Search Results (1,617)

The original ALPHA benchmark introduced a taxonomy-aware penalty for evaluating CWE-level vulnerability prediction in Python and proposed that the penalty could theoretically also serve as a training signal. This paper tests that proposal empirically, comparing three delivery mechanisms–supervised fine-tuning, a dual-head classification loss, and reinforcement learning with a dense reward derived from the normalised penalty. We find that supervised approaches consistently regress below the zero-shot baseline under distribution shift, while GRPO succeeds and generalises across independent training seeds. Using a default, non-test-informed configuration, GRPO reduces the mean cumulative ALPHA penalty of Qwen2.5-Coder-7B-Instruct on Security Hardening and Adversarial Testing (SVEN) by 13.5%. A further test-selected configuration reduces the mean penalty by 22.9% (27.9% in the best-observed run), reaching statistical parity with its 4.5× larger zero-shot teacher. Our experiments show that reinforcement learning outperforms the alternative delivery mechanisms we test. We conclude that delivery mechanism, rather than the specific hierarchical structure of the reward, is the primary variable—though whether taxonomy-aware structure contributes further remains an open question. Full article

Internet of Things(IoT)/edge intrusion-detection systems operate on distributed traffic and system-state data whose distributions vary across gateways, services, and attack conditions. We study a server-assisted federated setting in which a teacher reference is fitted from a permitted server-accessible training pool and explicitly distinguish this simulation assumption from fully decentralized deployment. The proposed framework evaluates progressive local training through boundary stabilization, confidence-weighted decision distillation, representation alignment, and validation-quality-aware aggregation. The evaluation uses a leakage-controlled protocol: server and client validation subsets are held out before federated training, update quality and early stopping use validation data only, and the final-test split is evaluated once. Results on NSL-KDD, CIC-IDS2017, Edge-IIoTset, and the ToN-IoT network dataset show competitive primary performance and stronger robustness in several severe label-skew settings. On the Telemetry of Things(ToN-IoT) with Dirichlet alpha = 0.1, the proposed method achieves 91.46 ± 5.54 F1, compared with 53.73 ± 49.00 for FedAvg and 53.77 ± 48.92 for FedProx. The results do not establish universal superiority or a universally optimal stage order: competing methods remain stronger in selected stable and attack-shift settings. The framework is therefore presented as a bounded, server-assisted robustness-oriented training strategy for heterogeneous IoT/edge intrusion detection. Full article

Background: Motivation for academic reading is a relevant component of initial teacher training. However, little is known about whether distinct motivational profiles exist in association with patterns of substance use and problematic behaviors among prospective teachers. Methods: A cross-sectional study was conducted with 289 students (77.5% female; aged 18–34) enrolled in Early Childhood Education and Primary Education degree programs. Motivation for academic reading was assessed using the Academic Reading Motivation Scale (EMLA), while substance- and behavior-related problems were evaluated using the MULTICAGE CAD-4 questionnaire. Latent profile analysis was used to identify academic reading motivation profiles, and multivariate and post hoc analyses were performed to examine differences in scores across profiles. Results: Three motivational profiles were identified, characterized by high, intermediate, and low levels of academic reading motivation. Significant differences between profiles were observed regarding problems related to alcohol, drug use, gambling, video games, and compulsive spending. Conclusions: Academic reading motivation profiles were associated with various screening indicators for substance use and problematic behaviors. Given the cross-sectional and self-report nature of the study, these findings should not be interpreted in terms of causality. The results highlight the potential value of jointly considering academic motivation and health-related behavioral indicators in future teacher training. Full article

This article presents a systematic review of recent literature on classroom pedagogy adopted in early and primary education to address linguistic diversity in early childhood and primary education. The study aims to answer the research question: ‘What does the literature say about current pedagogical practices and trends in addressing multilingualism in early and primary education?’ The review synthesizes evidence from a period of five years (2019–2024), utilising keywords such as “multilingualism,” “classroom pedagogy,” and “early and primary education.” From an initial screening of 806 studies, a total of 46 studies focusing on pedagogical practices in early childhood and primary education were included in this review. Findings highlight a shift towards multilingual pedagogies that recognise and build on learners’ linguistic and cultural strengths, particularly through translanguaging, multilingual and multimodal scaffolding, culturally responsive teaching, and practices that promote student agency, identity, and well-being. Emerging trends emphasise flexible use of learners’ full linguistic repertoires, digital and multimodal learning, and more formative and inclusive approaches to assessment, while highlighting the need for sustained teacher professional development and systemic changes to curriculum, assessment, and language policy to move beyond persistent monolingual norms. These findings have practical and policy implications, highlighting the need to strengthen teacher professional learning and develop curriculum, assessment, and language policies that embed multilingualism as a resource for learning. This synthesis contributes to a deeper understanding of how educational systems can adapt to and embrace linguistic diversity, ultimately improving educational outcomes for early and primary school-aged learners. Full article

