Digital Commons Network^™

Enabling Multi-Task Neural Network Inference On Heterogeneous Edge Devices, Redwanul Islam Arif

Master's Theses

Deep neural networks are increasingly required to run on the devices that generate the data. If such a device must perform more than one task, the standard practice is deploying one model per task, which makes memory grow linearly with task count, which is unacceptable when the entire budget is kilobytes. This thesis asks one question in three settings: how much capability can a network acquire without incurring deployment cost?

The first study takes an ImageNet-pretrained ResNet-18, sweeps the branch point across every residual stage and the classification-head depth across one, ten, and twenty layers, and deploys the resulting multi-head …

Spatial And Hydrologic Effects Of Climate Change And Urbanization On Freshwater Fish Communities In North Georgia Watersheds, Mayuko Mizutani

Master's Theses

Urbanization and climate change are altering streamflow and land cover across the southeastern United States, threatening freshwater biodiversity in rivers already burdened by impoundments and habitat loss. I investigated how changes in streamflow and land cover might affect fish community biodiversity in the Etowah River watershed of North Georgia. By combining a dataset of fish collections from 1999 through 2019 with contemporaneous land use and daily streamflow, I was able to model fish community responses to urbanization and climate-driven hydrologic changes and then forecast fish community structure under multiple CMIP6 global climate models and greenhouse gas emissions scenarios (SSP2-4.5 and …

Federated Learning For Early Medical Diagnosis: Enhanced Diabetic Retinopathy Detection In Smart Healthcare, Mohammad Nasajpour Esfahani

Master's Theses

This thesis investigates the role of federated learning as a privacy-preserving solution for modern healthcare challenges. In traditional machine learning, sensitive medical data must be centralized for model training, raising concerns about privacy, security, and regulatory compliance. Federated learning offers an alternative by allowing hospitals, clinics, and personal health devices to collaboratively train shared models without exchanging raw patient data. The study first explores how federated learning is being used across various healthcare domains, including cancer detection, medical imaging, and disease prediction— highlighting its potential to support secure collaboration across institutions. It addresses key benefits such as data privacy, scalability, …

A Phenomenological Study Of The Lived Experiences Of African American Female Leaders In Public K-12 Schools In Georgia, Tyisha Davis

Dissertations

The purpose of this phenomenological study was to explore how the intersectionality of race and gender shapes the leadership experiences of African American women in public K–12 schools. Despite comprising a significant portion of the educator workforce, African American women remain underrepresented in school leadership positions nationally, warranting deeper examinations of their professional pathways and daily realities. This qualitative study employed interpretative phenomenological analysis, guided by an intersectional framework, to understand the lived experiences of 10 African American female educational leaders across K–12 settings in Georgia. Principals, assistant principals, instructional specialists, and a superintendent engaged in semi-structured interviews analyzed through …

Perceptions Of Family Involvement And Its Influence On High School Attendance: Voices Of Families And Mathematics Teachers, Amber Phillips

Dissertations

Chronic absenteeism continues to be a growing concern in high schools, particularly in the years following the COVID-19 pandemic. Although family involvement has been widely studied in elementary and middle schools, less is known about how families and mathematics teachers experience school-based family involvement at the high school level. This qualitative case study explored those perspectives at a rural high school in Northwest Georgia. Individual semi-structured interviews were conducted with five families and five upper-level mathematics teachers, and a focus group was conducted with eight mathematics teachers. Interview and focus group transcripts were analyzed using structural and focused coding to …

Edge-Native Multi-Agent System For Latency-Critical Anomaly Detection And Qos Optimization In 6g Smart Manufacturing, Ndidi Nzeako Anyakora

All Dissertations

The increasing reliance on wireless connectivity in smart manufacturing places stringent demands on network latency, reliability, and adaptability that exceed the capabilities of static or threshold-based quality-of-service (QoS) control mechanisms. While 5G standalone networks support industrial deployments, emerging 6G environments are expected to introduce greater variability, tighter latency constraints, and increased performance degradation, necessitating autonomous, learning-driven control strategies.

