Digital Commons Network^™

Precision-Weighted Federated Learning, Jonatan Reyes, Lisa Di Jorio, Cecile Low-Kam, Marta Kersten-Oertel

Computer Science Faculty Publications

Federated learning (FL) using the federated averaging (FedAvg) algorithm has shown great advantages for large-scale applications that rely on collaborative learning, especially when the training data is either unbalanced or inaccessible due to privacy constraints. We hypothesize that FedAvg underestimates the full extent of heterogeneity of data when the aggregation is performed. We propose Precision-Weighted Federated Learning (PW) a novel algorithm that takes into account the second raw moment (uncentered variance) of the stochastic gradient when computing the weighted average of the parameters of independent models trained in a FL setting. With PW, we address the communication and statistical challenges …

Gradient-Based Model Shortcut Detection For Time Series Classification, Salomon Ibarra, Frida Cantu, Kaixiong Zhou, Li Zhang

Computer Science Faculty Publications

Deep learning models have attracted lots of research attention in time series classification (TSC) task in the past two decades. Recently, deep neural networks (DNN) have surpassed classical distance-based methods and achieved state-of-the-art performance. Despite their promising performance, deep neural networks (DNNs) have been shown to rely on spurious correlations present in the training data, which can hinder generalization. For instance, a model might incorrectly associate the presence of grass with the label "cat" if the training set have majority of cats lying in grassy backgrounds. However, the shortcut behavior of DNNs in time series remain under-explored. Most existing shortcut …

Nursing Students’ Attitudes Regarding Technical Standards For Entrance Into A Bachelor Of Science In Nursing Program, Jordyn Cook

Honors Theses

Admission into a Bachelor of Science in Nursing program requires students to meet specific criteria including a non-negotiable set of technical standards. Technical standards encompass physical abilities including twisting, bending, and handling objects of varying weights (Good Samaritan College of Nursing and Health Science, n.d.). While physical capabilities may hold significance in bedside nursing, the American Nurses Association does not establish specific physical qualifications (American Nurses Association, 2010). Due to admission criteria and technical standards, individuals with disabilities are rarely found among career nurses and nursing students (Leslie & Miller, 2020).

This study aims to determine nursing students’ attitudes towards …

Where Is Gender In The Historiography Of Lgbtq+ College Students?, Michael S. Hevel, Timothy Reese Cain

Counseling, Leadership, and Research Methods Faculty Publications and Presentations

This forum contribution considers the complexities and importance of gender in the historiography of LGBTQ+ college students. After a brief introduction, we focus on four key areas in the existing historiography: women's romantic relationships in the late nineteenth and early twentieth centuries, higher education leaders' enforcement of gender norms and purges of LGBTQ+ students through the mid-twentieth century, the gender dynamics within LGBTQ+ student organizations from the 1970s onward, and trans and other LGBTQ+ students' expansion of gender possibilities, including through the use of drag. In so doing, we argue that gender analysis is important to the historiography of LGBTQ+ …

The Influence Of Task Type And Learner Language Background On Writing Production And Anxiety: A Bayesian Linear Mixed-Effects Analysis, Mahmoud Abdi Tabari, Xinya Liang, Agnes Albert

Counseling, Leadership, and Research Methods Faculty Publications and Presentations

Despite growing interest in task-based language teaching (TBLT), limited empirical work has examined how different rhetorical task types influence second language (L2) writing development, especially in relation to affective variables, such as writing anxiety. Existing research in TBLT has largely focused on cognitive dimensions, often neglecting individual differences in learners' emotional responses. Moreover, Long's (2014) call to use first language (L1) data as a benchmark in TBLT remains underexplored, complicating the interpretation of L2 performance patterns. To address these gaps, we examined the impact of task type and writing anxiety on the written performance of 140 university students (70 L1 …

Real-Time Estimated Sequential Organ Failure Assessment (Sofa) Score With Intervals: Improved Risk Monitoring With Estimated Uncertainty In Health Condition For Patients In Intensive Care Units, Yan He, Qian Luo, Hai Wang, Zhichao Zheng, Haidong Luo, Oon Cheong Ooi

Research Collection Lee Kong Chian School Of Business

Purpose: Real-time risk monitoring is critical but challenging in intensive care units (ICUs) due to the lack of real-time updates for most clinical variables. Although real-time predictions have been integrated into various risk-scoring systems to aid monitoring, existing systems do not address uncertainties in risk assessments. We developed an enhanced risk monitoring framework based on commonly used systems like the Sequential Organ Failure Assessment (SOFA) score by incorporating uncertainties to improve the effectiveness of real-time risk monitoring in ICUs.Methods: This study included 5,351 patients admitted to the Cardiothoracic ICU in the National University Hospital in Singapore. We developed machine learning …

