Search Results (97)

Music education offers well-documented benefits for student learning; however, generalist teachers often report low confidence in integrating music into their lessons. This study applies Universal Design for Learning (UDL) principles to develop teaching resources that address teacher barriers to music integration. Using framework analysis, data collected from semi-structured interviews with ten trainee primary teachers in United Arab Emirates (UAE) British curriculum schools were mapped against UDL’s three core principles: engagement, representation, and action and expression. Despite recognising music’s holistic educational value in cognitive enhancement, memory retention, and student expression, participants reported significant barriers to integrating the subject into their lessons. These barriers included performance anxiety, a perceived lack of subject knowledge, and fear of student judgement. The barriers were most pronounced when faced with the prospect of teaching upper-primary students. Framework analysis revealed how these challenges align with the UDL’s core principles. These findings led to the development of five-step music resources, categorised into beginner and intermediate levels. Each step of the resources is designed to systematically address these identified barriers through UDL’s proactive and intentional design criteria. This demonstrates how teacher education can move beyond identifying barriers to creating structured solutions that support inclusive music integration while maintaining pedagogical authenticity. Full article

This study examines the relationship between learning environments and learning outcomes in acquiring Korean as a language target. We compare two learner groups residing in the United States: English-speaking learners of Korean in foreign language contexts versus Korean heritage speakers. Both groups share English as their dominant language and receive similar tertiary-level instruction, yet differ in their language-learning profiles. We measure two groups’ comprehension behaviour involving Korean subject−predicate honorific agreement, focusing on two conditions manifesting a mismatch between the honorifiable status of a subject and the realisation of the honorific suffix in a predicate. Results from the acceptability judgement task revealed that (1) both learner groups rated the ungrammatical condition as more acceptable than native speakers did, (2) Korean heritage speakers rated the ungrammatical condition significantly lower than English-speaking learners, and (3) overall proficiency in Korean modulated learners’ evaluations of the ungrammatical condition in opposite directions between the groups. No between-group difference was found in the infelicitous-yet-grammatical condition. Results from reaction time measurement further showed that Korean heritage speakers responded considerably faster than English-speaking learners of Korean. These results underscore the critical role of broad usage experience—whether through home language exposure for heritage language speakers or formal instruction for foreign language learners—in shaping non-dominant language activities. Full article

Higher Degree by Research (HDR) students in language and education disciplines, particularly those enrolled in thesis-only programmes, are increasingly expected to interpret complex statistical data. However, many lack the analytical skills required for independent statistical analysis, posing challenges to their research competence. This study investigated the pedagogical potential of ChatGPT-4o feedback and peer feedback in supporting students’ evaluative judgement during a 14-week doctoral-level statistical analysis course at a research-intensive university. Thirty-two doctoral students were assigned to receive either ChatGPT feedback or peer feedback on a mid-term assignment. They were then required to complete written reflections. Follow-up interviews with six selected participants revealed that each feedback modality influenced their evaluative judgement differently across three dimensions: hard (accuracy-based), soft (value-based), and dynamic (process-based). While ChatGPT provided timely and detailed guidance, it offered limited support for students’ confidence in verifying accuracy. Peer feedback promoted critical reflection and collaboration but varied in quality. We therefore argue that strategically combining ChatGPT feedback and peer feedback may better support novice researchers in developing statistical competence in hybrid human–AI learning environments. Full article

A 2D raster data representing building volumes of each grids are derived from 3D vector-format urban data for use in machine learning applications. Since the task is to explore patterns, i.e., urban heat islands, Gaussian blurring is implemented on these generated 2D raster data before the training process. This strengthens the visual capturing of spatial relationships, and as a result the correlation rate between air temperature and building volume data is also increased. After the model training, the prediction results are not simply evaluated with most widely used shallow metrics like the Mean Square Error (MSE), but thanks to the raster format of input and output results, some image similarity metrics such as Structural Similarity Index Measure (SSIM) and Learned Perceptual Image Patch Similarity (LPIPS) that are able to detect and consider spatial relations are used during the evaluation and interpretation process, because of their higher usefulness in mimicking human visual judgements. The trained models with Random Forest and XGBoost methods which are capable of predicting the spatial distribution of air temperature by using building volume information are compared. By doing so, this research aims to assist urban planners in incorporating environmental parameters into their planning strategies, thereby facilitating more sustainable and inhabitable urban environments. Full article

