Digital Commons Network^™

A Retrieval Augmented Approach To Improving Accuracy Of Biomedical Term Normalization By Large Language Models, Thanh Son Do

Graduate Theses/Dissertations

Ontology normalization is crucial in biomedical text processing, as it enables the mapping of medical expressions to standardized ontology terms and their corresponding identifiers. This thesis explores the feasibility of using large language models (LLMs) for ontology normalization, with a specific focus on the Human Phenotype Ontology and Gene Ontology. Prior research studies indicated that LLMs employing zero-shot learning tend to exhibit low accuracy and are prone to frequent hallucinations. We propose a retrieval augmented generation (RAG) approach to address these limitations and enhance normalization accuracy. We generated synthetic test sets of ontology-derived synonyms to evaluate normalization performance and developed …

Sparse Transformer For Anomaly Detection In Mobile Crowdsensing, Sanjeev Shrestha

Graduate Theses/Dissertations

Mobile Crowdsensing (MCS) is a sensing paradigm that leverages mobile devices to conduct a large-scale data collection. However, due to its openness and mobility nature, it is highly vulnerable to security issues such as injection attacks of malicious workers and fake tasks that can severely affect the platform’s normal functioning. To address this problem, the arrival of workers and task submission process is represented as a multivariate time series, and a two-stage framework is proposed. In the first step, we propose a novel transformer-based model, DozerAnomaly, that can efficiently detect anomalies in multivariate time series. We integrated a sparse attention …

Balancing Performance And Efficiency: An Autoencoder Approach To Api-Based Malware Detection, Selma Bouraoui

Graduate Theses/Dissertations

The constant evolution of malware presents a critical challenge to today's interconnected world. It poses an increasing threat on different scales, spanning from individuals, organizations to critical infrastructures such as government’s security. Hackers continuously develop new techniques to evade detection methods. When confronted with the high volume and variety of malware, conventional approaches tend to struggle to perform in robust, accurate and timely manner. This thesis explores the application of deep learning methods to improve malware detection and classification techniques. By analyzing API call sequences, the proposed approach leverages Autoencoders to compress high-dimensional malware data into more optimized representations that …

Clinicians In The Loop Of Medical Ai, W. Nicholson Price Ii

Articles

As medical AI begins to mature as a health-care tool, the task of governance grows increasingly important. Ensuring that medical AI works, works where it’s used, and works for the patient in the moment is a challenging, multifaceted task. Some of this governance can be centralized—in review by FDA or by national accreditation labs, for instance. Some must be local, performed by the hospital or health system about to use the product in their own, unique environment. But a large amount of governance is left to the individual provider in the room, the human in the loop who presumably knows …

Boden Lecture: Taxation Of Autonomous Artificial Intelligence, Reuven S. Avi-Yonah, Lucas Brasil Salama, Herbert Snitz, W. Robert Thomas

Articles

This Article proposes that tax can be a useful supplement to other measures to regulate Autonomous Artificial Intelligence (AAI) and limit its potential harmful effects. This proposal differs from command-and-control regulation of AAI along the lines of European Union legislation that may unduly limit the development of AAI. It also differs from existing proposals to tax AAI to generate revenue to help workers displaced by AAI programs, or to tax the data used by AAI The proposal is based on granting AAI programs like ChatGPT separate legal personhood, like corporate personhood, while incentivizing or requiring their corporate owner to place …

Multimodal Deception Detection Via Audio-Text Fusion With Deep Learning And Asr, Xiangyi Li

Master's Projects

Emotion detection plays a crucial role in human-computer interaction, enabling machines to recognize and respond appropriately to human emotional states. This project explores a two-stage approach to emotion detection using multimodal data, first predicting dimensional values (Arousal, Valence, Dominance) from textual and audio inputs, then mapping these representations to discrete emotion categories. We compare this approach with direct categorical classification using transformer-based language models like BERT, RoBERTa, and DeBERTa for text processing, alongside various audio feature extraction methods, including MFCCs and spectrograms. Using the IEMOCAP dataset, we evaluate both approaches across text-only, audio-only, and multimodal configurations. Our findings reveal that …

