Digital Commons Network^™

Towards Interpretable Transfer Learning With Limited Data, Youxiang Zhu

Graduate Doctoral Dissertations

Modern foundation models are typically trained with large-scale data to ensure good performance. However, certain tasks cannot scale with large amounts of data due to cost and practical constraints, resulting in limited performance. To address this, in this dissertation, I introduce model-task alignment, a general methodology for making a foundation model work well with tasks with limited data. The methodology comprises two parts: aligning downstream tasks with foundation models and aligning foundation models with downstream tasks. To study and validate this methodology, I first focus on speech-based dementia detection, a representative task with limited data, and then extend to general …

Adaptive Multimodal Smart Home Control On A Raspberry Pi 5 Using Hand Gestures, Voice Cues, And User Feedback, Vaibhav Bora

Theses

A real time multimodal smart home control system deployed on a Raspberry Pi 5 is presented. The system combines hand gestures, short voice cues, and proximity aware interaction to execute household commands such as light brightness control, fan speed adjustment, and stop or kill switch actions. Lightweight gesture and keyword spotting voice classifiers were trained offline and exported to TensorFlow Lite for efficient on device inference. For more natural spoken phrases, the system additionally integrates a locally deployed pretrained offline ASR component rather than a speech recognizer trained from scratch. Using a USB camera and microphone, the system operates fully …

Robustness Of Ai-Driven Histopathology Under Real-World Adversarial Examples, Ruchik N. Yajnik

Theses

This robustness of histopathology classification models under adversarial and real-world perturbations resembling clinical artifacts is being investigated.

Using whole-slide images from the CAMELYON17 cohort, four representative architectures—ResNet-18, ResNet-50, HIPT-2MLP, and ViT-B/16 —are benchmarked across controlled pixel-level distortions and artifact-like transformations. Adversarial methods include iterative Fast Gradient Sign, Projected Gradient Descent, Salt-and-Pepper noise, and the Adversarial Watermark—Stain Shift (AWSS). Three defense strategies—Randomized Smoothing, Adversarial Training, and an Artifact Detector—are evaluated for their ability to preserve diagnostic accuracy and model reliability. Structured perturbations consistently degrade performance, with transformer-based models showing the greatest sensitivity. The benchmark developed here offers a reproducible framework for …

Sound Detection Of Memory Leaks In Llvm Ir Programs Using Accumulation Analysis, Robert Blacha

Theses

Resource leaks occur when a limited resource such as memory is allocated by a program and needlessly held past the point of use. Leaks can lead to a degradation of services which can be specifically triggered with malicious behavior, for example abusing a memory leak in a program to cause a server to slow down and crash for a denial-of-service attack.

Prior work has demonstrated that accumulation analysis provides a sound detection of resource leaks with a working implementation for programs written in Java. While useful, current implementations are limited to programs written in Java, which has a garbage collector, …

Evaluation Of Medical Instrument Reliability For The Heath And Maddox Components Of The Accommodative-Vergence System, Jacob Chabuel

Theses

Binocular vision relies on the coordination of near response triad composed in part by the vergence and accommodative systems. Binocular dysfunctions linked to these systems impact the quality of life of those affected by complicating day-to-day tasks and leading to further health concerns. Clinical evaluations that diagnose dysfunctions and tailor therapies rely on subjective methods, creating a need for a reliable, quantitative, measurement tool. This is achieved using a haploscope, a tool that quantitatively measures and observes binocular vision to identify deficiencies within the eye. Modernization of a haploscope system, measurement of the Heath and Maddox components of the visual …

Investigation Of Environmental Influences On Multiple Sclerosis: A Study For Egypt’S Health, Noha Ezzaldin Alziny Ali

Theses and Dissertations

Background: Multiple sclerosis (MS) is a complex, chronic neuroinflammatory disease influenced by both genetic and environmental factors. While global research has identified potential environmental influences, most evidence comes from Western populations, creating a significant knowledge gap for regions like Egypt, which has witnessed a dramatic rise in MS prevalence over recent decades. This study hypothesized that Egyptian MS patients exhibit a distinct biological signature in their blood, characterized by an imbalance in essential and toxic trace elements (the metallome) and alterations in stress-response biomarkers associated with neuroaxonal injury and cellular stress. Objective: This study aimed to comprehensively profile and compare …

Improving Antihypertensive Medication Adherence Through Education And Follow- Up Phone Calls In A Primary Care Setting, Sarah Peterson