This dissertation proposed an intelligent multi-agent framework for real-time anomaly detection and QoS optimization in industrial wireless networks by integrating edge-based monitoring, unsupervised anomaly detection, and reinforcement learning-driven control. Network telemetry collected from a Firecell 5G standalone testbed served as an empirical …

Advancing The Hardware Realization Of Swinir Functions, Samuel Opoku Aboagye

All Dissertations

Image enhancement is a critical component of modern intelligent systems, including medical imaging, remote sensing, autonomous navigation, surveillance, industrial inspection, and lowlight vision applications. Recent transformer-based image restoration networks have demonstrated significant improvements over traditional convolutional approaches by leveraging self-attention mechanisms to capture long-range feature dependencies. The computationally intensive attention mechanism is typically executed prior to later convolution and reconstruction stages. Among these architectures, SWINIR has emerged as a state-of-the-art framework for image restoration and super-resolution. However, the computational complexity of transformer operations, present significant challenges for deployment on resource-constrained FPGA (Field Programmable Gate Array) platforms.

This dissertation investigated hardware-efficient …

Criticality In A Heterogeneous Neutron Transport Rod Model, Samuel Kaleb Crowford

All Graduate Reports and Creative Projects, Fall 2023 to Present

This work studies the stochastic behavior of neutron populations in a one-dimensional rod model using Monte Carlo simulation. The first part of this project reproduces the computational results of Dumonteil, Horton, Kyprianou, and Zoia (2025) by independently implementing the Monte Carlo algorithm described in their article, with the asymptotic behavior of the first moment analyzed in relation to the dominant eigenvalue and adjoint eigenfunction of the neutron transport operator. The model is then extended to a heterogeneous setting by introducing a central region where fission is suppressed. A global expectation over initial positions and directions is used to estimate the …

Effective Science Instruction For Elementary Aged Students With Significant Cognitive And Adaptive Delays, Rabecca L. Juarez

All Graduate Reports and Creative Projects, Fall 2023 to Present

In Utah, students begin participating in statewide end-of-year assessments in third grade, with science added in fourth grade. Students with the most significant cognitive and adaptive disabilities participate in the Dynamic Learning Maps® (DLM®) alternate assessment, which measures student performance in English language arts, mathematics, and science aligned to Essential Elements (EE). Despite participation in statewide accountability systems, students with significant cognitive disabilities in a Southwest Utah School District have consistently demonstrated low science performance on the DLM assessment.

The aim of this project was to collaboratively create science lessons for students who participate in the alternate assessment. This project …

Music And Middle School Mathematics, Cheryl L. Hawkins

All Graduate Reports and Creative Projects, Fall 2023 to Present

The purpose of this project was to see if music playing in the background while learning essential mathematics skills would make a difference in the learning process of middle school students. For nine weeks I collected data on the scores of the special education students relearning skills they had been previously taught in a general education mathematics class. The students took a baseline assessment, then had 30-40 minutes of instruction spread across two days. They then took another assessment to see what the change was. During the lessons and assessments there was one of three conditions happening, one was silence, …

History In The Social Studies Classroom: Developing Historical Literacy And Skills Through Inquiry Lessons, Nancy J. Thatcher

All Graduate Reports and Creative Projects, Fall 2023 to Present

This Plan B project has created a series of inquiry-based lesson plans and activities for World History classrooms that emphasize historical literacy and disciplinary thinking. Rather than producing a traditional research thesis, this project emphasizes applied historical scholarship by translating historical and professional methodology into three classroom ready instructional lesson plans that are aligned with the Utah Core Standards for Social Studies. The project reflects the History MA’s emphasis on disciplinary rigor while addressing the practical needs of public history and education with pedagogical best practices. This project responds to ongoing challenges in World History education, including constrained instructional time, …

Advancements In Modern Seismic Monitoring: Integrating Novel And Traditional Methods For Earthquake Detection, Characterization, And Structural Imaging, Marc Adrian Garcia

Open Access Theses & Dissertations

Modern seismic monitoring has been transformed by machine-learning methods that detect and locate earthquakes at scales manual analysis cannot reach. This dissertation develops, validates, and applies such workflows across three settings that span the range of modern monitoring problems: a major subduction-zone aftershock sequence, the tectonic questions that sequence can answer, and an urban region without any local monitoring at all. First, I construct a high-resolution catalog for the aftershock sequence of the September 8, 2017, Mw 8.2 Tehuantepec, Mexico earthquake, by integrating deep-learning phase detection with established location and relocation methods. The resulting catalog of 11,374 relocated earthquakes is …

