UTSA Runner Research Press (Univ. of Texas at San Antonio)
Not a member yet
6846 research outputs found
Sort by
Neighborhood Context, Divine Struggles, and Psychological Distress
No full textAlthough researchers have consistently linked structural neighborhood disadvantage with poorer mental health, they continue to search for underlying mechanisms. In this article, we test whether the association between structural neighborhood disadvantage and psychological distress is mediated by perceived neighborhood disorder and divine struggles. Using longitudinal national survey data from the Midlife in the United States (MIDUS 2 and 3) study (n = 2,083), we employed structural equation modeling and marginal models with unstructured covariances. Our mediation analysis confirmed the indirect effect of neighborhood structural disadvantage (concentrated socioeconomic disadvantage) on psychological distress (depression and anxiety) through perceived neighborhood disorder (perceptions of neighborhood safety and the built environment) and divine struggles (strained relations with God). Our analyses build on previous work by demonstrating that divine struggles may play a role in explaining why living in a neighborhood that is characterized by structural disadvantage and disorder is often associated with poorer mental health.Sociology and Demograph
Microbial Community Diversity of an Effluent Dependent Stream Ecosystem
No full textNutrient enrichment in freshwater ecosystems is a rapidly growing environmental crisis. Excess supply of these nutrients results in the increase of algal biomass production and the disruption of important ecosystem functions. In this study, we collected streambed rocks within 3 study reaches of the Cibolo Creek watershed (Boerne, TX, USA) monthly with different influences of wastewater discharges (1 control site upstream of discharge, 1 site immediately below discharge sites at a confluence, 1 site ~500-m downstream of discharge) from May 2021 to September 2021 (5 sampling periods). At each site, streambed rocks were sampled for algal biomass, total biofilm organic matter, and biofilm bacterial/archaeal communities via targeted (16S rRNA genes) amplicon Illumina MiSeq sequencing. Both chlorophyll <i>a</i> and biofilm OM was greatest in the most downstream location impacted by wastewater effluent. Overall, wastewater-fed sites (down-1, down-2) were depleted in Acidobacteria (p < 0.001), Deltaproteobacteria (p < 0.001) and Firmicutes (p < 0.001) and enriched in Cyanobacterial bacterial phyla. In addition, <i>Clostridium</i> was surprisingly a dominant Genus only in the upstream location that was not impacted by human wastewater effluent. This suggests that the upstream location may not be an adequate "reference" or that <i>Clostridium</i> is not a good fecal indicator for treated effluent in this region and may serve as a better indicator for wildlife waste. This data indicates that there is a complex relationship between autotrophic and heterotrophic biomass in this study and that other environmental factors (e.g. waterfowl and wildlife fecal contamination) besides wastewater effluent that are impacting microbial diversity shifts in this ecosystem.Integrative Biolog
Efficient and Direct Inference of Heart Rate Variability using Both Signal Processing and Machine Learning
Get PDFHeart Rate Variability (HRV) measures the variation of the time between consecutive heartbeats and is a major indicator of physical and mental health. Recent research has demonstrated that photoplethysmography (PPG) sensors can be used to infer HRV. However, many prior studies had high errors because they only employed signal processing or machine learning (ML), or because they indirectly inferred HRV, or because there lacks large training datasets. Many prior studies may also require large ML models. The low accuracy and large model sizes limit their applications to small embedded devices and potential future use in healthcare.
To address the above issues, we first collected a large dataset of PPG signals and HRV ground truth. With this dataset, we developed HRV models that combine signal processing and ML to directly infer HRV. Evaluation results show that our method had errors between 3.5% to 25.7% and outperformed signal-processing-only and ML-only methods. We also explored different ML models, which showed that Decision Trees and Multi-level Perceptrons have 13.0% and 9.1% errors on average with models at most hundreds of KB and inference time less than 1ms. Hence, they are more suitable for small embedded devices and potentially enable the future use of PPG-based HRV monitoring in healthcare.This research was in part supported by the National Science Foundation (NSF), under grants, 2155096, 2221843, 2215359, 2309760, and 2317117.Computer Scienc
Synthesis of Cellulose-Based Hydrogel—Nanocomposites for Medical Applications
