Digital Commons Network^™

Investigation Into Floral Structure Traits In Miscanthus, Wren Jenkins

Theses

Awns and callus hairs are significant dispersal traits of Miscanthus - understanding them will be essential in domesticating this crop. Due to their small size, these structures can be challenging to phenotype. To aid this, a method has been developed for automated length measurement via computational image analysis in Python. Lengths for both structures have been collected and processed for all plants in both the Mississippi State University and Alabama A&M University research fields. The measurements from both sites were compiled and mapped against the Miscanthus genome to create multiple Quantitative Trait Loci (QTL) maps of inflorescence trait-affecting genes. Chromosomes …

Why Believe An Unreliable Jailhouse Informant? : The Role Of Prosecutorial Vouching In Juror Decision-Making, Libbi A. Geoghagan

Theses

The present study asked jurors to explain why they believe jailhouse informant (JI) testimony, and investigated the role of prosecutorial vouching (i.e., the assumption that any witness called by the prosecution is credible) in jurors’ decision-making. A total of 89 juryeligible participants read a trial summary in which the JI was either absent, reliable (i.e., testified to consistent and accurate crime facts), or unreliable (i.e., accurate facts during trial, but inaccurate facts in initial statement). Primary results were that (a) in open-ended responses, participants assumed that the prosecution would not allow a dishonest or uncredible JI to testify (i.e., prosecutorial …

Subconscious Warnings As A Safety Mechanism Against Phishing Attacks, Henry Schulz

Theses

Phishing attacks are very common and considerably successful in stealing people’s information, but unfortunately normal security procedures fail to warn the user of their possible dangers. By considering how subliminal messages are very effective in informing people without actively disrupting their attention, a prototype of a browser extension was made by implementing subconscious messages and the elements of an effective warning, in an attempt to subconsciously warn users of possible phishing attempts, with minimal impact on browser performance.

The Effects Of Color, Framing, And System Certainty Information On Human-Machine Teaming, Bryanna S. Hamblin

Theses

This study examined the effects of socially learned predispositions, such as framing and color, on the use of certainty information from an automated teammate in the context of human-machine teaming. Participants were randomly assigned to a color condition (red, green, black) and saw a series of 30 certainty icons (0%, 25%, 50%, 75%, 100%) in each framing condition (success, failure). A 2 x 5 x 3 mixed subjects design was used to analyze the effects of each of these factors on the acceptance rates of the automation’s suggestion and the decision response time. The analysis found that color, framing, and …

Fault Tolerance In Krylov Subspace Methods, Sandesh Pandit

Theses

The increasing complexity of modern high-performance computing (HPC) systems, with their vast number of processing units and extensive memory hierarchies, introduces significant challenges to system resilience, particularly in the face of soft errors. Krylov subspace methods, which play a pivotal role in solving large-scale sparse linear systems and eigenvalue problems, are iterative in nature and thus susceptible to the propagation of soft errors. In this paper, we examine the fault tolerance properties of two widely-used Krylov subspace methods: the Lanczos method and the Bi- Conjugate Gradient (BiCG) method. Specifically, we analyze the impact of soft errors introduced during the critical …

Neural Acceleration Of Graph Partitioning, Vishvam Patel

Theses

Graph Partitioning is a critical problem in numerous scientific and engineering domains including social network analysis, VLSI design, and many more. Spectral methods are known to produce quality partitions while minimizing edge cuts for a wide range of problems. However, the computational cost associated with the calculation of the Fiedler vector, an eigenvector associated with the second smallest eigenvalue of the graph Laplacian, remains a significant bottleneck. In this paper, we present an neural acceleration approach to spectral bisection partitioning by replacing the traditional eigenvalue calculation with a simple artificial neural network model to approximate the fiedler vector. We demonstrate …

Application Of Machine Learning To Predict Ultrasound Wave Propagation In Biphasic Fluid–Solid Media, Jahnavi Dandamudi

