Computer Engineering Commons^™

Adversarial Robustness In Advanced Machine Learning Models Integrating Graph Neural Networks And Large Language Models, Mahmoud Nazzal

Dissertations

Artificial intelligence (AI) has achieved remarkable performances across various domains. In most real-world applications, data often takes relational forms, such as graphs and networks, or sequential forms, such as text and time series. As AI evolves, specialized models have emerged to handle these structures; Graph Neural Networks (GNNs) for relational mining and Large Language Models (LLMs) for sequential understanding. Despite their success, these models face challenges in security, robustness, and interpretability. GNNs excel in relational reasoning but are vulnerable to adversarial manipulation and lack interpretability, while LLMs are strong in linguistic reasoning and generalization yet struggle with relational data and …

Selected Artificial Intelligence Provisions In U.S. Fiscal Year 2025 National Defense Authorization Act, Bert Chapman

Libraries Faculty and Staff Presentations

The 2025 Fiscal Year National Defense Authorization Act contains multiple provisions relating to artificial intelligence (AI). These congressionally mandated provisions direct various sections of the Department of Defense (DOD) and individual U.S. armed service branches to execute congressional intent for AI policymaking. Examples of such intent include identifying and planning DOD's AI workforce, demonstrating AI biotechnology applications for national security, improving the human usability of AI systems, and establishing an AI security center. This presentation will note that reports on these initiatives must be prepared for relevant congressional oversight committees, and, in many cases, are in many cases, publicly released …

Measuring And Improving The Efficiency Of Python Code Generated By Llms Using Cot Prompting And Fine-Tuning, Ramya Jonnala

Masters Theses (Archived)

With the advanced AI technologies, the role of Large Language Models (LLMs) has grown rapidly for software development with generating the code that is functionally correct, solving complex problems, and debugging existing code. However, LLMs often produce inefficient code with unnecessary logic, hallucinated content, and errors. This research measures the efficiency of Python code generated by GPT-4o-Mini, GPT-3.5-Turbo, and GPT-4-Turbo models using execution time, memory usage, and maximum memory usage while maintaining correctness. Using EffiBench datasets on Google’s Vertex AI Workbench with different machine configurations, the study uses the seed parameter for consistency and optimization techniques like Chain-of-Thought (CoT) prompting …

Meshlet Rendering Using D3d12 Mesh Shading Pipeline, Anishva Bardhan

Programming Theses and Dissertations

Modern video games must render scenes with increasingly complex geometry. Technologies like Nanite in Unreal Engine 5 enable the handling of scenes with significantly higher object and triangle counts than ever before. This project draws inspiration from Nanite by operating on triangle clusters, allowing artists to focus solely on creating high-poly meshes. The primary objective is to implement fine-grained culling techniques on meshlets, combined with efficient meshlet instancing, to reduce render time and memory usage.

Meshlet instancing plays a crucial role in optimizing rendering performance by allowing multiple objects sharing the same geometry to be rendered efficiently. Instead of duplicating …

3d Multi-Threaded Ai Navigation With Pathfinding And Obstacle Avoidance, Jabari Belgrave

Programming Theses and Dissertations

In this thesis, I developed a 3D multi-threaded AI navigation system using my own custom-built C++ game engine. The system combines triangle-based A* pathfinding with real-time obstacle avoidance using a set of velocity-obstacle algorithms. It is designed to support large numbers of agents navigating complex environments while avoiding collisions. I created two main simulation modes: Navigation Mode, which integrates A* with ORCA to handle large-scale pathfinding and movement, and Obstacle Avoidance Mode, which allows direct comparison between VO, RVO, HRVO, and ORCA in a controlled test setting.

