Physical Sciences and Mathematics Commons^™

Potsdam: Pareto Optimization Targeting Security, Data, And Mediation, J Peter Brady

Dartmouth College Ph.D Dissertations

Given the growing amount and variety of data handled by modern systems, it is crucial to guarantee the accuracy and protection of input data without errors or malicious intentions. The need to improve security in software programs often conflicts with the assurance of maximum performance, making developers and maintainers hesitant to incorporate more testing.

LangSec (Language-Theoretic Security) is a security approach that treats input validation as a formal language recognition problem, ensuring that only well-defined, unambiguous inputs are processed to eliminate exploitable parsing flaws. This dissertation explores integrating LangSec principles with Pareto optimization to enhance safety and robustness in digital …

Extremal Trees For Random Walks, Ben Bridenbaugh

Mathematics, Statistics, and Computer Science Honors Projects

A random walk is a sequence of adjacent vertices that are chosen uniformly at random from the neighbors of the previous vertex. An access time is the average length of time that a random walk takes to reach a target probability distribution from a starting probability distribution, given an optimal stopping rule. This paper deals with characterizing the trees of diameter d and on n vertices that extremize three different types of access times.

Multitask Learning For Named Entity Recognition And Relationship Extraction, Adrienne D. Hembrick

Theses and Dissertations

Information Extraction (IE) is a fundamental task in Natural Language Processing (NLP), involving the identification of structured information from unstructured text. Two core components of IE—Named Entity Recognition (NER) and Relation Extraction (RE)—are widely used to extract key concepts and the relationships between them across various domains. However, the sequential dependency of RE on the output of NER makes it vulnerable to error propagation: inaccuracies in entity recognition can negatively affect downstream relation extraction.

To mitigate this issue, Multitask Learning (MTL) has been proposed as an approach that jointly models NER and RE, aiming to improve overall performance and reduce …

Leveraging Counterfactuals For Enhanced Natural Language Inference: A Multitask Knowledge Distillation Approach, Brian J. Davis

Browse all Theses and Dissertations

Natural-language inference (NLI) asks whether a hypothesis is entailed by, contradicts, or is neutral with respect to a premise. Modern transformers reach high raw accuracy on benchmarks such as SNLI, MNLI, and ANLI, yet they often rely on brittle lexical shortcuts and provide little insight into their decision process. This thesis shows that counterfactual-augmented knowledge distillation can simultaneously boost robustness and supply faithful, token-level explanations—without scaling model size. Four T5-v1_1 students (60M, 220M, 770M, 3B parameters) are trained under four curricula: (1) standard fine-tuning, (2) fine-tuning with free-text rationales, (3) multi-task distillation with naive counterfactuals, and (4) multi-task distillation with …

Searching For Cages In Graphs And Signed Graphs, Isaac Partee

Browse all Theses and Dissertations

The girth of a graph G is the minimum length of a cycle in G. A (k,g)-graph is a k-regular graph of girth g. A (k,g)-cage is a (k,g)-graph with the smallest possible number of vertices. For example, K4 is the unique (3,3)-cage, K3,3 is the unique (3,4)-cage, and the Petersen Graph is the unique (3,5)-cage. The search for cages particular values of (k,g) is an ongoing area of considerable research. For values of (k,g) where the number vertices in a (k,g)-cage is not known, covering graphs have recently been used for constructing progressively smaller and smaller (k,g)-graphs. The covering …

Optimizing Cloud Computing Resources For Operational Cost And Application Performance Using Machine Learning, Isaac K. Matthew

Browse all Theses and Dissertations

As AI-driven workloads accelerate the growth of cloud initiatives and spending, resource waste also increases due to persistent inefficiencies in cloud compute and infrastructure management. Overprovisioned resources and suboptimal configurations often lead to operational inefficiencies and unnecessary financial overhead. These challenges arise from the difficulty of anticipating resource demands in dynamic workloads and selecting suitable virtual machines to ensure optimal performance. Our research proposes a holistic, data-driven framework for managing cloud compute resources that reduces costs without compromising application performance. We integrate a predictive, model-driven, threshold-based autoscaling solution for cloud-native applications with an optimized instance right-sizing approach to select cost-effective …

Reducing Operator Training Time Through Virtual Reality: A Case Study On The Lpkf Protomat E44 Machine, Joshua C. Patel