No full textThis study focused on synthesizing a cellulose-based hydrogel nanocomposite as a green hydrogel by adding a microcrystalline cellulose (MC) solution to carboxymethyl cellulose sodium (CMC-Na) with citric acid as a cross-linker. Y<sub>2</sub>O<sub>3</sub> nanoparticles were incorporated during hydrogel preparation in different ratios (0.00% (0 mmol), 0.03% (0.017 mmol), 0.07% (0.04 mmol) and 0.10% (0.44 mmol)). FTIR analysis confirmed the cross-linking reaction, while XRD analysis revealed the hydrogels&rsquo; amorphous nature and identified sodium citrate crystals formed from the reaction between citric acid and CMC-Na. The swelling test in deionized water (pH 6.5) at 25 &deg;C showed a maximum swelling percentage of 150% after 24 h in the highest nanoparticle ratio. The resulting cellulose hydrogels were flexible and exhibited significant antibacterial activity against <i>Staphylococcus aureus</i> (<i>S. aureus</i>) and <i>Escherichia coli</i> (<i>E. coli</i>). The synthesized cellulose-based hydrogel nanocomposites are eco-friendly and suitable for medical applications.Biomedical Engineering and Chemical Engineerin
On Solutions of Two Post-Quantum Fractional Generalized Sequential Navier Problems: An Application on the Elastic Beam
Get PDFFractional calculus provides some fractional operators for us to model different real-world phenomena mathematically. One of these important study fields is the mathematical model of the elastic beam changes. More precisely, in this paper, based on the behavior patterns of an elastic beam, we consider the generalized sequential boundary value problems of the Navier difference equations by using the post-quantum fractional derivatives of the Caputo-like type. We discuss on the existence theory for solutions of the mentioned <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mo stretchy="false">(</mo><mi mathvariant="monospace">p</mi><mo>;</mo><mi mathvariant="monospace">q</mi><mo stretchy="false">)</mo></mrow></semantics></math></inline-formula>-difference Navier problems in two single-valued and set-valued versions. We use the main properties of the <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mo stretchy="false">(</mo><mi mathvariant="monospace">p</mi><mo>;</mo><mi mathvariant="monospace">q</mi><mo stretchy="false">)</mo></mrow></semantics></math></inline-formula>-operators in this regard. Application of the fixed points of the <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi mathvariant="sans-serif">&rho;</mi></semantics></math></inline-formula>-<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi mathvariant="sans-serif">&theta;</mi></semantics></math></inline-formula>-contractions along with the endpoints of the multi-valued functions play a fundamental role to prove the existence results. Finally in two examples, we validate our models and theoretical results by giving numerical models of the generalized sequential <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mo stretchy="false">(</mo><mi mathvariant="monospace">p</mi><mo>;</mo><mi mathvariant="monospace">q</mi><mo stretchy="false">)</mo></mrow></semantics></math></inline-formula>-difference Navier problems.Mathematic
Benchmarking of Energy and Latency of CNN Models on Raspberry Pi and Energy-Aware Weight Based Pruning of CNN Model
Get PDFDue to resource constraints, deploying Convolutional Neural Networks (CNNs) on edge devices like Raspberry Pi presents challenges to computational efficiency and energy consumption. This thesis investigates the performance and optimization of CNNs for edge applications. Initially, three popular CNN models?AlexNet, VGG16, and VGG8?are benchmarked on a Raspberry Pi 4, evaluating their inference time, energy consumption, and CPU utilization. Results highlight convolutional layers as the most resource-intensive components, emphasizing the need for targeted optimizations.
To address these challenges, the study explores energy-aware, weight-based pruning methods. Two pruning approaches are applied to AlexNet: one uniformly across all layers and another selectively targeting convolutional layers based on benchmarking data. The second approach significantly reduces model weight while maintaining acceptable accuracy after retraining, achieving a trade-off between computational efficiency and energy consumption. The pruned models demonstrated an 8% drop in accuracy with a significant improvement in average inference time for convolution layers, highlighting the viability of pruning for resource-constrained environments.
This work provides insights into designing lightweight and energy-efficient CNN models, enabling their deployment on low-power edge devices. It lays the groundwork for future research on optimizing advanced CNN architectures and exploring diverse datasets to enhance edge computing applications.Computer Scienc
A computational framework to examine public response of moral language on polarization following vigilantism incident
Get PDFIntroduction:
➢In August 2022, Kyle Rittenhouse fatally shot two men and wounded the third in Kenosha, Wisconsin.
➢Incidents of vigilantism provoke discourses related to moral values and may induce polarizing judgements in social media.
➢Moral foundations entail automatic gut-reactions of like and dislike when certain patterns are perceived in the social world, which in turn guide judgments expressed on social media.
➢The social media users could use moral languages of different vice and virtue. e.g., care (virtue) for the victim, harm (vice) the perpetrator
➢Such extreme judgements or attitude of right or wrong challenge the social cohesion in modern civil societies.
➢ We argue that understanding of polarization is incomplete without accounting for morality because moral foundations can predict attitudes on several social issues such as immigration, same-sex marriage and abortion.
Purpose (or Objective, Goal):
• How do public communicate moral language (virtue/vice) on social media following vigilantism incidents?