Doctor of Nursing Practice Final Manuscripts

Hypertension affects nearly half of adults and significantly increases the risk of stroke and cardiovascular disease; however, poor medication adherence remains a primary barrier to effective blood pressure management. The purpose of this evidence-based practice (EBP) project was to improve antihypertensive medication adherence among adult patients in a primary care setting. Guided by the Iowa Model of Evidence-Based Practice to Promote Quality Care, the project utilized a pre-test/post-test design at a private practice clinic. The sample included adults with a hypertensive diagnosis and active prescriptions. Medication adherence was defined as the Morisky Medication Adherence Scale (MMAS-8) scores, measured before and …

Closing The Gap: Integrating A Standardized Depression Screening In Primary Care, Cyndy Marie Oreas

Doctor of Nursing Practice Final Manuscripts

Objective: The purpose of this evidence-based practice (EBP) project is to implement the routine use of the Patient Health Questionnaire-9 (PHQ-9) screening instrument in the primary care setting to improve early diagnosis, treatment, and referrals to mental health.

Background: Depression is frequently underdiagnosed in primary care despite its significant impact on health outcomes. Inconsistent screening practices and a lack of standardized workflows contribute to missed opportunities for early identification and intervention. Routine PHQ-9 screening has been shown to improve detection and facilitate timely referral to behavioral health services.

Methods: A standardized PHQ-9 screening was implemented for adult patients aged 18 …

Reducing Preoperative Anxiety In Patients Undergoing Gynecologic-Oncology-Related Surgeries, Iman Sean Lavine-Richardson

Doctor of Nursing Practice Final Manuscripts

Elevated levels of preoperative anxiety are associated with adverse effects in the postoperative period. Evidence demonstrates the effectiveness of incorporating multimedia into preoperative education to alleviate preoperative anxiety. This project aims to decrease preoperative anxiety levels and increase satisfaction and feelings of preparedness in patients scheduled to undergo gynecologic-oncology-related procedures through the implementation of an evidence-based practice (EBP) informational video. The San Diego 8A’s EBP model was utilized to guide this project. Data was collected, anonymously, through online questionnaires which included the Anxiety Specific to Surgery Questionnaire (ASSQ) to assess anxiety levels and Likert-scale questions to assess satisfaction and feelings …

Implementation Of Symptom Management Standing Order Sets In An Ambulatory Palliative Care Clinic, Anna Manfreda

Doctor of Nursing Practice Final Manuscripts

Background: As outpatient palliative care expands, the need for timely, evidence-based symptom management has become increasingly critical. In many ambulatory settings, triage nurses are the first point of contact for patient symptom reporting; however, without standing orders, management may vary by provider availability and individual practice patterns, leading to potential treatment delays and inconsistencies in care from evidence-based recommendations.
Purpose: The purpose of this Doctor of Nursing Practice project was to improve the workflow and quality of gastrointestinal (GI) symptom management in an ambulatory palliative care clinic.
Framework: This project was guided by the Iowa Model Revised: Evidence-Based Practice to …

Enhancing Emergency Preparedness In California High Schools: A Policy Framework For Implementing Nata-Compliant Emergency Action Plans Through Multidisciplinary Task Forces, Jennifer Lu

Doctor of Nursing Practice Final Manuscripts

Abstract

Background: Sudden cardiac arrest (SCA) is the leading cause of death in young athletes, yet it is highly survival with proven measures: a structured emergency action plan (EAP), cardiopulmonary resuscitation (CPR) training, and access to an automated external defibrillator (AED) within 3 minutes can raise survival rates to 90%. Despite this, high schools show inconsistent compliance, largely due to the lack of fragmented oversight and lack of standardized practices

Purpose of Project: This proposal advocates for multidisciplinary task forces to enforce venue-specific, National Athletic Trainers Association (NATA)-compliant EAPs across California high schools, with the goal of improving emergency response …

Text Message Delivery Of Pre-Procedural Recommendations Prior To Lip Augmentation With Hyaluronic Acid Filler: An Evidence-Based Pilot Project, Jenna De Jonge

Doctor of Nursing Practice Final Manuscripts

Patients seeking lip augmentation procedures commonly experience self-resolving soft-tissue complications such as swelling and bruising. There are many widely accepted pre-procedural recommendations available to help patients minimize the risk of these undesirable outcomes. At a Medical Aesthetics practice in Southern California, lip augmentation patients often arrived to their appointment non-adherent to pre-procedural recommendations that were available on the practice website. Text message delivery systems have been used to improve patient adherence to healthcare instructions across a variety of healthcare settings. To address this gap, I implemented text message delivery of pre-procedural recommendations to lip augmentation patients prior to their appointment. …

Optimizing Lactation Support: Enhancing Patient-Centered Care And Breastfeeding Self Efficacy Through Data-Driven Survey Analysis, Jessica L. Doshay

Doctor of Nursing Practice Final Manuscripts

Objective: The objective of this project was to implement evaluation of an integrated lactation support model within a multi‑site pediatric primary care organization to assess patient experience and breastfeeding confidence associated with a lactation consultation.