Tedvil: Leveraging Transformer-Based Embeddings For Vulnerability Detection In Lifted Code, Gary Mccully, John Hastings, Shengjie Xu

Research & Publications

Ransomware and other malware inflict devastating financial and operational damage on organizations worldwide by exploiting deeply embedded, hard-to-detect vulnerabilities in their systems. Detecting these vulnerabilities in compiled code before malicious actors exploit them remains a critical challenge in cybersecurity. This research introduces TEDVIL (Transformer-based Embeddings for Discovering Vulnerabilities in Lifted Code), a novel framework which uses transformer-based embeddings to train neural networks to detect vulnerabilities in lifted code. The framework was implemented using bidirectional (BERT and RoBERTa) and unidirectional (GPT-1 and GPT-2) transformer-based models to generate embeddings for training Long Short-Term Memory (LSTM) neural networks to detect stack-based buffer overflows …

Predicting Human Steering From Optic Flow Using Deep Convolutional Neural Networks, Katie E. Bernard

Honors Theses

Human driving is a complex visuomotor task and the specific visual clues that guide it remain under investigation. While prior research has emphasized gaze-based strategies such as the Tangent Point and Future Path hypotheses, recent evidence highlights the potential role of optic flow, the visual motion pattern perceived during self-movement, as critical to steering ability. This thesis explores whether raw optic flow alone can support accurate predictions of human steering behavior. We trained a convolutional neural network to map optic flow vector fields to steering angles in a virtual reality driving simulation. The dataset, collected by Giguere et al., included …

Optimizing Radial Interfaces For Eye-Movement Authentication On Smartphones, Trey V. Tuscai

Honors Theses

Radial authentication interfaces offer privacy-preserving, calibration-free eye-movement authentication. While their effectiveness has been demonstrated on large displays, their performance on smartphones remains underexplored. This study investigates seven radial interface configurations on the iPhone 13, varying the number of radial indicators and password lengths to examine trade-offs between accuracy, security, and entry time. Through a controlled eye-tracking experiment with 27 participants, we evaluate each configuration’s performance and collect user prioritizations of the three factors. Our findings reveal that shorter passwords with fewer indicators improve speed and accuracy but reduce security, while longer configurations enhance security at the cost of usability. Based …

The Future Of Code Style: Learning With Gamified Online Tools, Jacob C. Tjaden

Honors Theses

High-quality code is universally pursued by software developers, and one of the most effective indicators of code quality is code style. However, code style is difficult to teach, particularly to introductory students and programmers who benefit most. In this project, we aim to investigate how online tools can improve and teach Python code style, as well as identify the role of gamification in the process. We build an online platform called Fishy that combines code style appraisal tools and utilizes gamification concepts. Our platform incorporates educational metrics such as a code analysis score and targeted quizzes to assess user performance. …

Applying Machine Learning Methods To Generate Understandings Of Differential Item Functioning In A Flu Knowledge Assessment, William L. Romine, Tanvi Banerjee, Derrick Cox

Computer Science and Engineering Faculty Publications

Current influenza trends, including the severity of the 2025 flu season and the prevalence of H5 bird flu in livestock, necessitate efforts to better understand how to educate students about its transmission. Although validated assessments of influenza knowledge exist, these have not been evaluated for affective and demographic biases. We explore differential item functioning (DIF) effects in four items focused on specific aspects of flu transmission derived from a validated influenza knowledge assessment. In doing so, we introduce and utilize a machine learning framework for exploration of DIF which offers greater flexibility than traditional statistical approaches in terms of studying …

Extraction Of Patient Subtypes Using Llm Generated Knowledge Graphs Integrated With A Transformer Architecture, Benjamin Holmes, Cogan Shimizu

Computer Science and Engineering Faculty Publications

Extracting patient subpopulations (clinically relevant cohorts of individuals who share overlapping symptoms, risk factors, or diagnostic criteria) from unstructured medical notes is an ongoing challenge due to the variability of clinical language and the complex nature of patient conditions. We demonstrate a pipeline that combines named entity recognition (NER), transformer embeddings, guided dimensionality reduction, and LLM-mediated knowledge graph integration to enhance patient extraction. The approach begins with NER using the UMLS metathesaurus [1] to extract clinical terms, followed by transformation into vector embeddings using a biomedical transformer. These embeddings are augmented with structured knowledge graph representations generated through an LLM-driven …

Ontology Population Using Llms, Sanaz Saki Norouzi, Adrita Barua, Antrea Christou, Nikita Gautam, Andrew Eells, Pascal Hitzler, Cogan Shimizu

Computer Science and Engineering Faculty Publications

No abstract provided.

