Digital Commons Network^™

Enhancing Fundraising Strategies In Higher Education Through Machine Learning, Laith Alatwah

Electrical Engineering Theses

This thesis presents a comprehensive application of machine learning techniques, namely Fine Gaussian SVM and RUS Boosted Trees, to enhance fundraising strategies in higher education institutions. Analyzing a rich dataset from Blackbaud Raiser's Edge NXT, spanning 2012 to 2022, the study focuses on donor profiles, including demographics, donation history, and engagement patterns. Key demographic insights include the increasing engagement of younger donors (20-29 age group) and significant contributions from older donors (70-99 age group). Geographical trends are also examined, revealing distinct patterns based on donors' city, state, and ZIP code. The Fine Gaussian SVM model demonstrates moderate discriminatory power, with …

Leveraging Generative Ai For Sustainable Farm Management Techniques Correspond To Optimization And Agricultural Efficiency Prediction, Samira Samrose

All Graduate Reports and Creative Projects, Fall 2023 to Present

Sustainable farm management practice is a multifaceted challenge. Uncovering the optimal state for production while reduction of environmental negative impacts and guaranteed inter-generational assets supervision needs balanced management. Also, considering lots of different factors (cost, profit, employment etc), the agricultural based management technique requires rigorous concentration. In this project machine learning models are applied to develop, achieve and improve the farm management techniques. This experiment ensures the resultant impacts being environment friendly and necessary resource availability and efficiency. Predicting the type of crop and rotational recommendations will disclose potentiality of productive agricultural based farming. Additionally, this project is designed to …

Neural Networks For Decisions Under Uncertainty, Edwin Tomy George

Open Access Theses & Dissertations

Neural networks are used in many real-world applications, ranging from classification tasks to medical diagnostics. For each task, a neural network is typically able to make predictions due to its ability to extract meaningful patterns from processing large amounts of data. Thus, given the increases in available data in recent decades, the performance of neural networks in making accurate predictions has greatly increased. However, this data often comes with ingrained uncertainties due to measurement errors or the inherent variability of individual data points. Neural networks can learn despite the errors in the overall data, but what if we want them …

Physics-Informed Machine Learning Methods For Inverse Design Of Multi-Phase Materials With Targeted Mechanical Properties, Yunpeng Wu

All Dissertations

Advances in machine learning algorithms and applications have significantly enhanced engineering inverse design capabilities. This work focuses on the machine learning-based inverse design of material microstructures with targeted linear and nonlinear mechanical properties. It involves developing and applying predictive and generative physics-informed neural networks for both 2D and 3D multiphase materials.

The first investigation aims to develop a machine learning method for the inverse design of 2D multiphase materials, particularly porous materials. We first develop machine learning methods to understand the implicit relationship between a material's microstructure and its mechanical behavior. Specifically, we use ResNet-based models to predict the elastic …

Personalized Driving Using Inverse Reinforcement Learning, Rodrigo J. Gonzalez Salinas

Theses and Dissertations

This thesis introduces an autonomous driving controller designed to replicate individual driving behaviors based on a provided demonstration. The controller employs Inverse Reinforcement Learning (IRL) to formulate the reward function associated with the provided demonstration. IRL is implemented through a dual-feedback loop system. The inner loop utilizes Q-learning, a model-free reinforcement learning technique, to optimize the Hamilton-Jacobi-Bellman (HJB) equation and derive an appropriate control solution. The outer loop leverages this derived control solution to generate parameters for the reward function, which are subsequently integrated into the HJB equation. The ultimate control policy is deduced from the final reward function obtained …

Investigating Bias In Mortgage-Rate Machine Learning Models, Will Kalikman

Computer Science Senior Theses

Banks and fintech lenders increasingly rely on computer-aided models in lending decisions. Traditional models were interpretable: decisions were based on observable factors, such as whether a borrower's credit score was above a threshold value, and explainable in terms of combinations of these factors. In contrast, modern machine learning models are opaque and non-interpretable. Their opaqueness and reliance on historical data that is the artifact of past racial discrimination means these new models risk embedding and exacerbating such discrimination, even if lenders do not intend to discriminate. We calibrate two random forest classifiers using publicly available HMDA loan data and publicly …

