Domain knowledge and data-driven framework for predicting the time of gas breakthrough during in-situ combustion

Waste management–related trust, acceptance, and reputation: A multidisciplinary big data analysis across knowledge domains

High-quality, error-free data is essential for developing reliable data-driven models, particularly in clinical decision support systems where inaccurate predictions can have serious consequences. While KGs offer a structured and semantically rich representation for clinical data, ensuring their consistency and correctness remains a challenge. Existing rule mining techniques provide solutions for the automatic extraction of logical constraints from KGs, but they often produce redundant or clinically irrelevant rules, especially when dealing with numeric or categorical literals such as age or lab values. KG constraints-rules intended to capture implausible or conflicting facts in the KG-can be used to spot semantic errors: facts that might conform to the underlying schema but contradict domain knowledge. In this work, we propose a novel framework for constraint learning in clinical KGs that identifies and transforms high-confidence rules into clinically plausible constraints. We propose two approaches, based on class disjointness and literal clustering combined with rule mining. We validate the clinical relevance of these generated rules using expert-curated constraints and large language models (LLMs). The results on the MIMIC-III clinical dataset show that rule filtering based constraint learning effectively preserves clinically meaningful rules that align with established medical knowledge. For numeric data, we achieve reliable value groupings through our clustering-based method, and the rules derived from these groupings were validated by LLMs. Their outputs confirm the clinical relevance of a portion of those discovered rules. By providing interpretable and scalable solutions to semantic inconsistencies in KGs, this study contributes to increasing the KG trustworthiness and its clinical usability.

Integrating data and domain knowledge for predictive intelligence: A comprehensive review of DKF-DPM in intelligent manufacturing

Calculating tourist sentiment ambivalence through aspect-level sentiment analysis: Infusing tourism domain knowledge into a pre-trained language model

DKF: Domain knowledge fusion in progressive incremental learning for multi-domain machine translation

Exploring approaches to computational representation and classification of user-generated meal logs.

Molecular representation, as one of the fundamental challenges in artificial intelligence-driven drug discovery, has attracted increasing attention due to its low cost and impressive speed while it is applied in molecular property prediction, drug molecule generation, drug-drug interactions, etc. Numerous models that integrate multi-modal representations have been proposed for molecular representation learning. However, existing methods have not yet considered the IUPAC International Chemical Identifier (InChI) as one of the multi-modal inputs. To address this issue, we propose InChINet, a self-supervised molecular representation learning framework that is pre-trained on 10 million unlabeled molecules. It leverages mutual information across the simplified molecular line input system (SMILES) and InChI. In addition, we present token reordering and token masking for SMILES. Combined with SMILES enumeration, these three strategies introduce domain knowledge and improve the model's stability against syntactic variations in SMILES representations. Benefiting from the introduction of InChI and augmentation strategies, InChINet achieves impressive performance on a wide range of downstream tasks, including molecular property prediction, drug-drug interaction (DDI) prediction, clustering analysis, zero-shot cross-lingual retrieval, and ablation study.

ABSTRACT Teacher shortages in Australia have led to a growing number of out-of-field (OOF) teachers delivering mathematics in secondary schools, raising concerns about the quality of instruction. This study investigates the professional learning priorities for OOF mathematics teachers from the perspectives of Queensland school leaders and teachers themselves. Drawing on conceptualisations of teacher knowledge, the research employed a two-phase design: interviews with seven school leaders and a questionnaire completed by 20 mathematics teachers. Data were analysed using a hybrid inductive – deductive thematic approach to identify key knowledge domains valued in practice and consequent opportunities for professional development. Findings indicate that school leaders prioritise Common Content Knowledge and elements of Knowledge of Content and Students and Knowledge of Content and Teaching, often assigning primary-trained teachers to junior secondary mathematics roles. Teachers reported lower confidence with senior mathematics content, teaching problem-solving and reasoning, and using inquiry-based approaches, likely connected with their specialised content and horizon knowledges. Insights inform professional learning strategies to strengthen mathematics knowledge for teaching among both experienced and OOF teachers.