The terrain is procedurally generated using Perlin noise, and this terrain data is …

Algorithms & Design Behind Autonomous Uavs And Ugvs Coordinated System, Aashish Dhakal

Honors Theses

Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicles (UGVs), when coordinated effectively, offer substantial potential for automating large-scale tasks—from search and rescue operations to precision agriculture. However, synchronizing these autonomous systems remains challenging, especially in time-sensitive missions requiring precision. This thesis investigates the design and algorithmic coordination of autonomous UAVs and UGVs, examining both single-vehicle scenarios and multi-agent (swarming) approaches. Using the Robot Operating System (ROS) as a communication backbone, I integrate GPS positioning with computer vision techniques through OpenCV, enabling accurate localization and object detection. During the development phase, I validate my methods using ArduPilot Software-in-the-Loop (SITL) simulations within …

Benefits And Applications Of Learning With Virtual Reality, Michael W. Timm

Honors Program: Senior Projects (Public)

Education is a fundamental pillar of society. It equips students for employment and interpersonal relations. Virtual reality (VR) has emerged as a transformative technology in the field of education. The aim of this paper is to synthesize existing research in order to determine the benefits of utilizing virtual reality in a variety of education settings, such as K-12 classrooms, universities, and workplace training. This paper observes significant benefits of virtual reality in constructivist and experiential learning, gamified learning, and tailored practice. This analysis also finds that virtual reality is advantageous for educational accessibility, particularly for absentee students and impoverished students. …

Adaptive Noise Estimation And Denoising With Deep Learning For Nmr Spectroscopy, Naveen Asokan

McKelvey School of Engineering Graduate Student Theses & Dissertations

Nuclear Magnetic Resonance (NMR) spectroscopy is a powerful analytical technique widely used for molecular structure elucidation in chemistry, biology, and medicine. However, spectral accuracy is often degraded by noise—particularly in low acquisition time settings—resulting in reduced resolution and obscured chemical features. While traditional noise reduction techniques such as signal averaging can improve spectral quality, they require longer acquisition times, limiting their utility in real-time and high-throughput applications.

This thesis presents a deep learning-based denoising framework designed to enhance the quality of complex-valued NMR spectra. The proposed model, built upon a U-Net architecture, incorporates both real and imaginary components of the …

Adaptive Oversight In Action: Proposing Context-Aware Governance For Ai In Cybersecurity, Russell S. Cunningham

Boise State Graduate Student Projects

Artificial intelligence is increasingly woven into cybersecurity operations, shaping everything from threat detection to automated incident response. While these technologies improve speed and scalability, they also raise urgent questions about governance, ethical use, and operational risk. Although frameworks such as the NIST AI Risk Management Framework, ISO/IEC 42001, the EU AI Act, IEEE's Ethically Aligned Design, and OECD AI Principles each offer structure, they tend to address only parts of the problem. Most focus on compliance or ethics but rarely both, and few are tailored to the high-pressure, risk-sensitive environments found in cybersecurity.

To bridge these gaps, this research proposes …

Gpu-Based Visual Effects System, Matthew Jaffe

Programming Theses and Dissertations

The objective of my thesis is to create a robust and efficient VFX system that can be used to edit and add particle effects to games. This system utilizes a compute shading pipeline to simulate millions of particles in real time. The behavior of particles is widely customizable through many different properties which can be manipulated changed over the lifetime of particles and introduce procedural randomness. There are many ways to customize the motion of the particles with various forces and collision. Additionally, particles can be rendered as billboarded quads, full meshes or partial meshes with different settings to further …

Modality Distillation Using A Sam-Guided Multimodal Teacher For Unimodal Wildfire Segmentation And Temperature Prediction, Michael N. Marinaccio

All Theses

Wildfires are one of the world’s most devastating natural disasters that affect the environment, communities, and more critically, humans that live in and around those communities. Due to the threat of large-scale destruction in landscapes and human inhabited areas, it has become increasingly more important to develop wildfire detection, management, and suppression strategies to mitigate and prevent these negative outcomes. Wildfire research encompasses many different areas. Most notably, the development of communication, navigation, remote sensing, and monitoring systems. In wildfire monitoring, limitations discovered in-ground and satellite observation have shifted the focus toward Unmanned Aerial Vehicle (UAV) based wildfire research, which …

From In-The-Head To In-The-World: Frameworks For Understanding And Applying Computational Thinking, Justin Olmanson, Gretchen K. Larsen, Azadeh Hassani