A Patch-Level Framework For Urban Vegetation Water Demand Estimation Using Remote Sensing And Deep Learning, Jesus Daniel Pereyra Manriquez

Open Access Theses & Dissertations

Urban water management in semi-arid regions requires an improved understanding of how vegetation and climatic conditions influence landscape water demand. Existing approaches often lack an integrated, spatially consistent framework to quantify this relationship at fine scales. This study proposes a patch-level framework to estimate relative landscape water demand by integrating vegetation coverage, vegetation condition, and atmospheric demand. Vegetation coverage is derived from high-resolution imagery obtained from the National Agriculture Imagery Program (NAIP) using a U-Net segmentation model with a MobileNetV2 backbone. A patch-based representation is used to ensure spatial consistency across the study area. Seasonal vegetation dynamics are captured using …

Domain Adaptation Of Facial Age Estimation For Law Enforcement Mugshot Repositories, Jorge Alejandro Pacheco Roque

Open Access Theses & Dissertations

Facial age estimation supports law enforcement via image-based, age-filtered queries, age-progressive re-identification, and bulk record labeling, where prediction accuracy determines if the resulting decisions can be trusted. State-of-the-art models excel on web imagery but incur higher error on mugshots due to domain shift between the professionally lit, filtered, and posed web photographs used during pre-training and the uniform backgrounds, uncooperative expressions, and decades of evolving capture technology found in mugshot collections. We address this gap by adapting SwinFace - a state-of-the-art multi-task Swin Transformer with public code and pretrained weights, trained on color face imagery for face recognition, facial expression …

"Fenomeno Frontera:" Family Language Policy As Lived Experience At The U.S.-Mexico Borderlands, Alejandra Sanmiguel-Lopez

Open Access Theses & Dissertations

Family language policy (FLP) research expands the understanding of how families' language ideologies, language management, and language practices shape children's bilingual development. However, limited research has examined these processes among transfronterizo families living along the U.S.-México border. This qualitative narrative inquiry explores how transfronterizo families describe and negotiate their family language policies and how these policies relate to their experiences in the borderland, including their relationships with schools. Guided by family language policy and borderlands theory, this dissertation employs Seidman's (2006) three-interview series with five transfronterizo mothers raising children enrolled in dual language programs in the El Paso, Texas, region. …

Dlc Thin Films For Semiconductor Applications, Balam Arturo Sotelo

Open Access Theses & Dissertations

The growing need for resilient semiconductor supply chains has increased the interest in developing local and accessible materials fabrication strategies. Since large-scale semiconductor manufacturing is complex to establish quickly, simpler deposition techniques may provide a practical route to start developing semiconductors materials. Diamond-like carbon (DLC) is a promising candidate in this context because its properties can be tailored through substrate selection and processing parameters.

This thesis establishes an initial framework for the study and fabrication of diamond-like carbon (DLC) materials in the Department of Physics at The University of Texas at El Paso. DLC thin films were deposited by magnetron …

Stabilized Continuous Galerkin Methods For Heat Transport In Fully Coupled Nonlinear Thermo-Poroelasticity In Heterogeneous Porous Media, Payel Dey

Open Access Theses & Dissertations

This thesis focuses on the numerical approximation of a coupled Thermo-Hydro-Mechanical (THM) model with advective heat transport in heterogeneous media. The governing equations consist of three coupled equations: the momentum balance, the fluid mass balance, and the energy balance. The primary unknowns are displacement, pressure, and temperature. In this system, the fluid pressure determines the Darcy flux, and this flux transports heat through the advective term in the energy equation. Since the Darcy flux is computed numerically from the flow equation, the conservation properties of the flow approximation can affect the temperature solution. This effect is especially important when thermal …

Modulating Linear Frequency Modulated Pulses To Send Communications Data In A Bistatic Sar Scenario, Jarren T. Worthen

All Graduate Theses and Dissertations, Fall 2023 to Present

Synthetic aperture radar (SAR) is the technology that enables the creation of images using radar waves, allowing images to be formed regardless of weather. In bistatic SAR a radar platform uses a radar pulse from a separate platform to form an image. It is important that these radar systems are able to communicate with each other. Rather than wasting energy and space flying with a separate communication system, the radar systems could use the already existing SAR system to send data between each other while still forming SAR images. The work of this thesis is to show how very simple …

