47011 research outputs found
Sort by
Optimizing super-feature selection for machine learning-enhanced spectroscopic analysis in biomedical research
Get PDFPurposeMachine-learning-powered label-free infrared spectroscopic methods offer significant potential for diagnostic and biomedical applications. However, their applications have been limited by spectral noise, where critical features are often obscured by overlapping bands and data redundancy. Although various feature selection methods have been proposed, many suffer from limitations in consistency and interpretability. To address these challenges, we introduce a novel multi-model machine learning approach that integrates five distinct algorithms to identify a set of “super-features”—spectral features consistently deemed significant across all models.Principal resultsThis novel workflow outperforms traditional algorithms, achieving superior classification accuracy (>99%) in distinguishing infected from healthy cells, despite using fewer spectral features. To ensure robustness and generalizability, we developed a comprehensive validation strategy that includes independent classifier evaluations, label randomization, and unsupervised analyses. Importantly, the identified super-features accurately differentiated infection states across multiple time points and enhanced the biological interpretability of infection-associated biochemical changes.ConclusionsThese findings highlight the potential of advanced multi-model feature selection techniques to enhance the diagnostic power of spectroscopic data in biomedical research, offering high accuracy and valuable biological insights into infection progression
Future opportunities for delivering primary care education – the political drivers of change
Get PDFGrowing patient demand, changing patient need, technological advancements and workforce challenges have resulted in a complex healthcare system in the United Kingdom (UK) that is unable to meet the needs of its populations. Consequently, reform is needed to ‘fix the broken National Health Service (NHS)’. In response, the UK government published its 10-year Health Plan, identifying strategic shifts to overcome the challenges, including from hospital to community, from sickness to prevention, and from analogue to digital. These changes in priorities for the NHS now provide an opportunity to reimagine the way in which not just general practice (GP) but education in primary care is both structured and delivered to undergraduate medical students. This article explores ideas for designing ‘fit for the future’ primary care education against this political backdrop. In doing so, the opportunities are discussed for training provided by the broader non-GP healthcare professional team, collaboration and co-creation of curricula with community partners or third-party healthcare organisations, and the role of heath informatics and technology enhanced learning for workforce development. Furthermore, the importance of consolidating and rationalising education curricula in parallel to growth and innovation is highlighted, to ensure they remain at the forefront of primary care whilst continuing to provide meaningful learning experiences
Buffering heat fluctuation of IGBT power electronic modules using phase change material-based thermal management for wind power generation
Get PDFThe insulated-gate bipolar transistor (IGBT) power electronic module is an essential component of the converter system for wind power generation. However, it suffers from the heat fluctuation induced by transient output power from the wind turbulence, which makes it vulnerable and easy to fail. In this study, we developed a phase change material (PCM)-based thermal management solution whose structure is simple and does not change the existing heat sinks. PCM absorbs most heat when the power loss surges, but its temperature stays stable owing to the solid-liquid phase transition. That reduces the peak junction temperature, and since the decrease in peak temperature is larger than that in the valley temperature, the heat fluctuation of chips is buffered. This solution decreases heat fluctuation by 8.7 °C at most and has the best buffering performance where heat fluctuation is largest. The melting fraction of PCM does not exceed 1 during the thermal management process, indicating that the PCM can work sustainably. Different types of PCM were examined, and the results indicate that PCM2’s melting fraction responds more sensitively to power loss and absorbs more power loss as the latent heat energy, thus has better buffering performance than PCM1. The lifetime of the IGBT module is also predicted and is extended significantly. The PCM-based thermal management solution developed in this study is a promising approach to improve the robustness and reduce the maintenance cost of the IGBT modules in wind power generation
Pressure sensing technology for remote control: Can we motivate users to stay on the learning curve?
