Search Results (477)

Background/Objectives: Communication decline is a hallmark of dementia, yet speech-language outcomes remain marginal in much of the virtual reality (VR) dementia literature. This evidence review synthesises empirical work on how VR has been used to support, train, or assess communication in dementia, positioning VR as a communication platform rather than only a cognitive tool. Methods: A structured search (2000–2025) across CINAHL, PubMed, PsycINFO, Scopus, and Web of Science was supplemented by reference list checking. Eleven empirical studies met eligibility criteria, spanning immersive and non-immersive VR used with people living with dementia, and VR-based communication training for caregivers, care staff, and clinicians. Findings were synthesised thematically through an explicit communication lens. Results: Evidence most consistently supports VR as a scaffold for communicative engagement and participation. Immersive and shared VR experiences commonly elicited increased verbal involvement, shared attention, and interactional responsiveness during or immediately after sessions, particularly when content was socially meaningful and appropriately paced. A second strand of work uses VR simulation to train communication partners, with participants reporting high acceptability and perceived improvements in confidence and strategy use, although behavioural transfer to real-world care is rarely measured. Assessment-oriented studies and stakeholder perspectives highlight VR’s potential to elicit functional behaviour in context and to complement clinic-based assessment, but communication validity is typically inferred rather than operationalised using standardised measures. Conclusions: VR shows early promise for dementia communication care, especially as an adjunct that structures interaction, supports participation, and scales communication training. Progress now depends on communication-specific intervention design, agreed outcome metrics capturing discourse and functional participation, and implementation studies addressing accessibility, cultural-linguistic diversity, and transfer to everyday care. Full article

Implicit discourse relation recognition (IDRR) addresses the classification of discourse relations between text segments without explicit connectives. Existing prompt-based methods for IDRR often rely heavily on predicting surface connectives as an indicator for the discourse relation, which is inherently limited by the capacity of pre-trained language models. Meanwhile, standard attention mechanisms in these models are easily distracted by task-irrelevant tokens. This paper proposes a Role-Focused Prompt Framework that addresses these limitations by introducing a role-centric perspective to IDRR. Our approach is built on two core innovations: (1) the incorporation of linguistically grounded semantic roles (e.g., Cause/Effect for Contingency relation) into IDRR, which directly captures the underlying argument structure that determines discourse relations, reducing reliance on connectives; (2) a focused prompt structure that condenses the input to its core semantic concepts (argument summaries, connective, and semantic roles), creating a high signal-to-noise environment for attention-based reasoning. Extensive experiments on Penn Discourse TreeBank 2.0 (PDTB 2.0) demonstrate that our framework achieves competitive results, providing complementary direction for IDRR research. Ablation studies validate that both innovations are essential to the framework. Our work demonstrates that incorporating linguistically grounded semantic roles and focusing on task-relevant concepts can effectively specialize pre-trained models for IDRR. Full article

The rapid growth of digital communication platforms has generated vast volumes of user-generated textual data and digital footprints, creating growing demand for scalable artificial intelligence systems capable of supporting evidence-based decision-making. This study proposes and evaluates a human–AI collaborative analytical pipeline for multi-class sentiment and aggression analysis of large-scale social media data (N = 15,064 messages) related to an urban infrastructure project. The proposed framework integrates standard NLP preprocessing, machine learning-based classifiers, temporal aggregation, and controlled large language model (LLM)-assisted classification within a structured analytical workflow that incorporates expert validation and oversight. A stratified manual validation procedure (n = 301) demonstrated substantial inter-annotator agreement (κ = 0.70) and stable multi-class classification accuracy (80%). The results indicate that combining sentiment polarity and aggression detection as complementary linguistic indicators improves sensitivity to shifts in discourse dynamics and enables early identification of emerging social tension. The study demonstrates the potential of human–AI collaborative analytical frameworks for transparent, interpretable, and predictive large-scale social media analysis in decision-support contexts. Full article