Department of Teaching, Learning, and Teacher Education: Faculty Publications

In the five decades since Papert coined the term Computational Thinking (CT), it has become a core framework for thinking about learning, problem-solving, design, and creativity. Although CT is most commonly, and initially, associated with the cognitive orientations involved in coding and learning to code, it also includes situated and critical processes related to computational problem solving. Herein we unpack ways researchers in different fields and points in time have organized CT. We include creative coding as a uniquely generative lens for rethinking CT, outlining its potential as an expressive, constructionist, and culturally situated practice. In doing so, we explore …

Ai-Powered Inspection: A Computer Vision System For Efficient Defects Detection In Underground Infrastructures, Rasha Alshawi

LSU New Orleans Theses and Dissertations

Undetected defects in culverts and sewer pipes pose significant risks to public safety, leading to infrastructure collapses, flooding, and transportation disruptions. Traditional manual inspections are time-consuming, costly, and prone to human error, while existing automated methods struggle with occlusions, irregular defect shapes, class imbalances, and high computational demands. To address these challenges, this dissertation develops advanced semantic segmentation systems that automate defect detection, significantly improving efficiency and accuracy.

This research introduces a series of innovative models designed to overcome these challenges in underground infrastructure inspection. Using dual-attentive mechanisms, sparsely connected blocks, and depth-separable convolutions, these models improve segmentation performance and …

Car Damage Detection Using Deep Learning, Rahul Varma Indukuri Sr.

Electronic Theses, Projects, and Dissertations

Growing vehicle usage has resulted in a notable increase in road accidents, so it is imperative to have effective systems for identifying and evaluating vehicle damage. This work aims to create a computer vision and deep learning-based automated car damage detection system. This project's main goal is to develop a model that, using visual cues, can categorize car photos as either damaged or undamaged.

The algorithm operates in two steps: first, determining whether the picture features an automobile; then, it classifies the state of the car—damaged or undamaged. We thus employ the InceptionV3 model for damage classification and the MobileNet …

Cart To Doorstep, Giridhar Yadav Nazarapur

Electronic Theses, Projects, and Dissertations

Due to hectic schedules and long working hours, many people find it challenging to allocate time for traditional shopping in today's fast-paced world. This often leads to a preference for online shopping, as it provides convenience and flexibility. However, despite its benefits, online shopping still lacks a seamless experience where customers can easily access a wide range of products, manage their shopping preferences, and make hassle-free payments—all from the comfort of their homes. To address this, the "Cart to Doorstep" project was developed. This web application aims to create an efficient and user-friendly online shopping platform that streamlines the shopping …

Urban-Rural Dynamics And Dui Fatalities In The Inland Empire: A Neural Network Analysis Of Traffic Safety Disparities, Armando Ceja-Lua

Electronic Theses, Projects, and Dissertations

This study examines the disproportionately high traffic fatality rates in California's Inland Empire region through neural network analysis of over 500,000 accidents (2013-2022). We argue that the Inland Empire's unique hybrid urban-rural landscape creates a multiplicative risk environment unlike other California regions. Our analysis reveals that San Bernardino County's fatality rate (1.920 per 100 million VMT) significantly exceeds neighboring regions, with alcohol-impaired driving fatalities (0.586) substantially higher than California's average (0.390). Neural network models (92% validation accuracy) identify pedestrian-involved collisions (correlation value 0.164) and alcohol involvement (0.075) as the strongest predictors of fatality in urban areas, while rural crash patterns …

Privacy Implications Of Data Collection In Android Automotive Os, Bulut Gözübüyük

All Theses

Modern vehicles have become sophisticated computational and sensor systems, as evidenced by advanced driver assistance systems (ADAS), in-car infotainment, and autonomous driving capabilities. They collect and process vast amounts of data through various onboard subsystems. One significant player in this landscape is Android Automotive OS (AAOS), which has been integrated into over 100 million vehicles and has become a dominant force in the in-vehicle infotainment (IVI) market. With this extensive data collection, privacy concerns have become increasingly crucial. The volume of data gathered by these systems raises questions about how this information is stored, used, and protected, making privacy a …

A System And Method For Measuring Spatially Varying Surface Appearances With A Study Of Feathers, Jessica Baron-Lis

All Dissertations

Real-world materials, particularly biological structures such as feathers exhibit complex appearances that vary spatially across their surfaces. The field of computer graphics provides a means of understanding such surfaces through material modeling which uses both analytical models and data acquired from light-surface interactions. There are many efforts within the past decade in measuring materials for graphics, but common limitations in these works include not accounting for spatially varying properties and reliance on neural networks and synthetic datasets.