Advancements in audio synthesis and manipulation technologies have reshaped applications such as personalised virtual assistants, voice cloning for creative content, and language learning tools. However, the misuse of these technologies to create audio deepfakes has raised serious concerns about security, privacy, and trust. Studies reveal that human judgement of deepfake audio is not always reliable, highlighting the urgent need for robust detection technologies to mitigate these risks. This paper provides a comprehensive survey of recent advancements in audio deepfake detection, with a focus on cutting-edge developments in the past few years. It begins by exploring the foundational methods of audio deepfake generation, including text-to-speech (TTS) and voice conversion (VC), followed by a review of datasets driving progress in the field. The survey then delves into detection approaches, covering frontend feature extraction, backend classification models, and end-to-end systems. Additionally, emerging topics such as privacy-preserving detection, explainability, and fairness are discussed. Finally, this paper identifies key challenges and outlines future directions for developing robust and scalable audio deepfake detection systems. Full article

This article analyses the coordination between professionals, students and technology in the communication and appropriation of know-how in healthcare simulations. To be successful, simulations require continuous interventions by professionals (in this case, operators and facilitators), who analyse, assess and reflect on the actions participants take as the simulation evolves. This study builds on interaction analysis of 30 video-documented (15 h) conversations between operators and facilitators in post-simulation discussions of outcomes. The specific focus of the analysis is the nature of work done by operators/facilitators as they analyse and evaluate simulations. The results show the multilayered nature of these analyses. The operators and facilitators show three prominent types of consideration. They (a) calibrate what they have observed, (b) monitor the progress of the scenario as an instructional event, and (c) comment on their own contributions as instructors/participants. All these considerations have evaluative elements, and the agentic nature of technologies, students and professionals is addressed. One general observation of interest is the ways in which simulations provide access to student learning, and how these activities become accessible for professional scrutiny and judgement. Full article

Alzheimer’s disease is the most common cause of dementia in the elderly population (aged 65 years and over), followed by vascular dementia, Lewy body dementia, and rare types of neurodegenerative diseases, including frontotemporal dementia. There is an unmet need to improve diagnosis and prognosis for patients with dementia, as cycles of misdiagnosis and diagnostic delays are challenging scenarios in neurodegenerative diseases. Neuroimaging is routinely used in clinical practice to support the diagnosis of neurodegenerative diseases. Clinical neuroimaging is amenable to errors owing to varying human judgement as the imaging data are complex and multidimensional. Artificial intelligence algorithms (machine learning and deep learning) enable automation of neuroimaging interpretation and may reduce potential bias and ameliorate clinical decision-making. Graph convolutional network-based frameworks implicitly provide multimodal sparse interpretability to support the detection of Alzheimer’s disease and its prodromal stage, mild cognitive impairment. In patients with amyloid-related imaging abnormalities, radiologists had significantly better detection performances with both ARIA-E (sensitivity higher in the assisted/deep learning method [87%] compared to unassisted [71%]) and for ARIA-H signs (sensitivity was higher in assisted [79%] compared to unassisted [69%]). A convolutional neural network method was developed, and external validation predicted final clinical diagnoses of Alzheimer’s disease, dementia with Lewy bodies, mild cognitive impairment due to Alzheimer’s disease, or cognitively normal with FDG-PET. The translation of artificial intelligence to clinical practice is plagued with technical, disease-related, and institutional challenges. The implementation of artificial intelligence methods in clinical practice has the potential to transform the diagnostic and treatment landscape and improve patient health and outcomes. Full article

This study aims to explore the reliability, construct validity, and content validity of the Child Learning and Developmental Playful Assessment Battery (Batería de Evaluación Lúdica del Aprendizaje y Desarrollo Infantil; BELADI), a quantitative instrument based on the authentic assessment and playful learning principles, the purpose of which is to assess infant learning and development through motor and competitive games as well as storytelling. The sample was composed of 113 children from Albacete (Spain) between 58 and 72 months of chronological age (M = 64.72; SD = 3.671). To explore the content validity, an expert judgement was carried out and the Content Validity Coefficient (CVC) was calculated. The reliability was analysed using the Cronbach’s alpha and McDonald’s Ω, and an exploratory factor analysis (EFA) was conducted. The results revealed high reliability indexes in each of the developmental domains, and the EFA included 11 items distributed in two factors for the psychomotor domain, 27 items grouped in three factors for the cognitive domain, and 20 items divided into four factors for the socioemotional domain. In conclusion, the study verifies the validity and reliability of the BELADI for the assessment of the infant learning and development through play, which may be used in research, education, and psychopedagogy. Full article