Get PDFLearning to use a novel human-in-the-loop control system is often a slow and frustrating process due to the need to understand new interaction paradigms. Unsurprisingly, research has commonly focused on identifying methods to accelerate such learning. In this paper we consider an alternative approach of motivating learners to persist with their learning. If learners are motivated to continue investing time in using a novel control system they will transition to proficiency, albeit at different timescales. Participants controlled the movements of a virtual robot in real-time by adjusting their movements whilst seated on a pressure sensing mat. In two experiments, participants played a game where their task was to move a virtual robot to collect targets as quickly as possible. Targets were only presented for a fixed duration such that participants received binary reward feedback dependent upon whether they collected a given target in time or not. This feedback was used to calculate each participant’s success frequency which was used as a proxy for their skill level and thus learning. Experiment 1 showed that participants could learn the control system but that their motivation to play the game decreased as the experiment continued. Experiment 2 investigated whether adapting task difficulty as a function of the participants’ current skill level (indexed by success frequencies) would increase the time participants chose to invest playing the game. Participants did not choose to play the game for longer when playing the game under this adaptive difficulty condition compared with fixed difficulty conditions. We conclude that most participants improved at using the pressure sensing mat for remote control but adapting the difficulty of the task to participant’s skill level did not increase the time they were willing to invest in playing it
β-catenin/TCF4/NANOG axis controls miR-302 transcription in colorectal cancer cells
Get PDFThe miR-302 cluster, a key pluripotency-associated non-coding RNA, has been implicated in stem cell homeostasis and tumourigenesis. However, its regulatory mechanisms in cancers, including colorectal cancer (CRC) remain poorly understood. Here, we demonstrate that the β-catenin/TCF4 complex significantly enhances miR-302 expression through direct promoter activation in CRC cells. We hypothesized that the β-catenin/TCF4 complex directly activates the miR-302 promoter and cooperates with NANOG in a transcriptional feedback loop sustaining stem-like traits in CRC cells. Using a combination of promoter-driven luciferase reporter assays, chromatin immunoprecipitation (ChIP), and molecular dynamics simulations, we identify a regulatory axis involving Wnt signalling and the transcription factor NANOG. Our data show that individual members of the miR-302 cluster activate the NANOG promoter, while both NANOG and β-catenin/TCF4 synergistically enhance miR-302 promoter activity, suggesting the presence of a positive feedback loop. Structural simulations further elucidate the binding interactions between TCF4, NANOG, and the miR-302 promoter, corroborating our experimental observations. Together, these findings position miR-302 as a downstream effector of Wnt/β-catenin signalling and an integral component of NANOG-mediated transcriptional networks in CRC stem-like cells. This work advances our understanding of non-coding RNA regulation in cancer and highlights potential therapeutic opportunities for targeting stemness-associated pathways
Location, housing and employment opportunities: Evidence from a randomized controlled trial with vulnerable youth in France
No full textHousing conditions, residential location, and employment are key determinants of individual welfare, particularly for vulnerable populations facing credit constraints and information frictions. We examine how housing assistance affects employment outcomes using a randomized controlled trial in France that provided vulnerable youth (aged 18-25) with both job search assistance and housing support, including rent guarantees. The program successfully improved housing conditions: beneficiaries experienced better accommodation stability, reduced precarious situations, and increased satisfaction with their housing. However, despite substantial social worker support, the program did not improve employment rates, contract types, or earnings. Strikingly, beneficiaries moved to neighborhoods with objectively worse employment opportunities and lower socioeconomic indicators, yet reported higher satisfaction with their residential areas. This apparent paradox reveals that beneficiaries appear to prioritize housing affordability and conditions over employment access. Our results suggest that successful interventions may need to explicitly balance housing improvements with maintaining access to employment opportunities
Steering through uncertainty: a systematic review of liability communication in autonomous vehicles
