53417 research outputs found
Sort by
A novel framework for fully-automated co-registration of intravascular ultrasound and optical coherence tomography imaging data
Get PDFREHEARSE-3D: A Multi-Modal Emulated Rain Dataset for 3D Point Cloud De-Raining.
Get PDFSensor degradation poses a significant challenge in autonomous driving. During heavy rainfall, interference from raindrops can adversely affect the quality of LiDAR point clouds, resulting in, for instance, inaccurate point measurements. This, in turn, can potentially lead to safety concerns if autonomous driving systems are not weather-aware, i.e., if they are unable to discern such changes. In this study, we release a new, large-scale, multi-modal emulated rain dataset, REHEARSE-3D, to promote research advancements in 3D point cloud de-raining. Distinct from the most relevant competitors, our dataset is unique in several respects. First, it is the largest point-wise annotated dataset (9.2 billion annotated points), and second, it is the only one with high-resolution LiDAR data (LiDAR-256) enriched with 4D RADAR point clouds logged in both daytime and nighttime conditions in a controlled weather environment. Furthermore, REHEARSE-3D involves rain-characteristic information, which is of significant value not only for sensor noise modeling but also for analyzing the impact of weather at the point level. Leveraging REHEARSE-3D, we benchmark raindrop detection and removal in fused LiDAR and 4D RADAR point clouds. Our comprehensive study further evaluates the performance of various statistical and deep learning models, where SalsaNext and 3D-OutDet achieve above 94% IoU for raindrop detection
Post-graduate pedagogy in neuroliberal times: affect, governance and volatility in higher education
Get PDFSituated within critical debates in education, this paper examines the challenges faced by teachers of post-graduate students through the emerging concept of neuroliberalism. The latter draws attention to how trends in brain science, affect studies and technology are reshaping the ways in which the roles and positionalities of HE practitioners are conceptualised. In a field inflected by long-standing debates about reflexion, neuroscience and educationalisation, neuroliberalism is used as an overarching governmental principle to bring together seemingly disparate strands of policy and practice. Drawing on analyses of qualitative data from interviews with teachers on post-graduate programs, we discuss how our interviewees’ concerns about HE practices signal a neuroliberal agenda focused on meta-cognition, affect and techno-hybridity. Positioning this study within wider debates about neuroliberal governance, we suggest that while the latter presents itself as more humanistic and less technocratic, its effects are politically regressive. It intensifies existing power asymmetries and consolidates structures that normalise and extract from an increasingly volatile model of HE practice and provision
Contribution of platelet-monocyte complexes to thrombo-inflammation and cardiovascular disease.
No full textThe crosstalk between platelets and monocytes has emerged as a key mechanism connecting thrombosis and inflammation. Platelet-monocyte complexes (PMCs) were discovered decades ago, with elevated levels associated with negative clinical outcomes across a range of diseases including cardiovascular disease. Despite their clinical significance, a causative role remains unproved with the specific physiological and pathological roles of PMCs in circulation remain poorly characterised. This thesis hypothesises that PMCs exhibit a more pro-inflammatory and pro-coagulative phenotype than monocytes alone. To explore this, PMCs were formed in vitro using platelet GPVI agonist collagen related peptide (CRP) in both whole blood and isolated cell preparations. Rapid changes in surface receptor expression were observed within 30 minutes using a novel 18-colour spectral flow panel which highlighted that receptors associated with adhesion (CD62L, CD11B), chemotaxis (CXCR4, CX3CR1) phagocytosis (CD36) and tissue factor (CD142) were all significantly higher expressed on PMCs compared to monocytes. Distinct transcriptomic differences between PMCs and monocytes were detected as early as two hours post stimulation, which became more pronounced by six hours. Pathway analysis of differentially expressed genes at six hours highlighted enrichment in pathways related to leukocyte migration, chemotaxis, differentiation, and wound healing, all indicative of a heightened inflammatory state. Functional assays supported these findings, identifying key differences in cytokine production, enhanced phagocytosis of Candida albicans by PMCs and a role in coagulation. Taken together, these data suggest a dual role for PMCs in both augmenting clot formation and supporting inflammation. The observed PMC phenotype aligns with an atherothrombotic pathophysiology and therefore could be contributing to the worsened clinical outcomes for patients, highlighting PMCs as a promising target for therapeutic intervention. Finally, the phenotypic approaches developed during this PhD will support future research to validate these findings in both localized haemostatic contexts and systemic thrombo-inflammatory conditions
Simplification, not deregulation: reframing financial regulation in the UK and the EU
