural dialogue, and ethically grounded engagement with global challenges. Full article

The fundamental purpose of education—preparing new generations to be contributing members of society—remains constant, yet achieving this has become increasingly complex amid multifaceted technological, cultural, economic, and social transformations. Educational leaders worldwide continuously seek innovative pedagogical models addressing diverse learner needs and rapid societal changes. However, this article challenges the assumption that educational quality requires constant novelty, arguing that solutions lie in the innovative integration of established pedagogical theories developed over the past 150 years by scholars such as Dewey, Vygotsky, Piaget, Feuerstein, Gardner, Freire, and others. The article’s primary objective is to encourage education leaders and teacher educators to reconceptualize innovation by prioritizing pedagogical integration over continuous adaptation to rapidly expanding domain-specific knowledge and emerging technologies. Accordingly, this article employs a conceptual synthesis of major pedagogical approaches to equip educators with theoretical foundations and practical tools to foster learner independence, critical thinking, and holistic development across cognitive, emotional, and social domains. It will also promote inclusion through a practical framework integrating pedagogical theories, addressing diversity from a dual perspective, recognizing that both teachers and learners bring unique characteristics, strengths, and needs. Moreover, developing independent learners requires empowering teachers to cultivate unique professional methodologies grounded in integrated pedagogical understanding, so that a shift from innovation-centered to integration-centered teacher education may serve as a sustainable path toward educational quality and academic excellence. Full article

Biological cognition depends on learning-structured representations in ambiguous environments. Computational models of structure learning typically frame this as an inference problem, but often overlook the temporally extended dynamics that shape learning trajectories under ambiguity. In this paper, we reframe structure learning as an emergent consequence of constraint-based dynamics. Informed by the literature on the role of constraints in complex biological systems, we develop a constraint-based approach to computational cognitive modelling and provide a proof-of-concept model. The model consists of an ensemble of components, each comprising an individual learning process, whose internal updates are locally constrained by both external observations and system-level relational constraints. This is formalised using Bayesian probability as a description of constraint satisfaction rather than epistemic inference. Representational structure is not encoded directly in the model equations, but emerges over time through the interaction, stabilisation, and elimination of components under these constraints. Through a series of simulations in environments with varying degrees of ambiguity, we demonstrate that the model reliably differentiates the observation space into stable representational categories. We further analyse how global parameters controlling internal constraint and initial component precision shape learning trajectories and long-term behavioural alignment with the environment. We discuss the formal relationship between the present approach and Bayesian inference accounts, and argue that a constraint-based approach offers a conceptually distinct foundation for relating computational models to biological systems. Full article

I revisit Benjamin’s text ‘The Work of Art in the Age of Its Technological Reproducibility’ and find it both problematic and richly suggestive. Carefully reading it, both with and against, I search it for continuities and breaks with ‘reproducibility’ today and hence insights into AI and its relationship to art. In so doing, I sketch some tentative conclusions about how such an investigation might assist us towards understanding what art is and isn’t, how the practice of art relates to our humanity and finally, though a thorough settling of accounts with AI and its boosters will require political change on a grand scale, how Benjamin’s ‘aura’ might offer a small but significant locus of resistance to the commodification and dehumanising drive currently occasioned by AI in the field of art. Full article

Background and Objectives: The term “appendiceal mucocele” is widely used in clinical practice, yet it remains diagnostically incomplete because it represents a preoperative clinical and imaging label rather than a definitive diagnosis. We examined the diagnostic and therapeutic pathway of patients managed under this working label over a 10-year period and assessed how final histopathology influenced subsequent management. Materials and Methods: We performed a retrospective single-center cohort study including patients with appendiceal mucinous lesions managed between January 2016 and December 2025. Eligible cases were identified from the institutional histopathology registry and classified on final histopathology as non-neoplastic mucoceles or neoplastic mucinous lesions, including low-grade and high-grade appendiceal mucinous neoplasms. Data were extracted on CT use, surgical strategy, margin status, extra-appendiceal/serosal mucin, chemotherapy, recurrence, and follow-up. Results: Eighteen patients met the eligibility criteria. Final histopathology showed two distinct endpoints: non-neoplastic mucoceles and neoplastic mucinous lesions. CT was the dominant preoperative investigation and most often supported the working impression of mucocele or suspected mucinous appendiceal lesion, although some neoplastic cases entered the pathway as presumed appendicitis or perforation. Appendectomy was the dominant index procedure, whereas primary right hemicolectomy was uncommon. Completion right hemicolectomy was performed only in the neoplastic subgroup, and management decisions were driven by a limited number of actionable histopathologic features, particularly margin status and extra-appendiceal/serosal mucin. Conclusions: This study highlights a real-world diagnostic and therapeutic pathway in which the term “mucocele” represents a preoperative label rather than a definitive diagnosis. Histopathology appeared to represent the key step in this pathway, rather than the preoperative imaging impression. Given the small sample size, these findings should be interpreted as descriptive and hypothesis-generating. In practical terms, suspected appendiceal mucocele should be approached through a standardized pathway involving safe index resection, structured pathology reporting using modern terminology, and risk-adapted escalation and follow-up when higher-risk features are present. Full article

