Computer Sciences Commons^™

Extending The Convolution In Graph Neural Networks To Solve Materials Science And Node Classification Problems, Steph-Yves Mike Louis

Theses and Dissertations

The usage of graph to represent one's data in machine learning has grown in popularity in both academia and the industry due to its inherent benefits. With its flexible nature and immediate translation to real life observed objects, graph representation had a considerable contribution in advancing the state-of-the-art performance of machine learning in materials.

In this dissertation proposal, we discuss how machines can learn from graph encoded data and provide excellent results through graph neural networks (GNN). Notably, we focus our adaptation of graph neural networks on three tasks: predicting crystal materials properties, nullifying the negative impact of inferior graph …

Predicting Material Structures And Properties Using Deep Learning And Machine Learning Algorithms, Yuqi Song

Theses and Dissertations

Discovering new materials and understanding their crystal structures and chemical properties are critical tasks in the material sciences. Although computational methodologies such as Density Functional Theory (DFT), provide a convenient means for calculating certain properties of materials or predicting crystal structures when combined with search algorithms, DFT is computationally too demanding for structure prediction and property calculation for most material families, especially for those materials with a large number of atoms. This dissertation aims to address this limitation by developing novel deep learning and machine learning algorithms for effective prediction of material crystal structures and properties. Our data-driven machine learning …

Flying By Ml Or Cnn Inversion Of Affine Transforms, Lloyd Van Warren

Theses and Dissertations

This dissertation describes how to automate the reading of dials, gauges, and instruments using machine learning (ML) methods. This process can be described as analog to digital conversion through an air gap without any direct electronic connection. The goal is to convert the values of existing instruments into digital values for control and monitoring without requiring any intrusion, changes, or upgrades to the instruments being observed. Images of instrument faces can be distorted by various kinds of noise, but this can be overcome using a deep learning convolutional neural network (CNN) approach similar to handwriting recognition. One advantage of this …

The Use Of Artificial Intelligence In Higher Education: A Study On Faculty Perspectives In Universities In Egypt, Farah S. Sharawy

Theses and Dissertations

Artificial Intelligence (AI) is an emerging technology that is transforming various aspects of society, including higher education. This paper examines faculty perspectives from five different institutions; The American University in Cairo (AUC), The German University in Cairo (GUC), The Arab Academy for Science and Technology (AAST), Ain Shams University, and Cairo University, on the use of AI in higher education in teaching and learning in Egypt, with all its challenges and resources available to support it, and how it can be used to achieve equity and accessibility. This research was conducted through a qualitative study using semi-structured one- on-one interviews …

Adversary Aware Continual Learning, Muhammad Umer

Theses and Dissertations

Continual learning approaches are useful as they help the model to learn new information (classes) sequentially, while also retaining the previously acquired information (classes). However, these approaches are adversary agnostic, i.e., they do not consider the possibility of malicious attacks. In this dissertation, we have demonstrated that continual learning approaches are extremely vulnerable to the adversarial backdoor attacks, where an intelligent adversary can introduce small amount of misinformation to the model in the form of imperceptible backdoor pattern during training to cause deliberate forgetting of a specific class at test time. We then propose a novel defensive framework to counter …

Computation Offloading Design For Deep Neural Network Inference On Iot Devices, Asmika Boosarapu

Theses and Dissertations

In recent times, advances in the technologies of Internet-of-Things (IoT) and Deep Neural Networks (DNN) have significantly increased the accuracy and speed of a variety of smart applications. However, one of the barriers to deploying DNN to IoT is the computational limitations of IoT devices as compared with the computationally expensive task of DNN inference. Computation offloading is an approach that addresses this problem by offloading DNN computation tasks to cloud servers. In this thesis we propose a collaborative computation offloading solution, in which some of the work is done on the IoT device, and the remainder of the work …

A Multimodal Immune System Inspired Defense Architecture For Detecting And Deterring Digital Pathogens In Container Hosted Web Services, Islam Khalil

