Physical Sciences and Mathematics Commons^™

Multimodal Learning For Visual Perception And Robotic Action, Taisei Hanyu

Electrical Engineering and Computer Science Undergraduate Honors Theses

Multimodal learning aims to weave information from images, language, depth, and other sensors into one coherent representation, much as people naturally combine sight, speech, and sound. Progress toward that goal is slowed by three gaps: vision encoders that cannot balance crisp object boundaries with global context, 3-D semantic maps that are computationally prohibitive for real-time, open-vocabulary queries, and vision-language-action pipelines that depend on large token pools with weak relational grounding.

We first introduce AerialFormer, a lightweight hybrid of convolutional and Transformer layers that captures long-range structure without sacrificing fine detail. On the large-scale iSAID benchmark it reaches 69.3% mean IoU, …

Analyzing Unmanned Aircraft System (Uas) Incidents From Nasa Asrs Data Using Unsupervised Machine Learning, Kacey Haws

Electrical Engineering and Computer Science Undergraduate Honors Theses

The NASA Aviation Safety Reporting System (ASRS) assembles voluntarily submitted aviation safety incident reports in their database to act on the information provided. This database allows the government, companies, and citizens to submit incident or situational reports to its database to discern recurring issues in the National Aviation System (NAS) so that the proper officials can act [1]. The narratives provided in these reports are text-based, resulting in large amounts of data to process. Previous work in the University of Arkansas Aerospace Systems Engineering and Transportation Laboratory (ASYST) lab involved parsing unmanned aircraft system (UAS) incident reports manually. While these …

An Introductory-Level Undergraduate Cs Course That Introduces Parallel Computing, Tia Newhall, Kevin C. Webb, Vasanta Chaganti, Andrew Danner

Computer Science Faculty Works

We present the curricular design, pedagogy, and goals of an introductory-level course on computer systems that introduces parallel and distributed computing (PDC) to students who have only a CS1 background. With the ubiquity of multicore processors, cloud computing, and hardware accelerators, PDC topics have become fundamental knowledge areas in the undergraduate CS curriculum. As a result, it is increasingly important for students to learn a common core of introductory parallel and distributed computing topics and to develop parallel thinking skills early in their CS studies. Our introductory-level course focuses on three main curricular goals: 1) understanding how a computer runs …

Identification And Environmental Impact Of 35 Different Cut Flower Retail Packages In The U.S.A., Mary Kathryn Gillespie

All Theses

Cut flowers represent 35-billion dollars of global spending with 10-million dollars per day in the U.S.A; each transaction involving packaging. Across the country, cut flower packaging is inconsistently sized, laminated, colored, and unlikely to be recycled. 35 unique cut flower packages were collected across 16 different marketplaces, representing 88% of top grocery retailers. Measuring the materiality of the predominately unlabeled and inconsistent packaging collected required calorimetry equipment, such as Fourier-transform Infrared Spectroscopy, Differential Scanning Calorimeter, and precision gauges. Results unveil five materials representing the cut flower collection: Oriented Polypropylene (OPP), High-Density Polyethylene (HDPE), Paper, Polyethylene Terephthalate (PET), and Nonwoven Polypropylene …

Diversity-Augmented Training For Generalizable Ai Agents, Wenjun Li

Dissertations and Theses Collection (Open Access)

Deep Reinforcement Learning (RL) has achieved remarkable success over the past decade, from superhuman performance in video games to real-world applications like robotics. However, RL models often lack generalization, making them unreliable when deployed in unfamiliar scenarios. For example, robots must adapt to varying terrains with different slopes and obstacles, yet standard RL training does not explicitly promote such adaptability. While various methods have been proposed to enhance RL robustness, achieving reliable generalization remains an open challenge.