Understanding how users perceive assistive robotic systems is critical for their successful adoption, particularly in rehabilitation settings where both patients and clinicians influence decision-making. While prior work has focused on technical performance and overall usability, affective responses such as trust, control, and perceived independence are often captured using coarse, single-score measures that overlook important nuances. This study analyzes focus group discussions with individuals with spinal cord injury to examine how users evaluate different aspects of assistive robot design. A hybrid aspect-based sentiment analysis approach is applied, combining lexicon-based and transformer-based methods to capture both interpretable and context-sensitive sentiment. The analysis separates sentiment across key dimensions, including independence, functionality, safety, control, cost, and data sharing. Participants expressed consistently positive views toward independence and functional support, while responses related to safety, control, and data sharing were more conditional. In particular, trust emerged as something that depends on transparency, user control, and the ability to override system behavior, rather than a fixed attitude toward the technology. These findings suggest that successful assistive robotic systems must balance autonomy with user authority and provide clear, adaptable mechanisms for control and data governance. Full article

Chemotherapy-induced peripheral neuropathy (CIPN) is a dose-limiting toxicity affecting 30–40% of patients treated with neurotoxic chemotherapy. Sensory symptoms arise from injury to dorsal root ganglion (DRG) neurons and their axons; yet, the underlying mechanisms remain incompletely understood. While human induced pluripotent stem cell (iPSC)-derived sensory neuron (iSN) monolayers have provided mechanistic insight, they lack the three-dimensional architecture and cellular heterogeneity of native DRG tissue. Here, we generated human iPSC-derived DRG organoids (iDRGOs) containing mixed neuronal and peripheral glial populations and established a quantitative neurite outgrowth assay to model chemotherapy-induced neurotoxicity in a 3D context. iDRGOs from three healthy donors were exposed to bortezomib, vincristine, or paclitaxel. All three drugs caused dose-dependent neurite outgrowth impairment without significant short-term changes in organoid size, consistent with early axonal injury. Vincristine reduced MAP2 levels when normalized to total protein, whereas bortezomib and paclitaxel showed divergent microtubule-associated responses compared to monolayer cultures. The developmental stage significantly influenced the baseline neurite outgrowth, highlighting the need for age standardization. These results establish iDRGOs as a physiologically relevant human platform that complements monolayer models for mechanistic studies and therapeutic screening in CIPN. Full article

Background: Subacute sclerosing panencephalitis (SSPE) is a chronic, progressive disease of the central nervous system (CNS) caused by persistent infection at this level with the wild measles virus. Its incidence is negatively correlated with measles vaccination coverage. The pathogenesis isn’t fully understood, but infection before the age of 2 is an important risk factor. Methods: This is a retrospective observational study conducted at the Louis Turcanu Emergency Children’s Hospital in Timisoara, Romania, based on the analysis of the medical records of patients diagnosed with SSPE between January 2021 and December 2025. We analyzed demographic and epidemiological factors, clinical and paraclinical findings, management, and outcomes. Results: Seven children were diagnosed during the study period, with a mean age of 8.4 years (range 7–11 years). Six of them had contracted measles during their first year of life, and one at the age of four. The mean latency period was 7.1 years (range 4–9 years). On admission, all patients presented symptoms consistent with clinical stage II, with periodic slow wave discharges on electroencephalogram (EEG). The initial brain Magnetic Resonance Imaging (MRI) was normal in two cases, while revealing varied abnormalities in all others. Despite complex treatment with isoprinosine and anticonvulsants, progressive cognitive and neurological deterioration continued in all patients. Conclusions: SSPE is a rare but serious, debilitating disease despite its complex, multidisciplinary care. Following a 10-year SSPE-free period, the reappearance of these pediatric cases constitutes a public health alert, unequivocally demonstrating the importance of measles vaccination. Full article

