36270 research outputs found
Sort by
Coordinator: Semantic-Coordinated Dynamic Fusion for Multi-View Clustering
No full textDeep multi-view clustering (DMVC) has become an important approach for analyzing complex data from multiple views. While existing deep methods perform well in extracting view-specific features, they often overlook high-level semantic alignment across views. This can lead to feature ambiguity and inconsistent results, particularly with heterogeneous data. To address this issue, we propose a novel semantic-coordinated dynamic fusion framework (Coordinator) to integrate semantic information into multi-view clustering. Specifically, Coordinator utilizes Large Vision-Language Models (LVLMs) to generate semantic representations and view encoders to extract view-specific features. To enhance consistency across heterogeneous views, Coordinator employs a cross-view alignment module to align representations of different views. Furthermore, a semantic-aware fusion strategy dynamically integrates visual and semantic features, beginning with semantic guidance and transitioning to data-driven view weighting. In order to promote collaborative optimization of the view encoders, Coordinator introduces a bidirectional feedback mechanism to refine the view encoders. Extensive experiments on six benchmark datasets demonstrate that Coordinator outperforms state-of-the-art methods, effectively bridging the modality gap by incorporating semantic knowledge into multi-view clustering.</p
Historical Institutionalism and Transnational Influence: Social Policy Responses to the Great Depression in the United States and Canada
No full textThe Great Depression was a turning point in the development of social programming in NorthAmerica. This paper explores the politics of social policy expansion during the Great Depressionin the United States and Canada through an analytical lens that combines the insights ofhistorical institutionalism and the analysis of transnational processes. Thus, the paper focuses onthe differences between these two cases as distinct entities as well as how they were shaped by,and influenced one another through, transnational processes. This combination of historicalinstitutionalism and close attention to transnational processes adds insight to the existingscholarship on the politics of welfare state development that tends to prioritize one perspectiveover the other. Empirically, the paper compares these politics in two discrete social policy areas:old-age pensions and unemployment insurance. The puzzle tackled is the following: why did theU.S. federal government create new programs in these areas during the 1930s while Canada didnot; and, simultaneously, how did these U.S. policy developments influence Canada during andbeyond that decade alongside other transnational trends and processes? As argued, the distincttiming and pace of reform and the role of courts explain why Canada did not implement majorpublic pension and unemployment programs during the Great Depression. This contrasts with theU.S., where the advent of the New Deal led to the creation of a federal pension program and adecentralized system of unemployment insurance. Over time, these policy innovations impactedwelfare state development in Canada through transnational influence.</p
Juno radio occultations reveal the structure of Jupiter’s cold northern polar vortex
No full textJupiter’s polar upper troposphere and stratosphere host a persistent cold vortex poleward of 65°N, but its detailed structure and dynamics have remained difficult to resolve.
The goal is to characterize the thermal structure and dynamics of the polar vortex using new and complementary remote sensing techniques.
We used a combination of high-resolution vertical profiles derived from Juno’s recent radio occultation measurements and mid-infrared imaging from the VLT/VISIR instrument. The former provided direct retrievals of temperature and density near and within the vortex, while VISIR imaging revealed spatial thermal contrasts across the region.
Our analysis confirms the presence of a steep meridional temperature jump at 65°N, of about 7±1,K at 100 mbar, which is consistent with a strong vertical wind shear and a prograde polar stratospheric jet reaching up to 80 -1 at the 10 mbar level. We find the atmosphere to be thermally stable above 0.55 bar, reaching a Brunt–Väisälä frequency of 0.025 in the mid-stratosphere. Thermal contrasts observed in the infrared data align with the vertical structures inferred from radio occultations, which validates the presence and extent of the cold vortex.