On Variations Of Isolation In Graphs, Geoffrey Boyer

All Dissertations

In 2015, Caro and Hansberg introduced a wonderful and natural generalization of the well-studied parameter of domination in graphs. For a graph $G$ and a family of graphs $\FF$, they define $S$ to be an $\FF$-isolating set if $G-N[S]$ contains no member of $\FF$ as a subgraph. The case where $\FF=\{K_1\}$ coincides with domination. The case where $\FF=\{K_2\}$ is now simply referred to as an isolating set. We strengthen known results about the isolation number of a graph, and explore variations of the parameter including the independent and total versions.

In particular for connected graphs of order $n$, a bound …

Navigating Safety: A Mixed-Methods Examination Of Supportive Workplace Practices In Maritime Contexts, Rebecca N. Pool

All Dissertations

Maritime Sailors operate in high-demand and high-risk environments that require sustained adherence to strict safety standards. Using the Conservation of Resources theory (COR), this dissertation examines how supportive workplace practices impact job stress and foster increased safety compliance among U.S. Navy personnel. Using a mixed-methods design, two studies were conducted with data from the Afloat Safety Climate Assessment Survey (ASCAS). Study 1 analyzed quantitative survey responses from active-duty Sailors (N = 15,517) nested in 1,033 departments nested within 122 ships to test a multilevel moderated mediation model linking social support, job stress, psychological safety, and safety compliance. Results indicated …

Exploring Halogen Bonding Interactions In Cocrystals And Deep Eutectic Solvents, Madhushi Bandara

All Dissertations

Halogen bonding is a noncovalent intermolecular interaction analogous to hydrogen bonding that occurs between an electrophile and a nucleophile. This is often observed in compounds containing heavy halogens such as iodine and bromine. When these halogens are covalently bonded to another atom or molecule, preferably to an electron-withdrawing moiety, an electrophilic region called a sigma hole is created on the halogen atom on the extension of the covalent bond.

The first study of this dissertation presents a comprehensive characterization of eight newly synthesized triiodide salts using single-crystal X-ray diffraction, powder X-ray diffraction, thermal analysis (DSC and TGA), and Raman spectroscopy. …

First Principles Calculation Of Electron-Phonon Coupling In Nonequilibrium Quantum Materials, Chendi Xie

All Dissertations

This dissertation combines first-principles computation and light-induced nonequilibrium analysis to understand and control electron–phonon coupling (EPC) and emergent many-body phenomena in quantum materials, in both equilibrium and nonequilibrium states.

Chapters 1 and 2 build the theoretical and computational foundation, beginning with electron and phonon self-energies, linewidths, and perturbation theory for EPC, followed by density functional theory (DFT) and density functional perturbation theory (DFPT), which serve as the first-principles workhorses used throughout. Chapter 3 reviews Bardeen-Cooper-Schrieffer (BCS) and Migdal-Eliashberg theory, including the BCS gap - transition temperature (T_c) relation and the strong-coupling generalization, and delineates the practical validity bound …

Interfaces To Enhance Performance In Human-Ai Teams Conducting Safety-Critical Military Operations: A Route Reconnaissance Use Case, Sarvesh Sawant

All Dissertations

Military teams frequently operate in hazardous and dynamic environments where they must recognize threats, act swiftly, and maintain awareness of their surroundings. To assist soldiers in identifying threats, analyzing data, and creating safer routes, Artificial Intelligence (AI) systems are being integrated into these teams. But only when humans comprehend what AI is doing, why it is doing it, and how to collaborate with it can AI be helpful in these high-stakes missions. To enhance human-AI collaboration during military route reconnaissance missions, this dissertation explores ways to improve human-AI communication. In a route reconnaissance mission, teams of soldiers must navigate through …

Universal Systems Simulation Via Constraint Hypergraphs With Applications To Digital Twins, John Morris

All Dissertations

The characterization of systems encompasses a variety of modeling frameworks designed to capture specific behaviors and components of various system domains. Whatever the framework, the core elements of a system representation are the information of the system and a description of how that information is related. The relations in deterministic systems are functions, which, when composed to form executable processes, can be used to simulate system data. A declarative modeling framework is one that encodes mechanisms for preparing these simulations within the model structure, allowing an external agent to form the execution processes required for a given context. To date, …

