Digital Commons Network^™

Perspectives On Interpretability For Neural Text Representations, Jia Peng Lim

Dissertations and Theses Collection (Open Access)

In this dissertation, we investigate interpretability in the three elements of learning neural text representations: inputs, passed into models, to produce probabilistic outputs. We emphasise perspectives as we present alternative novel methods to mine and organise meaning in this work.

Models. We initiate our investigation by examining Neural Topic Models (NTM), proposing an alternate angle of interpreting its word-topic distribution, producing better topic representations for interpretation. Our method maps the problem of finding these better interpretations to classical NP-hard graph problems, enabling examination of topic distributions in a composite manner. Next, we apply our previous findings to extract interpretations from …

Applications Of Cyrene For Pharmaceutical Synthesis, Anjali Mae Rabindran

Undergraduate Honors Thesis Collection

Cyrene is a cellulose-derived molecule that has gained interest in green chemistry as both a solvent and precursor for pharmaceutical intermediates. In this study, biocatalytic and chemical strategies were evaluated for converting Cyrene into these intermediates. Cyrene was carried through variations of a five-step sequence: reduction, reductive amination, esterification, tosylate salt formation, and Lewis-acid mediated acetal opening to generate ring-opened pyranose products (“deoxy sugars” and “deoxy amino sugars”). We successfully synthesized deoxy sugar intermediates; however, the production of stereochemically pure deoxy amino sugar intermediates was limited by unresolved diastereomeric mixtures. These findings highlight the promise of Cyrene as a renewable …

Base Running: A Lost Art In Baseball, Ethan York

Departmental Honors & Graduate Capstone Projects

In an era of baseball dominated by home runs and launch angles, the subtle art of baserunning is often overlooked, despite its measurable impact on winning games. Baserunning Runs (BsR) addresses this gap by quantifying the number of runs a player contributes through performance on the basepaths, capturing value beyond traditional metrics like stolen bases. This study constructs multiple regression models that predict BsR for Major League Baseball (MLB) players based on baserunning-related statistics. The primary objective is to examine the association between BsR and key predictors, including stolen bases (SB), extra bases taken (EB), and sprint speed (SS), while …

The Cobalt Curse: Cobalt’S Role And Risks, Marley Jackowitz

Student Theses 2015-Present

This paper examines the hidden environmental, geopolitical, economic, and social costs of the cobalt supply chain. Fueled by the demand for battery technology and electric vehicles, cobalt has become an essential element to the green energy transition. With the largest share of global cobalt deposits, the Democratic Republic of the Congo bears a disproportionate share of the industry’s harms. Cobalt mining in the DRC is marked by environmentally degrading practices and inhumane working conditions, compromising Congolese health, safety, and well-being. Chapter 1 discusses the environmental impacts of DRC cobalt mining and its adverse effects on public health. Environmentally degrading mining …

Strengthening Cyber Resilience In Critical Infrastructure: Lessons From Major Attack Case Studies, Jodi Barnes

Data Science Undergraduate Honors Theses

This comparative case study research paper analyzes the Colonial Pipeline attack, the Oldsmar Water Treatment Plant attack, and related case studies to identify past and current gaps in cyber resilience in critical infrastructure.  It provides insights into the importance of cybersecurity and opportunities to enhance protection in an increasingly digital world.  Findings include unsecure practices, limited communication between sectors, outdated technology, and weaknesses in security processes and employee training. These vulnerabilities are interconnected and are largely driven by limited funding within critical infrastructure systems, which restricts the ability to address them effectively.

Reimagining Less-Than-Truckload Pricing Development In Competitive Bid Environments With Artificial Intelligence, Lawson C. Levin

Data Science Undergraduate Honors Theses

This undergraduate thesis explores how data analytics and engineering judgment are used to support pricing decisions in the less-than-truckload (LTL) freight market. It’s based on an internship with ArcBest Corporation. It explains the company’s background, its role in the LTL market, and the responsibilities of a Pricing and Supply Chain Engineer within the Yield department.

