Search Results (6,243)

In contemporary debates on the United Nations Sustainable Development Goals, there is growing recognition that artificial intelligence (AI) may contribute meaningfully to SDG 2 (Zero Hunger), particularly by enhancing the efficiency of food aid distribution and resource allocation. However, such optimism must be critically situated within the broader institutional and ethical contexts in which AI operates. This study argues that the effectiveness of AI in conflict-affected settings is contingent not only on technical capacity but also on governance structures, ethical safeguards, and institutional trust, dimensions closely aligned with SDG 16 (Peace, Justice, and Strong Institutions). Using the Gaza Strip as a case study, this article demonstrates that AI-driven food assistance mechanisms may inadvertently reinforce structural vulnerabilities. Specifically, algorithmic targeting of aid risks deepening dependency, exacerbating digital exclusion, and weakening already fragile governance systems. The absence of robust data accountability frameworks further complicates these dynamics, raising concerns regarding transparency, fairness, and long-term sustainability. The findings caution against privileging technical efficiency at the expense of socio-political stability. Rather, they highlight that the sustainability of AI interventions in humanitarian contexts fundamentally depends on the credibility and legitimacy of institutions. Accordingly, this study proposes a conceptual model for AI in hunger relief and digital humanitarianism that integrates technical innovation with institutional accountability and social trust. This study presents a narrative review informed by structural searching that examines the influence of AI on food security interventions in fragile contexts. This analysis applies a combined ethical governance and sustainability lens to assess current applications and risks. This research advances a broader analytical framework that moves beyond purely technical interpretations of AI, emphasizing its role as a socio-political tool, through identifying five key pillars for sustainable AI governance: data sovereignty, algorithmic accountability, inclusive system design, community-led governance, and market integrity. Full article

Emerging technologies (ETs), such as AI and reality techniques, are reshaping educational measurement. However, existing studies remain dispersed and are rarely synthesized in ways that clarify how ETs participate in the evidentiary work of educational measurement. Guided by PRISMA 2020, we systematically reviewed 933 empirical studies published between 2016 and 2025 in formal educational settings. We coded studies by (a) grain size (micro, meso, macro), (b) Constructing Measures Theory building blocks (construct map, item design, outcome space, measurement model), and (c) ET category. Results showed a strong concentration at the micro level (88.88%) and in outcome space and measurement model work (86.80% combined), indicating that ET-enabled innovation has focused primarily on transforming performances into indicators and modeling those indicators for interpretation and decision-making. Learning analytics and educational data mining, machine learning and deep learning, and automated scoring and feedback systems were the dominant ET clusters. These findings point to an uneven development of ET-enabled educational measurement. Included studies also indicating recurring concerns about transparency, fairness, and governance are linked to the field’s main areas of ET-enabled concentration. We therefore argue for closer alignment among construct claims, evidence, modeling, and intended use, and offer implications for developers, researchers, and education practitioners. Full article

Recent research by Yucel and colleagues suggests that children perceive equality-based fairness violations (resources being distributed unequally) as less serious than prototypical moral harms, but that making the harmful consequences of unfairness salient shifts these judgments toward the moral domain. We examined whether merit-based fairness violations (someone receiving less than they earned) would similarly shift judgments toward the moral domain by making the injustice more salient. Replicating prior work, 4-year-old children (N = 62) rated prototypical moral violations as significantly more severe than equality-based fairness violations, which were rated as similar in severity to conventional violations. Contrary to predictions, merit-based fairness violations also showed this pattern: They were judged as less severe than prototypical moral violations and similarly severe as both equality-based fairness violations and conventional violations. Children also did not consistently group either type of fairness violation with moral or conventional violations. These findings contribute to a growing body of evidence that children’s (and adults’) perceptions of fairness—whether equality-based or merit-based—are more nuanced than previously thought and that unfairness may not spontaneously be treated like other, more prototypical moral norm violations. Full article