Theses and Dissertations

With the increased use of web technologies, microservices, and Application Programming Interface (API) for integration between systems, and with the development of containerization of services on operating system level as a method of isolating system execution and for easing the deployment and scaling of systems, there is a growing need as well as opportunities for providing platforms that improve the security of such services. In our work, we propose an architecture for a containerization platform that utilizes various concepts derived from the human immune system. The goal of the proposed containerization platform is to introduce the concept of slowing down …

A Machine Learning Model To Predict The Asd Traits And Challenging Behaviors In Children, Biswaranjan Senapati

Theses and Dissertations

A neurological disorder, along with several behavioral issues, may be to blame for a child's subpar performance in the academic journey (such as anxiety, depression, learning disorders, and irritability). These symptoms can be used to diagnose children with ASD, and supervised machine learning models can help differentiate between ASD traits and other behavioral traits in children, so that a correct diagnosis can be made. Autism usually shows up in a child within the first two years of life, but it can happen at any time. This research aims to create a model for predicting ASD traits and challenging behaviors in …

Pruning Ghsom To Create An Explainable Intrusion Detection System, Thomas Michael Kirby

Theses and Dissertations

Intrusion Detection Systems (IDS) that provide high detection rates but are black boxes lead
to models that make predictions a security analyst cannot understand. Self-Organizing Maps
(SOMs) have been used to predict intrusion to a network, while also explaining predictions through
visualization and identifying significant features. However, they have not been able to compete with
the detection rates of black box models. Growing Hierarchical Self-Organizing Maps (GHSOMs)
have been used to obtain high detection rates on the NSL-KDD and CIC-IDS-2017 network traffic
datasets, but they neglect creating explanations or visualizations, which results in another black
box model.
This paper offers …

Eddy Current Defect Response Analysis Using Sum Of Gaussian Methods, James William Earnest

Theses and Dissertations

This dissertation is a study of methods to automatedly detect and produce approximations of eddy current differential coil defect signatures in terms of a summed collection of Gaussian functions (SoG). Datasets consisting of varying material, defect size, inspection frequency, and coil diameter were investigated. Dimensionally reduced representations of the defect responses were obtained utilizing common existing reduction methods and novel enhancements to them utilizing SoG Representations. Efficacy of the SoG enhanced representations were studied utilizing common Machine Learning (ML) interpretable classifier designs with the SoG representations indicating significant improvement of common analysis metrics.

Monolithic Multiphysics Simulation Of Hypersonic Aerothermoelasticity Using A Hybridized Discontinuous Galerkin Method, William Paul England

Theses and Dissertations

This work presents implementation of a hybridized discontinuous Galerkin (DG) method for robust simulation of the hypersonic aerothermoelastic multiphysics system. Simulation of hypersonic vehicles requires accurate resolution of complex multiphysics interactions including the effects of high-speed turbulent flow, extreme heating, and vehicle deformation due to considerable pressure loads and thermal stresses. However, the state-of-the-art procedures for hypersonic aerothermoelasticity are comprised of low-fidelity approaches and partitioned coupling schemes. These approaches preclude robust design and analysis of hypersonic vehicles for a number of reasons. First, low-fidelity approaches limit their application to simple geometries and lack the ability to capture small scale flow …

Development And Evaluation Of An Automated Tactical Tillage Tool To Control Weeds In Row-Crop Production Systems, Grace Mccormick Friday

Theses and Dissertations

Weed control is an integral part of a successful overall production strategy in row- cropping systems and has the potential to reduce or eliminate yield losses that negatively affect profitability. Timely and correctly selected herbicide applications are the major keys for effective weed control in a majority of instances. However, there are negative factors that contribute to ineffectiveness and weed escape issues that currently lack viable options for management. Sparsely populated late-season weeds that emerge after lay-by herbicide applications and weeds that have become tolerant and resistant to traditional herbicide chemistries are of greatest concern. Historically, these weeds would have …