This dissertation focuses on improving the generalization capability of agents in three major settings: infinite horizon RL agents, finite horizon RL agents, and …

2025 May - Tennessee Monthly Climate Report, Tennessee Climate Office, East Tennessee State University

Tennessee Climate Office Monthly Reports

Hi All,

For May 2025, the main weather/climate stories were the continued pattern of severe weather impacting Tennessee as well as Chattanooga recording their wettest May in the 147-year weather history of the city. Starting with the record rainfall in Chattanooga, the Chattanooga airport weather station recorded a total of 12.02-inches of rainfall over May 2025, 8.08-inches above normal and 0.02-inches above the previous record wettest May in 1929. Many other areas of East Tennessee and the Cumberland Plateau also had well above normal rainfall this month, but they were emerging out of a drought from last fall and winter …

Statistical Designs For Learning User Preference And Experience, William S. Fisher

All Dissertations

For any organization, designing a new product or planning the features of a new service is a complex decision-making process which depends on understanding its users' preferences and the experiences they have with prototypes of the new product or service. Design of experiments (DOE) offers a framework to aid in the modeling and collection of data to learn user preferences and experiences to facilitate this decision-making process. Questions such as ``which of our newly proposed products is preferred most by our customers" and ``which version of our new service results in the highest monthly revenue" can be readily addressed by …

From The University To The Universe: Byu’S West Mountain Observatory, Amber Zendejas, Dan Broadbent

BYU West Mountain Observatory Project

Brigham Young University’s West Mountain Observatory (WMO) is a research facility that offers primarily undergraduate students valuable mentoring and hands-on learning opportunities. Established in 1981, the observatory was built in response to increasing light pollution in Provo, Utah, which began affecting research at the Orson Pratt Observatory (OPO) atop the Eyring Science Center on the BYU main campus. Before WMO was created, OPO served as the primary observatory for astronomical research at BYU. Construction plans for a new observatory began in the late 1970s, as university planners and astronomers sought a location with minimal light pollution, favorable elevation, relatively easy …

Using Gaussian Process Regression To Learn Thermodynamic Equations Of State With Uncertainty Quantification, Austen T. Lee

Chemical Engineering Undergraduate Honors Theses

This study investigates the use of derivative-informed Gaussian Process (GP) models to estimate thermodynamic behavior across temperature and density by building a Helmholtz-based equation of state. Argon, a stable monatomic gas, was chosen as a case study within the vapor region. The GP model was trained using values of experimentally measurable properties found by taking first and second derivatives of the original potential function. Results show that while the GP model offered uncertainty quantification and informed thermodynamic behavior, it predicted values that deviated from the ground truth depending on the property. The model exhibited high confidence in regions with substantial …

Deepfakes On Trial: Developing A High-Accuracy, Court-Admissible Ai Pipeline For Deepfake Detection In Corporate Fraud Litigation, Aiden J. Green

Honors College Theses

As deepfake technology advances, cybercriminals are increasingly using AI-generated videos and audios to impersonate executives and carry out sophisticated CEO fraud schemes. These synthetic forgeries target human trust and corporate communication systems, creating an urgent need for forensic tools capable of authenticating digital evidence with legal accuracy. This thesis presents a forensic-grade AI deepfake detection pipeline designed for this purpose, emphasizing courtroom admissibility, reproducibility, and evidentiary integrity. Built entirely with free, opensource tools, the framework combines metadata analysis, AI-powered spectrogram analysis, neural artifact detection, and facial manipulation recognition into a transparent workflow that accurately identifies synthetic media. It was trained …

Understanding University-Level Food Recovery Programs’ Approaches To Minimizing Food Waste While Combating Food Insecurity In The United States, Ella F. Flesher

Crop, Soil and Environmental Sciences Undergraduate Honors Theses

Food is wasted every day at each level of the food system from production to consumption. This surplus has significant negative impacts on the environment after food is sent to the landfill and allowed to rot. However, some surplus food can be repurposed and diverted from the landfill. This study served as an initial investigation to explore multiple food waste diversion programs and analyze their structure, best practices, and methods to divert leftover food. This study specifically aimed to compare and contrast food recovery programs by interviewing participants from three food recovery programs from different university institutions. With the key …

Land8fire: A Complete Study On Wildfire Segmentation Through Comprehensive Review, Human-Annotated Multispectral Dataset, And Extensive Benchmarking, Anh Tran

Data Science Undergraduate Honors Theses

Early and accurate wildfire detection is critical for minimizing environmental damage and ensuring a timely response. However, existing satellite-based wildfire datasets suffer from limitations such as coarse ground truth, poor spectral coverage, and class imbalance, which hinder progress in developing robust segmentation models. In this paper, we introduce Land8Fire, a new large-scale wildfire segmentation dataset composed of over 20,000 multispectral image patches derived from Landsat 8 and manually annotated for high-quality fire masks. Building on the ActiveFire dataset, Land8Fire improves ground truth reliability and offers predefined splits for consistent benchmarking. We evaluate a range of state-of-the-art convolutional and transformer-based models, …