Early identification of language and literacy disorders (LLDs) in multilingual children remains a challenge in linguistically diverse educational systems shaped by ongoing migration. In many contexts, including Greece and Cyprus, where LLDs have been poorly investigated, teachers lack screening tools that can reliably distinguish typical multilingual development from possible indicators of LLDs. This study presents the development and preliminary piloting of two teacher-report screening questionnaires for multilingual children aged 4–6 and 6–9 years, designed for use in everyday classroom settings to support early identification and referral. A structured multi-stage procedure guided development. First, items were derived from internationally established clinical markers of multilingual LLDs, covering oral language, phonological awareness, communication, literacy, and related cognitive domains. Second, a scoring framework was created to support consistent, referral-oriented interpretation across languages. Third, the questionnaires were reviewed by specialists in linguistics, education, and speech-language therapy. Fourth, pilot testing with teachers evaluated clarity, feasibility, and classroom relevance. Expert and teacher feedback indicated that the questionnaires are practical and support differentiation between multilinual language differences and potential underlying difficulties. Overall, this study introduces two promising cross-linguistic screening tools for educators in multilingual educational settings, currently undergoing psychometric validation. Full article

Remote sensing visual question answering (RS-VQA) is essential to intelligent Earth observation, as it supports interactive querying of high-resolution aerial images. Many existing methods struggle with fine-detail geospatial reasoning with remote sensing (RS) scenes due to RS scenes having intrinsic multi-scale object variance and pronounced spatial heterogeneity. The models tend to rely more on the linguistic prior than reasoning based on visual evidence. In this paper, we present PMA-VQA, a progressive multi-scale feature fusion with spatially adaptive attention, to embed the RS-VQA task in spatially based hierarchical feature integration. For hierarchical, multi-level, language-informed integration, we propose a spatial attention aggregation module (SAAM) and a progressive feature fusion and classification module (PFCM). The SAAM employs spatially adaptive gating to align cross-modal features with semantic context, while the PFCM integrates multi-scale representations across high-level semantic abstractions and low-level space. The experimental results on RS-VQA LR and HR benchmarks validate that PMA-VQA outperformed all competing methods in terms of accuracy and robustness. Evaluation of HRVQA further confirmed the effectiveness of the SAAM and PFCM across diverse RS scenes. Full article

The Automated Essay Scoring (AES) systems confront two fundamental challenges: opaque “black-box” decision-making that limits educator trust, and insufficient validation across linguistically diverse educational contexts. This study proposes IRT-AESF, an innovative framework that bridges educational measurement theory and artificial intelligence by integrating item response theory (IRT) with deep learning. The framework generates three theoretically grounded psychometric parameters: student ability, item difficulty, and item discrimination, which provide transparent and interpretable explanations for scoring decisions. We rigorously evaluated IRT-AESF through 5-fold cross-validation on three large-scale datasets comprising 41,328 authentic essays from English and Chinese educational settings, including both classroom assessments and high-stakes examinations. Results demonstrate statistically significant improvements over competitive baseline models, achieving an 8.4% relative increase in quadratic weighted kappa while maintaining robust cross-lingual performance. This research advances the development of transparent, trustworthy automated assessment systems that deliver not only scores but meaningful diagnostic insights for educational practice. Full article

In the post-pandemic higher education context, multilingual students, particularly those from widening participation backgrounds, continue to face academic, linguistic, and socio-emotional challenges that can limit their participation and sense of belonging. This study examines the use of Reflective Video Diaries (RVDs) facilitated through Microsoft Flipgrid as an inclusive pedagogical approach to support reflective engagement, communication, and socio-emotional development among multilingual undergraduate students. Adopting a qualitative iterative action research approach, the study was conducted within a UK university module and involved three cycles of implementation, reflection, and pedagogical refinement, capturing students’ lived experiences rather than measuring causal effects. Multiple methods, including RVDs, end-of-module reflective reports, an anonymous survey, and lecturers’ field notes, were deliberately combined to provide complementary perspectives on students’ experiences, allowing triangulation of data and enhancing the validity and richness of findings. Thematic analysis of this longitudinal dataset collected across the three action-research cycles explored how students experienced RVDs as a space for reflection, peer support, and engagement with learning. Findings indicate that Flipgrid-mediated RVDs functioned as a low-anxiety, flexible, and dialogic learning environment that enabled students to articulate challenges, share progress, and develop reflective awareness, confidence, and a sense of connection with peers and lecturers. Improvements in participation and reflective depth were more evident in later cycles, suggesting that benefits emerged through iterative pedagogical adjustment rather than by video technology alone. Both positive experiences and challenges are reported, providing a balanced account of engagement with the RVDs. The study underscores the potential of inclusive digital pedagogies to inform curriculum planning and policy implementation, supporting equitable learning opportunities and socio-emotional development. By conceptualizing RVDs as relational and inclusive pedagogical practices rather than technological interventions, and by demonstrating how reflective engagement developed across successive action-research cycles, this research contributes to understanding how reflective digital practices can support multilingual learners’ academic and socio-emotional development within socially just higher education contexts. Practical implications for designing inclusive reflective learning environments are discussed. Full article