Most of the internship was spent evaluating requests for proposals (RFPs), in which a negotiating third party provides a customer’s shipment data that must be cleaned, analyzed, and translated into a comprehensive pricing offer. Using the Data Science Analytics Process as a framework, this thesis …

Prescribing Company Action Through Machine Learning And Ai, Breck T. Husong

Data Science Undergraduate Honors Theses

The purpose of this research is to implement an OpenAI Reinforced Learning prescription-giving model for improving sales on a week-by-week basis. The data used comes from a segment of High Impact Analytics’s sales data that has been anonymized for proprietary reasons. The features among the data include inventory numbers, shipments in transit, total quantity and dollars of products sold each week for the past 2 years, all aggregated at the store-item-week level. In order to build this model, Tigramite, a causal discovery model combined with prediction models XGBoost, Linear Regression, Ridge Regression, Lasso Regression, Scikit-learn’s MLP, and Keras’s Neural Model …

A Comparative Machine Learning Framework For Identifying Ai-Generated Versus Real Celebrity Faces, Sidney Gehring

Data Science Undergraduate Honors Theses

The rapid advancements in the world of generative artificial intelligence has enabled the creation of highly realistic fictitious facial images, raising concerns about authenticity and bias in computer vision systems. This study investigates the capabilities of machine learning models to distinguish between real and artificially generated facial images across gender and race focusing on celebrity imagery. Four datasets were used against the classification model, each trained on images of a single celebrity within distinct demographic groups: White women, White men, Black women, and Black men. For each group, real images are paired with AI-generated counterparts designed to closely replicate the …

A Forecasting Framework For Distribution Center Capacity Utilization: An Applied Industry Study, Jordan J. Shortt

Data Science Undergraduate Honors Theses

This project develops and evaluates a predictive modeling framework for forecasting distribution center capacity utilization at Company Y, with monthly forecast horizons up to one year. Motivated by the operational challenges of seasonal demand volatility, promotional cycles, and the absence of a formally defined capacity metric, the study first constructs a historical capacity utilization measure from raw warehouse management system data — reconciling item volumes, location dimensions, and utilization factors across all DCs — which serves as the target variable for all modeling work. Four models are developed and evaluated against a naïve seasonal baseline: SARIMA, LightGBM, LSTM, and a …

Developing Tracking Compliance Standards For Inbound Freight: A Data-Driven Industry Application At O’Reilly Automotive, Jackson Endacott

Data Science Undergraduate Honors Theses

Visibility of inbound freight is critical for managing operational efficiency, yet many organizations lack standardized compliance metrics for third-party carriers to uphold, preventing them from utilizing tracking data to make data-driven decisions. During a summer internship with the Transportation Department at O’Reilly Automotive, data inconsistencies were addressed in the Transportation Management System (TMS), and that data was utilized to create tracking compliance standards for third-party carriers. Data populated from various sources within O’Reilly’s TMS was cleaned, validated, and utilized to create a Tracking Scorecard that evaluates message transmission rates, timeliness, and errors. This tool provides actionable insights to improve tracking …

Pyspqr: A Python Package For Density Estimation Using Deep Learning, Cameron Eddy, Reetam Majumder

Electrical Engineering and Computer Science Undergraduate Honors Theses

Splines are used for representing complex functions. In statistics, splines can be used for distributional shapes that are difficult to model by traditional parametric approaches. Ramsay (1) uses M-Spline bases to estimate continuous distributions. Semi-Parametric Quantile Regression (SPQR), developed by Xu and Reich (2), models conditional distributions where a neural network is used to estimate the basis function weights that depend on covariates. (3) implements a package for SPQR in R. We build on this by implementing a version of SPQR in Python with PyTorch. By using PyTorch, we can use more sophisticated deep learning architectures than those available in …

Towards Multi-Hop Retrieval Using Bipartite Question-Oriented Graphs, Micah Mccollum

Electrical Engineering and Computer Science Undergraduate Honors Theses

Accurately answering multi-hop questions requires full retrieval of multiple, interdependent passages and is a long-standing problem in the area of natural language question answering (QA). While retrieval-augmented generation (RAG) helps address single-hop questions, many retrievers presently focus on semantic similarity in a dense vector space, which is insufficient for handling multi-hop questions specifically. To ameliorate this, we propose constructing a bipartite question- oriented graph composed of hypothetically generated questions connected to passages at index time. The construction of the graph is guided by a large language model (LLM) to prioritize the formation of edges that signal whether a question can …