Theses

The objective of this thesis is to investigate the application of Machine Learning (ML) algorithms to predict ultrasound wave propagation in the cartilage tissue of the knee joint. The learning data necessary for the ML algorithm has been generated through finite-element method (FEM)-based simulations for solving the Biot theory equations governing the propagation of continuous ultrasound through the cartilage. Specifically, we computed the ultrasound-induced dilatations and displacements in the microscale cartilage that is represented as consisting of four zones, namely the chondrocyte cell and its nucleus, the pericellular matrix (PCM) that forms a layer around the chondrocyte, and the extracellular …

Efficient Lossless Compression Of Malaria Infected Image Datasets Using Vision Transformers And Deep Autoencoders, Md Firoz Mahmud

Theses

In the field of medical imaging, lossless compression techniques play a critical role in maintaining the fidelity of medical data, while improving the efficiency of data storage and transmission, especially in telemedicine applications. This thesis addresses the development of a novel machine learning-based method for lossless compression of images of erythrocytes infected with malaria. The proposed method has two stages. In the first stage, a Vision Transformer, a computer vision model that can learn long-range dependence, is employed to separate erythrocyte images into two categories: infected and uninfected. This classification stage is essential as it allows the subsequent compression approaches …

Performance Comparison Of Variable Center Body Configurations For A 25 Mm Rotating Detonation Rocket Engine, Kaito Jonathan Durkee

Theses

Rotating detonation rocket engines (RDREs) are propulsion devices that harness circumferentially traveling detonation waves, providing theoretical benefits compared to classical, deflagration-based combustors. Because of the integration challenges posed by typical annular (Center Body, CB) RDREs, a cylindrical (Center Bodiless, CBL) chamber geometry is attractive. To understand the corresponding trade-offs for small-scale devices, the performance and operational characteristics for a 25 mm with removable 15 mm center body are compared using gaseous methane and hydrogen with oxygen, operating at varying equivalence ratios and at a total mass flow rate of 0.076 kg/s and varying mass flow rates at an equivalence ratio …

170 Years Of San Francisco Sea Level Pressure, Brendan Heaven

Theses

This data-recovery project and subsequent data analyses address the construction of a continuous, long instrumental record of sea level pressure for San Francisco, California, USA and examines what the time series may reveal about atmospheric patterns through time. Atmospheric pressure is a physical component which is associated with both daily and seasonal weather phenomena and is one of the most robust of the meteorological variables. Observational data of surface pressure (SFP) and later sea level pressure (SLP) were accessed from digital archives, with a large part of these being manually keyed from original observations on paper. San Francisco, being near …

Deep Learning Time Series Prediction Strategies For Efficiently Emulating Noah Land Surface Model Soil Moisture Dynamics, Mitchell T. Dodson

Theses

This work examines the ability of deep learning time series generative models to accurately and efficiently emulate the hourly temporal dynamics of the Noah Land Surface Model (Noah-LSM) out to a 2 week forecast horizon, given atmospheric forcings and static parameterization provided by the second phase North American Land Data Assimilation System (NLDAS-2) framework. Results from multiple neural network architectures are compared alongside variations in prediction target, loss function characteristics, and model properties. The most performant model types are subsequently evaluated with respect to forecast distance, annual seasonality, and against a variety of regional scenarios, including several event case studies. …

Ai Enabled Scrabble Scoring System : Design And Implementation, Aditya Narayan Jha

Theses

The scope of this thesis is to present an AI-powered system for automatically scoring real-world Scrabble games using computer vision and deep learning. To achieve this purpose a DeepLabv3 model is trained to segment the board from top- down images, while an EfficientNet-B0 classifier identifies individual tile characters, including blank and empty tiles. The system continuously tracks the gameplay by detecting newly placed tiles and applies official Scrabble rules to compute scores, including crosswords, bonus tiles, and bingo logic simultaneously. The scoring engine supports multi-move gameplay, player turns, and endgame handling. All moves are visualized with a color-coded board and …

Subterranean Drought Classification And Ecological Risk In Karst Systems, Shaelyn Deal

