Digital Commons Network^™

Advancing Hate Speech Detection: Binary And Multiclass Approaches From Traditional Methods To Parameter-Efficient And Ontology-Guided Language Models, Mahmoud Abusaqer

Graduate Theses/Dissertations

The widespread proliferation of hate speech on social media platforms poses significant challenges for content moderation and user safety, requiring automated systems that are simultaneously accurate, efficient, and capable of fine-grained distinctions. This thesis investigates hate speech detection through five published manuscripts organized into two complementary threads: binary detection (hateful vs. non-hateful) and multiclass detection across demographic targeting categories. The binary thread progresses from a broad 38-model baseline spanning traditional machine learning, deep learning, and transformer architectures (where RoBERTa reaches 91.48% accuracy and CatBoost remains competitive at 88.60%) to parameter-efficient adaptation, in which Low-Rank Adaptation (LoRA) of large language models …

Foreword, Pedagogical Innovations In Computer Science Education, Helen Crompton

STEMPS Faculty Publications

[Introduction] Computer science education sits at a defining moment. Across schools and universities worldwide, computing is no longer a niche discipline reserved for a select few. It is a foundational literacy that shapes how learners understand the world, participate in society, and imagine their futures. At the same time, the rapid pace of technological change, particularly in artificial intelligence, data systems, and intelligent tools, has placed unprecedented pressure on educators to rethink not only what we teach, but why and how we teach it. This book arrives precisely when such reflection is most needed.

News Media Sentiment Toward Chinese Ai: A Comparative Analysis With Belt And Road Initiative Involvement And Public Opinion On China, Asya Vaisberg

Pomona Senior Theses

This study evaluates different countries' news media’s sentiment towards Chinese AI, between May 2023 and May 2024, by using Microsoft Azure NLP Sentiment Analysis. The results are then compared with the country’s public opinion on China and its involvement in the Belt and Road Initiative (BRI). For this study 12 countries have been selected which are USA, Australia, Pakistan, Peru, Russia, Romania, Italy, Greece, Portugal, Philippines, Brazil, and Egypt. For each one, GNews Application Programming Interface (API), which has access to more than 60,000 global news sources, was used to aggregate relevant news articles based on queried keywords. The collected …

An Llm-Driven System For Doctor-Patient Simulation, Akilan Amithasagaran

Computer Science Theses

Effective physician-patient communication is fundamental to clinical competence, yet traditional simulation-based training methods using standardized patients and high-fidelity manikins are costly, resource-intensive, and difficult to scale. This dissertation presents CLiVR (Conversational Learning system in Virtual Reality), an LLM-driven system that integrates large language models and 3D avatars to simulate doctor-patient interactions for medical communication training.

CLiVR addresses three key limitations in existing virtual reality medical training platforms. First, the system operates on standalone Meta Quest 3 hardware with realistic 3D patient avatars featuring synchronized lip movements and speech-based interaction. Second, CLiVR grounds LLM responses using a curated syndrome-symptom database, constraining …

Bail Reform, Large Language Model Risk And Reasoning, William Wyatt

CGU Theses & Dissertations

This dissertation contains three studies. Each asks how rules or language change the choices people and machines make when outcomes are uncertain. The first study, written with Kiran John, evaluates California’s 2020 cashless bail reform. We use propensity score matching on arrestee records from the windows before and after implementation, and we test whether the shift away from cash bail produced any effect on subsequent offending. It did not. Matched comparisons yield small, statistically insignificant differences across every window we examined. That null result cuts against both sides of the public argument. The reform did not drive a spike in …

The Role Of Education In Reducing Social Inequality: A Systems-Level Analysis Of Socio-Technical Infrastructures And Policy Governance, Aisling O'Shea, Batzorig Dashnyam, Ximena Quintanilla

Women's & Gender Studies Faculty Publications

Social inequality remains one of the most persistent challenges to global systemic stability, threatening the robustness of democratic institutions and economic sustainability. Education has long been theorized as the primary mechanism for social mobility and the mitigation of disparate life outcomes; however, its role within modern socio-technical infrastructures is increasingly complex and often contradictory. This paper provides a comprehensive systems-level analysis of the relationship between educational architecture and social stratification. By examining the structural trade-offs inherent in contemporary pedagogical deployment, the research evaluates how institutional governance, digital infrastructure, and policy mandates either facilitate or hinder the reduction of inequality. The …