Browse all Theses and Dissertations

This thesis presents the development of an immersive virtual reality (VR) simulation that replicates the operation of the LPKF ProtoMat E44 PCB milling machine. Aimed at reducing operator training time and improving procedural understanding, the simulation offers an interactive and realistic environment where users can safely engage with machine workflows and start-up sequences. The emphasis is on accurate representation, usability, and maintaining immersion to support intuitive learning. Although formal evaluation is outside the scope of this work, the system is designed to serve as a foundation for cost-effective, scalable training in technical and manufacturing contexts, offering a modern alternative to …

Scalable Real-Time Stream Clustering For Unbounded Text Streams, Nathaniel C. Crossman

Browse all Theses and Dissertations

Social media, AI systems, IoT sensors, and other platforms generate vast amounts of streaming data. Given this vast volume of information, techniques that can reduce and aggregate data into meaningful topics are essential. One such technique is the two-phase stream clustering approach. In the first, online micro-clustering phase, the system forms micro-clusters from the incoming data stream, incrementally merges new items into related existing micro-clusters, and prunes or fades micro-clusters as they become inactive, producing a constantly updating yet compact set of micro-clusters representing potential topics and subtopics of the stream. In the second, offline macro-clustering phase, these micro-clusters are …

Patient Subset Classification Using Encoded Embeddings And Knowledge Graph Retrieval-Augmented Generation, Benjamin A. Holmes

Browse all Theses and Dissertations

The widespread adoption of electronic medical records has created a vast reservoir of clinical data that can be leveraged to better understand how interventions relate to patient outcomes. Much of this information, however, exists as unstructured free-text, posing significant challenges for traditional statistical and machine-learning methods. Solving these challenges would allow the extraction of specific patient subpopulations (clinically relevant cohorts of individuals who share overlapping symptoms, risk factors, or diagnostic criteria), which could be used in precision medicine. Despite this promise, extracting these subpopulations from unstructured medical notes is an ongoing challenge due to the variability of clinical language and …

Generative Adversarial Networks (Gans) For High-Dimensional Biological Data, Harigovind Harikumar

Browse all Theses and Dissertations

This thesis investigates the application of Generative AI models, mainly Generative Adversarial Network (GAN) models to high dimensional and low sample size biological datasets like Motion Sickness, Breast Cancer, Crohn, and Melanoma. We utilized and compared three generative AI frameworks: Vanilla GAN, Wasserstein GAN (WGAN), Locality-Sensitive Hashing GAN (LSH-GAN) and Omics GAN. To address the challenges associated with high-dimensionality and low sample size, which was leading to very poor outputs of biological synthetic samples, we came up with an approach to stop the model when it reaches its saturation level. That is, we printed the loss plots to see where …

Pixmix Attack: Implementation And Evaluation Of A Novel Pixel Injection On Digital Video Port (Dvp) Interface In Embedded Camera Systems With Pcb Hardware Trojan, Sayed Md Tashfi Nowroz

Browse all Theses and Dissertations

Image sensors are at the heart of machine vision systems in robotics, industrial automation, and surveillance systems which ideally operate with minimal human supervision and only occasional maintenance. The image sensors convert visible light into electrical signals which are locally decoded to image on the printed circuit board (PCB) by an ordinary embedded processor System on Chip (SoC). This thesis investigates a critical vulnerability in such systems, targeting the communication protocol at the signal level during runtime. Specifically, it focuses on a novel attack in the Digital Video Port (DVP) protocol, possible to exploit with PCB-based hardware Trojans, to craft …

A Secure Ml-Assisted Framework For Resilient And Efficient Prediction Of Physiotherapy Sequence In Bilateral Carpal Tunnel Syndrome, Pratik Pandurang Kharat

Browse all Theses and Dissertations

Bilateral idiopathic carpal tunnel syndrome (CTS) is a neuromuscular disorder characterized by compression of the median nerve at both wrists, leading to symptoms such as pain, numbness, tingling, and muscle weakness. Unlike unilateral cases, bilateral idiopathic CTS presents distinct therapeutic challenges due to the simultaneous involvement of both hands and the lack of an identifiable underlying cause. This study explores the application of machine learning techniques to predict the optimal sequence of physiotherapeutic interventions Stretching followed by Myofascial Mobilization (S/M) or the reverse (M/S) in female patients with bilateral idiopathic CTS and right hand dominance. Data were drawn from a …

Learning Under Data Scarcity: Reasoning And Negative Distillation For Texts And Graphs, Calvin T. Greenewald