Modular Verification For Network-On-Chip Designs Using Probabilistic Verification And Assume-Guarantee Reasoning, Nicholas Waddoups

All Graduate Theses and Dissertations, Fall 2023 to Present

To satisfy increasing demands for computer chip performance in personal computing, mobile devices, and commercial server computing, a modern computer chip is constructed with tens (or hundreds) of small individual computing modules. Each of these modules must communicate with one other to share information about the running state of a computer. Historically, when chips were a few modules a simple communication method sufficied. However, as the number of modules in a chip grew, a more effecient method was needed in order to maintain performance across the system as a whole. A Network-on-Chip (NoC) design is the de-facto communication method for …

Magnetic Properties And Ultrafast Spin Dynamics In Quantum Materials, Trung Kien Mac

All Graduate Theses and Dissertations, Fall 2023 to Present

Modern electronics mostly work by moving electric charge, which generates heat and hence wastes energy, especially as devices become smaller and faster. An alternative is to use the spin of electrons (a tiny magnetic property) and related “valley” states in certain atomically thin materials to store and process information more efficiently. This thesis explores how magnetism and spin behavior can be created and measured in a family of ultra-thin materials known as van der Waals quantum materials. Three kinds of materials are studied. First, atomically thin semiconductors are shaped into narrow nanoribbons. Because of their narrow size, their edges can …

Theory, Implementation, And Practical Applications Of A Modern Parabolized Navier-Stokes Code, Logan B. Freeman

All Graduate Theses and Dissertations, Fall 2023 to Present

The Parabolized Navier-Stokes (PNS) method is a specialized computational approach for solving the equations that govern fluid flowing faster than the speed of sound, known as supersonic flow. By leveraging unique physical properties of supersonic environments—specifically that information cannot travel upstream—this method allows for significantly more efficient calculations than traditional approaches. Supersonic flows are important for many military and civilian applications including hypersonic weapons, high-speed passenger aircraft, and reentry. This thesis provides a comprehensive review of the two-dimensional PNS formulation, detailing the mathematical derivation and its practical implementation for solving flow problems. Through performance analysis and error assessment, this research …

Simulation Of Hypersonic Conditions In A Lab-Scale Environment Using A Hybrid Rocket With Nitrogen Cooling, Joshua R. Sorenson

All Graduate Theses and Dissertations, Fall 2023 to Present

Developing hypersonic vehicles requires testing under extreme heat, airflow, and pressure. Traditional facilities like shock tunnels and arc-heated wind tunnels are costly and slow to set up. This thesis proposes a simpler, low-cost method using a small-scale hybrid rocket motor as a gas generator. The rocket’s hot exhaust is directed onto the front edge of a scaled-down model that represents the nose of a hypersonic vehicle. This creates a realistic flow that matches both the intense heat energy and the pushing force experienced in high-altitude hypersonic flight. The setup allows easy measurement of motor performance, surface pressures, leading-edge temperatures, and …

The Augmented Matching Law, Matias A. Avellaneda

All Graduate Theses and Dissertations, Fall 2023 to Present

Every living organism, including humans, has a finite amount of time available: 24 hours in a day, 365 days in a year, and a certain number of years in its life. At every moment, such an organism will be engaging in one of a multitude of possible activities, such as eating, sleeping, or reading this dissertation. We use the term choice to denote the way in which the organism distributes its time among all these activities, and some of these choice patterns are more beneficial to the organism than others, showcasing the importance of understanding the mechanisms underlying choice. A …

Mining Time Series Shapelets And Association Rules For Solar Flare Prediction, Drew Watson

All Graduate Theses and Dissertations, Fall 2023 to Present

Solar flares are the largest explosions in the solar system; they are caused by changes in the Sun’s magnetic field. Strong solar flares can disrupt power systems, damage satellites, and interfere with radio communication, so improving flare prediction is important. This thesis develops a way to predict severe solar flares while also helping researchers understand why those predictions are made. The approach looks for short patterns in solar magnetic field data that are linked to future flare activity. It then studies how these patterns appear together and in what order they happen over time. By doing this, the research not …

Dbssnet: Dual-Branch Spectral-Spatial Network With Data-Driven And Knowledge-Guided Band Selection For Uav Hyperspectral Wheat Rust Detection, Subin Kim

All Graduate Theses and Dissertations, Fall 2023 to Present

Wheat rust is a serious plant disease that can reduce crop yield and quality. In practice, the disease is often noticed only after visible symptoms appear, when some damage may already be difficult to reverse. This thesis studies whether drone-based imaging can help detect wheat rust earlier and more reliably in field environments. 