Unboxing The Complicated Near Term Climatic And Geomorphic History Of Mars, Joshua Matthew Williams

Earth and Planetary Sciences ETDs

It has long been thought that glacial processes were unlikely within the tropical regions of Mars. However, growing evidence, including this work has identified and quantified relic glacial forms within the equatorial regions. These findings have major implications for understanding Martian climate history and its sensitivity to changes in insolation. As well, the presence of ice in the equatorial region of Mars has significant implications for the past global redistribution of the water ice in the Martian cryosphere. In this manuscript, I clarify and refine our understanding of the morphology of glacial features in the equatorial zone by applying novel …

Evaluating Neuroimaging Modalities In The A/T/N Framework: Single And Combined Fdg-Pet And T1-Weighted Mri For Alzheimer’S Diagnosis, Peiwang Liu

McKelvey School of Engineering Theses & Dissertations

With the escalating prevalence of dementia, particularly Alzheimer's Disease (AD), the need for early and precise diagnostic techniques is rising. This study delves into the comparative efficacy of Fluorodeoxyglucose Positron Emission Tomography (FDG-PET) and T1-weighted Magnetic Resonance Imaging (MRI) in diagnosing AD, where the integration of multimodal models is becoming a trend. Leveraging data from the Alzheimer's Disease Neuroimaging Initiative (ADNI), we employed linear Support Vector Machines (SVM) to assess the diagnostic potential of these modalities, both individually and in combination, within the AD continuum. Our analysis, under the A/T/N framework's 'N' category, reveals that FDG-PET consistently outperforms T1w-MRI across …

Toward The Integration Of Behavioral Sensing And Artificial Intelligence, Subigya K. Nepal

Dartmouth College Ph.D Dissertations

The integration of behavioral sensing and Artificial Intelligence (AI) has increasingly proven invaluable across various domains, offering profound insights into human behavior, enhancing mental health monitoring, and optimizing workplace productivity. This thesis presents five pivotal studies that employ smartphone, wearable, and laptop-based sensing to explore and push the boundaries of what these technologies can achieve in real-world settings. This body of work explores the innovative and practical applications of AI and behavioral sensing to capture and analyze data for diverse purposes. The first part of the thesis comprises longitudinal studies on behavioral sensing, providing a detailed, long-term view of how …

Code For Care: Hypertension Prediction In Women Aged 18-39 Years, Kruti Sheth

Electronic Theses, Projects, and Dissertations

The longstanding prevalence of hypertension, often undiagnosed, poses significant risks of severe chronic and cardiovascular complications if left untreated. This study investigated the causes and underlying risks of hypertension in females aged between 18-39 years. The research questions were: (Q1.) What factors affect the occurrence of hypertension in females aged 18-39 years? (Q2.) What machine learning algorithms are suited for effectively predicting hypertension? (Q3.) How can SHAP values be leveraged to analyze the factors from model outputs? The findings are: (Q1.) Performing Feature selection using binary classification Logistic regression algorithm reveals an array of 30 most influential factors at an …

Predicting Energy Expenditure From Physical Activity Videos Using Optical Flows And Deep Learning, Gayatri Kasturi

Theses and Dissertations

This thesis presents a novel approach for predicting energy expenditure of physical activity from videos using optical flows and deep learning. Conventional approaches mainly rely on wearable sensors, which, despite being widely used, are constrained by practicality and accuracy concerns. This proposal introduces a new strategy that utilizes a three-dimensional Convolutional Neural Network (3D-CNN) to evaluate video data and accurately estimate energy costs in metabolic equivalents (METs). Our model utilizes optical flow extraction to analyze video, capturing complex motion patterns and their changes over time. The results are good indicating potential for this method to be deployed in various healthcare …

Advancing Compact Modeling Of Electronic Devices: Machine Learning Approaches With Neural Networks, Mixture Density Networks, And Deep Symbolic Regression, Jack Robert Hutchins

Masters Theses

This thesis pioneers the integration of deep learning techniques into the realm of compact modeling, presenting three distinct approaches that enhance the precision, efficiency, and adaptability of compact models for electronic devices. The first method introduces a Generalized Multilayer Perception Compact Model, leveraging the function approximation capabilities of neural networks through a multilayer perception (MLP) framework. This approach utilizes hyperband tuning to optimize network hyperparameters, demonstrating its effectiveness on a HfOx memristor and establishing a versatile modeling strategy for both single-state and multistate devices.