Aboriginal and Torres Strait Islander adolescents living in rural communities do not have sufficient access to health promotion services. Community programs that respond to adolescent needs, highlight community strengths, and are locally tailored are needed. Set in Queensland (Australia), this study was cross-sectional and qualitative in design. Using implementation science and Aboriginal and Torres Strait Islander frameworks, this study aimed to identify community priorities for the co-design of a culturally appropriate, empowerment-focused nutrition program with rural Aboriginal and Torres Strait Islander adolescents. Through community yarning, the barriers, enablers, and opportunities for program implementation were explored within an Aboriginal and Torres Strait Islander community-controlled health organization. Ten adolescents, two parents/caregivers, eight healthcare staff, six community leaders, and four Elders participated. Thematic analysis identified six themes that outline community health priorities, contextualization to the local food environment, and the importance of cooking skills for empowerment and involving the family unit. Thematic analysis also explored community preferences for program evaluation. Themes were integrated with other knowledge sources to develop a program outline that is aligned with evidence-based practice and community voice. Implementation of the co-designed program is recommended and will be explored in partnership with the community through future research. Full article

Background/Objectives: The formulation development of Isotretinoin (ISN) is limited by its solubility and stability issues. This study aimed to characterise the BCS class II drug ISN, particularly the possible different solid state and formulate amorphous solid dispersion aiming for a supersaturation state. Methods: ISN’s physical states are investigated in its raw form, quench-cooled form, physical mixture with the polymer and corresponding solid dispersion form. Quench-cooled ISN was prepared in situ using DSC. Carrier stabilisation of ISN was attempted using the solid dispersion technique. Hereby, the solid dispersion of drug-polymer PVPVA at a ratio of 1:3 was prepared using the solvent evaporation method. Solid dispersion, physical mixture and raw ISN were characterised for the saturated solubility. Physical characterisation of the samples was performed using DSC, ATR-FTIR and a light microscope. Results: Two polymorphs of ISN (forms I and II) were found in the raw ISN, with form II being thermodynamically more stable. ISN possesses strong crystallinity and resistance to amorphisation under the applied quench-cooling condition without the presence of a carrier system. The conjugated polyene structure in ISN contributes to the polymorphic transformation and isomerisation. The incorporation of PVPVA in the solid dispersion system successfully improved the water solubility (sixfold) of ISN despite a combination of crystalline and amorphous components being present in the system. Conclusions: ISN is a class II drug crystal molecule. Taking advantage of solubility and possibility in the polymorphic transformation of ISN in a binary system, we concluded that ISN could potentially be formulated into its corresponding crystalline solid dispersion form. Full article

The integration of Unmanned Aerial Vehicles (UAVs) with the Internet of Things (IoT) is revolutionizing a wide range of applications. However, collecting massive sensing data from large-scale fields efficiently remains challenging, constrained by the limited energy of UAVs and sensing nodes. Existing schemes lack the computational intelligence of an Edge Server (ES) for deep coordination. To address this, we propose DC-CSAP, a novel “Edge-UAV-End” collaborative data collection framework. DC-CSAP introduces a systematic workflow orchestrated by the ES, which is operationalized through four dedicated collaboration mechanisms: (1) In our ES–UAV collaboration, we devise a two-phase path optimization algorithm that hybridizes Simulated Annealing (SA) with a convex-hull-inspired greedy method. (2) The ES–ISN collaboration features a prediction-based binary vector mechanism, transmitting only inaccurate data to slash communication overheads. (3) The UAV–ISN and (4) Inter-ISN protocols ensure efficient data exchange and aggregation. Extensive simulations validate that DC-CSAP outperforms benchmarks in terms of Correct Prediction Rate (CPR), energy efficiency, and UAV path length. Full article

