159370 research outputs found
Sort by
The first cadenza challenge: perceptual evaluation of machine learning systems to improve audio quality of popular music for those with hearing loss
Get PDFMusic is central to many people's lives, and hearing loss (HL) is often a barrier to musical engagement. Hearing aids (HAs) help, but their efficacy in improving speech does not consistently translate to music. This research evaluated systems submitted to the 1st Cadenza Machine Learning Challenge, where entrants aimed to improve music audio quality for HA users through source separation and remixing. The HA users (N = 53, ranging from “mild” to “moderately severe” HL) assessed eight challenge systems (including one baseline using the HDemucs source separation algorithm, remixing to original mixes of music samples, and applying National Acoustic Laboratories Revised amplification) and rated 200 music samples processed for their HL. Participants rated samples on basic audio quality, clarity, harshness, distortion, frequency balance, and liking. Results suggest no entrant system surpassed the baseline for audio quality, although differences emerged in system efficacy across HL severities. Clarity and distortion ratings were most predictive of audio quality. Finally, some systems produced signals with higher objective loudness, spectral flux and clipping with increasing HL severity; these received lower audio quality ratings by listeners with moderately severe HL. Findings highlight how music enhancement requires varied solutions and tests across a range of HL severities. This challenge provided a first application of source separation to music listening with HL. However, state-of-the-art source separation algorithms limited the diversity of entrant solutions, resulting in no improvements over the baseline; to promote development of innovative processing strategies, future work should increase complexity of music listening scenarios to be addressed through source separation
Do socioeconomic inequalities contribute to the high prevalence of child developmental risk in an ethnically diverse, socioeconomically disadvantaged population?:A Born in Bradford's Better Start (BiBBS) study
Get PDFBACKGROUND: Socioeconomic inequalities in child development are pervasive; however, less is known regarding the impacts of socioeconomic factors within and across ethnically diverse and socioeconomically disadvantaged populations. This study (1) describes the prevalence of children at risk of poor overall early child development; (2) investigates the relationship between individual indicators of socioeconomic position and early child development; and (3) investigates if the relationship between indicators of socioeconomic position and early child development varies by ethnic group. METHODS: This study uses data from a prospective birth cohort study, Born in Bradford's Better Start (BiBBS). Child development was measured with the Ages and Stages Questionnaire (ASQ) during routine health visiting appointments at age 2-years-old. Binary logistic regression investigated child development by key maternal socioeconomic indicators: maternal education, financial security, social status (measured via the MacArthur Scale of Subjective Social Status), and social support (measured via number of people to count on). RESULTS: 22% of the 2003 children with a valid developmental assessment were at risk of poor child development. Mothers who had a degree (OR=1.95, 95% CI 1.28 to 2.99), reported 'living comfortably' in financial security (OR=1.78, 95% CI 1.03 to 3.07) and had higher social status (OR=1.11, 1.02 to 1.22); all had higher odds of their child having a good development. Though socioeconomic gradients in maternal education and financial security were consistent across White British, South Asian and Other ethnic groups, both social support and social status had weaker relationships with child development for South Asian parents. CONCLUSION: A high proportion of children are at risk of poor development in this diverse, socioeconomically disadvantaged population. Higher socioeconomic position may protect against poor early development, and the mechanisms underlying this may differ by ethnicity. The findings underline the need for proportionate universal strategies to improve child development in such communities
The Cadenza lyric intelligibility prediction (CLIP) dataset
Get PDFThis paper presents CLIP, a dataset of 11,072 popular western music signals sourced from independent artists, accompanied by ground truth lyrics, and lyric intelligibility scores from listening tests. The dataset is designed to facilitate music information retrieval (MIR) research using machine learning. It was created to allow the development of algorithms to predict lyric intelligibility for the Cadenza ICASSP 2026 Signal Processing Grand Challenge. Currently, it is the only publicly available large-scale dataset for such a task. The music was sourced from the Free Music Archive (FMA) dataset and is unlikely to be familiar to listeners. We excluded tracks whose license did not allow derivative works and those that did not have English singing. Ground truth transcriptions were generated by seven native English speakers, resulting in 3700 excerpts of 5 to 10 words each from 1452 different songs. A hearing loss simulation was also applied to the stereo audio. This resulted in 11,100 music signals with no, mild or moderate hearing loss. This was done so more diverse hearing is represented in the dataset. Human transcriptions were then collected via an online listening experiment. Participants self-reported as having normal-hearing and being native English speakers. They listened to each music signal twice before transcribing each line. Final intelligibility scores were the ratio of matching words between the listening test responses and the ground truth transcriptions. The final dataset consists of audio, ground truth lyrics, intelligibility scores and associated metadata