The second approach explores the application of Mixture Density Networks (MDNs) to encapsulate the inherent stochasticity …

Learning Scene Semantics For 3d Scene Retrieval, Natalie Gleason

Honors Theses

This project presents a comprehensive exploration into semantics-driven 3D scene retrieval, aiming to bridge the gap between 2D sketches/images and 3D models. Through four distinct research objectives, this project endeavors to construct a foundational infrastructure, develop methodologies for quantifying semantic similarity, and advance a semantics-based retrieval framework for 2D scene sketch-based and image-based 3D scene retrieval. Leveraging WordNet as a foundational semantic ontology library, the research proposes the construction of an extensive hierarchical scene semantic tree, enriching 2D/3D scenes with encoded semantic information. The methodologies for semantic similarity computation utilize this semantic tree to bridge the semantic disparity between 2D …

Methods, Analyses, And Applications Of Multilayer Temporal Link Prediction In Networks, Xie He

Dartmouth College Ph.D Dissertations

Many applications stem from the possibility of accurately predicting links in various types of networks. In this thesis, we present methods, analyses, and applications for static, temporal, and multilayer networks. The first part of this thesis demonstrates how static network features serve as efficient and accurate predictors for link prediction in temporal networks. It includes an ensemble learning method we developed and presents experimental results on 90 synthetic stochastic block models and 19 real-world datasets. The second part closely follows, showcasing 20 different sampling methods and their effects on nine different link prediction algorithms for 250 real-world networks across 6 …

Exploring Tokenization Techniques To Optimize Patch-Based Time-Series Transformers, Gabriel L. Asher

Computer Science Senior Theses

Transformer architectures have revolutionized deep learning, impacting natural language processing and computer vision. Recently, PatchTST has advanced long-term time-series forecasting by embedding patches of time-steps to use as tokens for transformers. This study examines and seeks to enhance PatchTST's embedding techniques. Using eight benchmark datasets, we explore explore novel token embedding techniques. To this end, we introduce several PatchTST variants, which alter the embedding methods of the original paper. These variants consist of the following architectural changes: using CNNs to embed inputs to tokens, embedding an aggregate measure like the mean, max, or sum of a patch, adding the exponential …

Mri Image Regression Cnn For Bone Marrow Lesion Volume Prediction, Kevin Yanagisawa

Theses and Dissertations

Bone marrow lesions (BMLs), occurs from fluid build up in the soft tissues inside your bone. This can be seen on magnetic resonance imaging (MRI) scans and is characterized by excess water signals in the bone marrow space. This disease is commonly caused by osteoarthritis (OA), a degenerative join disease where tissues within the joint breakdown over time [1]. These BMLs are an emerging target for OA, as they are commonly related to pain and worsening of the diseased area until surgical intervention is required [2]–[4]. In order to assess the BMLs, MRIs were utilized as input into a regression …

Mitigating Safety Issues In Pre-Trained Language Models: A Model-Centric Approach Leveraging Interpretation Methods, Weicheng Ma

Dartmouth College Ph.D Dissertations

Pre-trained language models (PLMs), like GPT-4, which powers ChatGPT, face various safety issues, including biased responses and a lack of alignment with users' backgrounds and expectations. These problems threaten their sociability and public application. Present strategies for addressing these safety concerns primarily involve data-driven approaches, requiring extensive human effort in data annotation and substantial training resources. Research indicates that the nature of these safety issues evolves over time, necessitating continual updates to data and model re-training—an approach that is both resource-intensive and time-consuming. This thesis introduces a novel, model-centric strategy for understanding and mitigating the safety issues of PLMs by …

Using Pose Estimation Software To Predict Actions In Sabre Fencing, Micah Edwin Peters Ii