Research on academic achievement highlights the combined role of cognitive abilities and motivational beliefs. Grounded in the CHC framework, this study examined how three broad cognitive abilities—verbal, numeric, and spatial—and academic self-concept jointly predict achievement in Portuguese and mathematics. A sample of 3034 students from the third cycle (grades 7–9) and secondary education (grades 10–12) completed the BAC-AB cognitive battery and a validated academic self-concept scale. Using multigroup structural equation modelling, we tested whether the predictive patterns differed across educational stages. Academic self-concept emerged as the most consistent predictor across subjects and levels. Cognitive contributions displayed clear developmental differentiation: verbal ability was more strongly associated with Portuguese (and increasingly with Mathematics) in secondary education, whereas numeric and spatial abilities were comparatively more relevant for Mathematics in the third cycle. These patterns support the view that linguistic, quantitative, and visuospatial processes contribute to achievement in distinct and developmentally sensitive ways. Overall, the findings underscore the importance of instructional approaches that build on quantitative and spatial strengths in earlier grades while progressively supporting advanced verbal comprehension and reasoning in later schooling. Full article

Driven by the development of next-generation wireless networks and the widespread adoption of sensing, communication is shifting from traditional bit-level transmission to intelligent, rich interactions within our digital social system. However, existing speech semantic communication frameworks predominantly focus on textual accuracy, neglecting the critical affective information (e.g., tone and emotion) that is essential for natural human-centric interactions in the real world. To address this limitation, we propose the Heterogeneous Affective Speech Semantic Communication (HASCom) framework, designed for the robust transmission of highly expressive speech over complex wireless channels. Specifically, we design a heterogeneous dual-stream transmission architecture that decouples discrete phoneme-level linguistic content from continuous emotional embeddings. For discrete semantic information, we use reliable digital coding protected by Low-Density Parity-Check (LDPC) to guarantee strict recoverability. Conversely, for emotional features, we employ Deep Joint Source-Channel Coding (JSCC) analog transmission to prevent irreversible quantization errors and the cliff effect. Additionally, we develop a prior-guided diffusion reconstruction module at the receiving end. This module leverages a structural prior network to align the decoded semantics, which then steers the reverse diffusion process conditioned on the recovered affective features. Extensive experiments under both AWGN and Rayleigh fading channels demonstrate that HASCom significantly outperforms state-of-the-art baselines. Specifically, it achieves superior objective semantic similarity and subjective Mean Opinion Score (MOS) at low Signal-to-Noise Ratios (SNRs), while the JSCC transmission modules maintain an ultra-low inference latency of less than 0.1 ms, validating its high efficiency and robustness for practical deployments. Full article

Objective: Anxiety symptoms frequently occur alongside mood disorders and are associated with poorer clinical outcomes, highlighting the importance of early and accurate detection. This study evaluated the diagnostic accuracy and psychometric properties of the Anxious Distress Assessment Scale (ADS), a newly developed brief self-report instrument designed to detect anxious distress. Method: The study was conducted in two phases. Phase 1 involved the development of the ADS as a five-item instrument reflecting the DSM-5-TR anxious distress criteria. In Phase 2, 105 adults diagnosed with major depressive disorder (MDD) completed the ADS alongside the Generalized Anxiety Disorder-7 (GAD-7) and the Montgomery–Åsberg Depression Rating Scale (MADRS). Psychometric evaluation included internal consistency reliability (Cronbach’s α), analyses of convergent validity, and diagnostic accuracy assessment using correlation and receiver operating characteristic (ROC) analyses. Results: Anxious distress was highly prevalent, with 71% of participants meeting DSM-5-TR criteria. The ADS demonstrated strong diagnostic performance, with sensitivity of 88.0%, specificity of 90.0%, positive predictive value of 95.7%, and negative predictive value of 75.0%. ROC analysis yielded an area under the curve (AUC) of 0.97 (95% CI: 0.943–0.997), with an optimal cut-off score of ≥10. Internal consistency was excellent (Cronbach’s α = 0.897). Principal component analysis supported a unidimensional structure, accounting for 71.5% of the total variance, with all items loading above 0.80. The ADS also demonstrated strong convergent validity, correlating significantly with the GAD-7 (r = 0.82) and MADRS (r = 0.68). Conclusions: The ADS demonstrates promising psychometric properties, including strong reliability, meaningful convergent validity, and excellent diagnostic accuracy. Its brief format and direct alignment with DSM-5-TR anxious distress criteria support its potential utility as a practical screening tool in clinical settings. However, these findings should be interpreted in light of the study’s focus on English-speaking Malaysian adults with MDD recruited from a tertiary-care setting. Further validation across diagnostic groups, clinical contexts, and cultural and linguistic populations is warranted. Full article