Bytes, Banter, And The Bible: An Interdisciplinary Account Of Objective Meaning, Cameron Bonin

Senior Honors Theses

The claim that the Bible has objective meaning is contested in a postmodern world. This claim can be more persuasively defended when it is addressed by insights from multiple disciplines. In particular, the field of computer science is apt to illuminate the concept of meaning through its reflection on the nature of languages and its concern with the accurate transmission of information. By synthesizing insights from the field of computer science, such as that of Claude Shannon, with Nicholas Wolterstorff’s use of speech-act theory, the concept of meaning can be understood more clearly. Consequently, this synthesis assists in answering questions …

Estimating The Gender Wage Gap: A Comparative Analysis Of Different Estimators, Xinran Zhang

Mathematics, Statistics, and Computer Science Honors Projects

The gender wage gap between males and females has been well studied by labor economists. We take a multi-prong approach to evaluate three estimators —a regression-imputation estimator, a weighting estimator, and a doubly robust estimator—in estimating the gender wage gap. Using the Panel Study of Income Dynamics, we conduct an empirical study of the estimators’ performances. In a simulation study, we evaluate the properties of estimators and study whether bootstrapping is an appropriate measure of the uncertainty of each estimator. The findings show that while the estimators provide different results, the doubly robust estimator provides reliable and consistent results under …

Fostering Critically Conscious Lesson Planning In A Generative Artificial Intelligence Era, Derek Riddle, Paula Cristina Azevedo, Catharyn Shelton, Jaime Colwell, Jori Beck

Teaching & Learning Faculty Publications

Teacher candidates (TCs) use digital resources and social media to plan and develop learning material, and with publically accessible generative artificial intelligence (GAI), TCs are able to generate lesson plans within seconds rather than hours or days. While there is research on how to support TCs' evaluation of reliable digital media, there is little known on how to prepare TCs for GAI content. Using the complementary frameworks of Freire’s (1970) critical consciousness and Jonnasen’s (1991) theory of constructivism, this in-progress design based research seeks to develop an adaptable framework that addresses the evolving nature of technology, specifically GAI, and the …

Digital Infrastructure Development Through Digital Infrastructuring Work: An Institutional Work Perspective, Adrian Yeow, Wee-Kiat Lim, Samer Faraj

CCX Research

Being able to understand and characterize the digital infrastructure development (DID) process has become even more pressing today due to the rapid advent and implementation of new digital infrastructure (DI) in organizations as well as since the COVID-19 crisis. While information systems (IS) research has begun to recognize the institutional nature of such digital infrastructures, there remains a gap in our understanding of how such developments unfold from an institutional perspective. Through our field study of a digital infrastructure development project involving the implementation of an enterprise-wide electronic medical record system at a large US medical facility, we show how …

Ripples: An Automated Embedding Generation Algorithm For The Forward-Forward Algorithm, Spencer Connolly

All Graduate Theses, Dissertations, and Other Capstone Projects

The Forward-Forward algorithm (FF) is yet another novel invention by Geoffrey Hinton, the creator of the famous backpropagation algorithm (BP). Since its proposal, many papers have been published exploring its potential, and good progress has been made in increasing its viability. Though FF continually falls short of BP, its purpose is not to replace BP and preliminary research shows that there is plenty of room for growth. In this paper, we present a literature review for FF algorithms applied to Convolution Neural Networks (CNN) for image classification tasks and set the stage for applying FF to more complex datasets. The …

A Hybrid Deep Learning Model For Iot Network Anomaly Detection, Yonas Getachew Mulissa