No full textSimplification is not deregulation. Simplification represents a conscious effort to streamline and rationalise financial regulation without compromising prudential safeguards or conduct standards. It is a process of legal clarification and institutional learning aimed at making complex rulebooks coherent, proportionate and intelligible. The UK’s approach. In the post-Brexit landscape, the UK has embraced simplification as a strategic tool to sustain its position as a leading international financial centre. Through initiatives such as the “Smarter Regulatory Framework” simplification can operate as a principle of proportionate and clear supervision and regulation. The EU’s approach. The EU’s simplification agenda, as advanced in the 2024 Draghi Report, purports to consolidate the fragmented financial acquis into a coherent Single Rulebook, to restore the hierarchy between legislative levels, and to embed proportionality and digital accessibility in regulatory design. Simplification in the EU context seeks to render complex legislation more intelligible while preserving the robustness of the prudential and conduct regimes. However, simplification per se does not create growth. As the Draghi Report highlights, a more integrated EU capital market is a crucial catalyst for growth and competitiveness, while the lack of productivity and competitiveness requires attracting more investment through a “Savings and Investments Union”. The US’s approach. The US regulatory approach illustrates the risks of conflating simplification with deregulation. Since 2018, the relaxation of some prudential standards – justified in part in the name of innovation and market dynamism – has contributed to supervisory “blind spots” (for mid-sized banks) and liquidity mismatches culminating in the 2023 banking turmoil (Silicon Valley Bank). While the US model is characterised by speed and experimentation, it often redistributes rather than eliminates risk
Exploring the Role of Endothelial Cells in the Development of Diastolic Dysfunction Associated with Inflammatory Arthritis
Get PDFRheumatoid Arthritis (RA) is an immune mediated inflammatory disease that carries a two-fold increased risk of developing heart failure compared to the general population. Most cases of heart failure in the RA population present as heart failure with preserved ejection fraction (HFpEF). Although the precise mechanisms driving HFpEF are not fully understood, endothelial cell (EC) dysfunction and activation in the heart and bone marrow are thought to play a role. These processes involve alterations in nitric oxide (NO) bioavailability, elevated reactive oxygen species (ROS), increased expression of adhesion molecules, and heightened hematopoietic activity. Here, I have utilised the KBxN-F1 mouse model of inflammatory arthritis to investigate the mechanisms driving cardiac dysfunction in the context of inflammatory arthritis. KBxN-F1 mice develop diastolic dysfunction at around 6wks of age, 2 weeks after development of a chronic polyarthritis. To explore endothelial function in these mice, I used flow cytometry to examine endothelial and immune cell phenotype in the bone marrow, spleen, blood, lungs and hearts of KBxN-F1 mice over time. In comparison to control mice, KBxN-F1 mice exhibited an expansion of bone marrow endothelial cells, accompanied by increased CD62E expression and elevated ROS levels, which can contribute to hematopoietic stem cell recruitment and activity. In the spleen I also found an increase in ICAM-1 expression and an expansion of pro-inflammatory monocytes. In the heart, I found evidence of vascular rarefaction and elevated ROS levels in endothelial cells. Additionally, proteome profiling revealed increased expression of number of proteins of interest including ICAM-1 and CD105 in the endothelial cells of mice with diastolic dysfunction, further suggesting endothelial activation. However, there was no change in the activation status of EC in the lungs. Furthermore, I detected elevated levels of Ly6Chi pro-inflammatory monocytes in the bone marrow, spleen, blood, lungs and heart. In the lungs and heart, I found no change in macrophage count but a significant modulation with a switch towards a more inflammatory phenotype. The use of a CCR2 inhibitor treatment yielded no effect in the modulation of endothelial cell or immune phenotype across the bone marrow, spleen, blood, lungs or heart. My findings suggest that inflammation driven by arthritis expands the vascular niche in the bone marrow and enhances the production of pro-inflammatory monocytes. These monocytes are subsequently recruited to the heart via the activated vasculature. Elevated ROS levels likely contribute to vascular activation and dysfunction in both the heart and bone marro
Stochastic Sampling Techniques for Lattice Animals
No full textThis thesis investigates stochastic sampling methods for lattice polymer models, with a particular focus on branched configurations such as lattice animals and lattice trees. These models play a central role in statistical physics and combinatorics, yet their exponentially large configuration spaces and complex constraints pose significant chal- lenges for direct enumeration, motivating the use of Monte Carlo methods to obtain estimates of key quantities. We first evaluate the Generalized Atmospheric Rosenbluth Method (GARM) and show—through extensive empirical study—that GARM suffers from persistent sys- tematic biases and uncontrolled variance which are not corrected by na ̈ıve pruning and enrichment. To isolate and understand these failures, we introduce a tractable toy model of bi- nary trees, derive closed-form distributions, and demonstrate that biases persist even in the absence of excluded volume. We show that these effects stem from a mismatch between the sampling distribution and the true configuration space, particularly for complex branching topologies. Building on these insights, we propose atmospheric flattening, an adaptive reweighting scheme which enforces uniform sampling across positive and negative atmosphere bins. Numerical experiments confirm dramatic reductions in variance and sampling bias. Recognising the recent shift towards highly parallel computing architectures, we adapt GARM to a fixed-population framework inspired by particle filters. These methods are benchmarked on a range of models, and their performance is evaluated in terms of effective sample size and convergence quality. Our new sampler achieves improved scaling across multiple CPU cores while preserving sample diversity and effective sample size. Finally, to address settings where sequential growth may be impractical, we explore a Markov Chain Monte Carlo algorithm for lattice trees—adapting pivot and cut–and- paste moves—and use this to estimate growth constants. We then explore possible extensions of the SAWTree—an efficient data structure for self-avoiding walks—to branched polymers