Objective: The aim was to evaluate the diagnostic accuracy and temporal reproducibility of multimodal large language models (LLMs) in the image-based diagnosis of oral mucosal lesions. Materials and Methods: The study included 100 anonymized clinical photographs of oral mucosal conditions obtained from the archive of the Department of Oral Medicine, School of Dental Medicine, University of Zagreb. Images were categorized into four subgroups: physiological variations, benign mucosal lesions, oral potentially malignant disorders, and oral cancer (25 images each). Three multimodal LLMs (ChatGPT-5.1 Plus, Gemini 3 Pro, and Perplexity Pro) analyzed each image using an identical prompt and were required to provide a single most probable diagnosis based solely on visual features. To evaluate temporal reproducibility, the entire evaluation was repeated in three independent testing cycles conducted at one-month intervals. Diagnostic accuracy was compared using chi-square tests, while intra-model agreement across cycles was assessed using Fleiss’ kappa. Results: Gemini demonstrated the highest diagnostic accuracy, reaching 78% correct responses in cycles 2 and 3, significantly outperforming ChatGPT (55–57%) and Perplexity (28–31%) (p < 0.00001). Subgroup analyses showed similar trends, with Gemini achieving the highest accuracy across most lesion categories. Intra-model agreement across cycles was moderate for ChatGPT (κ = 0.525), fair for Gemini (κ = 0.338) and Perplexity (κ = 0.409). Gemini also showed the highest proportion of responses that remained correct across all three cycles (51%). Conclusions: Multimodal LLMs demonstrate promising diagnostic capabilities in the image-based assessment of oral mucosal lesions; however, variability in reproducibility highlights the need for cautious clinical implementation and further validation. Full article

The Mobile Reliable Opportunistic Routing (MROR) protocol improves data-forwarding reliability in Cognitive Radio Sensor Networks (CRSNs) through mobility-aware virtual contention groups and handover zoning. However, its heuristic decision logic is difficult to optimize under highly dynamic spectrum access and random node mobility. To address this limitation, we present DRL-MROR, a refined routing framework that incorporates deep reinforcement learning (DRL) to enable intelligent and adaptive forwarding decisions. In DRL-MROR, the secondary users (SUs) act as autonomous agents that observe local state information, including primary-user activity, link quality, residual energy, and neighbor-mobility patterns. Each agent learns a forwarding policy through a Deep Q-Network (DQN) optimized for long-term network utility in terms of throughput, delay, and energy efficiency. We formulate routing as a Markov Decision Process (MDP) and use experience replay with prioritized sampling to improve learning stability and convergence. The DQN used at each node is intentionally lightweight, requiring 5514 trainable parameters, about 21.5 kB of weight storage in 32-bit precision, and approximately 5.4k multiply-accumulate operations per inference, which supports practical deployment on edge-capable CRSN nodes. Extensive simulations show that DRL-MROR outperforms the original MROR protocol and representative AI-based routing baselines such as AIRoute under diverse operating conditions. The results indicate gains of up to 38% in throughput, 42% in goodput, a 29% reduction in energy consumed per packet, and an approximately 18% improvement in network lifetime, while maintaining high route stability and fairness. DRL-MROR also reduces control overhead by about 30% and average end-to-end delay by up to 32%, maintaining strong performance even under elevated PU activity and higher node mobility. These results show that augmenting opportunistic routing with lightweight DRL can substantially improve adaptability and efficiency in next-generation IoT-oriented CRSNs. Full article

Fully decoupled radio access networks (FD-RANs) achieve spectral efficiency and coverage flexibility for 6G via independent uplink (UL) and downlink (DL) base station operation, yet dynamic user mobility brings critical challenges to joint user association and resource allocation. Asymmetric interference and heterogeneous base station capacities cause persistent network unfairness, while uncoordinated mobility management triggers ping-pong handovers and heavy handover overheads. To resolve these intertwined problems, we propose a fully decoupled, mobility-resilient and fairness-guaranteed framework, which integrates short-term congestion pricing with the long-term Jain fairness index for equitable resource distribution and introduces a composite handover penalty with a strict physical hysteresis margin to block invalid handovers. We formulate the optimization problem as a novel Sliding-Window Hysteresis-Integrated Fairness Two-Layer Multi-Armed Bandit (SHIFT-MAB) model, embedding an exponentially weighted moving average (EWMA) sliding-window mechanism to track real-time channel fluctuations efficiently. Theoretical analysis confirms the model’s decoupling optimality, sublinear regret bound and fairness convergence. Extensive simulations show that SHIFT-MAB effectively suppresses invalid handovers, ensures high network fairness, optimizes system utility and achieves a superior handover–throughput trade-off. Full article