Browse all Theses and Dissertations

Modern machine learning (ML) models rely on large amounts of high-quality labeled data to achieve optimal performance. However, in many real-world domains, such as cyber security, acquiring sufficient labeled data is often infeasible due to cost, privacy concerns, and the rapid evolution of underlying phenomena. This challenge underscores the importance of learning under data scarcity. This thesis addresses this challenge by proposing distinct, modality-specific techniques for text and graph domains, which allow models to generalize effectively with minimal data. For text classification task, we incorporate distilled rationales from large language models and adversarial perturbations into the input space to improve …

Light Absorbing Aerosol Interactions With Warm Clouds, Shreya Joshi

Dissertations, Master's Theses and Master's Reports

Atmospheric particles such as biomass-burning aerosols, including black carbon - are critical in altering the Earth's energy balance through interactions with light and clouds. Biomass-burning aerosols also contain organic carbon, a diverse mix of organic compounds. As these aerosols age, they evolve by interacting with other atmospheric species, including cloud water, leading to changes in particle morphology and the development of coatings that impact their lifetime and ability to interact with clouds and light.

Some biomass-burning aerosols serve as cloud condensation nuclei. When cloud water droplets form, the properties of the nucleating particles can change. Many clouds evaporate before precipitating, …

Utilizing Biometrics And Blockchain For Enhanced Security Of Remote Patient Monitoring (Rpm) Data Sharing, Amaturrahman Raihanah Medlock

Dissertations, Master's Theses and Master's Reports

Advancements in Artificial Intelligence (AI) and Internet of Medical Things (IoMT) technologies have significantly revolutionized the conventional healthcare systems. Through the integration of smart devices, medical sensors, and communication technology, IoMT provides real-time patient’s monitoring data for healthcare providers, thus promoting accurate and timely clinical decisions for patient-centric care. The current healthcare sector is evolving to a connected ecosystem with connectivity and intelligence. While it also incurs increasing security and privacy concerns as integrating IoMT generated patient monitoring data into healthcare information systems. Both blockchain and biometrics are measures that have established reputable names in the security realm. When evaluating …

Quantifying So₂ Emissions From Explosive Events Of The 2021 La Soufrière Eruption Using Dscovr/Epic, Yuxi Jin

Dissertations, Master's Theses and Master's Reports

Volcanic eruptions release large quantities of gases, including sulfur dioxide (SO₂), aerosol precursors, and ash into the atmosphere. Quantifying the emission fluxes and temporal variability of SO₂ is crucial for understanding volcanic processes and assessing their impacts on air quality, climate, and aviation safety. Episodic vulcanian to sub-Plinian explosive eruptions are a common feature of many arc volcanoes, but the underlying causes of such episodic behavior remain poorly understood. The 2021 La Soufrière eruption (St. Vincent) featured up to 40 discrete eruptive events between April 9 and April 22, presenting a rare opportunity to analyze time-resolved volcanic SO₂ emissions during …

Methods In Statistics, Machine Learning, And Deep Learning For Combining Multi-Omics Dataset, Md Mutasim Billah

Dissertations, Master's Theses and Master's Reports

Transcriptome-wide association studies (TWAS) have emerged as a powerful strategy to bridge genome-wide association studies (GWAS) with gene regulatory mechanisms by integrating genotypic data with gene expression data. While early TWAS methods typically rely on linear models and single-tissue expression references, recent advances underscore the need for flexible, multi-tissue approaches that can capture heterogeneous regulatory architectures and tissue-specific expression patterns. This dissertation introduces a three‑part research project that advances multi‑tissue transcriptome‑wide association studies (TWAS) along complementary axes of methodology, statistical power, and modelling flexibility.

In chapter One, TWAS‑CTL introduces a two‑stage cross‑tissue learner that trains any user‑chosen single‑tissue imputers (STLs) …

Comprehensive Analysis Of Krauklis Waves And Long Period Seismic Events In Fluid-Filled Fractures Via Integrated Physical And Numerical Modeling, Sananda Ray

Dissertations, Master's Theses and Master's Reports

Long-period (LP) seismic events, commonly recorded in volcanic and hydrothermal environments, are widely attributed to the resonance of fluid-filled fractures and the excitation of Krauklis waves along the solid-fluid interface. These low-velocity, dispersive waves are sensitive to both fluid dynamics and fracture geometry, offering critical insights into subsurface fluid transport, fracture evolution, and seismic source processes. Despite robust theoretical treatments, experimental validation under realistic geological conditions has remained limited. To address this, I aimed to develop a series of laboratory-based experimental models that systematically investigate the key physical mechanisms driving LP signal generation. In the first set of experiments, I …