Feathers from modern birds present diverse appearances due to how light interacts with their unique hierarchical microstructures. Variations in those structures lead to …

Sensor Data Fusion For Air Quality Monitoring, Mirna Hesham

Theses and Dissertations

Since traditional air quality monitoring methods often rely on geographically sparse and costly air quality monitoring stations, image-based air quality method- ologies are recently offering a compelling alternative that utilizes images from sources like satellites, traffic cameras, and even smartphones to monitor pollution levels by using estimation models, image-processing techniques, and deep-learning models. In this thesis, we first conduct a systematic review, in which we categorize and discuss the existing literature work. Moreover, we introduce a novel, multi- modal dataset designed to address the limitations of existing datasets, which are restricted in size, geographical coverage, and fixed-scene imagery, impeding the …

Real Time Object Detection Using Yolo, Rohit Malik, Manisha Kumari, Sanghoon Lee

Symposium of Student Scholars

This project explores the implementation of real-time object detection using the You Only Look Once (YOLO) architecture. Leveraging its speed and accuracy, we developed a system capable of identifying and localizing multiple objects within live video streams. Our implementation focused on optimizing YOLO's performance for real-time applications, specifically addressing the trade-off between speed and accuracy.

We employed a pre-trained YOLO model and fine-tuned it on a custom dataset tailored to specific object classes. This fine-tuning process aimed to enhance the model's ability to recognize objects in our target environment. The system was implemented using Python and the OpenCV library, enabling …

A Comprehensive Analysis Of Recognition Of Hand Gestures Using Machine Learning, Shivani Shivani, Satinder Bal Gupta

Makara Journal of Technology

Hand gestures are a natural means of conveying information and thus, there is an increasing interest in utilizing gestures for communication with computers. This study focuses on systematically reviewing different machine learning algorithms while assessing their working mechanisms and accuracy. Articles were analyzed for comparing the performance of K-Nearest Neighbors (KNN), Random Forest (RF), Support Vector Machines (SVM), Naive Bayes (NB), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN). In accordance with input data, intricacy of gestures, processing resources, and real-time demands, the study shows that each technique has distinct advantages and disadvantages. RNN showed the best accuracy of …

Empowering Mental Support Health Through Ai Chatbot, Anish Ilapaka

ATU Scholars Symposium

The growing prevalence of mental health concerns worldwide underscores the urgent need for accessible, scalable, and supportive solutions. Artificial Intelligence (AI) has emerged as a promising tool in this domain, capable of delivering immediate and empathetic interactions to complement traditional methods of mental health care. This project introduces a conversational AI system to assist individuals experiencing mental health challenges. The proposed system is built on a LLaMA model fine-tuned with a dataset of 10,000 mental health-related dialogues; the system leverages advanced natural language processing and machine learning techniques for meaningful engagement. The core functionality of this tool lies in its …

Asymmetric Effects Of The Ebbinghaus Illusion On Relative Depth Judgments In A Perceptual Matching Task, Caden J. Thompson

Honors Theses

The Ebbinghaus illusion, also know as Titchner Circles, is a perceptual illusion that is typically presented as a two-dimensional set of disks. These disks are configured as a central disk surrounded by an annulus of smaller or larger disks. Numerous studies have found the illusion to have an effect on size perception, but fewer studies have analyzed its effects on depth judgments. This study utilized a head-worn virtual reality environment as well as a three-dimensional display to examine the effects of the Ebbinghaus illusion on relative depth judgments in a perceptual matching task. Findings indicated that Ebbinghaus configurations featuring an …

Balancergnn: Balancer Graph Neural Network For Imbalanced Datasets, Mallika Boyapati