Evaluating The Use Of Extended Reality Technology To Improve Marching Band Conducting Patterns, Nathan R. Fuhrman

Electrical Engineering and Computer Science Undergraduate Honors Theses

Conducting pattern consistency is an essential skill for marching band drum

majors, yet developing this consistency through individual practice remains diffi-

cult without real-time feedback. This thesis investigates if the use of extended

reality technologies can be used to enhance the conducting skills of novice drum

majors. Using the Meta Quest 3’s passthrough capability, the system overlays vi-

sual feedback elements — including a 3D pattern guide, path visualization, tempo

cues, and a real-time score — onto the user’s physical environment. A within-

subjects study with seven participants evaluated eight combinations of three binary

feedback variables: pattern guide visibility, tempo …

​Autonomous Agentic Orchestration For Physics-Aware Scientific Discovery: An Integrative Multimodal Framework For 2d Material Characterization​, Sankalp Pandey

Electrical Engineering and Computer Science Undergraduate Honors Theses

The advancement of next-generation semiconductor and quantum technologies relies on the scalability of the fabrication of two-dimensional (2D) van der Waals heterostructures. However, this process is severely bottlenecked by characterization workflows. Optical microscopy provides high-throughput imaging of 2D material flakes, but lacks the explicit physical priors required for the discernment of sub-nanometer thickness variations, such as distinguishing monolayers from bilayers. The use of computer vision models to automate the localization and characterization process of the flakes was proposed. As a part of this effort, we develop QuantumFlake, an open-source framework to streamline the integration and deployment of computer vision models …

Primitive Pythagorean Triples In Lean And Reduction Modulo Odd Prime Powers, Luke Biddle

Mathematical Sciences Undergraduate Honors Theses

Primitive Pythagorean triples (PPTs) are (a,b,c) triples that satisfy the Pythagorean theorem and share no other common factors outside of 1. This project examines these PPTs reduction modulo odd prime powers by combining proof writing and number-theoretical analysis with the process of verification and formalization in the Lean proof coding language. Using the parameterization of PPTs generated by using the unit circle with additional conditions, we investigate how these triples behave modulo  for odd primes , with emphasis on counting the number of elements in the set of PPTs (a,b,c) modulo pⁿ . By using cases based on initial …

Software Integration In Personal Healthcare Devices And The Patient User Experience, Yassine Chahid, Patrick Slattery

Publications and Research

This study examines the current landscape and future direction of medical device hardware and software integration, focusing on how each contributes to patient care. It begins by analyzing hardware focused medical devices, such as implantable tools patients may rely on to assist with their condition, alongside diagnostic and monitoring equipment used to treat conditions in a variety of medical areas (e.g. cardiovascular conditions). It then evaluates how software is currently integrated through embedded systems, data processing, and user interfaces that support real time monitoring and clinical decision making, and how this impacts quality of care for the patient whilst minimizing …

Course Insight Portfolio: Math 1060 Calculus Of One Variable I, Minh Ha, Dinaniaina Florence Rafidimanantsoa, Preston Sessoms, Rishabh Shukla

Publications

The following portfolio is a representation of the course MATH 1060, Calculus I, taught at Clemson University. The contents can be broken down into 3 major components.

• The first being an analysis of the students who may typically be in the course, the prerequisites required to take the course, and what paths students have after completing the course. This analysis is based upon the course curriculum at Clemson, and as a result will likely not align with that of other universities

. • The second major component is a map of the concepts discussed in MATH 1060 as well …

Course Insight Portfolio Math 1040: Precalculus And Introductory Differential Calculus, Jessica Ayers, Morgan Hayes, Meghan Kerrick, Uthman Rasaq, Drishty Singh

Publications

This portfolio includes four primary elements. First is a document titled “Where did they come from? Where are they going?” that provides a description of the types of students who are likely to take the course, followed by the “Course Concept Map” that visually depicts the flow of topics and possible misconceptions in MATH 1040. Additionally, the “Misconceptions Overlay” document lists articles that offer interesting discussion and further insight into the ideas students may bring to MATH 1040 or develop throughout the course. Two lesson plans are included that cover “Continuity and Introduction to Classifying Discontinuities” as well as an …