Enhancing Teacher Planning Practices To Improve Teacher Self-Efficacy Through Professional Learning Communities, Philip L. Price

All Dissertations

This improvement science dissertation investigated the impact of implementing Professional Learning Communities (PLC) on teacher self-efficacy at Parkway Elementary, a Title I school serving a diverse student population with historically lower academic performance. The study addressed the problem of low teacher self-efficacy, a critical factor influencing Tier 1 instructional quality and student outcomes. Utilizing the Plan-Do-Study-Act cycle, this research tested the theory that adopting a four-question framework within weekly collaborative grade-level meetings would measurably increase teacher self-efficacy. Seven teachers participated in the research, with data collected through the Teacher Sense of Efficacy Scale, semi-structured interviews, and PLC observations. Quantitative findings …

Exploring Glycosomes As Therapeutic Targets For Trypanosoma Brucei, Sabrina Sutton Pizarro

All Dissertations

Trypanosoma brucei is a protozoan parasite which causes human African trypanosomiasis and the livestock wasting disease, nagana. These parasites have highly specialized peroxisomes, called glycosomes, which house a variety of essential metabolic pathways, including glycolysis. Glycolysis is the sole means of energy production in the infective life stage of T. brucei making glycosomes and the glycolytic pathway promising targets for therapeutic strategies. In this work, we have studied the glycosome as a therapeutic target from multiple perspectives. The first portion of this work addresses drug trafficking. The efficacy of small-molecule inhibitors of glycosomal enzymes is restricted by the rate of …

Flexible Spatial Priors In Bayesian Neuroimaging: Gmrf, Nngp, And Deep Gmrf, Boyoung Hur

All Dissertations

Structural neuroimaging is essential for understanding neurological disorders such as Alzheimer’s disease, enabling accurate delineation of brain regions through image segmentation. Among various segmentation methods, multi-atlas-based approaches like label fusion have become leading techniques. In statistics, Bayesian hierarchical models for label fusion are increasingly favored for their ability to incorporate uncertainty and prior knowledge. Also, a key challenge in modeling neuroimaging data is spatial dependence among image voxels, making the choice of spatial prior critical—particularly in high-resolution settings where segmentation accuracy and computational efficiency are both essential.

This dissertation proposes fully Bayesian spatial hierarchical models that explore two flex- ible …

Thermal Management Of Power Electronic Converters In Electric Motor Drives Using Active Thermal Control And Active Cooling, S M Imrat Rahman

All Dissertations

The application of power electronic converters (PEC) in electric vehicles (EVs) has significantly increased due to their enhanced controllability and flexibility. Additionally, Permanent Magnet Synchronous Machines (PMSMs) are favored in EVs for their higher torque, efficiency, performance, and heat dissipation capabilities compared to asynchronous machines. The reliable operation of power converters in PMSM drives is crucial for EV operation, and this is driving interest in developing innovative and sustainable technologies that ensure the safe and reliable functioning of PECs. One of the main challenges in achieving reliable PECs is designing effective thermal management systems to control junction temperature and reduce …

Contributions To Statistical Modeling And Estimation Of Rainfall Intensity–Duration–Frequency Curves, Jiyun Huang

All Dissertations

Extreme rainfall can cause flooding, damage infrastructure, and create serious risks for communities. Engineers use Intensity-Duration-Frequency (IDF) curves to estimate precipitation extremes over different lengths of time, such as one hour or one day, and the average time one would expect wait for one of these events to occur. Several approaches exist for estimating IDF curves, but no single approach is known to be best in all cases. In this dissertation, I study statistical methods that aim to improve IDF curve estimation. I use the Canadian Regional Climate Model (CanRCM4) Large Ensemble, which contains 35 independent climate simulations. These simulations …

Surface Reaction And Transportation Of Radium-226 In Marine Systems, Jiusi Wang

All Dissertations

Historical uses of radium (Ra)-containing self-luminous paint, such as on gauges and dials in naval ships, have led to the potential for such paint, and thus Ra, to be released into the marine environment. Radium-226 (²²⁶Ra) behavior in marine systems is not fully described in the literature, presenting a knowledge gap in the process of evaluating potential ecological impact of Ra in marine systems. This study explores various mechanisms contributing to Ra environmental transport in marine systems on a laboratory scale, including leaching, sorption, and diffusion. ²²⁶Ra leaching from watch hand dials containing radioactive luminous paint is …