Generative Artificial Intelligence (GAI) is rapidly transforming higher education by introducing new mechanisms for supporting the development of advanced cognitive processes and research-related capabilities. This study examines how postgraduate students employ GAI to develop their cognitive and research talent, conceptualized here as higher-order academic skills such as analysis, synthesis, and critical reasoning, across six domains: literature review, theoretical development, research design, data analysis, academic writing, ethical use, and challenges encountered—signaled explicitly rather than listed line by line. We administered a validated multidimensional scale to 214 postgraduate students, and the results indicate a moderate overall use of GAI, with notably high involvement in practices that emphasize ethics and responsibility. Students reported clear cognitive benefits in tasks involving information processing, linguistic refinement, and conceptual clarification while showing caution toward delegating higher-order analytical or theoretical reasoning to AI systems. Key challenges included limited institutional training, concerns about data privacy and academic integrity, and difficulties evaluating the originality and reliability of AI-generated content. Inferential analyses indicated significant differences based on gender, academic level, and general technology proficiency, whereas no differences emerged across age groups, departments, or specializations. Overall, this study demonstrates how GAI can contribute to the development of higher-level cognitive skills and research competencies, with “moderate use” operationalized as consistent but selective engagement across domains, while underscoring the need for structured training, clear guidelines, and teaching approaches that foster the responsible and effective incorporation of AI within postgraduate research. The results highlight practical implications for higher education, including the importance of institutional training programs, governance frameworks for responsible AI use, and pedagogical models that foster critical engagement with GAI. Full article

Background/Objectives: The Diet Quality Questionnaire (DQQ) is a brief, food group–based instrument designed for globally comparable population surveillance of diet quality. We culturally adapted the DQQ for Poland and evaluated its internal structure and reliability in an adult cohort. Methods: Following forward–backward translation and expert review, the Polish DQQ was administered online to adult females. Internal structure was explored and test–retest reliability was assessed for total DQQ scores. Diet quality indicators (Dietary Diversity Score [DDS], NCD-protect, NCD-risk, and Global Dietary Recommendations score [GDR]) were summarized descriptively. Results: The average age in the cohort was 29.4 ± 13.6 years. A total of 296 respondents completed the survey; 100 completed the retest. Item-level test–retest reliability was good to excellent (Cohen’s kappa 0.72–1.00). Agreement for total scores was high with minimal bias (Bland–Altman bias 0.2, >95% of observations within limits of agreement) and there was no heteroscedasticity; Passing–Bablok regression indicated equivalence between the test and retest. Median (IQR) diet quality indicators were: DDS 6.0 (5.0; 7.0), NCD-protect 2.5 (1.5; 4.0), NCD-risk 2.5 (1.0; 4.0), and GDR 9.0 (7.5; 10.5). Eighty percent met DDS ≥ 5, while one-third consumed all five recommended food groups. Conclusions: DQQ-PL demonstrates high item-level stability and strong agreement for total scores, with structural findings aligning with its design as a non-latent, food group checklist for population monitoring. The Polish adaptation is feasible and reliable in the studied population (young adult women), supporting its potential use for rapid dietary surveillance pending broader validation. Full article