Translation and transformation of class through migration: rethinking social and spatial mobility across contexts
Get PDFThis issue of Current Sociology Monographs explores new conceptual opportunities at the intersection of migration studies and the sociology of class. The contributions examine migration as a distinct site where class is translated and transformed, using rich, empirically specific cases to show how social class is experienced and produced relationally, transnationally and temporally through the lives of migrants. Through this focus, the papers conceptualise class as a mobile process that travels with migrants, links different contexts and is translated through movement. Distinct from paradigms that tend to universalise a particular narrative or definition of class, this issue adopts a comparative and inductive approach aimed at exploring tensions between the universal and the particular. The contributions share a commitment to qualitative enquiry that attends to the particular, affective and experiential dimensions of class as shared and expressed by migrants, using these as an entry point to retheorise class ‘on the move’. Incorporating ethnography, life histories, intergenerational approaches and spatial mapping, the authors inductively examine how class shapes migrants’ lived experiences in diverse contexts. Some researchers follow participants across borders, whereas others trace the imaginative trajectories through which migrants recall and explain their movements. Contributions orient their analyses around relationality, transnationality and temporality, showing how class emerges as inherently processual and mobile, linking the personal and the global, the embodied and the institutional, and pasts and futures. Each link foregrounds translation and transformation as both an empirical process and a methodological sensibility, demonstrating how migration unsettles the taken-for-granted and provides opportunities to reconsider established theories of class
Modelling and metrics for optimal sizing of renewable power plants supplying green hydrogen generation systems
No full textThis paper presents a modular modelling approach for long-term analysis and design of renewable-powered hydrogen generation and storage facilities, encompassing both power generation and hydrogen system components. The proposed model can be used to integrate different sizes of solar and wind energy resources, different battery energy storage systems, a backup power source (if required), and main hydrogen system modules in power demand calculations. As a part of the paper’s novelty, the proposed modelling approach is modular and case study-free, which allows for generalisation to a variety of case studies. The expandability of the modelling method is strengthened by presenting a unified modelling framework for all modules required in modelling the system. As the second main paper’s contribution, a comprehensive set of performance metrics is proposed to support a multi-objective optimisation framework for optimal sizing of system components. Although the metrics focus on different technical and economic aspects, environmental issues can be covered using some metrics, like the grid share of total energy requirements for the hydrogen system. Both proposed modelling and sizing methods enable renewable power plant designers to evaluate different configurations and make informed decisions based on weighted performance criteria. The proposed model and sizing problem are implemented in a combined Editor and Simulink environment in MATLAB for a case study as a real feasibility study in the UK to operate a renewable-supplied hydrogen system, including a 1 MW electrolyser. Simulation results for the representative case study validate the model’s behaviour and its reliability through various primary output profiles, e.g., power profiles, and secondary outputs, e.g., met hydrogen demand and levelised cost of hydrogen. The proposed modelling and optimisation methods can easily be expanded for case studies with more technical data or different load demands, e.g., combined hydrogen, heat, and power
Comprehending C codes with LLMs: Effective comment generation through retrieval and reasoning
No full textSoftware maintenance requires substantial time for program comprehension. Code comments significantly improve understandability by providing a glass-box view of the code and are thus essential for maintainability. Prior work has analyzed comment attributes, built automated systems to detect irrelevant comments, and applied machine learning to generate meaningful comments. With the rise of large language models, comment generation has accelerated, particularly for Java and Python. In this paper, we present a first-of-its-kind framework for code comment generation in C, a language widely used in low-level tasks. We explore the effectiveness of few-shot learning, retrieval-augmented generation, and code structure based context modeling. Our work builds on prior field studies conducted across seven companies in India and the UK, resulting in a dataset of 20,206 human-annotated C comments rated for usefulness. By 2024, contributions from 40 academic teams and 50 hackathon groups expanded this dataset to 24,578 comments. We further introduce a reusable evaluation framework involving human experts and large language model evaluators, grounded in eight dimensions derived from four industry case studies. A subset of 11,797 comments has been annotated for the presence or absence of these dimensions, serving as both input for generation and evaluation. Our results show that GPT-4o mini-trained models produce comments most aligned with human-annotated ones, achieving a similarity score of 0.64, followed by Gemini 1.5 at 0.58. GPT-4.5 achieves the highest alignment with humans as an evaluator, while Llama-3.1-70b performs the lowest