Theses

Cave ecosystems depend on stable groundwater inputs, yet no standard method exists for detecting subterranean drought. This study develops a framework to define and monitor drought in karst cave systems by adapting surface-based hydrologic indicators. Using long-term water level data from Bobcat Cave in northern Alabama, we identified drought events through percentile-based thresholds and evaluated multiple surface metrics, including the Standardized Precipitation Index (SPI) and deep-layer soil moisture, as predictors of cave drought severity. Random Forest models supported the selection of key indicators and informed classification thresholds. These thresholds were then applied across the Middle Tennessee Elk watershed to map …

An Evaluation Of The Idm Car-Following Model : Simulation And A Dynamical Systems Perspective, Trevor Mcclain

Theses

The Intelligent Driver Model (IDM) is a car-following model that has existed since the early 2000s, existing as one of the most widely used follow-the-leader models to simulate safe and realistic driver behavior. In this work, the IDM is used to evaluate the suitability of ROS 2-Gazebo for vehicle simulation, showing that significant non-determinism is present within the simulation. Additionally, we evaluate the effects of delay and an inserted Followerstopper car following model on the traffic dynamics of the system. Specifically, the generation of stop-and-go waves in idealized traffic from a dynamical systems perspective is examined. We also show that …

Exploration Of Various In-Space Fuels With Differing Characteristic Timescales For Small-Scale Rotating Detonation Rocket Engines, Raj Tejas Dave

Theses

This thesis summarizes the performance of a small-scale rotating detonation rocket engine (RDRE) using methane or hydrogen as the fuel with oxygen as the oxidizer. Specifically, a 25 mm RDRE with a 5 mm annular width and 31.2 mm length is experimentally evaluated at varying equivalence ratio for a fixed total mass flow rate of 0.076 kg/s, as well as varying total mass flow rate for fixed equivalence ratios of 1.25 and 1.00. Experimental measurements for the hot-fire tests include thrust, chamber pressure, specific impulse, and detonation mode parameters. For both sensitivity studies, hydrogen outperforms methane with a peak thrust …

Evaluating The Accuracy Of Johnson-Cook Strain Rate Sensitivity For Fcc Aerospace Structural Materials, Sydney Carlota Resnick

Theses

Understanding aerospace materials’ strain rate sensitivity is crucial for accurate designs and analysis for structural performance in dynamic loading conditions, such as rocket launches, airplane collisions, and debris impacts. Three common aerospace face-centered cubic (FCC) metals are evaluated through various strain rate tests and FEA and simulation modeling. Quasi-static and intermediate tests are performed on a hydraulic load frame, and dynamic tests are run on a split-Hopkinson pressure bar, all in tensile motion. This covers the range of strain rates from 1x10-3 s-1 to 1x103 s-1. Results reveal little strain rate sensitivity at room temperature for A286 steel and AL7075 …

Towards Secure Runtime Auditing Of Remote Embedded System Software, Adam Ilyas Caulfield

Theses

Modern embedded systems are increasingly deployed in various settings to execute critical sensing/actuation tasks. At the lower end of the scale, they are designed with low-cost and energy-efficient microcontroller units (MCUs) that lack the common security features of general-purpose processors. To assess their state in a cost-effective manner, Control Flow Attestation (CFA) methods provide means for a trusted device operator (i.e., the Verifier -- Vrf) to detect attacks that compromise the integrity or runtime behavior of a remotely-deployed MCU (i.e., the Prover -- Prv). This is achieved by a root-of-trust (RoT) in Prv recording an authenticated trace (CFLog) of the …

Study Of An Evolutionary Sequence Of Nearby, High-Excitation Bipolar Planetary Nebulae With Alma And Hst, Paula M. Moraga Baez

Theses

Planetary nebulae (PNe) reveal the near-final stages of the evolution of intermediate-mass (1-8 solar mass) stars, immediately after such stars shed their envelopes through copious asymptotic giant branch (AGB) mass loss. PNe reveal products of intermediate-mass stellar nucleosynthesis just before they enrich the interstellar medium, and their central stars provide subjects for studying the origins of white dwarfs and interacting binary systems involving compact objects. The processes that shape planetary nebulae and determine their chemistries can be unveiled through multiwavelength observational studies of molecule-rich PNe ranging from young and rapidly evolving nebulae, such as NGC 7027, NGC 6302, and NGC …

