192815 research outputs found
Sort by
The pointer network for reward maximisation in multi-target space mission sequence selection
Get PDFMulti-target space mission scenarios such as asteroid rendezvous, debris removal or satellite servicing, require targeting several orbits in a single mission, often to be selected among a large set, and therefore choosing optimal sequences of these orbits to be visited. This paper demonstrates a reinforcement-learning-based framework for selecting the sequence of targets to be visited in large-scale multi-target mission optimisation problems. The sequence selection is a NP-hard combinatorial optimisation problem. The proposed method builds upon a neural network architecture for combinatorial optimisation originally developed for Euclidean problems, to produce estimates of the optimal sequence of targets in very short amounts of time. The neural network is trained using a policy-gradient reinforcement-learning approach. Once training is complete, the network can be evaluated in two ways: one of these (greedy decoding) produces solutions on average 15 % less optimal than Ant Colony Optimisation (ACO); the other (stochastic search) is on average 5 % less optimal than ACO, using an iterative process that is slower than greedy decoding but still orders of magnitude faster than ACO. The quality of the network’s solutions is shown both averaged over large amounts of problems, and demonstrated more closely on a few specific instances
Retrospective analysis of submissions to the World Reference Laboratory for foot-and-mouth disease: what can these data tell us about the role of small ruminants in disease epidemiology?
Get PDFEpidemiological studies can be useful to understand the dynamics of foot-and mouth disease (FMD) virus. Clinical signs of FMD in small ruminants are often difficult to spot, which may lead to an under appreciation of their epidemiological importance in the spread (and therefore control) of FMD. To elucidate the impact of FMD surveillance in small ruminants, retrospective analyses were performed using data collected by the World Reference Laboratory for Foot-and-Mouth Disease. The total number of samples included in these analyses was 32,802, with an average of 444 samples collected per year between 1958 and 2023. When samples were classified into domesticated species groups, the most represented group were Large Ruminant (n = 15021), followed by Small Ruminant (n = 1972), Pigs (n = 1486) and Wildlife (n = 294). Within the domesticated species group, 73.4 % of Pigs and 72.2 % of Large Ruminant samples were FMD virus positive, while Small Ruminant samples had significantly fewer FMD virus positive results (30.0 %). Of the positive samples within the small Ruminant group, serotype O accounted for 86.0 % of the records. These analyses highlight the relative contribution of FMDV positive samples from sheep and goats to global surveillance activities and the potential involvement of small ruminants in maintenance of serotype O. These findings emphasise the importance of these species in control strategies in endemic countries and the necessity to provide specific small ruminant guidelines for FMD diagnostic testing
Housing finance and inequality: mortgage debt in urban China
Get PDFMortgage credit has played a central role in driving housing market dynamics. The differentiation of the mortgage market is an important embodiment of housing inequality and a dimension of social stratification. Using the Chinese Housing Consumption Survey (CHCS) data, this study aims to understand the temporal dynamics of the mortgage market and factors influencing access to housing finance in urban China, which offers important insights into housing inequality in transitional economies. Since 2013, the use of mortgage loans to achieve homeownership has become the dominant source of housing finance in urban China, with intergenerational transfers remaining crucial in influencing the development of the mortgage market. Considerable variation is found in Housing Provident Fund participation, access to mortgages and the type of mortgage among different demographic and socio-economic groups. Institutional features, such as hukou status, hukou location, party membership and employer type are crucial for securing policy-based subsidies, influencing access to homeownership and therefore generating housing tenure inequality and housing wealth inequality. At the city level, state forces affect the acquisition of Housing Provident Fund loans, whereas market forces prompt reliance on commercial mortgage loans. The scrutiny of the housing finance systems contributes to the understanding of the formation of wealth inequality in transitional economies
Optimising human rabies vaccine supply chains: a modelling study
Get PDFBackground:
Rabies causes thousands of deaths annually in low- and middle-income countries. Despite effective vaccines for post-exposure prophylaxis (PEP), their expense, coupled with supply chain failures, leads to stockouts and preventable deaths. Investment by Gavi, the Vaccine Alliance, aims to improve access to post-exposure vaccines. We evaluate PEP demand in Tanzania and Kenya and examine stock management strategies for improving supply chains in Gavi-eligible countries.