This article examines how artificial intelligence and algorithmic systems are reconfiguring truth regimes in digital societies, introducing the concept of “Artificial Truth” to describe an emerging form of epistemic governance where knowledge production and validation become infrastructural functions of sociotechnical systems. The study develops an integrated theoretical framework combining Foucault’s notion of truth regimes, Bourdieu’s theory of symbolic capital and fields, and Actor-Network Theory’s constructivist approach. Through conceptual analysis, the article investigates how algorithmic recommendation systems, generative AI, and automated fact-checking operate as epistemic devices that actively shape what is recognized as credible, authoritative, and true in public discourse. The analysis reveals three fundamental transformations: (1) the restructuring of trust economies, with epistemic authority shifting from institutional expertise to platform-native capital based on engagement metrics and affective proximity; (2) the emergence of generative AI as an epistemic actor producing “synthetic truth” through linguistic fluency rather than propositional understanding; (3) the institutionalization of computational veridiction in algorithmic fact-checking systems that translate situated epistemic judgments into probabilistic classifications presented as neutral. These dynamics configure a regime where truth is evaluated less by correspondence with reality and more by computational plausibility and platform integration. The article’s primary contribution lies in providing a unified theoretical framework for understanding contemporary transformations of epistemic authority, moving beyond disinformation studies to analyze AI as an epistemic actor. By integrating classical sociological perspectives with Science and Technology Studies, it conceptualizes algorithmic systems as epistemic infrastructures that embody specific power relations, restructure symbolic capital economies, and distribute epistemic authority asymmetrically, with profound implications for democratic knowledge, citizen epistemic agency, and public sphere pluralism. Full article

Security organizations increasingly rely on cyber threat intelligence (CTI) sharing to enhance their resilience against cyberattacks. Indicators of Compromise (IoCs) play a critical operational role in CTI by providing malicious artifacts that support threat detection, incident response, and facilitate proactive defense. However, the rapid growth of social media as CTI sources, characterized by short-text content, poses significant challenges to automated IoC extraction, contextual interpretation, operational integration, and reliable verification. To address these challenges, this study proposes a comprehensive framework that integrates Large Language Models (LLMs) across multiple stages of the CTI pipeline. The framework leverages LLM-driven data augmentation, a hybrid classification model, and contextual summarization to enhance short-text understanding while supporting expert-in-the-loop validation for operational reliability. Extensive experimental evaluations demonstrate that LLM-driven data augmentation substantially improves model robustness and generalization while reducing false-positive alerts, achieving a precision of $98.87$%. Quantitative diversity analysis and expert-based human evaluation further confirm the linguistic quality and correctness of the generated augmented samples. In addition, IoC reports are validated using both reference-based and reference-free evaluation metrics that show strong alignment and high semantic adequacy. Moreover, a technology acceptance model was integrated with cybersecurity domain constructs to assess the acceptance factors of the proposed framework. Regression analysis showed that perceived usefulness, behavioral intention, trust in automation, and risk were the strongest predictors of actual use. These predictors are commonly interpreted as indicators of technology acceptance. Full article

Developing reliable Arabic question answering (QA) systems for Islamic fatwas requires datasets that capture the linguistic complexity and multi-step reasoning inherent in jurisprudential inquiries. However, the existing Arabic religious QA datasets primarily focus on direct retrieval or classification, often failing to address the multi-hop reasoning necessary for complex fatwa questions. To bridge this gap, we introduce MAFQA, a benchmark dataset specifically designed for multi-hop Arabic fatwa question answering. MAFQA was constructed from an extensive corpus of authentic fatwa records sourced from authoritative Islamic institutions. The dataset was developed via a semi-automated pipeline that integrates expert-guided identification of complex inquiries with a structured decomposition framework. This framework employs automated reasoning-pattern classification, semantic feature extraction, and template-guided annotation of subquestions and subanswers, followed by rigorous validation to ensure contextual grounding, logical coherence, and structural consistency. To evaluate the utility of the dataset, we conduct an extensive benchmarking study using Arabic-specialized, multilingual, and instruction-tuned language models across two primary tasks: question decomposition (QD) and generative question answering (QA). Performance is assessed using a comprehensive suite of lexical, semantic, relevance, and faithfulness metrics. Experimental results demonstrate that Arabic-specialized models consistently outperform their multilingual counterparts, with AraT5-base and AraBART achieving the highest performance in terms of lexical similarity, semantic alignment, and answer faithfulness. Full article