Background: Optimizing the patient experience is an indicator of healthcare quality. Improving breastfeeding outcomes remains a national priority. Structured lactation support delivered by International Board-Certified Lactation Consultants (IBCLCs) enhances breastfeeding outcomes, the impact of this is often undocumented in primary care settings.

Methods: This quality improvement project was conducted within the lactation department of a multicenter pediatric clinic in Southern California. Pediatric clinicians who were …

Implementation Of Comprehensive Checklist For Opioid Use Disorder Treatment In Primary Care, Sophia R. Madden

Doctor of Nursing Practice Final Manuscripts

Background: The escalating opioid crisis demands a comprehensive approach to Opioid Use Disorder (OUD) management. Patients often face comorbid mental illnesses, infectious diseases, concurrent substance use, and social determinants of health (SDOH)—all of which increase the risk of treatment attrition. Standardizing care through evidence-based clinical practice guidelines and tools is essential for improving treatment quality and consistency.

Purpose: This project aimed to increase comorbidity screenings and improve resource access for OUD patients at a San Diego Internal Medicine primary care office by implementing a comprehensive care checklist based on the American Society of Addiction Medicine (ASAM) guideline.

Methods: The ASAM …

Empowering Healing And Recovery: Adverse Childhood Experiences Screening In A Cognitive Behavioral Therapy Intensive Outpatient Program, Serena C. Lee

Doctor of Nursing Practice Final Manuscripts

Purpose: This evidence-based practice project implemented Adverse Childhood Experiences (ACEs) screening in an adult Cognitive Behavioral Therapy Intensive Outpatient Program (COG IOP) at San Diego's largest behavioral health hospital to identify ACE exposure, deliver supportive interventions, and enhance clinician competency in trauma-informed care.

Background: ACEs are potentially traumatic experiences occurring before age 18 that are associated with toxic stress and worsening health outcomes. Despite evidence linking ACEs to adverse outcomes and substantial healthcare costs, routine ACE screening is not standardized in behavioral health settings.

Methods: Guided by the San Diego 8A's Evidence-Based Practice model, eligible patients were offered ACEs screening …

Enhancing Tardive Dyskinesia Screening In An Outpatient Psychiatry Training Clinic: A Quality Improvement Initiative, Ariel Davalos

Doctor of Nursing Practice Final Manuscripts

Background: Tardive Dyskinesia (TD) is a debilitating, often irreversible movement disorder caused by prolonged exposure to dopamine receptor-blocking agents, affecting 15-30% of patients on long-term antipsychotic therapy. Despite endorsement of the Abnormal Involuntary Movement Scale (AIMS) as the gold standard for TD screening, the condition remains significantly under-diagnosed. Purpose: This quality improvement project evaluated whether a 45-minute educational intervention could improve PMHNP trainees’ confidence and competence in TD recognition and AIMS administration. Methods: Guided by the Iowa Model of Evidence-Based Practice, the project used a pre-post survey design with a planned one-month follow-up across two clinical training sessions. Ten PMHNP …

Edge Co-Occurrence Regularization For Node Classification, Kadir Altunel

Theses

We propose a simple yet effective regularization technique for node classification on graphs that leverages edge-based label co-occurrence patterns. We first train an MLP on node features to produce class probability distributions, then compute a fixed penalty matrix from edge-based co-occurrence statistics of these predictions. This penalty matrix, which captures unlikely class combinations on connected nodes, is then used to regularize GNN training without further updates. We evaluate this approach across multiple homophilic datasets (Cora, CiteSeer, PubMed, ogbn-arxiv) and heterophilic benchmarks (Chameleon, Squirrel, Actor, Roman-Empire) using three GNN architectures: GCN, GraphSAGE, and H2GCN. Results show consistent improvements on homophilic graphs, …

The Effects Of Osha State Plans And State Gdp On Occupational Injuries And Illnesses In The United States, Ahmad Abdel Rahman Mohammad Amoorah