Education for sustainable development (ESD) has gained increasing attention over the last few decades, as human innovation and prevailing consumer practices have driven significant environmental degradation, resource depletion, and social inequalities. ESD must extend beyond the transmission of knowledge; it must shape mindsets, behaviors, and systems. Accordingly, fostering sustainability competencies among primary and secondary school students has become an educational imperative. The aim of the study was to identify opportunities for developing eight key competencies for advancing sustainability transformations within obligatory and elective chemistry curricula in Slovenian lower and upper secondary schools. A qualitative content analysis of chemistry curriculum documents was conducted, focusing on general learning objectives, operational learning objectives, and associated learning outcomes. The results indicate that general learning objectives of the obligatory and elective chemistry curriculum for lower and upper secondary educational levels enable opportunities for the development of all key competencies for advancing sustainability transformation, whereas the extent to which these competencies can be developed through operational learning objectives and learning outcomes varies across chemistry curricula and educational levels. However, despite the existing possibilities in curricula, teachers play a crucial role in their implementation in classrooms. Therefore, it is essential to support teachers in recognising the importance of developing these competencies and meaningfully selecting didactic approaches that foster the development of all key competencies for advancing sustainability transformation within the subjects they teach during pre-service and in-service education. Full article

Spatial literacy and spatial thinking are fundamental skills for understanding space, interpreting geographical phenomena, and enabling active citizenship in a society increasingly driven by geospatial information. This review examines how digital tools and geospatial technologies contribute to cultivating spatial literacy and thinking within primary geography education. Following an integrative, PRISMA-informed review methodology combining searches of Scopus, Web of Science, and ERIC (n = 1373 records identified), this study draws on a corpus of 141 studies—comprising an original set of sources supplemented by newly identified, priority-ranked literature—examining digital maps, Geographic Information Systems (GIS), Web GIS, StoryMaps, collaborative mapping, digital games, and augmented and virtual reality in geography teaching. This review proposes an explicit conceptual model clarifying the hierarchical and complementary relationships among spatial skills, spatial thinking, critical spatial thinking, spatial literacy, and spatial citizenship, together with a four-dimensional typology distinguishing technological platforms, representational formats, pedagogical approaches, and learning activity types; both are proposed as synthesizing heuristics for organizing the reviewed evidence and are offered as a basis for future empirical validation rather than as established models. The findings suggest that pedagogically grounded digital integration strengthens spatial skills, shifts learning toward active, participatory models, and helps cultivate spatial citizenship, though the strength of this evidence varies considerably by tool and by educational level, with primary-specific support remaining comparatively limited for higher-order outcomes such as critical spatial thinking. Crucially, the effective integration of these technologies depends on structured teacher training—addressing self-efficacy and confidence-related barriers alongside technical skills—and the adoption of innovative instructional approaches. Ultimately, digital tools provide effective means of deepening geographical learning and strengthening spatial literacy in primary schools. Realizing this potential, however, depends on matching the selection and use of each tool to the specific dimension of spatial learning it is evidenced to support. Full article

The physical and multisensory conditions of classrooms influence students’ well-being, motivation, and learning, yet much of the available evidence comes from high-income contexts and from primary education, and it rarely evaluates integrated interventions in settings with fewer spatial and social resources. This study assessed the effect of a human-centered, multisensory architectural classroom intervention, developed through a participatory process with students and teachers, on students’ life satisfaction, school satisfaction, school motivation, and classroom satisfaction in a public school in Greater Concepción, Chile. A quasi-experimental repeated-measures design with three assessments (baseline, pre-intervention, and post-intervention) was used with students aged 12 to 14; the intervention was implemented in two classrooms of traditional typology (54.5 m² each) and analyzed as a single condition. The results showed significant improvements in classroom satisfaction and school motivation, whereas the more global dimensions, school satisfaction and life satisfaction, did not change. The effects followed a proximity gradient: they were largest for the dimension most directly linked to the modified space and attenuated toward the more distal dimensions. The findings indicate that classroom-scale interventions were associated with improvements in the school experience of adolescents. Moving toward a design that goes beyond the merely functional to incorporate the sensory dimension and elements that afford flexibility and choice within constrained spatial conditions constitutes a concrete pathway to improve that experience, particularly in resource-limited public schools. Full article