High-resolution geodetic velocities reveal role of weak faults in deformation of Tibetan Plateau
Get PDFUnderstanding the key mechanisms that control the tectonic deformation of the continents remains a fundamental challenge in geodynamics. We present a high-resolution geodetic velocity field of the Tibetan Plateau, which shows that a few major strike-slip fault systems separate regions of more uniformly distributed deformation. We suggest that focused strain on major fault systems is enabled by relatively low-viscosity ductile shear zones extending through the lithosphere beneath the seismically active fault planes. Simple model calculations show that high slip rates on the Kunlun Fault enable east-west extension to be distributed broadly across the relatively weak southern and central Tibetan Plateau. Activation of the Kunlun fault in the Miocene at the same time as the onset of rifting in the north-south grabens suggests a causal relationship
Multi-perspective machine learning MPML: a high-performance and interpretable ensemble method for heart disease prediction
Get PDFMachine Learning (ML) has demonstrated strong predictive capabilities in healthcare, often surpassing human performance in pattern recognition and decision-making. However, many high-performing models lack interpretability, which is critical in clinical settings where understanding and trusting predictions is essential. To achieve our objective, we proposed a Multi-Perspective machine learning framework (MPML) that combines established base classifiers with structured perspective-based design and interpretability pipeline. MPML organises features into meaningful subsets, or perspectives, enabling both global and instance-level interpretability. Unlike traditional ensemble methods such as Bagging, Boosting, and Random Forest, MPML delivers significantly higher-quality predictions across all evaluation metrics while maintaining a transparent structure. Applied to a heart disease dataset, MPML not only improves predictive accuracy but also provides detailed, accessible explanations for individual patient outcomes, advancing the potential for practical and ethical deployment of ML in healthcare
Embryonic upgrading in agri-food value chains:an analysis of small-scale environmental initiatives in the Mexican beef sector
No full textIn this paper, I investigate a series of recently introduced initiatives that aim to promote environmental sustainability governance in the Mexican beef industry. Drawing on the Global Value Chain (GVC) scholarship, I characterise these initiatives as environmental upgrading processes. The analysis begins by mapping the GVC to identify the Mexican beef chain's governance dynamics. This map is then used to investigate the extent to which chain actors are responding to a discernible external governance driver—specifically, a 'lead firm' or 'big buyer'—or if they are engaging with sustainability governance initiatives independently through bottom up or ‘horizontal’ governance processes. Based on a novel set of 30 key-informant interviews and triangulation with secondary data, I highlight the drivers for chain actors' engagement and leadership, as well as the role of non-firm actors in the environmental upgrading processes for Mexican beef. My findings show the extent to which environmental upgrading processes can and do occur in the absence of external pressure from downstream firms, while noting that in the specific case of Mexico these initiatives remain embryonic, their scalability uncertain and their environmental benefits are yet to be documented
The Michael Mason prize: development and feasibility testing of a complex intervention to improve adherence to fracture prevention medicine
Get PDFObjectives
Only 12% of people recommended fracture prevention medicines remain on treatment 1 year post fracture. The ‘improving uptake of Fracture Prevention treatments’ (iFraP) intervention aims to improve shared decision-making (SDM) about, and uptake of, osteoporosis medicines in Fracture Liaison Services (FLS). This paper details development and feasibility.
Methods
Intervention development was underpinned by (i) theories of SDM, medicines adherence and behaviour change; (ii) integrated findings from seven development studies; and (iii) extensive patient and clinician contribution, identifying key ‘needs’ to address and the intervention’s content, functionality and scope. Feasibility testing was conducted at one English FLS. Intervention consultations were observed and audio recorded. Interviews completed with FLS clinicians and patients explored perceived acceptability and feasibility.
Results
Intervention development identified patient and clinician unmet needs for personalized and evidence-based information about osteoporosis, its consequences, and its treatment within and after FLS consultations, to facilitate clinical and SDM about medicines. The prototype intervention (osteoporosis decision support tool, clinician skills training and information resources) was designed to meet identified needs and overcome barriers to use. Clinicians delivered the prototype iFraP intervention in 10 consultations with consenting patients. Findings demonstrated that the intervention was acceptable and feasible to deliver, with potential to improve patient outcomes. The intervention was refined to support implementation.
Conclusion
The multi-facilitated approach to intervention development and testing ensured that the iFraP intervention appears acceptable and feasible for use in UK FLS to support SDM about osteoporosis medicines. The iFraP trial will evaluate implementation, and cost and clinical effectiveness
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