No full textAutonomous vehicles (AVs) represent a potential technological transformation of transportation systems, however, incidents involving them have highlighted the complex challenge of assigning liability. While a growing body of literature addresses legal and technical liability, the communication of liability–how legal, moral, or financial responsibility for adverse outcomes is conveyed among stakeholders such as manufacturers, users, insurers and policymakers–remains a critical gap. This multidisciplinary systematic literature review analyzes 90 academic articles published between 2015 and 2024 across a range of disciplines to map the current state of liability communication. Specifically, it examines how liability is communicated: who or what is held accountable for potential harms, under what conditions and through what mechanisms. We find that liability communication is often reactive, inconsistent and poorly aligned with public understanding. Despite the development of expert legal and technical frameworks, communication practices frequently fail to bridge the gap between expert discourse and end-user comprehension. The analysis is organised across five key themes: governance challenges; safety concerns; ownership models; cross-country comparisons; and future AV deployment. Across all five, communication failures are consistently linked to ambiguous terminology and the absence of proactive, standardised protocols. Together, these themes contribute to a more nuanced understanding of how liability is communicated within the evolving AV ecosystem. They also highlight an urgent need for updated policies and more effective, stakeholder-oriented communication strategies. In response, this study offers a necessary reframing of the problem–calling for the development of stakeholder-centric communication practices capable of functioning even amid legal uncertainty. Addressing these challenges is essential not only for effective AV integration but also for ensuring that this transformation unfolds safely and equitably
Legitimizing Penality: Human Rights Discourses and Punitivism in Colombia
No full textDespite their benevolent appearance, transnational discourses of human rights have played a substantial role in fueling the deployment of coercive power. Yet, their impact on domestic systems, particularly in the Global South, remains largely unexplored. This article examines the interaction of these discourses with local practices, processes, and narratives in the Colombian context. I argue that, rather than being merely imposed or imported, notions of victims’ rights, anti-impunity, and security, have forged unique coercive pathways in domestic penal policy and constitutional law that transcend transnational obligations. On the one hand, they have played a key role in the justification of increased coercion throughout the criminal law as a whole, rather than being restricted to serious human rights violations. On the other hand, although the language of human rights has been instrumental in denouncing rights violations in prisons, it has also played a crucial role in legitimizing the imposition of punishment under conditions recognized as unconstitutional
Microwave depolymerisation of PMMA: Power density, char management, and reactor design for high-purity monomer recovery
Get PDFPoly(methyl methacrylate) (PMMA) is a versatile polymer valued for its optical and mechanical properties, durability and biocompatibility. However, its production carries a high carbon footprint, primarily due to the energy-intensive synthesis of its monomer, methyl methacrylate (MMA), while recycling rates remain below 10 %, largely due to limitations in current conventionally heated technologies. Microwave depolymerisation offers a promising route to address these challenges by enabling greener, more energy-efficient recovery of the monomer. However, due to the unique microwave heating mechanisms, innovation in microwave reactor design is required to realise this at industrial scale. This study defines the key processing parameters required for successful scale-up of microwave depolymerisation of PMMA into high-purity MMA. A semi-continuous microwave reactor was developed to explore these parameters under controlled conditions. Electromagnetic simulations were employed to optimise reactor performance and provide insight into the observed process behaviours. For the first time, the roles of power density and char accumulation in determining product quality during microwave depolymerisation of PMMA are quantitatively established, highlighting the need for reactor designs that enable high power density and effective char removal. These findings define key design principles for microwave reactor scale-up and represent a critical step toward economically and environmentally viable PMMA recycling within circular plastic systems
A systematic review of English medium instruction in higher education: An update of Macaro et al. (2018)
Get PDFThis paper updates Macaro et al.'s (2018) widely cited systematic review of English Medium Instruction (EMI) in higher education, synthesising research published between 2016 and 2023. With EMI research expanding exponentially, this review assesses global trends, language and subject learning outcomes, and language-related challenges in EMI contexts. Due to the growth in research, this update only addresses a subset of the original review questions, namely those focusing on language and content issues. A total of 196 empirical studies were systematically selected and analysed. Findings indicate continued growth in EMI research, with Asia and Europe as the most studied regions. Evidence on language learning outcomes suggests that EMI contributes positively to receptive skills and vocabulary, though productive skills yield mixed results. Regarding content learning, EMI appears not to negatively impact subject attainment, with some studies reporting comparable or even superior outcomes compared to L1 instruction. Language challenges persist, particularly in academic writing and transitions from L1 instruction, though they tend to subside with increased EMI exposure. Research gaps include limited longitudinal studies, underrepresentation of postgraduate EMI, and a lack of qualitative insights into sociolinguistic and pedagogical dimensions. The review highlights the need for a broader research agenda, incorporating EMI's impact on employability, sociolinguistic equity, and long-term academic success. It concludes with recommendations for future research and methodological advancements in EMI studies
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! 👇