Learning Design To Advance Human-Ai Collaboration In K-12 Education, Wing Sha Chan, Jinhee Kim, Seongryeong Yu, Rita Kay Detrick

STEMPS Faculty Publications

This chapter explores key components for designing effective Human-AI Collaboration (HAC) in K–12 education, addressing the current lack of theoretical and conceptual frameworks for structuring and implementing HAC in teaching and learning. It examines four essential areas: curriculum design, student and teacher–AI interaction, learning environments, and the evolution of HAC over time. The chapter introduces the concept of HAC in K–12 contexts, highlighting how humans and AI can leverage each other's strengths through co-evolutionary processes that foster mutual learning and collaboration. It reviews current HAC practices in schools and discusses their contributions to both teaching and learning. Finally, it presents …

Wearable Sensor-Based Phase Segmentation Analysis Of Front Crawl Swimming: A Scoping Review, Jonathan Simoes, Samuel Aylward, Daniel Hamze, Daniel James Goble, Daniel M. Russell, Joshua Haworth

Exercise Science Faculty Publications

Front crawl swimming stroke phase segmentation has historically relied on video analysis, but the development of wearable sensor technology has created new opportunities for automated phase segmentation. This scoping review mapped the available evidence on wearable sensor-based stroke phase segmentation methods in front crawl swimming, following PRISMA-ScR guidelines. A systematic search of SPORTDiscus, Web of Science, and IEEE Xplore conducted from January to June 2026, identified 15 eligible peer-reviewed studies published between 2000 and 2024. The review revealed an emerging field of research that has converged methodologically around inertial measurement units (IMUs) and the Chollet phase segmentation framework while remaining …

Tackling The Societal And Regulatory Challenges Of Emerging Technologies: A Case Study Of Deepfake, Jingyao Li

2026

Governing emerging technologies such as Artificial Intelligence (AI) poses enduring challenges for policymakers, industries, and societies. Early-stage governance is often hindered by limited understanding of technological implications, rapid innovation cycles, and resistance from powerful industry actors who favor minimal oversight. Yet, timely and effective governance is essential, as new technologies are most malleable in their formative stages. This dissertation examines how emerging technologies can be governed effectively by using deepfakes technology as a focal case. This dissertation comprises three interrelated studies.

The first paper reviews the literature on deepfakes and emerging technology governance, identifying the distinct characteristics of deepfake technology …

Do Developers Read Type Information? An Eye-Tracking Study On Typescript, Samuel W. Flint, Robert Dyer, Bonita Sharif

Research & Publications

Statically-annotated types have been shown to aid developers in a number of programming tasks, and this benefit holds true even when static type checking is not used. It is hypothesized that this is because developers use type annotations as in-code documentation. In this study, we aim to provide evidence that developers use type annotations as in-code documentation. Understanding this hypothesized use will help to understand how, and in what contexts, developers use type information; additionally, it may help to design better development tools and inform educational decisions. To provide this evidence, we conduct an eye tracking study with 26 undergraduate …

A Specification-Driven Framework For Self-Supervised Learning In Specialized Vision Domains, Mahmut S. Gokmen

Theses and Dissertations--Computer Science

Self-supervised learning (SSL) has emerged as a principled approach to visual representation learning that derives supervisory signal directly from unlabeled data, enabling foundation models to be trained at scale without manual annotation. Deployments in medical imaging and biometric recognition have demonstrated the potential of this paradigm, yet the assumptions that make SSL effective on natural image benchmarks fail systematically in specialized domains. Generic SSL pipelines encode a tacit assumption that the most informative correspondence is spatial proximity within a single acquisition. In specialized domains this assumption breaks at the level of the data-generating process: the signal that carries domain-specific information …

Counseling Professionals' Perspectives On Ai Integration In Education And Supervision: A Concept Mapping Study, Hank Crofford, Elif Bor, Gülşah Kemer

Counseling & Human Services Faculty Publications

The rapid emergence of artificial intelligence (AI) has raised important questions about how new technologies will shape professional norms and practices in counseling. The purpose of this study was to understand how counseling professionals expect AI to be integrated into counselor education and supervision (CES). Using a mixed‐methods concept mapping design, 31 participants generated and sorted statements about the potential roles, benefits, and concerns associated with AI in the profession. Participants represented diverse counseling roles, including counselor educators, licensed professional counselors, supervisors, master's‐ and doctoral‐level trainees, and other counseling‐related professionals. Standard concept mapping procedures were conducted using R, resulting in …