Beyond Algorithms: A User-Centered Evaluation Of A Feature Recommender System In Requirements Engineering, Oluwatobi Lasisi

Theses and Dissertations

Several studies have applied recommender technologies to support requirements engineering activities. As in other application areas of recommender systems (RS), many studies have focused on the algorithms’ prediction accuracy, while there have been limited discussions around users’ interactions with the systems. Since recommender systems are designed to aid users in information retrieval, they should be assessed not just as recommendation algorithms but also from the users’ perspective. In contrast to accuracy measures, user-related issues can only be effectively investigated via empirical studies involving real users. Furthermore, researchers are becoming increasingly aware that the effectiveness of the systems goes beyond recommendation …

Secure And Efficient Federated Learning, Xingyu Li

Theses and Dissertations

In the past 10 years, the growth of machine learning technology has been significant, largely due to the availability of large datasets for training. However, gathering a sufficient amount of data on a central server can be challenging. Additionally, with the rise of mobile networking and the large amounts of data generated by IoT devices, privacy and security issues have become a concern, resulting in government regulations such as GDPR, HIPAA, CCPA, and ADPPA. Under these circumstances, traditional centralized machine learning methods face a problem in that sensitive data must be kept locally for privacy reasons, making it difficult to …

Tornado Outbreak False Alarm Probabilistic Forecasts With Machine Learning, Kirsten Reed Snodgrass

Theses and Dissertations

Tornadic outbreaks occur annually, causing fatalities and millions of dollars in damage. By improving forecasts, the public can be better equipped to act prior to an event. False alarms (FAs) can hinder the public’s ability (or willingness) to act. As such, a probabilistic FA forecasting scheme would be beneficial to improving public response to outbreaks.

Here, a machine learning approach is employed to predict FA likelihood from Storm Prediction Center (SPC) tornado outbreak forecasts. A database of hit and FA outbreak forecasts spanning 2010 – 2020 was developed using historical SPC convective outlooks and the SPC Storm Reports database. Weather …

An Analysis Of Text-Based Machine Learning Models For Vulnerability Detection, Kollin Ryne Napier

Theses and Dissertations

With an increase in complexity of software, developers rely more on reuse and dependencies in their source code via code snippets. As a result, it is becoming harder to identify and mitigate vulnerabilities. Although traditional analysis tools are still utilized, machine learning models are being adopted to expand efforts and combat such threats. Given the possibilities towards usage of such models, research in this area has introduced various approaches which vary in usability and prediction. In generalizing models to a more natural language approach, researchers have opted to train models on source code to identify existing and potential vulnerabilities. Exploratory …

Utilizing Machine Learning Techniques To Predict Credit Card Payment Defaults, Madison Guerra

Theses and Dissertations

The question of accurately predicting credit card defaulters has been explored in numerous studies in the past. In these studies, the researchers utilized various machine learning theories and techniques to make the determination the extent of defaults. Unfortunately, some constraints were encountered, and the limitations that existed from the previous works have been discussed. This project attempted to address these issues with special attention given to more recently available data. Specifically, in this project, we looked at data provided by one Kaggle user, which utilized the data from the American Express credit card competition, which ranges from late March 2018 …

Topic Modeling And Future Prediction Of Aid Data In Development Studies Using Lda And Bert, Uttamasha Anjally Oyshi

Theses and Dissertations

This thesis presents a study on topic modeling and future prediction of aid data in development studies using LDA and BERT. The goal of this study is to explore the aid data from four sectors: Government and civil society, Government and civil society general, Conflict prevention and resolution, peace and security, and Women, and to identify the latent topics that are present in the data. The LDA and BERT algorithms were used for the topic modeling, and coherence scores were computed to evaluate the quality of the models. The results of the study show that the LDA and BERT models …

Modeling User Behavior For Cyber Security With Formal Methods And Agent Based Simulation, Hamdah Albalawi