Neuro-Evolution For Multivariate Time Series Anomaly Detection, Aryan Jha

Theses

Given the exponential growth of data in various sectors, the timely identification of anomalies has become crucial in mitigating risks such as financial losses and data breaches. This research focuses on advancing multivariate time series anomaly detection through the innovative application of neuro-evolutionary methods. Utilizing the Evolutionary eXploration of Augmenting Memory Models (EXAMM) algorithm, the study automates the design and optimization of Long Short-Term Memory (LSTM) autoencoder networks, enabling enhanced forecasting capabilities in recurrent neural network architectures. Leveraging these networks along with a One-Class Support Vector Machine (OCSVM) to generate boundary thresholds, this project delivers a robust and scalable procedure …

Predicting Airbnb Prices In Dubai Using Machine Learning, Hamad Younus Altamimi

Theses

The Airbnb rental market in Dubai is dynamic and full of varying factors like location, equipment, and demand time shortages, which makes it challenging to have fair pricing for the different levels of Airbnb properties in Dubai. The goal of the project was to construct a powerful machine-learning model for the forecast of the Airbnb rental price in Dubai The method used by the project was systematic from the beginning to the end stages, which included the data collection from Kaggle, the data preprocessing which involved the handling of missing values, outlier elimination, data type conversion, and exploratory data analysis …

Resilient Power System Load Frequency Control, Ghadeer Ahmed Almahri

Theses

Standard frequency and voltage levels need to be maintained when the electric power system is running in order to keep it in a satisfactory state of operation. Between AC power generation and use, there should be a balance between active and reactive power. However, due to variations in active and reactive power requirements, the frequency and voltage provided may fluctuate from their rated values. Any unwelcome change in frequency or voltage can have an immediate effect on how the power system operates and may even cause harm to linked devices. These parts are intended to operate at certain rated frequencies …

Domain-Specific Customization For Improving Speech To Text, Zubin Mario Tobias

Theses

The advent of transformer-based models has revolutionized natural language processing, bringing remarkable improvements in tasks like automatic speech recognition (ASR). Inspired by these advancements, this thesis explores the optimization of a transformer-based ASR model to improve transcription accuracy in educational settings, particularly for lecture content. The goal of this research is to provide real-time, high-accuracy captions that enhance accessibility for all students, while offering a cost-effective solution for educators. To assess the potential of domain-specific fine-tuning, Whisper-small underwent two phases of fine-tuning. In the first phase, it was finetuned on care- fully selected, publicly available datasets: SpeechColab’s Gigaspeech-XS [39], AMI …

Predictive Policing - Leveraging Cctv Data And Ai For Crime Hotspots, Maitha Ahmad

Theses

This thesis explores fine-grained sentiment analysis in the context of police social media posts, leveraging hybrid approaches to manage the rapid flow of big data and mitigate its negative impact on youth. As social media becomes a primary medium for public interaction, understanding sentiment within police-related posts is crucial for law enforcement agencies to gauge public opinion and address community concerns effectively. The rapid dissemination of information and the potential for negative sentiments to spread swiftly pose significant challenges, particularly for young audiences. The research begins with a comprehensive review of existing sentiment analysis techniques and their applicability to large-scale …

Determinants Of Profitability In Uae Banks (2019-2023): An Analysis And Forecast For The Next Five Years, Rashed Aljanahi

Theses

This study aims to analyze the factors that led to the banks' profitability in UAE from 2019 to 2023 based on factors such as capital adequacy, asset quality, and liquidity. With these factors and Return on Assets, the study seeks to determine the link between the variables affecting financial performance in the banking sector. The study relies on quantitative data from UAE banks to include statistical analysis and construct predictive models. The relationships between the variables are established using Multiple Linear Regression. At the same time, Random Forest Algorithms explore Non-Linear Interactions and Time Series Analysis using ARIMA models to …