Evaluating Hybrid Quantum-Classical Models For Image Classification In The Nisq Era, Utku Binkanat

Theses and Dissertations

This thesis presents a systematic empirical evaluation of quantum machine learning performance under noisy intermediate-scale quantum (NISQ) era constraints. Through 670 controlled experiments, it evaluated quantum kernel support vector machines and variational quantum classifiers against classical baselines on MNIST binary and multiclass classification tasks with systematic variation of problem difficulty, feature dimensionality (4, 8 qubits), and training set size (n ∈{100, 250, 400, 500, 2000, 4000}). Statistical rigor was ensured through five random seeds per condition and comprehensive significance testing. During the testing with binary datasets, classical methods (SVM, logistic regression, k-NN, neural networks) achieved 85.9% to 99.6% accuracy with …

A Pipeline For Creating Obfuscated Binary Samples To Train Ai-Powered Detection Models, Luka R.H. Wilmink

Theses and Dissertations

The analysis of binary files is a critical component of antivirus software and is one of the most important tools for incident response teams across the industry. In the field, malware is often obfuscated, a practice in which the compilation process is transformed with different techniques to hinder decompilation and reverse engineering. Artificial Intelligence and Machine Learning techniques can assist, but models need to be trained on well constructed datasets first. This paper outlines a pipeline for creating such a dataset and builds a proof-of-concept machine learning classification model. All associated data and code are supplied in the project GitHub …

Understanding Phishing Susceptibility Through Expert Consensus Using Digital Marketing Parallels And A Machine Learning-Based Implementation, Mansoor Ahmad

All Graduate Theses, Dissertations, and Other Capstone Projects

Phishing remains one of the most effective attack vectors for gaining unauthorized access to organizational systems, yet defenders often lack systematic methods to assess their exposure before an attack. This study develops a framework that uses machine learning to encode the collective expertise of cybersecurity practitioners into a portable phishing susceptibility assessment tool, with the goal to help security teams proactively identify patterns, prioritize awareness training, and strengthen detection controls. The study surveyed 27 practitioners with extensive experience in social engineering, red teaming, penetration testing, and threat analysis to identify which factors most influence phishing susceptibility. Practitioners provided quantitative ratings …

Ai Vs. Genai: Combating Llm Generated Prescription Fraud With Transformer Based Detection Models, Ankitha Vokkaleri Shankarappa

Selected Full-Text Master Theses 2021-

The rapid evolution of Large Language Models (LLMs) has introduced a sophisticated new vector for healthcare fraud: the generation of high-fidelity, synthetic medical prescriptions. Traditional fraud detection systems, which rely on rule-based engines and basic statistical anomalies, are increasingly ill-equipped to identify these AI-generated forgeries that mimic the structural and clinical logic of authentic records. This thesis presents a robust detection framework using Transformer-based architectures to distinguish between human-authored Medicare Part D prescriptions and fully synthetic records generated by GPT-4.

The research was conducted across two distinct phases: an initial pilot study using 4,000 samples and a rigorous validation stress …

Erosion Of Trust In Online Information, Tirth Desai

A with Honors Projects

Researching how AI spreads misinformation and impacts trust in information.

A Comprehensive Survey On Facial Expression Generation: From Gans To Llm-Guided Multimodal Models, Murad Hasan, Rabab Abdelfattah, Kareem Abdelfatah, Mostafa Fouda, Ahmed Sherif

Faculty Publications

Facial expression generation (FEG) has emerged as a vital area in human–computer interaction, virtual avatars, and affective computing, aiming to synthesize natural and expressive facial behaviors across diverse interaction contexts. This survey presents a comprehensive analysis of recent advances in FEG, organized into six key paradigms: speech-driven expression generation, facial reaction generation, face video generation, facial animation, avatar-based generation, and text-driven expression generation. We review a wide range of model architectures, including VQ-VAEs, Generative Adversarial Networks (GANs), 3D Morphable Models (3DMMs), Transformers, and diffusion-based approaches, and compare their performance using commonly adopted evaluation metrics such as Fréchet Distance (FD), Peak …

Flood Risk Prediction System For Irish River Networks, Fernando Adalberto Naatz Heringer, Hugo Rodrigues De Freitas

ICT

Flooding costs the Irish economy more than 300 million euros annually; however, the hydrometric and meteorological data needed to predict dangerous river rises is already collected and freely available. This project applies machine learning to a decade of hourly observations from OPW Station 09001 on the River Liffey at Leixlip and Met Eireann Casement Aerodrome to build a flood risk prediction system with a 72-hour forecast horizon.