Theses

Workplace safety in the United States is governed by a system in which some states operate under federal OSHA oversight, while others administer their own OSHA-approved safety and health plans. This duality raises questions about whether state-level control leads to variation in safety and health outcomes. Economic conditions have also been shown to influence workplace injury rates, yet the interaction between safety and health regulations and local economies remains underexplored. Understanding these relationships is critical to protecting workers in occupational settings. The effects of OSHA state-plan status and state-level economic activity on occupational injury and illness rates across the United …

Deep Learning Approaches For Ctenophore Identification And Tracking, Anagha Bharadwaj

Theses

Ctenophores are translucent marine organisms with nearly invisible tentacles and pose significant challenges due to their transparent morphology and ambiguous structural features. This research addresses the classification and tracking of these organisms and evaluates the performance of current computer vision models under sparse-data environments.

A dataset from the NJIT Life History Lab consisting of microscopic laboratory videos and photographs of different growth stages is used to train and assess a number of convolutional neural network designs, including VGG16, ResNet, BioCLIP2, YOLO, and DeepLabCut. Additionally, a web-based interface is developed to evaluate expert-labeled ground truth with the model's performance.

The findings …

Polarimetric Terahertz Imaging For The Measurement Of Birefringence In Plastic, Rachel Cohen

Theses

Birefringence offers a promising way to observe stress concentration in materials such as glass and plastic, and thereby to identify weaknesses. Polarimetric imaging can be used to measure the birefringence of material, so long as the material is transparent to the light being used for the imaging. In this research, 2D Terahertz imaging was investigated as a means of measuring the birefringence of plastics that are opaque to visible light but transparent to THz radiation, for the eventual purpose of analyzing the residual stress present. In order to do so, two separate terahertz cameras were characterized for potential use in …

Autonomous Exploration Of An Environment With Static Obstacles Using A Ppo Agent, Brandon Knight

Theses

Research in autonomous exploration has created many effective algorithms that have been tested and proven to work in many different virtual and physical environments. Many optimizations have also been developed to reduce computational effort and increase exploration speed.

However, despite optimizations, these algorithms can still require considerable computational effort and time to explore even small environments. To obtain further improvements in computation and exploration speed, a reinforcement learning agent using actor-critic style proximal policy optimization (PPO) is trained to explore various environments efficiently, then compared to an algorithm using contemporary exploration methods.

Testing is performed in virtual environments with ideal …

Pypitfall: Dependency Chaos And Supply Chain Vulnerabilities In Python, Jacob Mahon

Theses

Python software development heavily relies on third-party packages. Direct and transitive dependencies create a labyrinth of software supply chains. While it is convenient to reuse code, vulnerabilities within these dependency chains can propagate through dependencies, potentially affecting downstream packages and applications. PyPI, the official Python package repository, hosts many packages and lacks a comprehensive analysis of the prevalence of vulnerable dependencies. PyPitfall, a quantitative analysis of vulnerable dependencies across the PyPI ecosystem, is introduced. The dependency metadata of 378,573 PyPI packages is analyzed. 4,655 packages that explicitly require a known vulnerable package version and 141,044 packages that permit a vulnerable …

Retinomorphic Mid-Wave Infrared In-Sensor Processing Engine: Device-To-Architecture Co-Design, Hemalatha Nagaraju

Theses

Conventional frame-based CMOS image sensors acquire full-frame pixel data at discrete time intervals, resulting in substantial spatial redundancy and loss of temporal information between frames. The repeated conversion and transfer of redundant pixel data increases bandwidth and power consumption in machine vision systems. Retinomorphic sensing architectures address these limitations by enabling programmable, analog-domain processing directly at the sensor interface. A compact behavioral model of the PbSe device is developed in HSPICE based on calibrated TCAD simulation data to capture gate-controlled photocurrent modulation under varying illumination and gate bias conditions. Error analysis is performed to quantify the deviation between TCAD-generated photocurrent …

Increasing Access To Medications For Opioid Use Disorder For All Justice-Associated Populations, Edgar Rodriguez

Doctor of Nursing Practice Final Manuscripts

The purpose of this paper is to support a federal mandate requiring universal access to medication of opioid use disorder for all the justice-involved populations. The policy proposal aims to standardize evidence-based treatment across correctional systems, improve treatment initiation during incarceration, strengthen continuity of care after release, reduce overdose mortality, and address disparities in access for racial and economic groups and rural communities. This policy paper calls for the increase of access to medications of opioid use disorder across the United States for all justice-involved population. The opioid crisis has created a public health problem in the United States by …