Course Insight Portfolio: Math 1080, A Gentle Introduction To Infinite Series, Igor Luis Aureliano, Leyanis Falcon Hernandez, Ross George, Fnu Monika

Publications

This lesson aims to introduce the concept of an infinite series through geometric representations and partial sums to help students develop an intuitive understanding of convergence. A common misconception students may have when first encountering infinite series is the belief that “a sum of infinitely many things must be infinite”. Through the study of a convergent geometric series, students will investigate how the limit of partial sums can produce a finite value, thereby addressing this misconception and developing a more accurate understanding of infinite series.

Multi-Population Sufficient Dimension Reduction, Xuerong Meggie Wen, Yuexiao Dong, Li Xing Zhu

Mathematics and Statistics Faculty Research & Creative Works

A novel dimension-reduction method is introduced for multi-population data. The approach conducts a joint analysis that exploits information shared across populations while accommodating population-specific effects. Unlike partial dimension reduction methods, which identify related directions across all populations, or conditional analyses conducted independently within each population, the proposed two-step procedure leverages cross-population information to enhance estimation accuracy. The methodology is demonstrated through simulations and two real-data applications.

Modeling Hot Jupiter Atmospheres For The Classification Of Low Abundance Molecules, Mika Brown

Physics Undergraduate Honors Theses

Transmission spectroscopy is a powerful method for studying exoplanet atmospheres. With the recent launch of the James Webb Space Telescope (JWST), several molecules have been successfully detected in hot Jupiter atmospheres at an unprecedented level of precision with transmission spectroscopy. To better understand current detection capabilities and limitations, this project used machine learning techniques to determine which molecules are currently identifiable at the spectral resolution and wavelength range of JWST’s NIRSpec instrument in PRISM mode. A large synthetic dataset of transmission spectra was first generated using the petitRADTRANS radiative transfer Python package. The resulting spectra were then used to train …

Environmental Sustainability Assessment Of Field Phosphate Monitoring Methods For Creeks, Quintino Medi

All Theses

Creeks serve as conduits for water, nutrients, and contaminants to larger water bodies such as rivers and lakes, and their health is a key indicator of overall watershed water quality condition. Unlike rivers, creeks are under-monitored due to their small size, hydrological variability, and limited accessibility. Many technologies exist for monitoring phosphorus in surface waters. Technologies may be classified into two general categories: i) proxy real-time sensors, which record generic water quality parameters, and ii) chemical reagent-based assays, which quantify soluble reactive phosphorus (SRP) in batch samples. There is a lack of studies that compare the features and benefits of …

High Throughput Phenomics Pipeline For Pulse Crop Nutritional Breeding, Amod Udayanga Madurapperumage

All Dissertations

Dry pea (Pisum sativum L.), lentil (Lens culinaris Medik.), and chickpea (Cicer arietinum L.) are major pulse crops valued for their high nutritional composition and importance to global food systems. Pulses are rich in carbohydrates, protein, and essential minerals, making them ideal whole foods and critical contributors to food and nutrition security. Due to these advantages, pulse breeding programs are increasingly focusing on enhancing nutritional traits, such as protein quality, amino acid balance, and micronutrient density, through the process of biofortification. However, improvement of agronomic traits remains equally essential. Characteristics such as plant height, standability, stress tolerance, …

A Low-Rank Solver For The Stokes–Darcy Model With Random Hydraulic Conductivity And Beavers–Joseph Condition, Yujun Zhu, Yulan Ning, Zhipeng Yang, Xiaoming He, Ju Ming

Mathematics and Statistics Faculty Research & Creative Works

This paper proposes, analyzes, and demonstrates an efficient low-rank solver for the stochastic Stokes-Darcy interface model with a random hydraulic conductivity both in the porous media domain and on the interface. We consider three interface conditions with randomness, including the Beavers–Joseph interface condition with the random hydraulic conductivity, on the interface between the free flow and the porous media flow. Our solver employs a novel generalized low-rank approximation of the large-scale stiffness matrices, which can significantly cut down the computational costs and memory requirements associated with matrix inversion without losing accuracy. Therefore, by adopting a suitable data compression ratio, the …