A two-phase modelling pipeline was built following the CRISP-DM framework. Phase 1 compared LSTM and GRU recurrent neural networks on the task of forecasting the river level in metres 72 hours ahead. LSTM …

Machine Learning System For Cereal Yield Prediction In South America., Victor Gabriel Oliveira, Kelvin Henrique Ferreira Dumas

ICT

This project developed a machine learning solution to predict cereal yield in South America using FAOSTAT agricultural data. The objective was to estimate yield in kg/ha for rice, maize, and wheat based on variables such as country, crop type, year, harvested area, producer price, pesticide use, and nutrient indicators.

The project followed the CRISP-DM methodology, covering business understanding, data understanding, data preparation, modelling, evaluation, and deployment. Several regression models were tested, including Random Forest, Gradient Boosting, AdaBoost, XGBoost, KNN, and SVR.

The models were evaluated using MAE, MSE, RMSE, and R² score. The results showed that tree-based ensemble models performed …

A Comparative Analysis Of Supervised Machine Learning Techniques For Predicting It Incident Resolution Time, Ammad Hussain, Hussnain Yaqoob

ICT

IT support teams are facing the persistent challenge in estimating how long an incident will take to resolve. The lack of reliable predictions, planning decisions around staffing, escalation and user communication are largely guesswork. This project set out to address that gap by building a machine learning solution capable of predicting incident resolution time from details available at the point a ticket is logged.

Using the UCI IT Incident Event Log dataset, which contains 141,712 records from a real service management system, the project followed the CRISP-DM framework across six phases from business understanding through to deployment. After thorough data …

A Comprehensive Study Of Sarima And Xgboost Models For Short-Term Traffic Volume Forecasting At A Melbourne Intersection (Victoria), Australia, Douglas Vinicius Dierings, Felipe Fontanive Marques

ICT

Traffic congestion is an increasing challenge in modern cities, impacting transportation efficiency and urban sustainability. This project focuses on short-term traffic volume forecasting, using 15-minute Sydney Coordinated Adaptive Traffic System (SCATS) data collected from the State of Victoria, Australia. The study compares two forecasting approaches: SARIMA, a statistical time-series model, and XGBoost, a decision-tree machine-learning model. Historical traffic data from 2022 to 2024 was analysed to identify traffic patterns, seasonal trends, and peak traffic periods. The models were evaluated for their effectiveness in handling urban traffic behaviour. The results of this study aim to support smarter traffic management, infrastructure planning, …

Characterizing Cyber Intrusions In Critical Infrastructure Networks Using Discrete-Event Simulation, Lawrence M. Dilworth

Dissertations, Master's Theses and Master's Reports

Over the past two decades, cybersecurity compliance frameworks such as the North American Electric Reliability Corporation Critical Infrastructure Protection (CIP) have introduced prescriptive measures for protecting power system networks, emphasizing restricted access, segmentation, and minimizing routable exposure. While effective for baseline cyber hygiene, these approaches do not capture system-level risks or adversarial propagation across interconnected infrastructure. In contrast, Cyber-Informed Engineering (CIE), advanced by Idaho National Laboratory, embeds security in system design by considering threat vectors and physical constraints.

Despite CIP guidance, many deployments rely on IP-routable, bidirectional communication that enables handshaking, allowing adversaries to infer reachable targets. This work presents …

The Excess Path Length Distribution: A Stochastic Model For Sample-Based Path Planners, Chaz B. Cornwall

Dissertations, Master's Theses and Master's Reports

Through random sampling, sample-based path planners enable autonomous agents to quickly find paths without human intervention. However, due to the paths' randomness, sample-based path planners currently require additional verification, partially nullifying agents' ability to act autonomously. I set out to characterize this uncertainty so humans know what to expect from these path planners and know how to alter the path planner to desired specifications. To ensure the results are theoretical as well as practical, I first create a stochastic model of path length uncertainty using the trade-off between sampling time and optimality. By leveraging this model, my proposed algorithm reduces …

Using Automotive Lidar To Reduce The Energy Consumption Of An Ego Autonomous Vehicle, Logan P. Schexnaydre