• How does public’s vigilantism-related moral language (virtue/vice) on social media affect polarization?Information Systems and Cyber Securit
Explainable Artificial Intelligence Methods for Intelligent Fault Detection in Inverter-Based Distribution Systems
Get PDFThe use of artificial intelligence (AI) and machine learning (ML) for intelligent fault detection in electric distribution systems (DS) has recently gained significant interest owing to the impact of the ongoing integration of distributed generation (DG) resources including inverters. While these approaches demonstrate superior fault detection accuracy over traditional means, they are dominated by the use of complex, black-box models such as shallow and deep neural networks. The nature of these models limits the ability of protection engineers to understand the underlying mechanism behind the fault detection process and hinders the practical adoption of these models in the DS protection schemes. This research aims to close this gap and proposes model and data-based explainability via explainable AI (XAI) to enhance trust and understanding of the models used for fault detection in inverter-based distribution systems. Artificial neural networks (ANNs) are developed and trained to detect the presence of a fault and its location in a simulated distribution system with inverter-based generation. The model explainable AI method demonstrated the ability of Shapley values to measure the relative feature importance for the input features of the fault type and location ANN classifiers. The novel data-based explainable AI method developed showed the ability of decision trees to provide data-driven explanations behind each of the predictions for each ANN classifier. Holistically, this work highlights the benefits and limitations of two distinct XAI techniques that will enable a greater transparency and understanding for the next generation of AI driven DS fault detection schemes.Electrical and Computer Engineerin
Green Places, Active Places: An Assessment of Park Quality, and Community Health in San Antonio, Texas
No full textThis thesis will investigate the relationship between park quality, park accessibility, physical activity, and mental health status. The association between human health and urban green spaces has previously been investigated in the public health and urban planning fields utilizing broad measures of greenery such as NDVI or urban tree canopy. This research provides a nuanced approach in determining how the quality and accessibility of green spaces may provide a better understanding on its association to human health. As a result, this study will assess the association between community health and urban green space while considering park accessibility, park quality, and park usage variables. Park quality and accessibility variables were collected on parks owned and managed by the City of San Antonio. Additionally, park visitation figures were extracted utilizing Advan Research Monthly Patterns to further understand how park quality and park accessibility might play a role in influencing park visitation at the census block group level. Results from this study did not identify significant relationships with respect to park visitation. Community health variables had significant relationships with park accessibility and park quality such as walkability and park tree canopy. Furthermore, cumulative green space variable operationalized through block group tree canopy also had a significant relationship with community health variables.Urban and Regional Plannin
Chemical Species Transport in Buoyancy Dominant Flows: Computational Fire Modeling and Applications
No full textThis work aims at using a combination of different numerical models and tools to better understand the phenomena of chemical species transport in buoyancy driven flows. Chemical transport can be highly complex multi-scale coupled non-linear problem from ( down to ) scales. A large class of geophysical flows maintain some form of buoyant or convection driven dynamics not limited to buoyant plumes and jets, density currents that directly impact and influence trajectories and bulk flow of chemical species. Wildfire in particular are of relevance (not strictly from a scientific inquiry), but from an environmental degradation and damage, economic loss and a potential for loss of life. Wildfires exhibit a wealth of challenges and difficulties involved for forecasting and modeling. Not limited to the incomplete and not fully understood physics involved with the evolution dynamics of fire, but resolving all the necessary physics (down to centimeters for combustion and turbulence properties but as large as kilometers to resolve the planetary boundary layer) is an intractable problem. Even solving a simplified version of this problem may take weeks to days, where timely reliable decisions and applications need to be made, for instance in safely evacuating a residential community or carrying out a prescribed burn. The objective here is to address these type of problems with a variety of tools designed for specific use cases. The dissertation is organized as follows, (1) developing an unsteady entrainment formulation for buoyancy driven plumes using the weather research and forecast model - large eddy simulation - buoyancy plume in a convective planetary boundary layer, the unsteady entrainment formulation is unique in that it incorporates time-dependence, and we express a relationship with dimensionless parameters such as the Reynolds and Froude numbers and the unsteady entrainment factor; (2) apply a new blackbox multigrid method in the Los Alamos National Laboratory's quick urban and industrial diagnostic wind model. We show good improvement using energy norm analysis of the backward error and residuals. The model simulated wind measurements are also validated against field experiment data obtained from energetic material and testing center near Socorro, New Mexico in November 2019; we report reasonable agreement (of a normalized absolute deviation score at best of 0.36) for cases during where there is little to no influence from solar radiative heating and cooling; and (3) present fire toxicity results and tenability with the Fire Dynamics Simulator large eddy simulation tool for a sequence of upcoming fire scenario experiments at Worcester Polytechnic Institute. We report temperature, mass-flux measurements in a simulated compartment to inform sensor placement during the experiment, future work will be to incorporate measurements from said experiment and report and model adjustment and improvement for fire toxicity The problems addressed in this dissertation maintain and used different tools for solving different aspects of fire modeling using analytical, numerical, and experimentally obtained data. It is only by the synthesis and feedback of this triplicate can any progress be made, or 'if you ain't burning you ain't learning'.Mechanical Engineerin