Modular Category Optimization For Substitutability: An Item-Level Approach, Medhansh A. Sankaran

Data Science Undergraduate Honors Theses

This thesis examines substitutability within Walmart apparel as a foundation for modular category optimization. Using large-scale item-level data, I develop an attribute-based framework that aggregates products to the fineline level, constructs a structured feature space, and identifies candidate substitute relationships through similarity-based matching within relevant merchandise groupings. The results show that Walmart item master data contains sufficient structure to support scalable substitute generation across a high-variety assortment. However, substitutability is not uniform: many item pairs exhibit high similarity but low observed demand transfer, indicating that structural similarity alone does not guarantee substitution. To address this, the framework is positioned within …

Recursion, Regurgitation, And Regeneration: Testing Limits And Revealing Biases Of Generative Ai Models Through Multimodal Feedback Loops, William Donnell-Lonon

Data Science Undergraduate Honors Theses

Contemporary generative AI systems such as OpenAI's GPT-4o and DALL-E models embed complex priors about society, reality, and history shaped by training data distributions, social alignment procedures, legal constraints, and safety regulations. This study uses a "telephone game" methodology to investigate how embedded social, political, and visual biases propagate and reveal themselves through iterative multimodal generation loops, where image captioning and text-to-image models are chained in successive feedback cycles.

Using CLIP similarity metrics, facial recognition algorithms, semantic drift analysis, and qualitative content observations, I tested how image subject matter affects the rate and quality of semantic and visual shift, identity …

Escaping The Promotion Trap: A Machine Learning Framework For Brand Equity Preservation In Beverage Cpg, Lucas P. Jones

Data Science Undergraduate Honors Theses

When companies acquire beverage brands, they typically value them based on total sales revenue. This traditional approach treats all sales equally over time, whether they are driven by genuine consumer demand or temporary discounts. This is important because while promotions can boost short-term sales, they tend to erode brand value over long periods of time. The measurement problem extends to acquisitions, where buyers lack the tools to distinguish real consumer demand from artificial promotional inflation.

This thesis develops a framework to separate genuine baseline demand from promotional dependence using Nielsen scanner data covering 189 beverage brands across 188,304 weekly observations …

A Spatial Analysis Of Streetlights In The City Of Sugar Land, Samuel J. Trout

Data Science Undergraduate Honors Theses

The purpose of this paper is to analyze patterns between public safety and streetlighting for the City of Sugar Land, TX so that they may better protect their citizens.  The data involved come from the City of Sugar Land’s public works division and include type and location for all the attributes. The method of doing so involved visualizing the patterns of streetlights and their closest light readings to visualize which streetlights are underperforming using the Shiny package in R. Statistical tests were also used to quantify the association between lighting, crime occurrence, and crosswalks. From this, and the literature review, …

Estimating Body Size And Biogeography In The Genus Alligator, Madelyn M. Turala

Electronic Theses and Dissertations

Several studies have examined correlations between total body length (TL) and skeletal measurements in the American alligator (Alligator mississippiensis), with some measurements serving as stronger predictors of TL than others. Femur length (FL) has been found to be tightly correlated with TL in A. mississippiensis, and this relationship has been used to estimate body size of extinct crocodilians. This thesis demonstrates that the humerus, radius, ulna, and tibia are also strongly correlated with TL and useful for body size estimation in A. mississippiensis, and thus, can be applied to related extinct crocodilians. Additionally, species distribution modeling …

Adsorption Of Pfas On Bridged Functionalized Organosilica Materials, Elisha Lawerh Kabutey

Electronic Theses and Dissertations

PFAS are hazardous contaminants that have a devastating impact on human health and the environment. Their hazardous nature has led to the development of various adsorption methods for removing these contaminants from water sources. In this study, functionalized organosilica materials were synthesized from bis[3-(trimethoxysilyl)propyl] amine using the sol-gel method. The surface amino groups of the organosilica were converted into amine hydrochloride groups. Their adsorption properties were evaluated using salts of perfluorooctanoic acid, perfluorooctanesulfonic acid, and perfluorobutanesulfonic acid. Results showed excellent adsorption capacity of the materials. Adsorption of PFAS leads to particle agglomeration and flotation of the spent material. A column …