Unlike an ordinary color photograph, a hyperspectral image records reflected light at many narrow wavelengths. These measurements can reveal useful information about plant condition, but they are also high dimensional, noisy, and difficult to analyze when only a limited number of labeled field samples are available. To …

Viability Assessment Of Bovine Embryos: A Public Dataset And Deep Learning Baselines, Erfan Khayyati

All Graduate Theses and Dissertations, Fall 2023 to Present

Improving the success rates of cattle breeding is essential for sustainable agriculture, global food security, and high-quality livestock production. Currently, determining whether a lab-grown bovine embryo is healthy enough for a successful pregnancy requires highly trained experts to manually evaluate days of continuous time-lapse video footage. This process is not only incredibly time-consuming but also highly subjective; human reviewers often suffer from visual fatigue when tracking subtle, microscopic cellular changes over a seven-day period, leading to significant disagreement among even top experts on an embryo’s true potential. Furthermore, assessing bovine embryos is notoriously difficult due to their dark, lipid-dense cellular …

Divergent Belowground Strategies Of Savanna Grasses Under Varying Fire Regimes, Abigail C. Schmidt

All Graduate Theses and Dissertations, Fall 2023 to Present

Savannas are unique ecosystems, where trees and grasses live together across the landscape, not quite forming full forests or open grasslands. The ability for trees and grass to coexist is largely the result of widespread disturbances, such as fire. Fire in savannas is largely thought of in how it maintains plants above ground, however, fire also plays a key role in ecosystem and plant processes belowground. First, fire can change how soil nutrients, namely phosphorus, are cycled. Phosphorus is a key nutrient that plants need to function, and plants acquire phosphorus through their roots. Frequent fire can enhance phosphorus availability, …

Multispecies Weed Detection Using Unmanned Aerial Vehicles And Deep Learning Object Detection Models In Utah Forage Crop Corn Field, Utsav Bhandari

All Graduate Theses and Dissertations, Fall 2023 to Present

Weeds cost global agriculture over $32 billion annually and reduce crop yields by nearly one-third. Current weed control relies heavily on spraying herbicides uniformly across entire fields, leading to herbicide-resistant weeds, environmental harm, and increasing costs for farmers. A smarter approach is detecting weeds from the air and spraying only where they are present. This could dramatically reduce herbicide use while protecting crop yields. This thesis developed a complete system for identifying weeds in corn fields using drone-captured images and artificial intelligence. Working in commercial forage corn fields in Cache Valley, Utah, high-resolution aerial images were collected and used to …

Spatial Prediction Under Uncertainty: Methodological And Computational Advances In Bayesian Maximum Entropy, Kinspride K. Duah

All Graduate Theses and Dissertations, Fall 2023 to Present

Environmental decisions such as infrastructure design, water management, and snow load estimation depend on spatial data that are often incomplete or uncertain. In many cases, measurements are not exact values but ranges, reflecting limitations in data collection methods. Traditional mapping techniques typically simplify these uncertain measurements, which can lead to less accurate predictions. This dissertation introduces improved statistical tools for making spatial predictions when data are uncertain or partially known. By utilizing a framework called Bayesian Maximum Entropy (BME), this research demonstrates how exact measurements and range-based data can be combined in a mathematically consistent way. The work demonstrates that …

Unifying And Expanding Global And Local Variable Importance Methods For Explainable Machine Learning, Kelvyn K. Bladen

All Graduate Theses and Dissertations, Fall 2023 to Present

Machine learning methods are powerful analytical tools used across all scientific disciplines and many other fields of investigation for prediction and inference from diverse data sources. Despite their broad applicability, machine learning methods are often highly complex and difficult to interpret. Developing a greater understanding of which variables most influence a response is essential for increasing the interpretability of these models and supporting informed decision-making. This research focuses on improving how we evaluate the importance of these variables.

One common approach is to shuffle the values of a variable and see how much the model accuracy gets worse. Another approach …