CCAC Theses and Dissertations

The rapid expansion of Internet of Things (IoT) networks has heightened the need for intelligent, automated Anomaly Detection (AD) systems to identify sophisticated and evolving cyber threats. This study designed, implemented, and evaluated a broad range of deep learning models—including Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNNs) (Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional Gated Recurrent Unit (BiGRU), Bidirectional Long Short-Term Memory (BiLSTM)), Transformer-based architectures, Autoencoders, and hybrid combinations—to address the challenge of multiclass anomaly classification in IoT traffic. Using two benchmark datasets, IoT-DS-2 and CIC-IoT-2023, we conducted extensive experiments to assess classification performance, training efficiency, and …

Development Of A Phishing Risk Exposure Taxonomy On Mobile Devices In The Healthcare Industry, Christopher P. Collins

CCAC Theses and Dissertations

Phishing emails accessed on mobile devices present substantial risks to healthcare organizations when employees often operate under high cognitive load and with limited cybersecurity training. Despite widespread security awareness initiatives, healthcare workers continue to engage with phishing content on mobile platforms, posing threats to organizational data. Given the high value of healthcare data and the increasing sophistication of phishing schemes targeting healthcare professionals, there is a pressing need to enhance their ability to recognize phishing indicators on mobile devices.

This study developed and validated a Healthcare Workers Phishing Risk Exposure (HWPRE) taxonomy, designed to classify healthcare workers based on their …

Architectural Technical Debt Migrating Object Oriented Systems To Modular Architectures, Lionel Standridge

CCAC Theses and Dissertations

Architectural Technical Debt (ATD), a subset of Technical Debt (TD), arises when outdated architectural decisions present significant challenges to the maintenance and evolution of legacy object-oriented monolithic systems. These systems tend to have tightly coupled components and rigid dependencies, making it difficult to scale, adapt, and modernize. This dissertation investigated strategies for managing ATD during the transition from monolithic architectures to modular systems. By identifying the root causes of ATD in a legacy objectoriented system and evaluating various decomposition strategies, this research proposed a framework to guide practitioners in reducing ATD and improving a system’s modularity. Using quantitative metrics, the …

Enhanced Network Anomaly Detection Using Machine Learning Models, Ousmane Barry

CCAC Theses and Dissertations

This dissertation investigates enhanced network anomaly detection using Machine Learning (ML) models. The study addresses two distinct classification problems: binary classification and multiclass classification. In the binary classification task, network traffic data is categorized as either "normal" or "abnormal," where abnormal includes all non-normal traffic. Leveraging the balanced nature of the dataset, this study develops optimized models that achieve consistently high classification performance. Key metrics, including precision, recall, and F1 scores, are used to ensure robust evaluation and reliable detection across all classes.

For multiclass classification, only classes present in both training and test datasets are included to ensure meaningful …

Empirical Analysis Of Political Districting Splitability Via Uniform Spanning Trees In Polynomial Time, Brooke C. Feinberg

Scripps Senior Theses

This work expands a recently proven conjecture that a polynomial fraction of all uniform spanning trees (USTs) are splittable into k balanced partitions on grid graphs to real-world political districting plans. We investigate whether similar structural properties hold for the planar dual graphs of U.S. counties (cnty) and tracts (t), using Wilson’s algorithm to generate uniform random spanning trees and Breadth- First Search (BFS) to check for splitability into balanced partitions. Our empirical findings suggest that real-world districting plans can be split into 2-balanced, connected partitions in a fraction of polynomial time. This result highlights the potential for scalable redistricting …

A Vision Transformer Based Assistive System For Dermatological Diagnosis In Systemic Lupus Erythematosus, Syeda Lamima Farhat

All Graduate Theses, Dissertations, and Other Capstone Projects

Systemic Lupus Erythematosus (SLE) is a complex and often underdiagnosed autoimmune disease that affects multiple organs and presents with a wide range of symptoms-ranging from fatigue and joint pain to life-threatening organ damage. One of its most visible and diagnostically significant indicators is the Butterfly Malar Rash (BMR), a distinctive facial rash that often resembles other common dermatological conditions like rosacea, acne, eczema, and fifth disease. This overlap can lead to misdiagnosis or delayed detection, especially in busy clinical environments. To assist dermatologists in distinguishing BMR from similar facial rashes, this study explores the development of an AI-powered image classification …