Machine Learning-Enabled Safety-Critical Model Predictive Control For Uncertain Dynamical Systems, Hossein Nejatbakhsh Esfahani

All Dissertations

This dissertation explores the interactions between Model Predictive Control (MPC), Safety-critical Control, and Artificial Intelligence (AI)/Machine Learning (ML) methods, with a particular focus on Reinforcement Learning (RL) and Bayesian Optimization (BO). We then leverage AI/ML to address several challenges in the control design and safe operation of uncertain dynamical systems. In many applications, ranging from autonomous vehicles and robotics to energy systems and industrial processes, ensuring safety is as essential as satisfying control objectives. The use of Control Barrier Functions (CBFs) within the MPC framework recently has emerged as a powerful tool to guarantee safety by enforcing constraints in optimal …

Legitimacy, Learning, And Launch: Exploring Relationships, Community, And Connection In An Undergraduate Research Experience, Randi J. Sims

All Dissertations

Undergraduate research experiences (UREs) are critically important methods for integrating undergraduate students into legitimate scientific practices. While widely celebrated for producing positive outcomes such as increased self-efficacy, research skills, and STEM retention, less is known about the processes through which undergraduate mentees become integrated members of their research groups and the broader scientific community. This dissertation presents findings from both a systematic literature review and a quantitative ethnographic study of a biological sciences URE at a southeastern R1 predominantly white institution (PWI). Grounded in communities of practice and legitimate peripheral participation frameworks, this study explores how undergraduate researchers build relationships, …

Reverse (Bio)Engineering: A Machine Learning Approach To Optimize Baseball Pitcher Health And Performance, Robert C. Moore

All Dissertations

Ball tracking systems are becoming ubiquitous in sport, creating an unprecedented opportunity for big data applications to optimize human health and performance. These applications are especially common in baseball, a sport known for analyzing ball flight data to quantify performance. Analysts routinely use ball flight data to identify the attributes of top performing pitchers, finding that the best pitchers throw with optimal combinations of release speed and spin to precise locations. However, for certain pitchers, the throwing motion required to produce optimal ball flight places exceedingly high biomechanical load on the elbow, and consequently injury rates continue to rise. This …

Virtual Coaching: Ushering In A New Era In School Improvement, Tygere Alisha West Charles

All Dissertations

This research study examined virtual coaching as a vehicle for principal professional development. The study focused on coaching a principal in a rural, low-performing, high-poverty South Carolina school district to improve support for teachers and students. This study explored virtual coaching and is rooted in improvement science as measured by plan-do-study-act (PDSA) cycles (Bryk et al., 2017). It explored how virtual coaching supported a rural principal’s professional development needs.

Principals must be transformational leaders who balance their support needs and accountability while meeting the needs of the diverse stakeholders they serve. Ongoing professional development is an essential component in helping …

Hybrid Learning For Rough Terrain Navigation Of Actively Articulated Wheeled Vehicles, Dhruv Mehta

All Dissertations

Conventional wheeled ground vehicles have been used for rough terrain navigation in the recent years. They consist of a chassis connected to wheels through passive, semi-active, or active suspension systems. However, their fixed configurations limit mobility and maneuverability, constraining their ability to autonomously navigate diverse and rough terrains. Autonomous Ground Vehicles (AGVs) face significant challenges in this regard, including varying terrain roughness, soil hardness, and obstacle crossing.

To address these limitations, Actively Articulated Wheeled Vehicle (AAWV) architectures have recently emerged, offering real-time geometric adaptability. AAWVs have chassis and wheels connected via articulated serial or parallel linkages. However, increased articulation introduces …

Data-Driven Discovery Of Finite-Dimensional Koopman Operator For Modeling And Control Of Uncrewed Ground Vehicles, Ajinkya Joglekar

All Dissertations

This dissertation advances data-driven modeling and adaptive control techniques for Uncrewed Ground Vehicles (UGVs), with a focus on autonomy in mission-critical and safety sensitive environments. UGVs are deployed across a wide spectrum of domains, from structured manufacturing shop floors to unstructured off-road terrains, including planetary exploration, precision agriculture, and disaster response. These platforms, operating in dull, dirty, and dangerous conditions, demand autonomy that is both adaptable and robust. While traditional model-based control methods offer interpretability and robustness, they struggle with unmodeled dynamics, parameter variations, and integration of high-dimensional sensing. Conversely, modern machine learning approaches can directly exploit sensory data but …