Abstract With LLMs increasingly used as conversational assistants, this study compares the sarcasm detection performance of five LLMs and 323 humans in German, conducting a binary classification task of 100 soccer-related tweets. The analysis includes metrics of individual vote and majority vote for both groups, examining how voting type affects classification accuracy. Human participants were further grouped by their level of soccer knowledge to assess the effect of domain knowledge on detection accuracy. Among the LLMs, Mistral Large achieved the highest accuracy (83%), nearly matching the human average (83.3%). Significant performance differences were observed among LLMs and between LLMs and humans, with the largest gap between Gemini 1.0 Pro (64%) and the human majority vote (97%). Majority vote significantly improved human classification accuracy by 13.7 percentage points, and domain knowledge was positively correlated with sarcasm classification accuracy. Humans also performed similarly across sarcastic and non-sarcastic tweets, whereas LLMs frequently misclassified tweets containing negative lexical markers as sarcastic. These results highlight the importance of ensemble strategies and domain knowledge in sarcasm detection tasks and underscore the need for further research into polarity-related classification biases in LLMs.

Purpose The Mental Capacity Act (2005; MCA) entered into force in 2007, some 18 years ago. Since its implementation, several authors have identified gaps in knowledge and confidence in different professional groups. There has, however, been very little research considering this question in relation to Clinical Psychologists, and many wider surveys are now dated. The purpose of this study was to assess the knowledge and confidence in working with the MCA in a group of practicing Clinical Psychologists. Design/methodology/approach A bespoke survey assessed knowledge and confidence in key domains of the MCA specific to the role of a Clinical Psychologist, including a scenario-based vignette and qualitative responses to address additional training needs. A total of 58 Health and Care Professions Council-registered Clinical Psychologists responded to the survey. Findings In the scenario, Clinical Psychologists demonstrated good knowledge of applying principles of the MCA to a psychology-specific scenario. However, confidence seemed lower, and not all psychologists considered the MCA as very important to their work. Research limitations/implications Research limitations include using a bespoke survey and likely self-selection bias which may mean knowledge and confidence is over-represented. A number of different areas for potential training were identified by the participants. Originality/value Whilst various studies (Marshall and Sprung, 2016; Scott et al., 2020) have explored the knowledge and confidence of the MCA in different professional groups, there has been limited reported data on the knowledge and confidence of a psychological workforce.

Patient knowledge is pivotal to inflammatory bowel disease (IBD) management, yet educational opportunities vary widely worldwide. This scoping review maps existing evidence on patient knowledge, explores regional and demographic differences, and identifies gaps to guide culturally adaptable interventions. Following PRISMA-ScR guidelines, PubMed and Embase were searched up to December 31, 2025, for studies assessing patient knowledge in IBD. Eligible adult and pediatric studies were independently screened and synthesized by two reviewers across key knowledge domains, including disease basics, treatment, complications, colorectal cancer (CRC) risk, surgery, vaccination, and diet. From 8464 records, 53 studies met inclusion criteria. Validated tools, including CCKNOW and IBD-KNOW, predominated in adult studies, whereas IBD-KID and IBD-KID2 were used in pediatric populations. Knowledge levels varied widely: correct understanding of anatomy (36%-68%), risk factors (16%-85%), CRC risk (24%-78%), and surgery (13%-16%) was frequently suboptimal. Only 5%-51% recognized azathioprine as an immunosuppressant. Awareness of vaccination (39%-78%) and dietary relevance (23%-65%) remained limited. Gastroenterologists were the primary information source (42%-96%). Higher knowledge was associated with both sociodemographic factors (female sex, younger age, higher educational attainment) and disease-related characteristics (longer disease duration, prior exposure to biologics or surgery, Crohn's disease phenotype). Patients in Europe and North America had higher awareness than those in Asia and the Middle East, probably due to better healthcare access and patient education infrastructure. Global IBD knowledge and awareness remain inadequate and uneven across regions and domains. Updated, culturally appropriate and adaptable assessment tools and multidisciplinary, technology-enabled educational strategies are needed to enhance IBD literacy.

Abstract This paper stands at the intersection between specialised translation and terminology. Terms and their relations are rather central in specialised texts and become even more important when texts are transferred into another language, i.e. within specialised translation. The analysis discusses the needs of specialised translation and how they could be covered through a terminological representation method in order for it to be a useful tool for translators of specialised texts. The needs of specialised translation are presented thoroughly following an analysis divided into levels of equivalence. A graphic representation of two axes of a Cartesian coordinates system is used to depict the progressive saturation of the needs at several levels of equivalence in specialised translation using an adequate terminology representation. It is argued that in order to provide a good translation of a specialised text, translators do not just need equivalents for the terms involved. They also need a good domain representation that will help them increase their domain knowledge, translate the text sufficiently up to the highest levels of equivalence and suggest new translations for terms in case of non-equivalence. Moving from terminological resources that mainly focus on terms as linguistic units to resources that consider terms not just as linguistic but also as domain-specific units can be the key to the saturation of the needs of specialised translation at all levels of equivalence.