These findings offer a quantitative analysis of the thermal structure and the dynamical behavior of Jupiter’s polar atmosphere and demonstrate the diagnostic power of combining radio occultation and thermal infrared techniques in planetary atmospheric studies.</p
Performance of the Leicester risk assessment and Leicester practice risk scores for assessing the risk of undiagnosed type 2 diabetes or prediabetes in diverse populations: protocol for a systematic review of published validations and updates
No full textBACKGROUND: Approximately one million adults in the UK are estimated to have undiagnosed type 2 diabetes mellitus (T2DM), with a further 5.1 million adults with nondiabetic hyperglycaemia (NDH) that does not meet the threshold for a diabetes diagnosis. The Leicester Risk Assessment score (LRA) and Leicester Practice Risk score (LPR) are diagnostic risk prediction models that estimate an individual's risk of undiagnosed T2DM and NDH, developed for use in community and primary care settings respectively. The LRA is also used as a prognostic model; neither model has been updated since development. This study will systematically review all applications of these models as diagnostic and prognostic tools and any published updates to evaluate their performance in different populations. This review has been registered with PROSPERO (CRD420251005841). METHODS: We will implement a citation search strategy to search Scopus, Web of Science and Google Scholar, restricted to full text, English language papers. Eligible papers will validate, update or modify either model. Data will be extracted using a form based on the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist; missing information will be sought from authors or estimated from other available information where possible. Meta-analysis of predictive performance measures will be completed if sufficient data exist. Subgroup and sensitivity analyses will be used to explore between-study heterogeneity and risk-of-bias impact. DISCUSSION: This review will identify studies that have implemented, modified or validated the LRA and LPR for the risk of undiagnosed T2DM and NDH in different populations. This will allow summary measures, including level of uncertainty, of model performance to be calculated, making this highly relevant to individuals and stakeholders who recommend and implement these models. Review conclusions will also inform the potential update and recalibration of the models. This will ultimately lead to improved outcomes through earlier diagnosis and management.</p
<b>Comparative Governance of Newborn Whole Genome Sequencing: Ethical, Legal, and Social Considerations in the UK and Nigeria (</b><b>Poster presented at the</b><b> International Conference on Newborn Sequencing 2025 , ICoNS '25</b><b>, London, UK, 23–24 October 2025).</b>
No full textIntroduction:The expansion of whole genome sequencing (WGS) into newborn screening raises complex ethical, legal, and social implications (ELSI), particularly concerning consent models, long-term data storage, privacy, equity, and the child’s evolving autonomy and medical rights. These challenges become more pronounced when viewed across jurisdictions with vastly different healthcare infrastructures and legal capacities. This study compares regulatory and governance frameworks for newborn WGS in the United Kingdom (UK) and Nigeria to explore context-specific ELSI challenges.Methods:A systematic literature review and legal doctrinal analysis were conducted, focusing on peer-reviewed publications, policy documents, regulatory guidelines, and grey literature relating to newborn WGS governance in both countries. The study used a reflexive balancing framework to analyse and compare stakeholder positions, existing policies, and regulatory gaps.Results:Preliminary findings highlight significant contrasts. The UK demonstrates structured regulatory oversight through integrated ethical review processes, evolving consent models (including dynamic consent), and detailed genomic data stewardship guidelines. Conversely, Nigeria exhibits regulatory underdevelopment, fragmented oversight mechanisms, limited infrastructural support, and ethical pluralism that complicates uniform governance. Cross-cutting themes include the tensions between parental authority and the long-term interests of the child, as well as issues related to equitable access to genomic technologies.Discussion and Conclusion:This comparative analysis demonstrates the importance of culturally responsive, jurisdiction-specific governance models that address both universal ELSI concerns and local healthcare realities. The findings suggest that international genomic screening efforts would benefit from adaptable ethical frameworks that respect global diversity while upholding foundational principles of justice, autonomy, and beneficence. The study offers actionable insights for policymakers, healthcare professionals, and global genomic governance bodies working towards ethically sustainable integration of WGS into newborn screening programmes worldwide.</p