Honors College Theses

Fencing is a combat sport that uses three different swords: epee, foil, and sabre. Due to its fast-paced nature and employment of right of way, sabre fencing is often considered the most difficult of the three to learn. Computer vision and pose estimation software can be used to lower the barrier of entry to sabre fencing by identifying the different actions in sabre fencing. This project focuses on using open-source software to design a program that can identify the sabre parries as well as the main sabre movements. This program could be used to help newer fencers and spectators better …

Applications Of Genetic Algorithms To Chess, Elliot M. Harris

Senior Projects Spring 2024

This thesis discusses the use of genetic algorithms to tune the parameters of a chess engine, resulting in a significant increase in playing strength. The design of the genetic algorithms builds on the 2008-2011 work of David-Tabibi et al. and Vázquez-Fernández et al. The overwhelmingly positive result presented in this thesis not only suggests a promising potential for genetic algorithm use to improve computer chess, but also supports the efficacy and potential of applying genetic algorithms to a broader set of use cases.

Machine Learning And Natural Language Processing For Crossword Puzzles, Finn Brennan

Senior Projects Spring 2024

Senior Project submitted to The Division of Science, Mathematics and Computing of Bard College.

Towards Algorithmic Justice: Human Centered Approaches To Artificial Intelligence Design To Support Fairness And Mitigate Bias In The Financial Services Sector, Jihyun Kim

CMC Senior Theses

Artificial Intelligence (AI) has positively transformed the Financial services sector but also introduced AI biases against protected groups, amplifying existing prejudices against marginalized communities. The financial decisions made by biased algorithms could cause life-changing ramifications in applications such as lending and credit scoring. Human Centered AI (HCAI) is an emerging concept where AI systems seek to augment, not replace human abilities while preserving human control to ensure transparency, equity and privacy. The evolving field of HCAI shares a common ground with and can be enhanced by the Human Centered Design principles in that they both put humans, the user, at …

Secure And Privacy-Preserving Federated Learning With Rapid Convergence In Leo Satellite Networks, Mohamed Elmahallawy

Doctoral Dissertations

"The advancement of satellite technology has enabled the launch of small satellites equipped with high-resolution cameras into low Earth orbit (LEO), enabling the collection of extensive Earth data for training AI models. However, the conventional approach of downloading satellite-related data to a ground station (GS) for training a centralized machine learning (ML) model faces significant challenges. Firstly, the transmission of raw data raises security and privacy concerns, especially in military applications. Secondly, the download bandwidth is limited, which puts a stringent limit on image transmissions to the GS. Lastly, LEO satellites have sporadic visibility with the GS, and orbit the …

Ensemble Classification: An Analysis Of The Random Forest Model, Jarod Korn

Williams Honors College, Honors Research Projects

The random forest model proposed by Dr. Leo Breiman in 2001 is an ensemble machine learning method for classification prediction and regression. In the following paper, we will conduct an analysis on the random forest model with a focus on how the model works, how it is applied in software, and how it performs on a set of data. To fully understand the model, we will introduce the concept of decision trees, give a summary of the CART model, explain in detail how the random forest model operates, discuss how the model is implemented in software, demonstrate the model by …

Flexible Attenuation Fields: Tomographic Reconstruction From Heterogeneous Datasets, Clifford S. Parker

Theses and Dissertations--Computer Science

Traditional reconstruction methods for X-ray computed tomography (CT) are highly constrained in the variety of input datasets they admit. Many of the imaging settings -- the incident energy, field-of-view, effective resolution -- remain fixed across projection images, and the only real variance is in the detector's position and orientation with respect to the scene. In contrast, methods for 3D reconstruction of natural scenes are extremely flexible to the geometric and photometric properties of the input datasets, readily accepting and benefiting from images captured under varying lighting conditions, with different cameras, and at disparate points in time and space. Extending CT …

Advanced Mathematical Graph-Based Machine Learning And Deep Learning Models For Drug Design, Farjana Tasnim Mukta