Methods:
We fitted negative binomial distributions to five years of bite patient data from Tanzania (6646 patients, 20 districts) and Kenya (199,112 patients, 47 counties) to parameterise simulations of post-exposure vaccine demand under WHO-recommended intramuscular (IM) and intradermal (ID) regimens. We compared simulated vaccine use, stockouts, and the impact of stock management strategies across the observed range in demand.
Results:
Bite patient incidence varied dramatically; demand surges far exceeded monthly averages (in 6 % of months exceeding 3× average monthly bite patient presentations) and were most extreme in low-incidence settings. ID vaccination reduces vial use by >55 % and reduces stockout risk. Under ID vaccination vial savings are greatest in high-throughput settings, whilst risk mitigation is maximised in low-throughput settings. Decentralizing PEP to more facilities improves access, though reduces vial-sharing opportunities and so increases vial use. Resilient supply chain strategies were identified according to patient throughput, allowing for adaptation to changing demand.
Conclusions:
ID vaccination reduces vial use and stockouts, even in low-throughput settings. Tailoring stock management—through adjusted alert thresholds and restocking volumes—can simplify the integration of rabies vaccines into essential immunisation supply chains, improving their availability and preventing unnecessary deaths. However, logistical trade-offs must also be considered
Inflammation's impact on the interaction between oligodendrocytes and axons
Get PDFOligodendrocytes are responsible for myelination of axons and providing trophic and metabolic support to the myelinated axon. They also interact with immune effector cells, including microglia and T cells, hence, are involved in CNS immune regulation. Given the crucial roles for oligodendrocytes and myelin in axonal function and maintenance, dysfunction, whether through cell death, myelin injury and loss, or failure in normal myelin formation, impairs neurological function. In diseases such as multiple sclerosis, the leukodystrophies and viral infection, neuroinflammation is an important effector of myelin injury, having secondary consequences for the myelinated axon. In this review we discuss the role of oligodendrocytes in health and inflammatory disease, with a focus on the interplay between inflammation and oligodendrocyte-axon interactions
FetaFix: Automatic Fault Localization and Repair of Deep Learning Model Conversions
No full textConverting deep learning models between frameworks is a common step to maximize model compatibility across devices and leverage optimization features that may be exclusively provided in one deep learning framework. However, this conversion process may be riddled with bugs, making the converted models either undeployable or problematic, considerably degrading their prediction correctness.