The Inverse Scattering Transform For The Nonlinear Schrödinger Equation, Ivan Casas-Rocha

Honors Undergraduate Theses

The Nonlinear Schrödinger (NLS) Equation, iψ_t + 1/2 ψ_xx ± |ψ|²ψ = 0, is a nonlinear partial differential equation which is used to model several physical phenomena including nonlinear effects inside optical fibers and the formation of rogue waves in shallow water. It is particu- larly difficult to study solutions to this equation due to the nonlinearity, and the nonlinearity leads to incredibly interesting solutions not found in linear PDEs such as solitons. In this thesis, we highlight two methods of obtaining solutions to the (NLS) equation: the Inverse Scattering Transform and the Dressing Method. Furthermore, …

Comparative Efficacy Of Hallucinogens In Treating Mood Disorders Through A Meta-Analysis Of Symptom Reduction, Dosage, And Duration, John Marco D.F. Muniz

Honors Undergraduate Theses

Background: Hallucinogens including psilocybin, lysergic acid diethylamide (LSD), ketamine, N,N-dimethyltryptamine (DMT) (as ayahuasca), have re-emerged as potential rapid-acting treatments for mood disorders. We conducted a meta-analysis of placebo-controlled trials evaluating their efficacy in depression and anxiety disorders. Methods: A systematic review identified 12 trials (Total ≈ 670) meeting inclusion criteria (randomized, placebo-controlled). Data on Cohen’s d and Hedges’ g effect sizes for depression- and anxiety-related outcomes were extracted. We computed pooled effect sizes (weighted by sample size and inverse variance), performed subgroup analyses by drug, diagnosis, follow-up duration, and outcome measure type, and assessed heterogeneity (I², Q …

Modified Equations Of Conformal Symplectic Exponential Time Differencing Methods, Taylore M. Keesler

Honors Undergraduate Theses

Planetary orbits, pendulums, and hurricanes are everyday examples of nonlinear systems, often studied using differential equations. However, their exact solutions can not always be computed, and thus, we use numerical methods to approximate their solutions. Certain methods are better suited to preserve special properties of the system like energy and geometry. Through previous numerical simulations, a conformal symplectic method proves more effective in this preservation. To understand why, we find the modified equation of a nonlinear system with damping, which is a differential equation for which the numerical solution is exact. Through backward error analysis, we obtain these modified equations, …

A Time Series Analysis Of The Macroeconomic Indicators, Mia Houston

Honors Undergraduate Theses

Understanding inflation—particularly across regions and categories—is crucial for effective policymaking, strategic business decisions, and safeguarding vulnerable populations, as it highlights the diverse drivers and impacts of price changes within the economy. This has become increasingly crucial in recent years between the volatile inflation conditions introduced by the COVID-19 pandemic, energy price shocks, and renewed trade tensions and tariffs. This thesis analyzes 77 U.S. monthly inflation time series from 2003 to 2023 using two forecasting approaches: an elementwise Seasonal Autoregressive Integrated Moving Average (SARIMA) model and a Factor-Augmented Vector Autoregressive (FAVAR) model. The data obtained from the Bureau of Labor Statistics …

Solvability Of Stochastic Linear-Quadratic Optimal Control Problems Under Partial Stabilizability Conditions, Al-Sadh Rahman Imadh

Honors Undergraduate Theses

Optimal Control Theory, a branch of Control Theory, is applicable in fields such as engineering, operations research, and economics. Stochastic Optimal Control deals with noisy systems and data using Ito’s formulation. Given a noisy system and a cost functional, the goal is to find a control that will minimize the cost. This thesis focuses on linear quadratic stochastic optimal control, and we explore state equations that are not stabilizable. We first address measurability concerns arising from the semigroup property of the state trajectory. The notions of partial stability and partial stabilizability are introduced, and we formulate their corresponding Lyapunov and …

Theoretical Analysis Of Cnns For Automatic Seizure Detection In Eeg Signals, Jackson T. Small