Abstract This article examines the role of expertise in the emergence and evolution of new diplomatic issue areas. Contemporary diplomatic practice increasingly requires coordination across knowledge domains. Nowhere is this more evident than in issues involving fast evolving technology, such as artificial intelligence or cyber issues. Examining the case of cyber diplomacy, we show how diplomats have adapted to these concerns. Contrary to claims that growing demands for issue-specific expertise erode the traditional diplomatic monopoly over authoritative knowledge, the article shows that traditional diplomats have expanded their role. Rather being displaced by technical experts, diplomats act as key points of passage through which knowledge enters the diplomatic arena. Their ability to mediate between competing knowledge claims is becoming a defining feature of diplomatic practice in complex issue areas. We further demonstrate how early framings of cyber diplomacy produced lasting effects diplomatic practice, prioritizing great power politics and skewing debates towards the international security dimensions of cyber issues. The article thus contributes to diplomatic studies by showcasing how early issue framings institutionalize particular practices and constrain adaptation over time, offering broader insights into the tensions between institutional stability and technological change in the governance of digital technology.

Several teachers experience difficulties teaching programming in school mathematics. While the Technological Pedagogical Content framework (TPACK) has previously described links between pedagogical, content and technological knowledge for incorporating technology in competencies for teaching mathematics, these links must be reevaluated after new programming elements have been introduced in the Norwegian national curriculum. Using teachers’ self-reported knowledge, 127 answers were analysed through confirmatory and exploratory factor analysis. Results show strong loadings for technological knowledge, but weak associations to pedagogical knowledge, indicating a separation with programming constructs. Our findings challenge the notion of programming as merely a technological component, suggesting programming should be considered a partially separate domain in TPACK.

The present study aimed to design and standardize a valid and reliable knowledge assessment tool to measure livestock farmers’ understanding of climate change impacts and adaptation strategies in the Jammu division. The instrument encompassed major knowledge domains, including weather pattern observation, causes of climate change, climate advisory services, impacts on animal behavior, health, reproduction, fodder and water resources, productivity, economic consequences, and adaptation strategies such as housing management, feed and water management, breed improvement, disease prevention, seasonal migration, and financial risk mitigation. An initial pool of 162 items was generated through literature review and expert consultation. Based on standard item selection procedures, 103 items were shortlisted, of which 84 were retained following expert relevancy testing. Item analysis was conducted on a sample of 20 livestock farmers, resulting in the selection of 61 items that satisfied acceptable psychometric criteria, with difficulty indices ranging from 30 to 80 and discrimination indices between 0.30 and 0.55. Reliability was established using the split-half method (r = 0.813) and further adjusted using the Spearman–Brown prophecy formula (r = 0.837). Internal consistency was confirmed through Cronbach’s alpha (α = 0.85, p < 0.05). Content validity was ensured through systematic expert evaluation. The finalized 61-item standardized knowledge test demonstrates strong psychometric properties and provides a robust tool for researchers, extension personnel, and policymakers to assess knowledge gaps, guide training interventions, and support evidence-based climate-resilient livestock development initiatives.

The rapid digital transformation of industrial systems in the 21st century has led to an exponential growth in data generated by manufacturing processes and end-user products, particularly in the automotive sector. While this big data creates new opportunities for monitoring and diagnostics, it also introduces significant challenges related to system complexity, scalability, and nonlinearity, as well as the increasing shortage of experienced domain experts. These challenges motivate the adoption of intelligent, automated fault and anomaly detection techniques capable of operating reliably under real-world conditions. The primary objective of this paper is to provide a comprehensive and structured review of the anomaly detection methodologies for automotive applications, with particular emphasis on intelligent fault diagnosis, tolerance, and monitoring architectures. To this end, the paper systematically categorizes existing approaches, including model-based, data-driven, and hybrid techniques, and analyzes their underlying principles, data requirements, computational complexity, and applicability to safety-critical systems. Based on this analysis, the paper highlights current limitations, open research challenges, and emerging trends, including the integration of machine learning and artificial intelligence with domain knowledge and control-oriented frameworks. The main contribution of this work is a unified perspective that supports researchers and practitioners in selecting, designing, and deploying effective anomaly detection solutions for next-generation automotive and cyber-physical systems.