Evaluating Aspect-Based Sentiment Analysis In Healthcare Drug Reviews Across Machine Learning, Deep Neural Networks, And Transformer Models, Eun Soo Park

All Graduate Theses, Dissertations, and Other Capstone Projects

Sentiment analysis has become a critical area of research in Natural Language Processing (NLP), enabling insights from unstructured text. Within this field, Aspect-Based Sentiment Analysis (ABSA) plays a practical role in domains such as healthcare, where patients drug reviews often contain diverse opinions across multiple aspects, including overall comments, perceived benefits, and side effects. However, aspect-level classification remains challenging due to class imbalance, subtle sentiment expression, and the limitations of traditional models. This research investigates the performance of three modeling paradigms: traditional machine learning (SVM, SVC, and XGBoost), deep learning (CNN-BiLSTM), and transformer-based approaches (DistilBERT sentence-pair classification). Using the UCI …

Identifying All Matches Of A Rigid Object In An Input Image Using Visible Triangles, Abdullah N. Arslan

Faculty Publications

It has been suggested that for objects identifiable by their corners, every triangle formed by these corner points can serve as a reference for detecting other corner points. This approach enables effective rigid object detection, including partial matches. However, when there are many corner points, the implementation becomes impractical due to excessive memory requirements. To overcome this, we propose a new algorithm that leverages Delaunay triangulation, considering only the triangles generated by the Delaunay triangulation to reduce the complexity of the original approach. Our algorithm is significantly faster and requires significantly less memory, offering a viable solution for large problem …

Predicting Lung Cancer Severity Using Machine Learning Algorithms: Enhanced By Statistical Analysis, Esin Bilgin

Theses, Dissertations and Culminating Projects

Cancer is a serious and severe cause seen in every region of the world and severely affects the quality of life and life span. Among the various types of cancer, lung cancer is one of the most critical, having a fatal impact on life. While medical imaging techniques, laboratory results, and biomarkers play a significant role in diagnosis and prognosis, clinical studies are also crucial in monitoring the progression of cancer and identifying diagnostic and prognostic factors. The findings demonstrate satisfactory accuracy, and the analysis incorporates statistical data with machine learning techniques. These findings play a pivotal role in supporting …

Brain Tumor Classification Using Deep Learning: Custom Cnn Vs. Resnet50, Rayen Bentemessek

ICT

The project presents a deep learning solution to classify brain tumors through MRI images. Following the CRISP-DM framework, two Convolutional Neural Network (CNN) models were developed and evaluated, a custom CNN designed from scratch and a pretrained ResNet50 that was transfer learned and fine-tuned. Both models were assessed using standard performance metrics such as accuracy, precision, recall and F1-score. Despite the higher test accuracy achieved by the custom CNN, further interpretability indicated inconsistent attention to the actual tumor regions also known as shortcut learning. On the other hand, ResNet50 showed more reliable and clinically relevant focus which supported its selection …

Gramseq-Dta: A Grammar-Based Drug-Target Affinity Prediction Approach Fusing Gene Expression Information, Kasul Debnath, Pratip Rana, Preetam Ghosh

Computer Science Faculty Publications

Drug–target affinity (DTA) prediction is a critical aspect of drug discovery. The meaningful representation of drugs and targets is crucial for accurate prediction. Using 1D string-based representations for drugs and targets is a common approach that has demonstrated good results in drug–target affinity prediction. However, these approach lacks information on the relative position of the atoms and bonds. To address this limitation, graph-based representations have been used to some extent. However, solely considering the structural aspect of drugs and targets may be insufficient for accurate DTA prediction. Integrating the functional aspect of these drugs at the genetic level can enhance …

A Personal Interview With William Patry: His Thoughts On Music, Ai, And Copyright

IP Theory

No abstract provided.