Geriatric pharmacotherapy is usually challenged by physiological senescence. For instance, progressive declines in organ function and alterations in body composition can complicate drug disposition. However, conventional pharmacometrics models commonly have limited capacity to map these high-dimensional, nonlinear relationships. In this review, we are examining the recent shift toward integrating machine learning (ML) with mechanistic pharmacokinetic (PK)/pharmacodynamic (PD) models to improve the accuracy and precision of dosing. Machine learning approaches like Random Forest and XGBoost consistently provided more accurate exposure predictions and significantly more efficient computational workflows than conventional methods. Nevertheless, concerns such as “black box” transparency and the potential of algorithmic bias toward specific patient demographics are challenging. It is important to incorporate explainability tools like SHAP, and adopting FAIR data principles is crucial for achieving professional trust and ensuring site-specific generalizability. Full article

Recent advances in mass spectrometry, data-independent acquisition, proteoform-resolving workflows, and multi-omics integration have significantly expanded the scale and scope of proteomics. However, the reuse and translational application of these datasets are limited by inconsistent standards, insufficient metadata, and inadequate computational interoperability. Proteoform-centric approaches provide higher molecular resolution by capturing intact protein variants and patterns of post-translational modification. Computational methods, including selected applications of machine learning and large language models (LLMs), are increasingly used for tasks such as spectral prediction and pattern discovery in clinical proteomics datasets. Despite these advancements, FAIR (Findable, Accessible, Interoperable, and Reusable) data practices, proteoform biology, and AI analytics are often pursued independently. This work presents an integrated framework for next-generation proteomics in which standardization and FAIR (Findable, Accessible, Interoperable, and Reusable) principles establish machine-actionable foundations for proteoform-resolved analysis and computational inference. It examines community efforts to promote data sharing and interoperability, as well as strategies for characterizing proteoforms using bottom-up, middle-down, and top-down approaches. It also highlights emerging AI and ML applications within the proteomics workflow. The framework emphasizes the importance of treating proteoforms as primary computational entities and adopting FAIR practices during data collection to enable reproducible and interpretable modeling. Finally, it introduces an architectural model that integrates FAIR infrastructures and proteoform resolution. In addition, practical recommendations for making AI-ready proteomics, including a minimal community checklist to support reproducibility, benchmarking, and translational scalability, are provided. Full article

Global warming and increasing heatwaves elevate the risk of exertional heat illnesses, particularly heatstroke, in endurance athletes and outdoor workers. This study developed and validated a wearable heatstroke warning system integrating physiological and environmental monitoring with a real-time web dashboard. The wrist- and finger-worn prototype comprised an ESP32 microcontroller and heart rate (MAX30101), skin temperature (MAX30205), ambient temperature and humidity (SHT31), and galvanic skin response (Grove-GSR v1.2) sensors with dual acoustic–visual alerts and WiFi transmission. Fifteen triathletes (18–39 years) completed 30 min of cycling in a climatic chamber: 0–15 min at 24 ± 1 °C, 70 ± 10% RH, and 16–30 min at 27 ± 1 °C, 90 ± 10% RH, with the workload rising from 40%HRmax by 10% every 10 min. Heart rate, estimated core temperature, ambient temperature, relative humidity, and GSR were recorded every 30 s and compared with standard devices using Spearman correlation (p = 0.01) and Wilcoxon signed-rank tests (p < 0.05). Heart rate, skin temperature (used a linear model to calculate core body temperature), ambient temperature, and humidity sensors showed fair–very good validity (r = 0.692, 0.995, 0.994, 0.952), while GSR was low (r = 0.298). No significant differences were observed for heart rate, skin temperature, and humidity (p > 0.05), but body temperature (p = 0.003) and GSR (p < 0.001) differed. The system showed promising validity for real-time heatstroke risk monitoring, with further refinement needed for skin temperature and GSR sensing. Full article