Exploring The Robustness Of The Effect Of Evo On Intention Valuation Through Replication: Supplemental Material, Yesugen Baatartogtokh, Kaitlyn Cook, Alicia M. Grubb

Computer Science: Faculty Publications

Supplemental material for the paper: "Exploring the Robustness of the Effect of EVO on Intention Valuation through Replication"

Characterization Of Residual Mineral Content Of Bauxite Stockpiles In Saline County, Arkansas: An Xrd Analysis Approach, Andrew Kaze Sezibera

Electronic Theses and Dissertations

The Saline Mining District of central Arkansas once led U.S. alumina production and hosts extensive bauxite and nepheline syenite deposits. With the depletion of highgrade ore, renewed interest in bauxite stockpiles has focused on critical mineral potential. Previous X-ray Fluorescence studies have identified critical elements in significant concentrations, including uranium (U), thorium (Th), gallium (Ga), molybdenum (Mo), and lanthanides (Ln). This study examines five stockpiles (SP1, Clay, SP2, SP3, and SP4) in Section 20 of the Aluminum Company of America (ALCOA) Saline County bauxite mine, using Xray Diffraction (XRD) on 25 samples to characterize mineral phases and compare them with …

Pastoral Land Condition: Kimberley Land Condition Standards, Robert Sudmeyer, Rick Fletcher, Kathryn Ryan, Matthew Fletcher, Charles Martin, Philip Thomas, Karen Holmes, Karyn Reeves, Wayne Fletcher

Natural resources published reports

This report outlines the conceptual basis and quantitative land condition standards that define acceptable pastoral land condition for the Kimberley, Western Australia (WA), to meet the regulatory requirements of the Land Administration Act 1997 and the Soil and Land Conservation Act 1945.

The Department of Primary Industries and Regional Development (DPIRD) develops regional-level pastoral land condition standards for WA to implement the Framework for sustainable pastoral management – Revised edition, which is designed to address the Office of the Auditor General’s recommendations to improve the management of pastoral lands in WA. The application of these standards is expected to …

Design, Testing, And Characterization Of A Dc-Discharge Miniature Gridded Ion Thruster, Ishaan Mishra

Senior Project - Physics and Optical Engineering

Gridded Ion Thrusters are a satellite propulsion system that generate a Low- Temperature Plasma and electrostatically accelerates the ions to produce thrust at a high specific impulse. This work presents the design, testing, and initial characterization of a 3- cm diameter DC-discharge ring-cusp Gridded Ion Thruster that utilizes Argon as a propellant. The test infrastructure including vacuum chamber, propellant feed system, and power supplies are detailed, and operation of the ion thruster described. The material selections and design details have been presented, along with characterization based on steady-state propellant and power inputs described. In addition, a Heaterless Hollow Cathode design …

Accelerating Knowledge Graph And Ontology Engineering With Large Language Models, Cogan Shimizu, Pascal Hitzler

Computer Science and Engineering Faculty Publications

Large Language Models bear the promise of significant acceleration of key Knowledge Graph and Ontology Engineering tasks, including ontology modeling, extension, modification, population, alignment, as well as entity disambiguation. We lay out LLM-based Knowledge Graph and Ontology Engineering as a new and coming area of research, and argue that modular approaches to ontologies will be of central importance.

Education In The Era Of Neurosymbolic Ai, Chris Davis Jaldi, Eleni Ilkou, Noah Schroeder, Cogan Shimizu

Computer Science and Engineering Faculty Publications

Education is poised for a transformative shift with the advent of neurosymbolic artificial intelligence (NAI), which will redefine how we support deeply adaptive and personalized learning experiences. The integration of Knowledge Graphs (KGs) with Large Language Models (LLMs), a significant and popular form of NAI, presents a promising avenue for advancing personalized instruction via neurosymbolic educational agents. By leveraging structured knowledge, these agents can provide individualized learning experiences that align with specific learner preferences and desired learning paths, while also mitigating biases inherent in traditional AI systems. NAI-powered education systems will be capable of interpreting complex human concepts and contexts …

Deleting Values May Either Increase Or Decrease Variance, David L. Farnsworth

Articles

The impact upon variance when a value is deleted is addressed. It is shown that the cutoff for the deleted value yielding an increase or a decrease in variance is approximately one standard deviation from the mean for a univariate random variable with equally distributed probability on a finite set of elements and for a univariate set of observations. The influence of truncation of the domain for such a discrete random variable and for observations is considered.