Theses and Dissertations

Despite society’s positive outlook, technology poses real cyber security threats. Technology’s benefits can sometimes make it difficult to believe that potential threats lurk behind every device and platform. As cybercrime rises, we have come to rely increasingly on flawed devices and services. As cyberattacks become more prevalent, security professionals are committed to developing more robust and dependable security solutions. In cyber security, human error is regarded as the weakest link since all technical security solutions are vulnerable to human error. Among other human characteristics, risk-taking, logical decision-making, extraversion, and gender can significantly affect cyber security. However, there still is the …

Towards Privacy-Preserving Social Media Networks: Protecting The Facial Privacy Of Images Uploaded On Social Media, Ahsi Lo

Theses and Dissertations

Since the 2000s, social media has allowed individuals the ability to communicate online. As the popularity of social media increased, the sharing of information such as pictures increased as well. Recently, there have been privacy concerns about the information shared online such as cases where third parties were able to gain access to users’ information without being given explicit access through scraping or other means. When user images are scraped from social media, there is a risk that these individuals can be identified o✏ine. Bystanders, who may be captured in images also run this risk of identification. This research investigates …

Analysis Of Post-Translational Modifications (Ptm) Crosstalk, Amit Das

Theses and Dissertations

Mass spectrometry-based proteomics is a powerful tool for identifying post-translational modifications (PTMs) across the proteome. O-GlcNAcylation and phosphorylation are two PTMs that play crucial roles in regulating cellular processes, including cardiac contractile function. Dysregulation of these PTMs has been implicated in the development and progression of diabetic cardiomyopathy. In this study, we aimed to investigate the interplay between O-GlcNAcylation and phosphorylation in healthy and type 2 diabetic hearts, with a specific focus on the functional relationships between these PTMs and their potential therapeutic implications.

Utilizing mass spectrometry data, we identified and quantified specific PTMs on myofilament proteins, uncovering 1354 O-GlcNAcylated …

Self-Supervised Representation Learning For Motion Time Series: A Case Study In Activity Recognition, Luis Carlos Garza Perez

Theses and Dissertations

In this thesis we will learn about what contrastive learning and time series are and understand the differences between supervised and self-supervised frameworks in machine learning. In addition, we will describe how the newest and most efficient self-supervised learning framework for visual representations to this date works, called SimCLR, which was originally developed to obtain useful vector representations from static images. We will also explain what TS2Vec is, and how a combination of both approaches can be applied to the concept of a time series, and still be able to extract a vector representation of the subject described by the …

The Effect Of Cybersecurity Training On Government Employee’S Knowledge Of Cybersecurity Issues And Practices, Juan Jaime Saldana Ii

Theses and Dissertations

There is an ever-pressing need for cybersecurity awareness and implementation of learning strategies in the workplace to mitigate the increased threat posed by cyber-attacks and exacerbated by an untrained workforce. The lack of cybersecurity knowledge amongst government employees has increased to critical levels due to the amount of sensitive information their agencies are responsible for. The digital compromise of a government entity often leads to a compromise of constituent data along with the disruption of public services (Axelrod, 2019; Yazdanpanahi, 2021). The need for awareness is further complicated by agencies looking to cater to a digital culture looking for a …

Problems In Algorithmic Self-Assembly And A Genetic Approach To Patterns, Andrew Rodriguez

Theses and Dissertations

As it becomes increasingly harder to make transistors smaller, replacements for traditional silicon computers become sought after. To study the computing power of these potential computers, various theoretical models have been proposed, such as the abstract Tile Assembly Model (aTAM) and chemical reaction networks (CRNs). This thesis compiles research in various models such as the aTAM, Tile Automata, and CRNs. This work shows an investigation of covert computation in the aTAM and an evolutionary algorithm to approximate solutions to the pattern self-assembly tile set synthesis (PATS) problem. Next, optimal state complexity for building squares in Tile Automata is shown along …