Dissertations

Addressing imbalanced datasets is challenging due to machine learning models' inclination to learn the majority class. Graph construction plays a major role in determining how Graph Neural Networks (GNNs) perform on imbalanced datasets. In this research, we introduce the BalancerGNN framework to tackle highly imbalanced datasets, demonstrating its effectiveness in fraud detection as one of the case studies. This framework is designed to work for any binary node classification dataset with significant class imbalances. This research addresses the following questions: i) How effective are feature engineering techniques in the case of imbalanced datasets? ii) How do graph representation learning and …

The National Cybersecurity Teaching Coalition: Expanding Cybersecurity Education Opportunities, Paul Wagner, Melissa Dark, Robert Honomichl, Filipo Sharevski, Sandra Leiterman

Journal of Cybersecurity Education, Research and Practice

The increasing prevalence of cybersecurity threats and the shortage of qualified professionals necessitate innovative solutions for cybersecurity education at all levels. Despite the expansion of post-secondary cybersecurity programs, employer dissatisfaction with graduates and a lack of standardized introductory curricula highlights the need for structured secondary education pathways. The National Cybersecurity Teaching Coalition (NCTC) and its National Cybersecurity Teaching Academy (NCTA) address this gap by equipping high school educators with the necessary knowledge and credentials to teach cybersecurity effectively. NCTA offers an 18-credit cybersecurity graduate certificate program to ensure teachers are competent and confident to develop and teach cybersecurity curriculum with …

Visualizing Chattanooga’S Freeway Accidents: An Interactive Dashboard Built On Us National Laboratory Data*, Joshy Kasahara

Campus Research Month

Despite the availability of a freeway accident dataset collected by Oak Ridge National Laboratory, National Renewable Energy Laboratory, and Tennessee Department of Transportation (TDOT), there is no interactive visualization of the data that is easily accessible to the public. Consequently, the local community's awareness of accident trends is very limited. The contribution of this research project is a dashboard that allows the visualization of traffic accidents patterns in Chattanooga, Tennessee, using datasets from national laboratory researchers. By creating an interactive web dashboard with animated and color-mapped geographical map, the project seeks to enhance community awareness of accident trends.

Looking Good: The Math Behind Computer Vision*, Corbin Weiss

Campus Research Month

Exploring the mathematical foundations of a Multilayer Perceptron (MLP), a foundational approach to computer vision. Then expanding this understanding to create a visualization of the representation of reality in the MLP.

The Impact Of Tariffs On Auto Parts Trade With China, Canada And Mexico: Ai-Driven Strategies For Supply Chain Optimization, Katie Cerda, Layla Dickerson, Riley Gibson, Oluwabunmi Sanusi

Posters - 2025

U.S. tariffs (7.5-25%) on auto parts from China, Canada, and Mexico are severely disrupting the automotive industry, a key global economic driver. These tariffs dramatically increase production costs and vehicle prices, potentially by up to $12,200 per vehicle (CBS News, 2025; MarketWatch, 2025). These tariffs necessitate major supply chain adjustments, leading to inefficiencies (MIT Sloan, 2024). Supplier diversification, while intended to mitigate tariff impact, extends lead times and shrinks profit margins (XenonStack, 2025). The industry's complex supplier network is now highly vulnerable, compelling companies to seek more adaptable strategies. AI-driven technologies like predictive analytics and route optimization offer potential solutions …

Cult: Virtual Tourian, Ian Poll

Posters - 2025

Introduction: The Blank Shepherd Building is an important place at St. Mary’s University where students research, invent, and work together. But not everyone can visit it. This virtual tour solves that problem by using games and technology to bring the building to life.

Clinical Use Of Sit2stand Ai Application For Kinematic Analysis In Prosthesis Users., Samerial Brown

Posters - 2025

Biomechanical analysis is a tool to evaluate prosthetic and orthotic patient's. These tools offer the clinician capability of understanding the mechanism of injury, gait deviation or prosthesis problem. Video based analysis require expensive hardware, software, and training which sometimes costs $40-100,000.

The recent advent of artificial intelligence (AI) has opened up the possibility of acquiring high speed human motion video analysis using low-cost hardware and open-source machine learning algorithms. Still, free assessments like the Sit2Stand test is a current clinical outcome measure which assesses ability of a patient to stand and sit as fast as possible 5x. The faster the …