Theses and Dissertations--Mathematics

Drug discovery is a highly complicated and time-consuming process. One of the main challenges in drug development is predicting whether a drug-like molecule will interact with a specific target protein. This prediction accelerates target validation and drug development. Recent research in biomolecular sciences has shown significant interest in algebraic graph-based models for representing molecular complexes and predicting drug-target binding affinity. In this thesis, we present algebraic graph-based molecular representations to create data-driven scoring functions (SF) using extended atom types to capture wide-range interactions between targets and drug candidates. Our model employs multiscale weighted colored subgraphs for the protein-ligand complex, colored …

On Vulnerabilities Of Building Automation Systems, Michael Cash

Graduate Thesis and Dissertation 2023-2024

Building automation systems (BAS) have become more commonplace in personal and commercial environments in recent years. They provide many functions for comfort and ease of use, from automating room temperature and shading, to monitoring equipment data and status. Even though their convenience is beneficial, their security has become an increased concerned in recent years. This research shows an extensive study on building automation systems and identifies vulnerabilities in some of the most common building communication protocols, BACnet and KNX. First, we explore the BACnet protocol, exploring its Standard BACnet objects and properties. An automation tool is designed and implemented to …

Enhancing Scanning Tunneling Microscopy With Automation And Machine Learning, Darian Smalley

Graduate Thesis and Dissertation 2023-2024

The scanning tunneling microscope (STM) is one of the most advanced surface science tools capable of atomic resolution imaging and atomic manipulation. Unfortunately, STM has many time-consuming bottlenecks, like probe conditioning, tip instability, and noise artificing, which causes the technique to have low experimental throughput. This dissertation describes my efforts to address these challenges through automation and machine learning. It consists of two main sections each describing four projects for a total of eight studies.

The first section details two studies on nanoscale sample fabrication and two studies on STM tip preparation. The first two studies describe the fabrication of …

Data Driven And Machine Learning Based Modeling And Predictive Control Of Combustion At Reactivity Controlled Compression Ignition Engines, Behrouz Khoshbakht Irdmousa

Dissertations, Master's Theses and Master's Reports

Reactivity Controlled Compression Ignition (RCCI) engines operates has capacity to provide higher thermal efficiency, lower particular matter (PM), and lower oxides of nitrogen (NOx) emissions compared to conventional diesel combustion (CDC) operation. Achieving these benefits is difficult since real-time optimal control of RCCI engines is challenging during transient operation. To overcome these challenges, data-driven machine learning based control-oriented models are developed in this study. These models are developed based on Linear Parameter-Varying (LPV) modeling approach and input-output based Kernelized Canonical Correlation Analysis (KCCA) approach. The developed dynamic models are used to predict combustion timing (CA50), indicated mean effective pressure (IMEP), …

An Unsupervised Machine Learning Algorithm For Clustering Low Dimensional Data Points In Euclidean Grid Space, Josef Lazar

Senior Projects Spring 2024

Clustering algorithms provide a useful method for classifying data. The majority of well known clustering algorithms are designed to find globular clusters, however this is not always desirable. In this senior project I present a new clustering algorithm, GBCN (Grid Box Clustering with Noise), which applies a box grid to points in Euclidean space to identify areas of high point density. Points within the grid space that are in adjacent boxes are classified into the same cluster. Conversely, if a path from one point to another can only be completed by traversing an empty grid box, then they are classified …

Improveing F-Beta Score In Classifying Shark Data Into Shark Behaviors, Ibrahim M. Ali

CGU Theses & Dissertations

One metric used to measure classification performance in machine learning is F-beta score. The objective in this thesis is to improve the average F-b score computed in classifying shark data into shark behaviors, namely; Resting, Swimming, Feeding, and Non-Directed Motion (NDM). Synthetic Minority Oversampling Technique (SMOTE) and Adaptive Synthetic Sampling (ADASYN) are utilized to balance the data, from which pre-processed Fast Fourier Transform (FFT), Walsh-Hadamard Transform (WHT), and Autocorrelation (AC) features are extracted then classified using Convolutional Neural Network (CNN) and K-Nearest Neighbors (K-NN). All the combinations of the two balancing techniques, the three feature types, and the two machine …