Protein Structural Changes Of Envelope Glycoprotein In Human Immunodeficiency Virus Type 1, Abigail S. Bridges

Chemistry & Biochemistry Undergraduate Honors Theses

The envelope glycoprotein (Env), present on the viral surface of human immunodeficiency virus type 1 (HIV-1) facilitates host cell infection, making it a critical target for immune responses and vaccine development. Env is able to transition through three prevalent prefusion conformational states, state-1 (pre-triggered), state-2 (closed), and state-3 (open). By transitioning through these conformation states, Env can successfully evade immune responses and initiate viral entry into the host cell. This study uses molecular dynamics (MD) simulations to analyze the conformational dynamics of the mature, cleaved Env glycoprotein in the state-2 conformation (PDB ID: 4ZMJ). Simulations were conducted with five replicates, …

Earthscape Ai Dataset V1.1, Matthew Massey, Abdullah-Al-Zubaer Imran

Research Data--KGS

EarthScape v1.1 Dataset Download Links:

esv1p1_smokeset.zip (197 MB)

esv1p1_splits.zip (2.5 MB)

esv1p1_usa_ky_data.zip (1 MB)

esv1p1_usa_ky_hardin_howevalley.zip (26.3 GB)

esv1p1_usa_ky_hardin_sonora.zip (26.3 GB)

esv1p1_usa_ky_warren.zip (166 GB)

Surficial geologic mapping is essential for understanding Earth surface processes and supporting applications in engineering, hazard assessment, and resource management. However, traditional mapping approaches are labor-intensive, limiting spatial coverage and introducing subjectivity. To address these challenges, we introduce EarthScape, a multimodal, AI-ready geospatial dataset designed for surficial geologic mapping and Earth surface analysis. EarthScape integrates high-resolution aerial RGB and near-infrared (NIR) imagery, digital elevation models (DEM), multi-scale terrain features derived from DEMs, and hydrologic and infrastructure vector …

The Impact Of Irrigation On Agricultural Land Values In The United States Southeast And An Analysis Of Farmland Prices In South Carolina, Christopher L. Thomas

All Theses

Land is a key component of the agriculture industry in the United States. According to the United States Department of Agriculture, farm real estate (land and buildings) accounted for 82.8% of total US farm sector assets in 2022, consequently making land the largest source of wealth in a farmer’s investment portfolio and the primary source of collateral used for farm operating loans. The Southeastern region of the United States has seen rapid farmland appreciation and intense population growth in recent years. This research investigates the rising farmland prices across the Southeast from both a regional and state-level perspective. Chapter Two …

Modeling Dna Repair In Escherichia Coli Using A Boolean And Stochastic Framework, Gabrianne Ivey

All Theses

DNA can be damaged through both internal and external sources. Therefore, cells have created methods to repair DNA damage. In Escherichia coli, the system responsible for DNA repair is termed the SOS response. This system consists of more than 50 genes and contains three main repair pathways: nucleotide excision repair, translesion synthesis, and homologous recombination. The response is initiated when DNA lesions result in the accumulation of single-stranded DNA (ssDNA). The protein RecA is activated by binding to ssDNA and is then denoted RecA*. RecA* assists in the auto-cleavage of LexA which is the primary repressor protein involved in …

Evaluation Of Pre-Trained Vision Language Models In Challenging Contexts, Kankan Zhou

Dissertations and Theses Collection (Open Access)

The rapid advancement and proliferation of pre-trained vision-language models (VLMs) have heralded a new era in the realm of artificial intelligence (AI), opening up unprecedented opportunities and challenges alike. This dissertation sets forth on an ambitious and comprehensive journey to critically evaluate the performance and limitations of pre-trained VLMs, particularly in complex and challenging contexts that test the bounds of their capabilities. Our focus is twofold: to rigorously assess the extent of bias embedded in these models, and to meticulously scrutinize their reasoning abilities, highlighting parallels and disparities between machine and human cognition.