The lessons learned from reviewing national risk assessments to modernise the Australian Standard for the post-mortem inspection and disposition judgement of beef, sheep, goat, and pig carcases are discussed. The initial risk profiles identified priorities for quantitative assessments. Broadly, the main difficulty encountered was the paucity of quantified performance for the current inspection. Resolving this involved acquiring gross abnormality data representing regional production/proportional abattoir volumes, the range of gross abnormalities appearing nationally, proportional occurrence at carcase sites, and seasonality to enable the comparison of procedures. The methodologies followed the Codex Alimentarius Commission’s risk assessment guidelines and are fully documented in the associated publications. The evidence and discussion are provided for the associated challenges experienced, including preventing contamination, the use of food chain information to support amendment, inspection as a part of industry Quality Assurance programmes, and opportunities to improve inspector training. The criteria considered by the Competent Authority for the determination of the equivalence of alternative post-mortem inspection techniques included comparisons of public health risk, non-detection rates for gross abnormalities, and microbial contamination resulting from inspection activities, as appropriate. Most of the gross abnormalities detected arose from animal health and welfare conditions affecting wholesomeness and did not present as food safety hazards. The non-detection rates between the current and alternative inspection (observation) were negligible. A quantitative risk assessment for Cysticercus bovis was conducted. Carcases with multiple gross abnormalities predominantly reflected historic infections (prior septicaemia), where trimming achieved wholesomeness unless they were cachexic. Full article

With the popularity of the Internet and the increase in the level of information technology, cyber attacks have become an increasingly serious problem. They pose a great threat to the security of individuals, enterprises, and the state. This has made network intrusion detection technology critically important. In this paper, a malicious traffic detection model is constructed based on a decision tree classifier of entropy and a proximal policy optimisation algorithm (PPO) of deep reinforcement learning. Firstly, the decision tree idea in machine learning is used to make a preliminary classification judgement on the dataset based on the information entropy. The importance score of each feature in the classification work is calculated and the features with lower contributions are removed. Then, it is handed over to the PPO algorithm model for detection. An entropy regularity term is introduced in the process of the PPO algorithm update. Finally, the deep reinforcement learning algorithm is used to continuously train and update the parameters during the detection process, and finally, the detection model with higher accuracy is obtained. Experiments show that the binary classification accuracy of the malicious traffic detection model based on the deep reinforcement learning PPO algorithm can reach 99.17% under the CIC-IDS2017 dataset used in this paper. Full article

Nursing homes (NHs) are crucial for de-hospitalization and addressing the needs of non-self-sufficient individuals with complex health issues. This study investigates the patient safety culture (PSC) in NHs within a northern Italian region, focusing on factor influencing overall safety perceptions and their contributions to subjective judgements of safety. A cross-sectional study was conducted on 25 NHs in the Autonomous Province of Trento. The Nursing Home Survey on Patient Safety Culture (NHSPSC) was utilized to assess PSC among NH staff. Multilevel linear regression and post hoc dominance analyses were conducted to investigate variabilities in PSC among staff and NHs and to assess the extent to which PSC dimensions explain overall perceptions of PS. Analysis of 1080 questionnaires (44% response rate) revealed heterogeneity in PSC across dimensions and NHs, with management support, organizational learning, and supervisor expectations significantly influencing overall safety perceptions. Despite some areas of concern, overall safety perceptions were satisfactory. However, the correlation between individual dimensions and overall ratings of safety was moderate, suggesting the need to enhance the maturity level of PSCs. Promoting a shift in PSC could enhance transparency, prioritize resident safety, empower nursing staff, and increase family satisfaction with care provided in NHs. The support provided by management to PSC appears essential to influence NH staff perceptions of PS. Full article