In this paper, we propose an automated approach for fault localization and repair, FetaFix, during model conversion between deep learning frameworks. FetaFix is capable of detecting and fixing faults introduced in model input, parameters, hyperparameters, and the model graph during conversion. FetaFix uses a set of fault types (mined from surveying common conversion issues reported in code repositories and forums) to localize potential conversion faults in the converted target model and then repair them appropriately, e.g., replacing the parameters of the target model with those from the source model. This is done iteratively for every image in the dataset, comparing output label differences between the source model and the converted target model until all differences are resolved. We evaluate the effectiveness of FetaFix in fixing model conversion bugs of three widely used image recognition models converted across four different deep learning frameworks. Overall, FetaFix was able to fix 462 out of 755 detected conversion faults, either completely repairing or significantly improving the performance of 14 out of the 15 erroneous conversion cases
DIGITtally - a new tool for simplifying and streamlining Drosophila melanogaster meta-analysis
Get PDFDrosophila melanogaster has one of the deepest research bases within the life sciences, with a wealth of high-quality tissue- and cell type-specific transcriptomic data available. However, integrating large datasets derived from disparate sources is not trivial. We have designed a broadly applicable solution to this problem in the form of the Drosophila Interesting Genes in Individual Tissues-tally (DIGITtally) system. It is freely available online at www.digittally.org. DIGITtally is customizable and hypothesis-free, allowing meta-analysis across the Drosophila research space along with analysis of conservation in other species, querying 10 data sources for seven indicators of tissue-specific activity. We have applied DIGITtally to a pertinent question within entomology—that is, whether a specific pattern of gene expression underlies the transporting activity of epithelial tissues (an ‘epitheliome’). By using DIGITtally to survey gene expression throughout the tissues comprising the D. melanogaster alimentary canal (salivary gland, midgut, Malpighian tubules, and hindgut), we have verified the existence of a specific ‘epithelial’ vacuolar-type ATPase configuration
Generating realistic single-cell images from CellProfiler representations
No full textHigh-throughput imaging techniques acquire large amounts of images efficiently. These images contain rich biological information including cellular processes. A common method to analyze them is to encode them into quantitative representation vectors. Generally, there are two ways to extract cell biological information into representations, hand-crafted and machine-learning. Although representations obtained from machine learning models often demonstrate commendable reconstruction performance, they lack biological interpretability. In contrast, hand-crafted representations have clear biological meanings, making them easily interpretable. However, the capability of hand-crafted representations to generate realistic images remains uncertain. In this work, we propose a CellProfiler to image (CP2Image) model capable of directly generating realistic cell images from CellProfiler representations. The proposed model is demonstrated to be robust to different architectures, including ResNet, InceptionNet and Transformer. We also show that the biological information is well preserved during the generation process. The changes in certain CellProfiler features will reflect the corresponding changes in the generated single-cell images. In addition, the CP2Image model can generate conditional phenotypes, which will ultimately help diagnostics and drug screening
All-optical reconfigurable photonic crystal-microring hybrid cavities with high-efficiency tuning via the mechanical Kerr effect
Get PDFAll-optical control of silicon photonic integrated devices is crucial for on-chip applications such as signal processing, computing, and switching. A key limitation of current integrated devices is high power consumption, stemming from the weak nonlinear effects of silicon. An alternative nonlinear effect in deformable platforms is the mechanical Kerr effect (MKE), which arises from the optical gradient force (OGF) generated by highly localized optical fields that can deflect freestanding waveguides near a dielectric substrate. In this work, we present a hybrid optomechanical cavity design, driven by OGF, which integrates a compact microring resonator (MRR) with a radius of 10.08 µm and a quadratically tapered photonic crystal nanobeam cavity (PCNC). This design results in two distinct types of resonant modes, enabling mode-dependent wavelength routing. Due to strong localization and intensity enhancement, the tuning range and efficiency are significantly improved compared to conventional MRRs. An experimental 1.98 nm redshift is achieved in the probe PCNC mode, corresponding to a tuning efficiency of 142 GHz/mW. Additionally, substantial mode splitting is observed due to the mode-dependent tuning capability of the device. This design holds great potential for wavelength routing applications, particularly in advanced all-optical tunable optical filtering systems
Understanding and creating learning spaces in research with primary children in Tanzania
Get PDFEducation research is largely done on rather than with children and their experiences of learning have often been neglected in the global South. This paper presents the approach taken to increase participation and visibility of children in a comparative case study of school, home and neighbourhood learning spaces in Tanzanian primary schools. I outline the creative, spatial methods developed to enable children to share their perceptions and experiences of learning, including arts-based methods; go-alongs; and photo-elicitation interviews. I highlight the possibilities that each method opened up but also flag limitations and how these were mitigated through multiple methods. The paper extends the discussion of the benefits of multiple methods and highlights the value of multiple sources of data within methods and of penetrating multiple layers of interpretation. Finally, the paper highlights how the creation of new research spaces offered additional insight into and opportunities for children’s learning beyond home and school
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! 👇