Using Generative Models For Data Augmentation In Medical Imaging, Nilesh Kumar

Theses

Obtaining large amounts of labelled data in medical imaging is hindered by issues like privacy and cost, therefore limiting the gains that can be achieved from Deep Neural Networks (DNNs). A lot of work has shown data augmentation to be an effective strategy for expanding training data leading to more robust models. However, the lack of data is not always uniform across datasets e.g., some classes in a dataset may be underrepresented. This is particularly problematic because the overall good performance of a model on a dataset can easily mask the under-performance of underrepresented classes or instances. Naively employing data …

Exploration Of The Iterative Design Process: Integrating A Requirements Matrix And Hazard Analysis Toolsexploration Of The Iterative Design Process: Integrating A Requirements Matrix And Hazard Analysis Tools, Blaine Edlefsen

Theses

The iterative design process is applied to multiple industry sectors from software development to product manufacturing. The cyclical approach integrates the steps of design, prototyping, and testing toward the realization of a user's specified needs. The use of a requirements traceability matrix and hazard analysis documentation that supports the definition of the iterative design process is explored. These design project management tools are applied to a case study involving workholding design for a Wire Electrical Discharge Machining (Wire-EDM) manufacturing process. This industry-relevant case study is conducted in collaboration with New Jersey Precision Technologies, Incorporated (NJPT). The design requirements involve both …

Automated Segmentation Of The Ulnar Nerve In Mri Using Deep Learning Techniques, Akhil Nagulapalli

Theses

Cubital Tunnel Syndrome (CuTS), a condition caused by compression of the ulnar nerve, results in numbness, tingling, pain, and even muscle atrophy, affecting fine motor skills and diminishing patient quality of life. Accurate diagnosis of CuTS is challenging, as current diagnostic methods—including clinical exams, nerve conduction studies, and unaided MRI—often lack the precision to reliably identify the nerve and detect compression in its early stages. Deep learning-based segmentation offers a promising solution, enabling precise and automated identification of nerve structures in MRI images, which could significantly improve diagnostic accuracy and support timely intervention.

A novel deep learning model for segmenting …

I Hear A New World, Katie Christ

Theses

“I Hear a New World” is a documentary short film that reflects on the relevance and persistence of nostalgia and tradition during a time of rapid technological advancement and societal change. The film highlights the memories and musings of 8 college students, who were all Class of 2020 high school seniors and whose senior traditions were inevitably brought entirely online due to the Covid-19 global pandemic to varied degrees of success. In rambling, thoughtful delivery, the Gen Z interviewees provide insight into their pre-pandemic internet obsession that led to their teenage reclusivity and how this perspective changed greatly following the …

Amorfe: An Ai-Integrated Interactive Brainstorming Interface, Ipshita Pal

Theses

Amorfe is an interactive multi-touch digital interface that reimagines brainstorming and collaboration using artificial intelligence (AI). Designed to make ideation more dynamic, the platform allows participants to explore, merge, and refine ideas in real time. AI is a creative catalyst, offering unpredictable outcomes and streamlining the brainstorming process. This makes Amorfe engaging, intuitive, and inclusive, encouraging active participation. The platform’s potential was demonstrated through a hypothetical scenario: “How do we create more green spaces on the RIT campus?” Four participants explored various brainstorming phases through touch interaction, leading to the creation of a prototype simulation video and a promotional video. …

Dissonance, Shreya Talegaonkar

Theses

Cognitive dissonance represents a profound psychological phenomenon where conflicting beliefs create mental tension. My thesis, Dissonance, transforms this abstract concept into an immersive multimedia installation that visually explores the human mind's intricate journey of reconciling internal contradictions. Using Cinema 4D, Adobe AfterEffects, and Resolume Arena, the project creates dynamic 3D animations projection-mapped across an 8x8x6 cubic feet space. Carefully designed blackout curtains and strategic seating immerse participants in a transformative narrative of psychological exploration. Audiences navigate through four distinct phases: Consistency, Chaos, Turmoil, and Adaptation. Each phase invites viewers to experience the unsettling yet profound journey of internal conflict resolution, …