This study identifies two stage-specific configurations of CLIL teacher well-being in Andalusia: an ecology of proximity in Primary Education and a disciplinary-departmental ecology in Secondary Education. Drawing on 14 semi-structured interviews with seven Primary and seven Secondary teachers, the study used a deductive–inductive ecological analysis combining within-case, cross-case, cross-stage, and subsystem-coupling procedures in QDA Miner. Participants described professional well-being not merely as a subjective psychological state but in terms of professional viability: the perceived ecological enactability of pedagogically legitimate and sustainable practice. Primary teachers’ accounts foregrounded sustained pupil contact, tutorial responsibility, family expectations, classroom routines, and adaptation for younger learners. Secondary teachers’ accounts foregrounded subject specialisation, departmental organisation, heterogeneous language proficiency, assessment ambiguity, and the Primary–Secondary transition. Across both stages, participants described compensatory overfunctioning—additional care, coordination, material redesign, translation, and assessment work undertaken to keep CLIL operational when implementation conditions were perceived as insufficient. Protected coordination time, stage-specific training, shared assessment frameworks, appropriate materials, and transition structures emerged as potential regulatory leverage points. The study advances a phase-sensitive systems interpretation of teacher well-being by identifying a fast balancing loop of frontline adaptation and a slower reinforcing stabilisation trap through which restored continuity may conceal persistent underprovision. Full article

Consistency verification of drilling reports is critical for engineering quality control because a single data item may be distributed across reports with different formats, units, and page structures. Existing retrieval-augmented generation methods remain sensitive to retrieval and parsing errors in such documents, whereas ultra-large models impose substantial local computing and memory costs. This study proposes a lightweight tool-augmented framework based on dual-teacher distillation and sparse mixture-of-experts (MoE) modeling. Qwen3-235B-A22B serves as the primary teacher and Qwen3-30B-A3B as the assistant teacher. Their tool-use and task-planning capabilities are transferred to a sparse MoE student upgraded from a Qwen3-1.7B dense backbone through trajectory pruning, sample decomposition, and token-level Kullback–Leibler (KL) distillation. The student adopts an eight-expert Top-2 routing architecture. Experiments on 1000 drilling reports containing 30,127 verification instances show an F1 score of 58.0 ± 0.5%, with file-level, location-level, and exact-match accuracies of 66.5%, 55.2%, and 45.0%, respectively. The model contains 9.1B total parameters and 2.8B activated parameters, and reaches a latency of 12.1 ms per forward pass and a memory footprint of 18.4 GB under bfloat16 (BF16) precision. The reported F1 score characterizes the end-to-end verification task rather than an autonomous safety decision capability. The framework is intended to support evidence localization, anomaly prioritization, and expert review in local deployment settings. Full article

This exploratory, classroom-based mixed-methods study examined a five-phase, AI-supported multimodal literacy sequence implemented with 25 Grade 4 students over four weeks. The intervention combined interactive digital reading, collaborative narrative writing, AI-supported comic imagery, multimedia editing, and peer reflection. Quantitative analyses compared students’ baseline and post-intervention comprehension scores and summarized responses to a 12-item post-intervention questionnaire. Qualitative data from open-ended responses, classroom observations, and student artifacts were analyzed thematically. Students’ mean comprehension scores increased from 63.81 (SD = 9.70) at baseline to 75.37 (SD = 7.63) after the intervention, representing a significant within-group change, t(24) = 5.49, p < 0.001, dz = 1.10. Questionnaire responses were generally positive, particularly for original story writing, digital comic creation, and the addition of movement and sound. Three qualitative themes were identified: AI-supported representations as scaffolds for narrative comprehension, as catalysts for authorial identity and creative agency, and as mediators of collaborative meaning-making. Because the study used a single-group design and the comparability of the parallel assessment forms was supported through expert content review rather than large-scale psychometric equating, the findings should be interpreted as exploratory evidence of association rather than as causal effects of AI. The study nevertheless illustrates how developmentally appropriate, teacher-mediated AI activities may support multimodal literacy participation in primary education. Full article

Research exploring the relationship between emotional intelligence and workplace outcomes is predominantly variable-centered and treats the dimensions of emotional–social competencies as separate from one another, thereby losing sight of the intra-individual interplay of these competencies and the potential heterogeneity of their effects. This study examines teachers’ emotional–social intelligence using a person-centered approach. Using a sample of 587 Hungarian teachers from 26 public education institutions, we conducted latent profile analysis based on Bar-On’s five composites and then used unbiased three-step procedures (BCH; categorical BCH) to examine the antecedents of profile membership, as well as perceptions of school organizational climate and dominant conflict management style. Four distinct profiles emerged, primarily differing at the global level—a uniformly low profile (12%), a medium-low profile (25%), a dominant, balanced medium-high profile (53%), and a uniformly high profile (9%). Participants with higher profiles perceived the school’s organizational climate as more supportive (η² = 0.07) and exhibited more adaptive conflict management patterns (less avoidance, more problem-solving; χ²(12) = 37.24, p < .001). A compositional (log-ratio) reanalysis of the complete conflict-mode profiles, reported as the primary conflict analysis, showed that the relative weight of problem-solving rises across profiles and plateaus between the modal and the highest profile, that avoidance declines with profile elevation, and that an apparent surplus of accommodating in the highest profile was an artifact of dominant-style coding rather than an indication of a broader repertoire. At the same time, profile membership was essentially independent of job position, career stage, and gender, and the relationships with outcomes were of modest to moderate strength. The results extend person-centered, profile-based intelligence research to the emotional–social domain and identify promising directions for research on teacher professional development. Full article