Towards Wetter Peatland Landscapes for Climate, People and Nature
No full textEnabling lowland peatland rewetting and paludiculture in UK landscapesThis Policy Briefing Paper was produced by Zoe Lipkens from the University of Leicester, with contributions from Professor Caroline Upton, Dr Jacqueline Kirk and Professor Susan Page, with support from the University of Leicester Institute for Policy. This Policy Briefing Paper draws on research from the University of Leicester on the perception, implementation and development of lowland peatland policy in the Northwest of England. These insights inform a series of policy recommendations aiming to support the transition to wetter peatland systems, and contribute to UK government peatland rewetting and restoration goals. </p
Self-Harm in Primary-Aged Children: Insights from Children, Parents, and Educators
No full textRates of self-harm in England continue to rise, with growing evidence that children under 12 years are engaging in such behaviours. While self-harm in adolescents and adults is well-researched, limited empirical understanding exists regarding how it is understood and experienced in primary-aged children. This gap is significant, as parents and educators, often on the frontline of responding, are identified in research as feeling ill-equipped to provide effective support.This inquiry adopted an exploratory, multi-method qualitative design to address this gap. Data were generated through: (1) focus groups with educators (n = 26; 25 females, 1 male), parents (n = 19; 17 mothers, 2 fathers), and children aged 7-11 years old (n = 24; 15 girls, 9 boys); (2) a triad of interviews with a child who was known to self-harm, their parent, and their Special Educational Needs Coordinator (SENCO); and (3) naturally occurring data comprising of 85 online counselling transcripts from children aged 7-11 years old seeking help for self-harm. Data were analysed using Template Analysis.Findings indicate that self-harm in primary-aged children is recognised by children, parents, and educators, but is often surrounded by uncertainty regarding its definition, meaning, and boundaries. Across the datasets, self-harm was most often understood as a means of managing difficult emotions. Children’s accounts revealed a range of contextual influences, including family adversity, peer relationships, academic performance, and perceptions of support. Educators and parents emphasised the lack of access to specialist services for this age group and the need for bespoke training in schools.By centring the voices of children alongside those of parents and educators, this inquiry extends understanding of self-harm in younger populations and highlights the need for clearer, developmentally informed frameworks and policies for recognition and response. Recommendations focus on improving early identification, fostering supportive environments in schools, and strengthening adults’ capacity to provide timely, sensitive, and effective help for primary-aged children who self-harm.</p
A Tree Without Roots: An Exploration of Femicide in Jamaica Through a Decolonial Feminist Lens
No full textThis thesis explores femicide in Jamaica and evaluates the effectiveness of women’s rights laws in preventing it. Using a socio-legal methodology, it examines Jamaican societal and cultural norms to identify the historical and structural factors contributing to high femicide rates and to propose strategies for prevention. It argues that femicide in Jamaica is a complex phenomenon rooted in the country’s colonial history, where the legacies of slavery, racial and gender hierarchy continue to shape social relations. The thesis emphasises the importance of naming these killings femicide, as recognition can foster public awareness and collective action. International and regional women’s rights instruments such as the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW) and the Inter-American Convention on the Prevention, Punishment, and Eradication of Violence Against Women (BDP), to which Jamaica is a party, are examined as potentially valuable for preventing femicide in Jamaica. The thesis highlights that neither of the Committees overseeing these Conventions, as reflected in their communicative documents to Jamaica, nor Jamaica, in its interpretation of these instruments, acknowledges Jamaica’s colonial past and its ongoing legacies on Jamaica’s socio-cultural norms. This omission suggests a failure to contextualise contemporary GBVAW, including femicide, within Jamaica’s historical and structural inequalities, thereby limiting the effectiveness of these instruments in addressing the deeper socio-cultural roots of femicide in Jamaica.For the purposes of this thesis, empirical qualitative research was conducted through interviews with senior members of the Jamaican Constabulary Force (JCF) as well as representatives from other governmental and non-governmental organisations involved in the prevention of GBVAW and femicide. Overall, the thesis findings suggest that Euro-colonial legacies contributed to the development of Post-Traumatic Slave Syndrome (PTSS), a