Leveled Homomorphic Encryption Schemes: Noise And Precision Control, Kyle Yates

All Dissertations

Homomorphic encryption allows for computations on encrypted data without exposing the underlying plaintext, enabling secure and private data processing in various applications such as cloud computing and machine learning. In this thesis, we conduct a comprehensive worst-case noise analysis for three prominent leveled homomorphic encryption schemes: Brakerski-Gentry-Vaikuntanathan (BGV), Brakerski-Fan-Vercauteren (BFV), and Cheon-Kim-Kim-Song (CKKS). We propose modifications to these schemes and their residue number system (RNS) variants, ensuring fresh encryption noise falls under a constant bound. For BFV and BGV, we design and prove parameter conditions which guarantee certain homomorphic circuit evaluations return ciphertexts containing noise within a fixed bound. For …

An Experimental Evaluation Of A Gamified Learning Approach To Address Bullying Knowledge, Perceptions, And Self-Efficacy Among Elementary Students: Playing Yozgoo In Albania, Sashen Rexhepaj

All Dissertations

This study evaluated the effectiveness of Yozgoo's web-based gamified educational intervention in enhancing elementary school students’ knowledge, attitudes, and self-efficacy related to bullying prevention in Albania. Utilizing a true experimental pre-test and post-test control group design, the study involved 120 students (ages 10-12) equally divided by gender and randomly assigned to either an experimental group (n = 60) or a control group (n = 60). The experimental group participated in Yozgoo gameplay sessions over six to eight weeks, while the control group engaged in standard activities without exposure to the intervention.

Three validated measures were administered to assess outcomes: …

Integrating Dft And Machine Learning To Predict Structural Properties In High Entropy Alloys, Nathan Linton

All Dissertations

In the past decade, a paradigm shift in the design of metal alloys has been observed. These new alloys are commonly referred to as high entropy alloys (HEAs), multi-principal element alloys (MPEAs), or complex, concentrated alloys (CCAs). In contrast to conventional alloys, which consist of one main element (for example 80%) with other elements in small amounts, HEAs are made of four or more main elements ranging from 5 to 35% each element. Due to the large presence of multiple elements, HEAs have shown substantial material property improvements over conventional alloys such as steel. For example, they have high ductility …

Digital Reflections: Evaluating Body Dissatisfaction In Xr Through Eye- And Body-Tracked Virtual Humans, Deyrel Diaz

All Dissertations

In an era where digital and physical realities increasingly intertwine, the perception of body image is undergoing a significant transformation. Traditional understandings of body dissatisfaction, long studied in relation to psychological distress and eating disorders, are now being reshaped by technologies such as Virtual Reality (VR), Augmented Reality (AR), and Artificially Intelligent (AI)- generated media. These technologies have introduced novel ways of experiencing and interacting with the human form, raising critical questions about their impact on self-perception and internalization of beauty standards.

As virtual representations become more prevalent in entertainment, social media, and interactive platforms, it is becoming more crucial …

Atomic-Scale Engineering Of Dilute Pd–Cu Alloys For Selective Electrocatalytic Cox Reduction, Zehua Jin

All Dissertations

The excessive release of carbon dioxide (CO2) into the atmosphere has become a major cause of global warming and ocean acidification, threatening the sustainability of human life on Earth. To address this challenge, scientists are exploring ways to convert CO2 into useful fuels and chemicals instead of simply capturing and storing it. One promising approach is electrocatalysis, which uses electricity—potentially from renewable sources—to drive chemical reactions that transform CO2 under mild, environmentally friendly conditions.

This research aims to understand how the arrangement of atoms in metallic catalysts affects the conversion of CO2 and carbon monoxide (CO) into valuable products such …

Cosmic Duets: A Search For Binary Supermassive Black Holes In Merging Galaxies, Sagar Adhikari

All Dissertations

Galaxies are vast cosmic islands of stars, dust, and gas. They come in various shapes and sizes. They might look static for the timescales we are used to, but they are dynamic and collisional systems that can merge with other galaxies to make bigger galaxies. Most galaxies, including the Milky Way, host a central supermassive black hole (SMBH) with a mass greater than a million (sometimes a billion) times the mass of the Sun. When galaxies merge, their SMBHs form a binary system before ultimately merging. These cosmic duets are of great interest to astronomers and astrophysicists as they are …