Background/Objectives: Recently the European Association of Urology (EAU) guidelines presented the EAU Guidelines bot to assist urologists in the reading of the guidelines; however, there is a lack of up-to-date external validation. The aim of our study is to assess the accuracy, completeness, and clarity of the guidelines bot in urethral strictures. Methods: A total of 117 questions based on the EAU urethral strictures guidelines recommendations were developed. Each question was input to the EAU guidelines bot and the response was assessed by two expert urologists to assess the accuracy, completeness, and clarity. Moreover, 10 simple clinical cases were input. A 5-point Likert scale was used as a score and, in case of discrepancies, a third urologist was queried. Accuracy, completeness and clarity were assessed per chapter and per grade of recommendation. All questions and answers were recorded in an Excel file. Results: Overall 117 questions were developed. In terms of accuracy, 111/117 (95%) were defined as accurate (scores 4–5), 4/117 (3%) presented a fair accuracy (score 3), and 2/117 (2%) were deemed not accurate. In terms of completeness, 93/117 (80%) were defined as complete (scores 4–5), 22/117 (19%) presented a fair completeness (score 3), and 2/117 (2%) were deemed not complete. Finally, in terms of clarity, 104/117 (89%) were defined as clear (scores 4–5), 13/117 (11%) presented a fair clarity (score 3), and 0/109 (0%) were deemed not clear. When comparing strong and weak recommendations, no differences were recorded. Overall the answers to simple clinical cases were in line with the guidelines with good accuracy, completeness and clarity scores. Conclusions: The EAU guidelines bot represents an accurate tool for urethral stenosis guidelines. Some fine-tuning is needed to improve readability and clarity. Full article

Background/Objectives: Whole-blood viscosity (WBV) is increasingly used in cardiovascular risk assessment; however, inter-device comparability may depend on shear-rate definition. We performed a paired comparison of two scanning capillary viscometers to evaluate shear-dependent analytical agreement and its impact on clinical classification. Methods: In 300 identical blood samples, WBV was measured using Rheovis 2000A and Hemovister. Systolic WBV was defined at 300 s⁻¹ for both devices (shear-matched), whereas clinically defined diastolic WBV corresponded to 1 s⁻¹ for Rheovis 2000A and 5 s⁻¹ for Hemovister. Agreement was assessed using linear regression and Bland–Altman analysis. Hematocrit tertiles were examined as effect modifiers. Clinical agreement was evaluated using quadratic weighted Cohen’s κ. Results: Across matched shear rates (1000 to 1 s⁻¹), Hemovister yielded consistently higher WBV values than Rheovis 2000A, with statistically significant inter-device differences at all shear levels except 1000 s⁻¹. The magnitude of bias increased progressively as shear rate decreased, reaching −8.34 mPa·s at 1 s⁻¹. Under shear-matched systolic conditions (300 s⁻¹), the mean difference was −0.25 mPa·s (limits of agreement −1.72 to 1.22). In contrast, under clinically defined diastolic conditions (1 vs. 5 s⁻¹), the mean difference was 14.54 mPa·s (3.93 to 25.15), increasing across hematocrit tertiles. Clinical agreement was fair for systolic (κ = 0.31; 95% CI 0.24 to 0.39) and moderate for diastolic WBV (κ = 0.44; 95% CI 0.37 to 0.51). Notably, among samples classified as high by Hemovister, 72.8% (systolic) and 54.0% (diastolic) were reclassified as normal by Rheovis 2000A. Conclusions: Inter-device agreement in WBV measurement is strongly shear-dependent. Although numerical divergence increases at low shear, categorical concordance may remain moderate when device-specific reference thresholds are applied. Harmonization of shear definitions and reference frameworks may therefore be essential for consistent cross-platform interpretation. Full article

Inclusive education (IE) aims to promote meaningful participation and a sense of belonging for all learners. However, limited research has examined how students with learning difficulties (LDs) experience inclusion in everyday school life. This study explored how primary school students with mild LDs perceive their participation, relationships with teachers and peers, and the role of inclusive classes (ICs) within mainstream Greek primary education. A qualitative design was adopted, and data were collected through semi-structured interviews with ten Grade 6 students receiving support through ICs. Transcripts were analyzed using thematic analysis. Findings indicated that participation was associated with perceived competence in academic tasks, with language-based activities frequently described as cognitively demanding and stressful. Belonging was predominantly felt through peer acceptance and supportive teacher practices rather than solely through classroom placement. The ICs were perceived as providing individualized support and emotional safety, although some ambivalence regarding withdrawal from the mainstream classroom was reported. Students stressed the need for flexible assessment and clearer instructional guidance to enhance fairness and participation. Overall, the findings show that inclusion is experienced as a dynamic interaction between academic accessibility, interpersonal relationships, and supportive learning environments. They also underline the importance of incorporating student voice into inclusive practice. Full article