Teacher resilience (TR) is essential for teacher well-being, professional commitment, and educational quality. Although recognition, authority, and resilience have received considerable scholarly attention, limited research has examined how recognition and professional authority interact to support TR. Guided by Self-Determination Theory (SDT), this exploratory qualitative study investigated these dynamics through semi-structured interviews with 12 purposively selected teachers (≥3 years’ teaching experience), 12 purposively selected Primary 4–6 students (aged 10–13), and 12 purposively selected parents from selected urban public primary schools in Kigali, Rwanda. Data were analyzed using thematic analysis. Findings suggest that TR is a relational process shaped by interactions among recognition, authority, and supportive professional relationships. Recognition was perceived to foster teachers’ sense of relatedness through student success, parental trust, collaborative engagement, and collegial affirmation, while authority was associated with competence and autonomy through pedagogical expertise, fairness, and care. Based on these findings, the study proposes a preliminary SDT-informed interpretive framework in which recognition and authority function as interconnected relational resources that may support TR. By integrating recognition and authority into a single explanatory framework, this study contributes to TR scholarship and offers insights for strengthening teacher well-being and professional sustainability in resource-constrained and evolving educational contexts. Full article

This study examines pre-service teachers’ self-efficacy in teaching real-world problems and mathematical modelling. Using a quantitative design, it investigates variation in self-efficacy between primary and special education pre-service teachers across different competence facets. The findings reveal a differentiated, domain-specific pattern of self-efficacy rather than uniform differences between groups. Special education pre-service teachers reported higher self-efficacy in pedagogical content knowledge and task design and differentiation, whereas no significant differences emerged for situation-specific skills and instructional design and enactment. The study contributes to a more differentiated understanding of modelling-related self-efficacy by showing that domain specificity may also operate within a specific area of mathematics teaching. It highlights the relevance of considering different competence facets in teacher education for inclusive mathematics classrooms. Full article

Primary CLIL teacher wellbeing remains underexplored in France and Spain. This exploratory, context-specific qualitative study examined the factors that ten primary teachers, five in France (Montpellier and Lille) and five in Andalusia, associated with their wellbeing. Bronfenbrenner’s ecological framework, enriched with the ontosystem and the flourishing/floundering distinction, was used as an interpretive device. The interviews were analysed through reflexive thematic analysis. Participants’ accounts converged around perceived gaps in CLIL-specific training, time, coordination, and human and material resources, while differing in how they interpreted programme stability and recent policy change. The analysis tentatively developed a compensatory pedagogy of affection and a broader pattern of professional delegitimation and identity-related loss. Explicit bilingual professional shame appeared in one account. It also identified native-speaker-oriented linguistic self-demand and an emerging pattern of AI-assisted material adaptation across six accounts. We offer a provisional, context-specific tensional ecological framework to organise the relationships participants described between conditions perceived as structural pressures and personal, relational, formative, and material resources. The framework is intended as a heuristic and hypothesis-generating contribution rather than a validated causal or longitudinal model. The findings identify areas for institutional review and further research. Full article

The integration of Information and Communication Technologies (ICT) into face-to-face compulsory education has accelerated globally, but the conceptual definition and pedagogical application of Digital Education (DE) remain unclear. This systematic review explores how DE is defined, implemented, and assessed in in-person primary and secondary school settings. Guided by PRISMA methodology, 33 peer-reviewed empirical studies published between January 2012 and June 2025 were selected from Scopus and Web of Science, and through reference checking. Methodological quality and potential sources of bias were appraised using the Mixed Methods Appraisal Tool (MMAT), and that appraisal informed the interpretation of the evidence. The results indicate significant ambiguity surrounding the term DE, which is often conflated with ICT use or digital competence. While many studies report positive outcomes in student engagement and basic skill acquisition, few articulate clear operational frameworks or pedagogical strategies for effective implementation. DE is predominantly applied to low-cognitive-demand tasks and is underused in promoting deeper learning. Key barriers include insufficient teacher training, fragmented educational policies, and infrastructural limitations. Nonetheless, when digital tools are intentionally combined with pedagogical goals—such as personalized instruction, learner autonomy, and content creation—DE demonstrates potential to improve motivation and support achievement. The findings suggest that well-designed DE can contribute to educational sustainability by strengthening learning opportunities, supporting equity and inclusion, and advancing Sustainable Development Goal 4 (Quality Education), but they also highlight the need for coherent policies, teacher training, and equitable access to digital resources. Full article