The Importance Of Socialization: Family, School, Peer, And Media Influence On White Identity Development, Miandria Erickson

Honors Projects

This study aims to explore the different aspects of socialization that can impact one’s White racial identity development, in order to analyze why White individuals struggle to describe their racial identity. Acknowledging an individual’s White identity is an imperative aspect to understanding diversity and race, and yet, White individuals will often have difficulty recognizing and explaining their racial identity. This qualitative study used data collected from semi-structured interviews of 60 young White college participants (ages 18-25) in two different geographical locations, the Northwest and Southeast of the U.S.^[1] Several themes were extracted from these interviews, among them were family, …

Self-Reported Gum Bleeding, Oral Health Behaviors, And Cardiovascular Risk Indicators Among Adults Aged 45-70 In A Community Dental Setting., Kimberleigh Edwards

Population Health Leadership and Analytics Theses

Cardiovascular disease remains a leading cause of morbidity and mortality among adults in the United States, particularly during midlife and older adulthood. Periodontal inflammation and poor oral hygiene behaviors have been linked to increased cardiovascular risk through systemic inflammatory pathways. This exploratory cross-sectional study described patterns of self reported gum bleeding, oral hygiene behaviors, and cardiovascular risk factors among 29 adults aged 45 to 70 at a community dental clinic using the researcher-developed Oral and Heart Health Study Questionnaire. Fourteen participants reported both gum bleeding and at least one cardiovascular risk factor. Participants who reported inconsistent flossing and infrequent professional …

Working Out Your Self-Image: Strength, Self-Esteem, And Human Flourishing, Kingston K. Christie

Honors Projects

This project integrates empirical research in exercise science with theological anthropology to examine the psychological effects of resistance training and the meaning of strength within a Christian framework of human dignity and flourishing. The empirical component investigated the effects of a four-week progressive resistance training intervention on global self-esteem and physical self-concept among undergraduate university students with moderate-to-low baseline self-esteem. Eleven participants (N = 11; 20.15 ± 1.52 yrs) completed the Rosenberg Self-Esteem Scale (RSES) and Physical Self-Description Questionnaire Short Form (PSDQ-S) before and after the intervention, with muscular strength assessed via submaximal testing across the squat, bench press, deadlift, …

Clinician Perspectives On Fistula Mental Health, Victoria K. Leonard

Doctoral Dissertations

Background – Obstetric fistula is a childbirth injury caused by prolonged labor that leads to stillbirth and incontinence, spurring social exclusion and isolation. These layers of trauma put women with fistula at great risk for psychological suffering, which has profound negative socioeconomic impacts on them, their families, and communities. This study captured treatment as usual at Comprehensive Community Based Rehabilitation in Tanzania (CCBRT), the country’s largest provider of fistula care.

Method – Improving holistic fistula treatment requires engaging the clinicians who care for women with fistula. This study aimed to investigate the training, beliefs, and treatment approaches of nurses and …

Navigating Trust: The Impact Of Social Media Influencers On Gen Z Health Behaviors, Abigail Stelling

Honors Thesis

Social media plays a defining role in the daily lives of Gen Z, influencing how people communicate, access information, and make decisions. More than a form of entertainment, social media is a powerful source that shapes identity, perceptions, and behaviors. As the first generation to grow up fully immersed in a digital world, Gen Z is uniquely impacted by the content and interactions that occur on social media platforms. Within this digital landscape, social media influencers have emerged as powerful figures. These individuals use large online platforms to promote lifestyles, products, and behaviors. Influencers often have the ability to shape …

Debatrix, Vivienne Lu, Huy Ngo, Luke Ponssen, Jonathan Preiss, Ryan Rani

Computer Science and Engineering Senior Theses

Developing public-speaking skills remains a persistent challenge in formal education, constrained by limited instructional time and the lack of scalable, individualized feedback. Existing automated tools address only narrow aspects of this problem, offering text-based coaching against rigid rubrics that fail to capture argument quality, evidence use, or real-time rebuttal skill. This thesis presents Debatrix, an AI-powered platform that enables K-12 students, university learners, and independent self-studiers to debate an intelligent opponent. The system combines automatic speech recognition, large language model-driven rebuttal generation, and a multi-dimensional rhetorical analysis engine that evaluates argument structure, evidence integration, and persuasive technique. Users receive explainable …