100039 research outputs found
Sort by
Interpreting Neural Networks for Particle Tracing in Fluid Simulation Ensembles: An Interactive Visualization Framework
Get PDFUnderstanding the internal mechanisms of neural networks, particularly Multi-Layer Perceptrons (MLP), is essential for their effective application in a variety of scientific domains. In particular, in the scientific visualization domain their adoption has recently shown to be a promising tool to predict particle trajectories in fluid dynamics simulation and aid the interactive visualization of flows. This research addresses the critical challenge of interpretability of such models.
While interpretability has been extensively explored in fields like computer vision and natural language processing, its application to time series data, particularly for particle tracing (or prediction of trajectories), has not garnered sufficient attention.
The overarching objective of this thesis is to augment the interpretability of MLP networks through interactive and comparative visualization of model ensembles. We aim to contribute to address the challenges associated with the ”black-box” nature of neural networks in this specific context. Our primary contribution lies in the development of a comprehensive visualization tool that integrates multiple linked views, including gradient visualization, particle trajectories, layer-wise activation, and weights visualization. This tool facilitates a more profound understanding of the intricate relationships between model components and model predictions.
In particular, the proposed framework provides a user-friendly interface for comparing different models trained to predict particle trajectories in fluid dynamics simulation ensembles.
This tool not only aids in understanding the MLP network behaviour, but also serves as a practical resource for researchers and practitioners wanting to analyze and use similar models.
Finally, we test our framework using a variety of different 2D flows with different degrees of complexity. This helps understanding the effectiveness of the tool in providing insights about which components of the model affects a particular prediction and also what the network is learning at different training epochs
Raising Awareness of the Link Between Coal Mining and Mental Health
Get PDFMental health is a concern in Carbon and Emery Counties of Utah, particularly in its association with coal mining. An event was held to raise awareness about how the industry impacts mental health and resources to address it. Results from pre- and post-surveys show the usefulness of the event to attendees
Effects of Extracellular Vesicle Transfer on the Immunological Acceptance of the Fetal Allograft
Get PDFBuilding Resilience: Incentivizing Agricultural Water Conservation in Utah Amid Climate Uncertainty and Prior Appropriation Water Rights
Get PDFSupplementary files for Impact of Snow Accumulation on Structural Integrity: Present and Future Perspectives
No full textEvaluating the impact of weight exerted by settled snow (i.e., snow load) on structures poses numerous statistical challenges, including missing data, biased distribution parameters, and the influence of climate change. This dissertation aims to address challenges related to the use both direct and indirect measurements of snow load (or equivalently, snow water equivalent), as well as the anticipated impact of climate change on future extreme snow loads. The first paper within this dissertation investigates short-term snow loads by comparing various techniques for estimating extreme values of short-term snow accumulations. Additionally, the first paper includes a comparative analysis of short-term and long-term snow accumulations, revealing significant differences in snow load accumulation patterns across geographical regions. The second paper focuses on bias correction in the scale parameter of the generalized extreme value distribution describing extreme snow loads in situations where the snow load is estimated indirectly using snow depth data. The bias correction is accomplished using bootstrap techniques when some of the snow load data is only approximated, rather than directly measured. We demonstrate the effectiveness of our approach in correcting scale parameter bias, as evidenced by simulation studies and real-life snow data. In the third paper, we incorporate the effects of climate change in the snow load estimation process and discuss the implications of considering the effects of climate change in snow load design. Our findings indicate that most locations in the United States have a reduced risk of snow-induced structural failure in a future climate. However, other locations appear to have an increased risk of structure failure, though there is no agreement among climate models as to which areas are at increased risk. Together, these interconnected papers refine methods for characterizing extreme snow accumulations and address the statistical complexities of estimating design snow loads for both current and future conditions