We initiate our exploration with a targeted …

Global Sporadic-E Prediction And Climatology Using Deep Learning, J. A. Ellis, Daniel J. Emmons, M. B. Cohen

Faculty Publications

Sporadic-E (E_s) is an ionospheric phenomenon defined by strong layers of plasma which may interfere with radio wave propagation. In this work, we develop deep learning models to improve the understanding of E_s, including the presence, intensity and height of the layers. We developed three separate models. The first, building off earlier work in (J. A. Ellis et al., 2024, link in AFIT Scholar, 10.1029/2023sw003669), includes only the main features from radio occultation (RO) measurements. The second adds to that time, date, location, geomagnetic and solar indices, solar winds, x-ray flux, weather and lightning. A …

Full-Stack Web Applications: Industry Standard Frameworks, Libraries & Technologies, Yassine Chahid, Patrick Slattery

Publications and Research

This research explores emerging full-stack web development technologies across front-end, back-end, and DevSecOps domains. It evaluates modern tools including Django, React, and TypeScript—focusing on their key features such as compile-time error checking—through to the development of a web application. By examining documentation for the frameworks Node.js, Next.js, Tailwind CSS, and others, along with the deployment tools Docker and Git for version/release control, the study analyzes how these innovations speed up development, improve existing practices, and have often replaced older technologies. Cloud solutions for tasks such as authentication and deployment will also be evaluated, along with various web-application technology stacks and …

Development Of Novel Pharmaceutical For Optimization Of Opioid Overdose Treatment, Larae Kram

Celebrating Scholarship and Creativity Day (2018-)

Although opioids are highly effective drugs for pain management, their overuse and addictive nature has caused an opioid epidemic, resulting in many cases of opioid addiction and overdose. There are two current opioid overdose drugs that are FDA approved: naloxone and nalmefene. These drugs have a higher affinity to the active binding pocket on the opioid receptor than opioids do and displace the opioids, preventing any further binding of the opioid and halting the opioid’s effects. However, naloxone is not active long enough to outlast the duration of the opioid and nalmefene causes severe withdrawal symptoms. The development of more …

Performance Optimization Of A 1:10 Scale Radio-Controlled Car Through Aerodynamic And Mechanical Enhancements, George Kerner

Celebrating Scholarship and Creativity Day (2018-)

Changes in the performance of a 1/10 scale radio-controlled car due to various mechanical and aerodynamic modifications were investigated. The car’s mechanical performance was evaluated by measuring lap times on custom figure-eight and oval tracks to assess handling. A straight 60-foot track was used to measure acceleration times. Mechanical modifications included a low-center-of-gravity chassis, a higher power motor, and wider street tires to optimize suspension, drivetrain, and grip. Aerodynamic enhancements included the integration of front and rear spoilers, a diffuser, and an undertray all designed using Autodesk Fusion. All aerodynamic components were 3D printed and mounted to the vehicle for …

Does The Transit Industry Understand The Risks Of Cybersecurity And Are The Risks Being Appropriately Prioritized?, Scott F. Belcher, Terri Belcher, James Grimes, Lusa Holmstrom, Andy Souders

Mineta Transportation Institute

The intent of this study is to assess the readiness, resourcing, and capabilities of public transit agencies to detect, identify, be protected from, respond to, and recover from cybersecurity vulnerabilities and threats. This study is an update of the 2020 Mineta Transportation Institute (MTI) study, “Is the Transit Industry Prepared for the Cyber Revolution? Policy Recommendations to Enhance Surface Transit Cyber Preparedness.” In the previous study, the authors found that the transit industry was ill-prepared for cybersecurity attacks. Unfortunately, after four years and the development of new, and often free, resources, the situation has not markedly improved. In fact, this …

Glacial-Interglacial And Millennial-Scale Changes In Nitrous Oxide Emissions Pathways And Source Regions, J. A. Menking, J. E. Lee, E. J. Brook, J. Schmitt, L. Soussaintjean, H. Fischer, J. Kaiser, A. Rice

Physics Faculty Publications and Presentations

During the transition from the Last Glacial Maximum (LGM) to the Holocene, the atmospheric N2O mole fraction increased by 80 nmol mol−1. Using ice core measurements of N2O isotopomer ratios, we show that this increase was driven by increases in both nitrification and denitrification, with the relative partitioning between both production pathways depending on the assumed isotopic end‐member source signatures. Similarly, we also attribute a 35 nmol mol−1 N2O mole fraction increase during the Heinrich Stadial 4/Dansgaard Oeschger 8 (HS4/DO8) millennial‐scale event to increases in both N2O production pathways. In contrast, the 25 nmol mol−1 N2O mole fraction decrease during …