MODUL University Dubai

As AI models tackle increasingly complex problems, ensuring reliable human oversight becomes more challenging due to the difficulty of verifying solutions. Approaches to scaling AI supervision include debate, in which two agents engage in structured dialogue to help a judge evaluate claims; critique, in which models identify potential flaws in proposed solutions; and prover-verifier games, in which a capable 'prover' model generates solutions that must be verifiable by a less capable 'verifier'. Evaluations of the scalability of these and similar approaches to difficult problems benefit from datasets that include (1) long-form expert-verified correct solutions and (2) long-form flawed solutions with annotations highlighting specific errors, but few are available. To address this gap, we present FindTheFlaws, a group of five diverse datasets spanning medicine, mathematics, science, coding, and the Lojban language. Each dataset contains questions and long-form solutions with expert annotations validating their correctness or identifying specific error(s) in the reasoning. We evaluate frontier models' critiquing capabilities and observe a range of performance that can be leveraged for scalable oversight experiments: models performing more poorly on particular datasets can serve as judges/verifiers for more capable models.

Virtual reality (VR) devices are increasingly being utilized within operating theaters and intensive care units where appropriate sanitation is vital to ensure that patients do not unnecessarily acquire hospital‐associated infections. The morphology of VR devices in conjunction with the variety of materials and internal components provides challenges to their repurposing. This study aimed to evaluate the microorganisms remaining on VR headsets following sanitation by laboratory staff in a medical education anatomy teaching facility. The external components and internal facial interface were swabbed and separately cultured on four AGAR plates (Horse Blood, Nutrient, bile Esculin, and Mannitol Salt). Colonies were counted, sampled, pooled and subsequently processed for shotgun metagenomic sequencing. A higher number of colonies were present on surfaces closest to the eyes and facial interface compared to the external components. Metagenomic analysis identified 27 pathogenic bacteria including 4 “ESKAPE” pathogens (Enterobacter sp., Staphylococcus aureus, Klebsiella spp. and, Escherichia coli) and numerous organisms associated with ocular infections. A broad range of antimicrobial resistance genes were identified conveying resistance to Methicillin, Aminoglycosides, Macrolides, Tetracyclines, and Polymixins. Further research is required to ensure that current sanitization practices of VR head mounted displays are appropriate within high‐risk hospital settings.

[This corrects the article DOI: 10.3389/fpubh.2025.1713092.].

Madrasah as an Islamic educational institution faces challenges in implementing Islamic values for student character building. Although religious activities have been carried out routinely, they have not effectively touched the core of students’ spirituality as reflected in their daily behavior. The gap between the ideality of madrasah as a producer of noble character and the reality of value shifts highlights the urgency of structured and sustainable Islamic value implementation studies. This study employs a qualitative approach with ethnographic-phenomenological methods at Madrasah Ibtidaiyah Nasy’atul Muta’allimin Candi, Sumenep, Madura. Research subjects include the madrasah principal, vice principals, teachers, school committees, parents/guardians, and surrounding community. Data were collected through in-depth interviews, participatory observation, and documentation, then analyzed through data reduction, data presentation, conclusion drawing, and verification stages. Data validity was ensured through credibility, transferability, dependability, and confirmability. The findings reveal that Islamic values implementation is based on three main foundations: faith, sharia, and ihsan, manifested through habituation of greeting-smiling-salutation values, honesty, trustworthiness, and responsibility. Implementation strategies include teacher exemplary behavior, madrasah culture, project-based learning, and cross-curricular integration, reinforced with specific practices such as teachers welcoming students, reciting Asmaul Husna, congregational dhuha prayer, and intensive parental cooperation. This study concludes that Islamic character building requires continuous habituation through real behavioral examples from teachers and parents. Recommendations are directed to madrasah education practitioners to develop structured, organized, and dynamic value implementation concepts according to each institution’s factual context, and to future researchers to explore the long-term impacts of habituation strategies on madrasah students’ character.

Large Language Models (LLMs) show potential in medical document generation, but ensuring reliability requires extensive expert involvement, limiting clinical applications. To address this challenge, we developed an LLM-based evaluation framework with three progressive Chain of Thought (CoT) strategies: Qualitative (expert persona), Quantitative-qualitative (error analysis), and Insight-integrated (expert reasoning). This framework captures nuanced evaluation patterns while maintaining efficiency. When tested on 33 LLM-generated Emergency Department records across five criteria, our Insight-integrated approach demonstrated strong correlation with expert evaluations (r = 0.680, p < .001), outperforming both Qualitative (r = 0.524) and Quantitative-qualitative (r = 0.630) approaches. Our findings suggest that LLM-based evaluation frameworks can align with expert assessments as useful tools for validating medical documentation in clinical settings.

Materials with sufficient strength and stiffness can transfer nonlinear design loads without damage. The present study compares crack propagation speed and shape in rock-like material and sandstone when subjected to seismic acceleration. The nonlinear extended finite element method (NXFEM) has been used in numerical simulation. It assumes the model has a pre-existing crack at 0° from the horizontal. The mechanical properties of the model, crack propagation shape, and crack speed were selected as the main parameters. The nonlinear stress and strain along the crack have been compared in two simulated models. NXFEM and Artificial Neural Networks (ANNs) were used to predict the displacement. The simulation results illustrate that the materials’ crack propagation mechanism and mechanical properties control the stress, strain, and displacement at the selected points in the model. In addition, crack propagation in materials is related to elastic-plastic stresses and strains along the crack path. The speed and shape of the crack are associated with the mechanical properties of the materials. The prediction of crack paths helps to understand failure patterns. Comparison of the seismic response of the rock-like material with sandstone helps to assess the stress, strain, and displacement levels during cracking. This study’s findings agree with the literature report and field observations.

Abstract Sarawak, a state in Malaysia has long been the key contributor to probable and proven (2P) Petroleum reserves, in which the state bearing and account to 60.8% of total National reserves. Known for its untapped reserves and with sturdy relationship with partners and investors to spur growth to national economy for decades, now it just the right time for the State to focus on its social contribution to Sarawak's young talents in the energy industry.