Time-Series Feature Selection for Solar Flare Forecasting
Get PDFSolar flares are significant occurrences in solar physics, impacting space weather and terrestrial technologies. Accurate classification of solar flares is essential for predicting space weather and minimizing potential disruptions to communication, navigation, and power systems. This study addresses the challenge of selecting the most relevant features from multivariate time-series data, specifically focusing on solar flares. We employ methods such as Mutual Information (MI), Minimum Redundancy Maximum Relevance (mRMR), and Euclidean Distance to identify key features for classification. Recognizing the performance variability of different feature selection techniques, we introduce an ensemble approach to compute feature weights. By combining outputs from multiple methods, our ensemble method provides a more comprehensive understanding of the importance of features. Our results show that the ensemble approach significantly improves classification performance, achieving values 0.15 higher in True Skill Statistic (TSS) values compared to individual feature selection methods. Additionally, our method offers valuable insights into the underlying physical processes of solar flares, leading to more effective space weather forecasting and enhanced mitigation strategies for communication, navigation, and power system disruptions
Review of Family and Justice in the Archives: Historical Perspectives on Intimacy and the Law
Get PDFReview of Family and Justice in the Archives: Historical Perspectives on Intimacy and the Law
Dietary Milk Phospholipids Increase Body Fat and Modulate Gut Permeability, Systemic Inflammation, And Lipid Metabolism in Mice
Get PDFThe study aimed at how dietary milk polar lipids affect gut permeability, systemic inflammation, and lipid metabolism during diet-induced obesity (DIO). C57BL/6J mice (n = 6x3) were fed diets with 34% fat as energy for 15 weeks: (1) modified AIN-93G diet (CO); (2) CO with milk gangliosides (GG); (3) CO with milk phospholipids (MPL). Gut permeability was assessed by FITC-dextran and sugar absorption tests. Intestinal tight junction proteins were evaluated by Western blot. Plasma cytokines were measured by immunoassay. Body composition was assessed by magnetic resonance imaging. Tissue lipid profiles were obtained by thin layer chromatography. Hepatic expression of genes associated with lipid metabolism was assessed by RT-qPCR. MPL increased the efficiency of converting food into body fat and facilitated body fat accumulation compared with CO. MPL and GG did not affect fasting glucose or HOMA-IR during DIO. MPL increased while GG decreased plasma TG compared with CO. MPL decreased phospholipids subclasses in the muscle while increased those in the liver compared with CO. GG and MPL had little effect on hepatic expression of genes associated with lipid metabolism. Compared with CO, MPL decreased polar lipids content in colon mucosa. Small intestinal permeability decreased while colon permeability increased and then recovered during the feeding period. High-fat feeding increased plasma endotoxin after DIO but did not affect plasma cytokines. MPL and GG did not affect plasma endotoxin, adipokines and inflammatory cytokines. After the establishment of obesity, MPL increased gut permeability to large molecules but decreased intestinal absorption of small molecules while GG tended to have the opposite effects. MPL and GG decreased mannitol and sucralose excretions, which peaked at d 45 in the CO group. MPL decreased occludin in jejunum mucosa compared with CO. GG and MPL did not affect zonula occludens-1 in gut mucosa. In conclusion, during DIO, milk GG decreased gut permeability, and had little effect on systemic inflammation and lipid metabolism; MPL facilitated body fat accumulation, decreased gut permeability, did not affect systemic inflammation
Faculty Senate Executive Committee Minutes November 18, 2024
Get PDFCall to Order Approval of Minutes University Business Faculty Senate Business Information EPC Report Library Budget Faculty Evaluation Committee Annual Report USUSA Annual Report Old Business New Business Adjourn: 4:25 p
Faculty Senate Executive Committee Minutes October 21, 2024
Get PDFCall to Order Approval of Minutes University Business Faculty Senate Business Information EPC Report Report Academic Freedom and Tenure Annual Report Athletic Council Annual Report Library Advisory Council Annual Report Old Business New Business Adjourn: 3:28p