Background: Inhaler device changes at hospital discharge should address patient capacity yet often reflect routine. We evaluated the appropriateness of these decisions and their impact on clinical outcomes. Methods: In this prospective observational study (N = 480), we assessed patient technical capacity using a composite of critical errors, inspiratory flow, adherence, and knowledge. We stratified patients into ‘Need-Positive’ and ‘Need-Negative’ cohorts to quantify patterns of clinical inertia and over-adjustment. Multivariable models identified predictors of decision-making and associations with 30-day outcomes. Results: Device changes were primarily determined by the pre-admission device class (spacers: aOR 0.52; 95% CI 0.28–0.96; p = 0.037) and by the patient’s treatment pathway rather than by clinical need. This disconnect generated two types of errors: 38.3% of Need-Positive patients (n = 214) experienced clinical inertia (no corrective action), while 36.8% of Need-Negative patients (n = 266) underwent over-adjustment (unnecessary switching). Inertia perpetuated errors in patients with need, whereas over-adjustment was associated with the emergence of new errors in patients without need. Successful mismatch resolution was associated with a significantly lower 30-day readmission rate (12.1% vs. 32.5%; OR 0.48; 95% CI 0.26–0.88; p = 0.017). Conclusions: Discharge prescribing is driven more by habit than by objective assessment, leading to widespread missed opportunities for correction. Implementing evidence-based protocols to identify and resolve patient–device mismatches may represent a high-impact strategy to reduce readmissions and associated healthcare use. Full article

Background: Poor clinical control in Chronic Obstructive Pulmonary Disease (COPD) is prevalent, yet the interplay of disease severity, modifiable factors, and clinician perception remains poorly understood. This study aimed to determine the frequency of poor control, identify its independent determinants, and characterize the heterogeneity of the poorly controlled population receiving maintenance inhaled therapy with various devices in primary care. Methods: In a multicenter, cross-sectional analysis of 988 patients from the Study SIMPLIFY, clinical control of COPD was classified using the objective RADAR score. We used multivariable logistic regression and Machine Learning (Random Forest with SHAP analysis) to identify determinants of poor control (RADAR ≥ 4) and k-medoids cluster analysis to characterize the poorly controlled subgroup (n = 452). Results: Nearly half the cohort (45.7%, n = 452) had poor clinical control. Agreement between physician-assessed control (five categories) and RADAR classification was 49.3%, with overestimation in 34.0% and underestimation in 16.7% of cases (Cohen’s κ = −0.081; weighted κ = −0.037). The strongest independent determinants were the exacerbator phenotypes (eosinophilic aOR 6.85; non-eosinophilic aOR 4.91). Key modifiable factors included active smoking (aOR 1.92), lower TAI-12 adherence score (per point; aOR 0.96), high dosing frequency (≥4 inhalations/day; aOR 1.54) and high inhaler burden (≥3 devices; aOR 1.84). Machine learning analysis identified clinical phenotype and adherence behavior as the top two scale-independent predictors of poor control. Cluster analysis of the poorly controlled group revealed five reproducible and clinically meaningful phenotypes (C0–C4), primarily separated by treatment complexity, comorbidities, and adherence. Conclusions: Poor clinical control is common and critically under-recognized in primary care patients with COPD on maintenance inhaled therapy. This is driven by a profound clinician perception gap and a failure to address key modifiable determinants, such as high dosing frequency, regimen complexity, and poor adherence, which likely drives therapeutic inertia. Our findings underscore the need to integrate objective tools to unmask poor control and highlight the importance of treatment simplification. The identification of distinct clinical phenotypes provides a roadmap toward a more personalized, evidence-based standard of care. Full article

The application of AI in financial planning services has the potential to enhance universal access to financial services. However, AI still faces common consumer mistrust and resistance, hindering the long-term sustainability of AI-powered financial planning. This research aims to explain why consumers resist AI in financial planning and the mechanisms that lead to this resistance and negative customer behavior. This research developed a conceptual model by integrating the S-O-B-C framework with Innovation Resistance Theory, AI ethical risks, and social influence that influence AI mistrust and intention to resist, which lead to negative outcomes such as negative word-of-mouth and customer disloyalty in the context of digital financial planning services in Thailand. The research collected data from a sample of 420 persons and the data was analyzed using PLS-SEM. The research identified social influence and the risks associated with AI transparency and accountability as primary factors contributing to AI mistrust, whereas privacy risk serves as a more fundamental catalyst for resistance. This resistance contributes to negative word-of-mouth and leads to customer disloyalty. It emphasizes that developing sustainable AI financial advisors must go beyond technically secure design to transparent, accountable, and socially legitimate governance to maintain long-term relationships with customers in the digital financial system. Full article