In recent years, classification and identification of Earth’s surface materials has been a challenging research topic in the field of earth science and remote sensing (RS). Although deep learning techniques have achieved some results in remote sensing image classification, there are still some challenges for multimodal remote sensing data classification, such as information redundancy between multimodal remote sensing images. In this paper, we propose a multimodal remote sensing data classification method IFF-Net based on irregular feature fusion, called IFF-Net. The IFF-Net architecture utilizes weight-shared residual blocks for feature extraction while maintaining the independent batch normalization (BN) layer. During the training phase, the redundancy of the current channel is determined by evaluating the judgement factor of the BN layer. If this judgment factor falls below a predefined threshold, it indicates that the current channel information is redundant and should be substituted with another channel. Sparse constraints are imposed on some of the judgment factors in order to remove extra channels and enhance generalization. Furthermore, a module for feature normalization and calibration has been devised to leverage the spatial interdependence of multimodal features in order to achieve improved discrimination. Two standard datasets are used in the experiments to validate the effectiveness of the proposed method. The experimental results show that the IFF-NET method proposed in this paper exhibits significantly superior performance compared to the state-of-the-art methods. Full article

The study investigates the efficiency of integrating Machine Learning (ML) in clinical practice for diagnosing solitary pulmonary nodules’ (SPN) malignancy. Patient data had been recorded in the Department of Nuclear Medicine, University Hospital of Patras, in Greece. A dataset comprising 456 SPN characteristics extracted from CT scans, the SUVmax score from the PET examination, and the ultimate outcome (benign/malignant), determined by patient follow-up or biopsy, was used to build the ML classifier. Two medical experts provided their malignancy likelihood scores, taking into account the patient’s clinical condition and without prior knowledge of the true label of the SPN. Incorporating human assessments into ML model training improved diagnostic efficiency by approximately 3%, highlighting the synergistic role of human judgment alongside ML. Under the latter setup, the ML model had an accuracy score of 95.39% (CI 95%: 95.29–95.49%). While ML exhibited swings in probability scores, human readers excelled in discerning ambiguous cases. ML outperformed the best human reader in challenging instances, particularly in SPNs with ambiguous probability grades, showcasing its utility in diagnostic grey zones. The best human reader reached an accuracy of 80% in the grey zone, whilst ML exhibited 89%. The findings underline the collaborative potential of ML and human expertise in enhancing SPN characterization accuracy and confidence, especially in cases where diagnostic certainty is elusive. This study contributes to understanding how integrating ML and human judgement can optimize SPN diagnostic outcomes, ultimately advancing clinical decision-making in PET/CT screenings. Full article

The United Nations (UN) Sustainable Development Goals (SDGs) offer an opportunity to improve soil science education on sustainability because they provide specific context to educate faculty and students from various disciplines, including Science, Technology, Engineering, and Mathematics (STEM) about SDGs. Soil science is a STEM discipline with a wide range of applications in the SDGs. The objectives of this study were to use a matrix approach (framework for presenting options for discussion and implementation) to integrate SDGs into an existing introductory soil science course taught to undergraduate students from different STEM fields (environmental and natural resources; wildlife biology; and forestry). The course was enriched with a lecture on SDGs and students were asked to link soil properties and class activities to specific SDGs. A post-assessment survey revealed an increase in students’ familiarity with SDGs, and their relevance to soil properties and course activities. Students acknowledged the importance of soils and individual actions for achieving the SDGs. There was an overall increase in student familiarity (+59.4%) with SDGs. Most students agreed (46.7%) and strongly agreed (23.3%) that the course activities were an effective way to learn about SDGs with examples from soil science. Identified learning gaps in subject matter found through the surveys on SDGs were clarified during later classroom discussions. The advantage of this teaching approach is that it seamlessly integrates SDGs with existing course materials while relying on students’ critical thinking skills to effectively analyze soil science information and form a judgement on how it relates to SDGs. Full article

As the complexity of banking technology systems increases, the prevention of technological risk becomes an endless battle. Currently, most banks rely on the experience and subjective judgement of experts and employees to allocate resources for technological risk management, which does not effectively reduce the frequency of technology-related incidents. Through an analysis of mainstream risk management models, this study proposes a technology-based risk assessment system based on machine learning. It first identifies risk factors in bank IT, preprocesses the sample data, and uses different regression prediction models to train the processed data to build an intelligent assessment model. The experimental results indicated that the Genetic Algorithm–Backpropagation Neural Network model achieved the best performance. Based on assessment indicators, indicator weight values, and risk levels, commercial banks can develop targeted prevention and control measures by applying limited resources to the most critical corrective actions, thereby effectively reducing the frequency of technology-related incidents. Full article