Nonlinear Mathematical Transformations For Improved Image And Signal Recovery Using Artificial Neural Network, Haoran Chang

Theses and Dissertations

Medical imaging plays a vital role in modern healthcare, enabling clinicians to diagnose and treat a range of conditions. However, image acquisition and processing can be challenging because they can often be hindered by motion blurring, leading to inaccurate results. To address that, this dissertation proposes a novel approach based on nonlinear mathematical transformations and artificial neural networks (ANN). The dissertation begins with an introduction to Nuclear Medicine and the problem of motion blur in image reconstruction. A background on Medical Imaging techniques, including the Radon transform and Image Reconstruction methods such as Filtered Back Projection and Iterative Reconstruction methods …

Results And Simulation Of Active Self-Assembly, Robert M. Alaniz

Theses and Dissertations

Self-assembly is the process by which simple elements in a system organize themselves into more complex structures based on a set of rules that govern their interactions. With many ways to create self-assembling systems, new models of abstraction have also arisen to handle specific mechanisms and procedures. We explore several open problems in the seeded model of Active Self-Assembly, Chemical Reaction Networks, and Surface Chemical Reaction Networks, proving new results while developing a robust simulation environment, AutoTile, to help build and test these results.

Assessing The Effect Of Atmospheric Turbulence On Long-Range Face Recognition Accuracy, Muskan Jain

Theses and Dissertations

Recent investigations have demonstrated that it might be challenging to identify faces in the images taken using a long-distance camera. A face seems blurry in these images because of the presence of atmospheric turbulence. To examine how atmospheric turbulence impacts face biometrics, we establish a simulated environment that exhibits different degrees of turbulence. We employed the Rytov Variance, which relies on the distance and refractive index, C2 n, to get various turbulence levels. We used the LRFID dataset to carry out the study, which is a collection of photos and videos taken in the field and in a controlled setting. …

Towards Explainable Ai: Predicting Linear And Nonlinear Feature Relations Of Datasets, Nimmy Chhaganbhai Patel

Theses and Dissertations

Artificial Intelligence (AI) makes critical decisions in an opaque way without explaining the reasoning behind them. Decision support systems are often built as black boxes. This has generated interest in Explainable AI (XAI), an area of research that explains AI algorithms and provides more insight into their internal decision-making process. Advancements in XAI have enhanced machine learning (ML) models’ interpretability, explainability, and transparency. Additionally, there has been speculation on whether it is possible to predict the type of dataset used in the model by analyzing the results. In this research, we proposed a methodology for determining the linearity or non-linearity …

From Point Estimates To Predictive Distributions In Machine Learning Models - A Statistical Importance Sampling Framework, Giuseppina Carannante

Theses and Dissertations

In this thesis, we leverage powerful statistical frameworks for optimal sequential estimation and tracking in non-linear and non-Gaussian dynamical models, which enjoy proven (asymptotic) optimality properties. Initially, we build upon our previous work, which employed first-order Taylor series approximation to propagate the first two predictive moments, to derive Bayesian encoder-decoder networks. This work introduced the notion of dense, pixel-level uncertainty map that is crucial in fields, such as autonomous vehicles and medical segmentation. We then extended the Bayesian framework to an ensembling scheme based on ensemble Kalman Filtering (EnKF). While EnKF represents the predictive distribution with an ensemble of draws, …

Semantics-Based Data Security Models, Theppatorn Rhujittawiwat

Theses and Dissertations

In this dissertation, we studied how an adversary could attack databases and how the system could prevent or recover from such an attack. Our motivation to improve the current security capabilities of database management systems. We provided better recovery capabilities of database management systems by incorporating data provenance. We also expand our study to express security and privacy needs of data in the Internet of Things (IoT) environments such as a smart home environment. For this, we proposed a stream data security model to theoretically represent the data in the IoT network. We built a dynamic authorization model on our …