Honors Undergraduate Theses

Epilepsy is a common brain disorder where neurons in the brain rapidly fire, causing recurring seizures. The brain activity during a seizure can be detected by electroencephalogram (EEG) signals; however, this process is not only labor-intensive and time-consuming but is also subject to inter-rater variability, with a study showing only moderate agreement when diagnosing patients, even among experts. Convolutional Neural Networks (CNNs) are often proposed to detect seizures automatically, achieving high performance. The focus on performance comes at a cost of losing interpretability, leaving the model as effective but seen as a ’black box’. This thesis confronts the interpretability knowledge …

Archival Stellar Flare Follow-Up In Evryscope Observations, Tyler J. Thomas

Honors Undergraduate Theses

M dwarfs, the most abundant stars in the galaxy, are prime targets in the search for habitable exoplanets. However, their high flare activity may compromise planetary atmospheres, with Coronal Mass Ejections (CMEs) increasing this risk. Despite this, the relationship between M dwarf flares and CMEs is not well understood. This study uses archival data from Evryscope North and South, crossmatched with Sloan Digital Sky Survey (SDSS) and Gaia data, to investigate this relationship. This crossmatching demonstrated an abundance of stars within the 20 arcsecond Evryscope uncertainty. The analysis also revealed a modest excess of K1 and F9 type stars but …

Generating Negotiations For Iago, Kylee R. Weener

Honors Undergraduate Theses

Negotiation is a complex field that can benefit from introducing artificial intelligence (AI); doing so would benefit researchers as they try to deepen their understanding of human-human and human-agent negotiation. Investigating how large language models (LLMs) can generate negotiation dialogue with emotional context would bring agents closer to acting more human. This study explores how fine-tuning and prompt engineering can achieve this goal and the possibilities for an AI that fills these criteria to be included in the Interactive Arbitration Guide Online platform (IAGO). Doing so will make the negotiation interactions in IAGO feel more complex and natural, allowing researchers …

Quantized Average Agreement Algorithms With Error Correction For Digraphs, Shuaib A. Mughal

Honors Undergraduate Theses

Multi-agent systems have become more and more prevalent as technology increasingly gets integrated into our daily lives. Some of these technological systems are large in size; for example, the smart grid where multiple devices are used to monitor and control different aspects of the energy grid. Another example is a team of autonomous systems deployed for a specific task. When these systems are spatially distributed, an important component of distributed algorithms is the ability for the agents to reach consensus on the global state of the system. Reaching agreement enables the spatially distributed agent make decisions or determine the next …

Cloud Condensation Nuclei Activity Of Fresh And Aged Phenolic Acid Aerosol Particles, Emily A. Nortmann

Honors Undergraduate Theses

In recent years, the incidence of wildfires has considerably increased, with fire seasons starting earlier and ending later than usual due to changes in snowpack, precipitation, and temperature, which are attributed in part to climate change. Wildfires are responsible for the emission of a variety of gases and organic aerosol particles in the atmosphere with profound impacts on air quality, visibility, and human health. The aerosol particles produced by wildfires can contribute to climate change by scattering or absorbing solar radiation, affecting the radiative balance of the planet through cooling or warming effects, or by acting as cloud condensation nuclei …

Role Of C4 Resources In Isotopic Variability In Diet Among Children From Kellis 2 Cemetery, Dakhleh Oasis, Egypt, Faith R. Hendrix

Honors Undergraduate Theses

Using stable carbon isotope analysis, this study investigates dietary diversity in children buried at the Kellis 2 Cemetery (c. AD 50–450) in Egypt's Dakhleh Oasis. From the analysis of δ¹³C isotope values in hair keratin and bone collagen, the study reconstructs short-term and long-term dietary signals in juvenile and adult subjects. The aim is to clarify the role of C₄ plants—particularly millet—in weaning and childhood diets in a Romano-Christian Egyptian village context. A total of 631 segmented hair and 54 bone collagen samples were analyzed from 127 juveniles and 97 adults. Juvenile individuals (i.e., under 15 years biological age) showed …

Partisan Divides In Environmental Spending Attitudes: A Two-Level Hierarchical Analysis, 1973-2022, Jordan Lipner

Honors Undergraduate Theses

Public attitudes toward environmental spending have become increasingly divided along party lines, with sharp shifts over the past five decades. This thesis updates and expands on Johnson and Schwadel’s 2019 study by applying a two-level hierarchical linear model to General Social Survey data updated to include data from 2015-2022, capturing how political affiliation, education, race, and economic context interact with broader political and economic contexts to shape environmental attitudes over time.
The results show that political affiliation remains the strongest and most reactive predictor of environmental spending attitudes. Republican respondents are significantly more likely to oppose environmental spending, especially under …