Dissertations, Master's Theses and Master's Reports

There is significant potential to reduce the energy consumption of the transportation sector through autonomous vehicles. Prior work on autonomous vehicle energy efficiency focuses on the whole system or the control subsystem. Yet, the sensing and processing components, which have direct and indirect effects on net energy use, are less explored. This dissertation fills this gap by modeling and evaluating these effects for lidar sensors, which provide high-resolution spatial data at the cost of high power and processing demands. I apply lidar to the energy-saving tasks of automated vehicle following and road surface profiling. For automated vehicle following, I model …

Real Time Waste Classification Using Deep Learning: Comparing Mobilenetv2 And Resnet 18 With Transfer Learning And Fine Tuning, Abdul Moaiz

ICT

Waste contamination is a major problem all over the world. The Environmental Protection Agency reports that over two thirds of waste found in general household and commercial bins could have been placed in the recycling or organic waste bins instead in Ireland, with food waste and plastics being the most common misplaced items. This project presents a deep learning solution to classify nine categories of waste from camera images in real time with the goal of helping users sort waste correctly at the source. Following the CRISP DM framework, two convolutional neural network architectures were trained and compared, MobilenetV2 and …

Bitcoin Prediction System Usind Machine Learning Techniques, Carolina Azevedo De Castro

ICT

This project investigated the use of machine learning techniques to predict short-term Bitcoin price direction using historical market data obtained from Yahoo Finance. Following the CRISP-DM methodology, the dataset was analysed, prepared, and transformed through feature engineering techniques including moving averages, volatility indicators, return measures and price position metrics. Multiple classification algorithms were evaluated, including Bayesian Classification, K-Nearest Neighbour, Decision Tree, Random Forest, Logistic Regression, Support Vector Machine and XGBoost. Several optimisation strategies were also tested, including feature selection, hyperparameter tuning, feature scaling and class weight balancing. Results showed that predicting short-term Bitcoin price movements remains challenging, with most models …

Beyond The Lace Index: Benchmarking Machine Learning Architectures And Explaining 30-Day Hospital Readmission Risk With Shap Analysis, Carl E. Hughes Iii

Williams Honors College, Honors Research Projects

Unplanned 30-day hospital readmission remains a fundamental challenge in US healthcare, associated with increased risk to patient recovery and representing an estimated $52.4 billion in annual expenses (Beauvais et al., 2022). While the rigorously validated LACE index serves as the clinical standard for readmission modeling, its linear structure and four explanatory variables lack the complexity to capture the high-dimensional and interactive nature of patient risk. This study utilizes an admission granularity level cohort of the MIMIC-IV database to develop and compare machine learning architectures against the baseline LACE index. Due to the imbalanced prevalence of readmission, the penalized logistic regression, …

Designing Ai Systems To Support A Productive-Failure-Based Learning: Insights From Adult Learners On Ai Applications And Ai System Design Principles, Jinhee Kim, Xi Lin, Seongryeong Yu, Rita Detrick

STEMPS Faculty Publications

Emerging capabilities of generative artificial intelligence (GenAI) offer significant potential to support productive failure (PF)-based learning, which engages adult learners (ALs) in exploring problems before instruction and learning from their initial attempts. However, the effective use of AI to support multifaceted areas of PF-based learning, including problem generation, exploration, consolidation, and knowledge assembly, is limited. Furthermore, AI design principles to support PF-based learning remain under-researched. This study, therefore, aims to investigate ALs’ perceptions of AI applications in enhancing PF-based learning and to explore the essential design principles of AI systems for PF-based learning. To achieve these aims, the study conducted …

Towards Breach Hypothesis Based Predictive Autonomous Cyber Defense Ecosystem: Proactive Prevention Of Imminent Threats And Productivity Losses, Yogesh Chavarkar

Master's Theses and Doctoral Dissertations

As cyberattack tools and techniques get sophisticated and persistent, reactive cybersecurity has been unable to effectively prevent breaches and compromises. Organizations and institutions with large complex environments have wide vulnerable exposure with higher chances of a cyberattack. This also increases overall security and financial risk. Possibility of repetitive attacks from cyber threats increases, too, despite the use of standard defense measures. Security tools and system vulnerabilities often need manual patching and updates to keep the environment secure and functioning effectively. Productivity suffers from manual trade-offs in keeping the systems secure from adverse impact. Persistent high-severity cyberattacks, despite continued reactive defensive …