Abstract An increasing number of research studies indicate that student learning outcomes improve when humanistic perspectives are integrated into science, technology, engineering, and math (STEM) education. Practical guidance is currently limited on how to effectively integrate science and humanities; therefore, we developed an integration model and process applicable to a STEM Futures framework, which integrates three knowledge domains, foundational, humanistic, and meta knowledge, into STEM coursework. Here, we describe the rationale for developing integrated coursework, our chosen method of integration (case studies), the goal of the case studies, and the process of developing these integrated case studies. A team of instructors from three universities and one historian of science developed, tested, refined, and taught case studies in science courses at Colorado State University, Southern Illinois University Edwardsville, and University of Texas at El Paso. Importantly, case studies were focused on the reincorporation of Black, Indigenous, and people of color (BIPOC) into histories of science. The historian of science provided expert advice on examples for integration and how to approach difficult conversations around BIPOC in agriculture. Focusing on the collaboration between one STEM instructor and the science historian, we describe here the collaborative process of STEM and humanities faculty members designing case studies, the importance of collaboration between the instructors, and the key components of the content and questions written to meet the goals of the course.

This study reports the design, implementation, and descriptive evaluation of “Codici del Futuro”, a STEM-oriented education pathway developed by the Italian National Research Council (CNR) to promote students’ interest in science and awareness of research-related careers and addressed to local high school students. The programme involved 167 high school students organised in 10 groups and combined an orientation session with hands-on workshops delivered in CNR research facilities (chemistry, biotechnology, artificial intelligence, eXdended Reality/Augmented Reality (XR/AR), and game design). The chemistry workshop will be described as a case study. The study addresses two research questions: (RQ1) What group-level outcomes (participation, engagement, interest, behaviour) are observed across the multidisciplinary pathway? (RQ2) What post-activity satisfaction and short-term knowledge outcomes are observed in the chemistry workshop as an embedded case study? Group-level outcomes were assessed through a facilitator-based evaluation grid, using four single-item indicators rated on a 10-point scale and including field notes. The chemistry case study included an anonymous post-activity questionnaire (satisfaction, prior experience, and an eight-item knowledge test). Results documented high levels of engagement, interest, and appropriate behaviour across groups, whereas participation showed greater variability. In the chemistry case study, students reported high satisfaction and moderate post-activity knowledge scores, with differences across knowledge domains. Overall, findings provide descriptive evidence on student responses within a research-centre-based, multi-workshop STEM pathway.

Background/Objectives: Early tracheostomy seems favorable in prolonged ventilated patients after surgery. Hence, predicting tracheostomy after cardiac surgery is essential. Recently proposed prediction models aim to support this decision-making process, but their diagnostic validity across other patient populations remains uncertain. Methods: A retrospective single-center study was performed at a university hospital. The patient sample included consecutive patients between 2010 and 2020 who underwent cardiac surgery. Patients who underwent tracheostomy after cardiac surgery were assigned to the intervention group. Control group patients, who had not undergone tracheostomy, were randomly assigned to the group. An existing model was evaluated by receiver operating characteristics curve analysis. Four sets of risk features were chosen depending on results from regression analysis, lasso regularization, random forest or clinical domain knowledge. Newly developed models were created using machine learning methods: random forest, naïve Bayes, nearest neighbor and deep learning. Multiple models were trained with either feature set and then assessed using confusion matrices on an independent test set. Results: A total of 4744 patients were included in this study. One-hundred and eighteen patients were included in the tracheostomy group. Diagnostic accuracy of the existing model showed insufficient discrimination (area under the curve (AUC) = 0.57). Likewise, newly developed models also showed overall poor diagnostic discrimination across all feature sets and algorithms. Conclusions: This study shows the diagnostic limitations of retrospective clinical data for the diagnostic prediction of tracheostomy, thereby informing the design of future prospective diagnostic studies. Training new models should not rely on retrospective data alone. Instead, prospective data collection and integration of physiological or imaging-based diagnostics could likely contribute to the development of a good classifier.