In the joint operation of multiple virtual power plants, after day-ahead optimal dispatch is completed, some participants may engage in speculative behaviors such as misreporting profit contribution data to obtain greater benefits during profit distribution, thereby undermining fairness. To address this issue, this paper constructs a profit distribution model designed to prevent speculation. An improved Nash bargaining equilibrium algorithm based on a third-party trading intermediary is proposed to curb speculative actions. Furthermore, a dual-layer monitoring mechanism centered on profit deviation is established, which can effectively identify both single-day speculative behaviors and long-term systematic speculative trends, thereby triggering verification procedures. This forms a closed-loop management mechanism for speculation prevention—“detection, monitoring, analysis, verification”—ensuring fair profit distribution among participants within virtual power plants. Case study results demonstrate that the proposed method achieves an average deviation of only 2.32% compared to the profit distribution outcome under non-speculative conditions. In contrast, commonly used methods such as the Shapley value method, nucleolus method, and Nash–Harsanyi bargaining solution exhibit an average deviation as high as 18.44%. The research presented in this paper enables the detection of speculative behaviors among participants and facilitates verification, significantly enhancing the fairness and rationality of profit distribution. Full article

Recurring socio-temporal events, such as Ramadan in Middle Eastern cities, introduce pronounced non-stationarity in urban traffic demand. During these periods, daytime traffic volumes typically decline, while congestion becomes more severe in the evening around the Iftar (fast-breaking) period and persists into late-night hours, making conventional fixed-time signal plans less effective. An additional challenge is that demand is not only time-varying, but also unevenly distributed across competing movements: attempts to prioritize high-volume phases can inadvertently cause excessive delays—or even starvation—on lower-demand approaches. To address these issues, this study presents an adaptive, regime-aware traffic signal control framework that combines predictive modeling with constrained optimization. Short-term phase-level delays are forecast using Long Short-Term Memory (LSTM) models, and a Model Predictive Control (MPC) scheme then determines the green time allocation at each control cycle through a receding-horizon strategy. The optimization explicitly represents phase interactions by including constraints that prevent excessive delay in competing movements, thereby yielding a balanced and operationally realistic control policy. The approach is validated with one-minute-resolution TomTom delay data from a signalized intersection in Jeddah, Saudi Arabia, covering both Normal and Ramadan conditions. The LSTM models show stable predictive performance, achieving root mean square errors (RMSEs) of 19.8 s under Normal conditions and 17.1 s during Ramadan. In general, the results show that the proposed framework cuts total intersection delay by about 0.3% to 2.8% compared to standard control strategies. Even though these total-delay improvements are small, they come with big drops in delay for lower-demand phases (about 12–20%) and keep the delay increases for higher-demand phases under control. This shows that the method makes the whole process more efficient by fairly spreading out the delay instead of just making one phase better on its own. The results show that combining forecasting with constrained optimization is a strong and useful way to handle changing traffic demand. This is especially true during times of high demand when flexibility, stability, and fairness across movements are all important. Full article

Customer loyalty is a critical asset for the restaurant industry, yet the mechanisms linking the physical environment, price perception, and satisfaction remain underexplored in emerging Latin American gastronomy markets. This study examines the relationships among three servicescape dimensions—décor and artifacts, spatial layout, and ambient conditions—price perception, customer satisfaction, and loyalty in full-service restaurants in northern Peru (Chiclayo, Trujillo, and Piura). A cross-sectional survey was administered to 310 diners, and the proposed model was tested using partial least squares structural equation modeling (PLS-SEM) with 10,000 bootstrap resamples. Results supported seven of nine direct hypotheses and three of four mediation hypotheses. Décor and artifacts and ambient conditions significantly predicted both price perception and satisfaction, while spatial layout showed no significant effect on any path. Price perception partially mediated the effect of décor and ambient conditions on satisfaction, and satisfaction partially mediated the relationship between price perception and loyalty. The satisfaction–loyalty path yielded the largest effect size (β = 0.708, f² = 0.798). Serial chain analyses revealed that the physical environment shapes diner loyalty through sequential cognitive and evaluative mechanisms. These findings offer actionable insights for hospitality managers seeking to enhance gastronomy destination competitiveness through strategic servicescape investment. Full article