manifestation of intergenerational trauma. PTSS is a concept developed by Dr DeGruy to describe the adaptive survival responses and psychological patterns resulting from centuries of enslavement. These enduring effects, combined with entrenched colonial legacies, perpetuate cycles of violence and contribute to Jamaica’s high prevalence of femicide. To effectively address femicide in Jamaica, the thesis proposes a multidimensional decolonial feminism praxis that draws on Pan-Africanism concepts such as revolutionary decolonisation and re-Africanisation. This is a critical progressive framework that rejects colonial hierarchies while reclaiming and revaluing non-Western knowledges and communal practices. This approach represents a critical return, resisting conservatism and dismantling oppressive colonial structures and gendered hierarchies to envision egalitarian and community-centred relations. Revolutionary decolonisation and re-Africanisation are central to this framework because they seek to revitalise African cultural values and collective practices that were suppressed under colonial rule and draw on positive Western concepts to create a new normal. Based on this notion, the thesis argues that women’s rights law, including CEDAW and the BDP, can serve as valuable Western tools to Jamaica’s femicide prevention measures. However, they must be critically assessed and adapted to ensure that their interpretation and implementation address local realities, engage with colonial legacies and align with a decolonial feminist praxis, without supplanting it.</p
State of charge estimation considering cell-to-cell differences using two-stage transfer estimation
No full textAccurate state of charge (SOC) estimation for lithium-ion batteries is a vital challenge hindered by inherent cell-to-cellvariations and the lack of labeled data in target cells. To address thelimitations, this paper proposes an unsupervised two-stage transferestimation (TSTE) framework, integrating a fuzzy long short-termnetwork (FLSN) that utilizes residual connections, parallel LSTMbranches, and a mixed-Gaussian fuzzy block to model the nonlinearand noisy dynamics of the battery. TSTE operates in two distinctstages. First, it pre-trains a source-domain SOC estimator and anerror corrector designed to map the discrepancy between source andtarget domain measurements. Second, a cyclic adversarial trainingscheme adapts the pre-trained SOC estimator for the target domainby delivering domain-invariant representations. This adaptationprocess is performed without any target-domain SOC labels.Experiments on self-collected and public datasets demonstrate thesuperiority of TSTE. Under diverse operating conditions, initialSOCs, current noise levels, temperatures, and the distinct batterytype, the proposed method outperforms comparative approacheswith an overall mean absolute error of 0.82%. Ablation studiesverify the contributions of each architectural component of TSTE.Overall, TSTE provides a robust and high-precision solution forunsupervised, cell-to-cell SOC estimation, effectively addressing thechallenge of label scarcity.</p
Exploring Novel Deep Learning Frameworks for Blood Cell Classification: From Algorithmic Design to Data Enhancement and Efficient Adaptation
No full textDeep learning has been widely adopted in medical image analysis due to its ability to learn informative representations. However, its application to blood cell classification remains challenging. In particular, the limited availability and low quality of annotated blood cell datasets can lead to overfitting when training parameter-heavy models, thus hindering generalization and downstream diagnostic utility. Blood cell analysis presents unique challenges, including high inter-class similarity, subtle morphological differences, and a high dependency on expert annotations. These factors make the task particularly sensitive to overfitting and call for domain-specific adaptations. This thesis aims to address the issue of data scarcity and model overfitting in blood cell classification through a comprehensive strategy spanning algorithm design, data-level enhancement, and practical deployment. At the algorithmic perspective, a parameter-efficient model named ReRNet is proposed, in which the reliance on large pre-trained backbones is reduced by incorporating lightweight randomized neural layers. At the data perspective, generative adversarial network (DLBCNet) are proposed to synthesize realistic blood cell images, mitigating the lack of diversity and annotation in existing datasets. At the application level, an adapter-based method (LKA) with large kernel attention is introduced to effectively capture global contextual features while keeping computational cost manageable. The proposed methods are evaluated not only on blood cell datasets but also validated across other medical imaging modalities to demonstrate broad adoption. Overall, this work presents a targeted approach to improving the reliability and applicability of deep models in low resource medical scenarios, with a specific focus on blood cell classification.</p