Computer Sciences Commons^™

Optimizing Deep Neural Networks Performance: Efficient Techniques For Training And Inference, Ankit Sharma

Graduate Thesis and Dissertation 2023-2024

Recent advances in computer vision tasks are mainly due to the success of large deep neural networks. The current state-of-the-art models have high computational costs during inference and suffer from a high memory footprint. Therefore, deploying these large networks on edge devices remains a serious concern. Furthermore, training these over-parameterized networks is computationally expensive and requires a longer training time. Thus, there is a demand to develop techniques that can efficiently reduce training costs and also be able to deploy neural networks on mobile and embedded devices. This dissertation presents practices like designing a lightweight network architecture and increasing network …

Deep Learning Approaches For Automatic Colorization, Super-Resolution, And Representation Of Volumetric Data, Sudarshan Devkota

Graduate Thesis and Dissertation 2023-2024

This dissertation includes a collection of studies that aim to improve the way we represent and visualize volume data. The advancement of medical imaging has revolutionized healthcare, providing crucial anatomical insights for accurate diagnosis and treatment planning. Our first study introduces an innovative technique to enhance the utility of medical images, transitioning from monochromatic scans to vivid 3D representations. It presents a framework for reference-based automatic color transfer, establishing deep semantic correspondences between a colored reference image and grayscale medical scans. This methodology extends to volumetric rendering, eliminating the need for manual intervention in parameter tuning. Next, it delves into …

Material Appearance Modeling For Physically Based Rendering, Alexis Benamira

Graduate Thesis and Dissertation 2023-2024

Photorealistic rendering focuses on creating images with a computer that imitates pictures of reallife scenes as faithfully as possible. To achieve this, rendering algorithms require incorporating accurate modeling of how light interacts with various types of matter. For most objects, this model needs to account for the scattering of the light rays. However, this model falls short when rendering objects of sizes smaller or comparable to the wavelength of the incident light. In this case, new phenomena such as diffraction or interference are observed and have been characterized in optics. Digital rendering of those phenomena involve different light representations than …

A Systematic Review Of Cryptocurrencies Use In Cybercrimes, Kieran B D Human

Graduate Thesis and Dissertation 2023-2024

Cryptocurrencies are one of the most prominent applications of blockchain systems. While cryptocurrencies promise many features and advantages, such as decentralization, anonymity, and ease of access, those very features can be abused. For instance, as documented in various recent works, cryptocurrencies have been frequently abused in many different forms of cybercrime. Despite the plethora of works on measuring and understanding the abuse of cryptocurrencies in the digital space, there has been no work on systemizing this knowledge by comprehensively understanding those contributions, contrasting them based on their merit, and understanding the gap in this research space.

This thesis initiates the …

Machine Learning Algorithms For Molecular Signature Identification With High-Throughput Genome Sequencing Data, Jiao Sun

Electronic Theses and Dissertations, 2020-

Powered by the high-throughput genomic technologies, the RNA sequencing (RNA-Seq) method is capable of measuring transcriptome-wide mRNA expressions and molecular activities in cells. Elucidation of gene expressions at the isoform resolution enables the detection of better molecular signatures for phenotype prediction, and the identified biomarkers may provide insights into the functional consequences of disease. This dissertation research focuses on developing advanced machine learning algorithms for mining large-scale RNA-Seq data in cancer transcriptome analysis. A platform-integrated model for transcript quantification (IntMTQ) is developed to improve the performance of RNA-Seq on isoform expression estimation. IntMTQ provides more precise RNA-Seq-based isoform quantification, and …

Algorithms And Variations On The Positional Burrows-Wheeler Transform And Their Applications, Ahsan Sanaullah

Electronic Theses and Dissertations, 2020-

In this dissertation, we develop algorithms and variations on the Positional Burrows-Wheeler Transform (PBWT). The PBWT is a data structure that stores M binary strings of length N while allowing efficient search. We develop the dynamic-PBWT (d-PBWT). The d-PBWT is a variation of the PBWT that allows its relevant algorithms to run with unchanged time complexity, but also allows efficient insertion and deletion of haplotypes. We provide insertion and deletion algorithms on the PBWT with average case O(N) time complexity. We also improve upon the query algorithms for the PBWT. Durbin described a set maximal match query algorithm on the …

Methodologies For Evaluating Interaction Cues For Virtual Reality, Xinyu Hu

Electronic Theses and Dissertations, 2020-

Virtual reality (VR) games and educational systems commonly employ interaction cues to provide information on how to take appropriate actions at particular moments. Interaction cues can be employed for different purposes, such as informing the user to look, go, pick, and operate. Additionally, different types of interaction cues can directly affect usability and user experiences. In our early research, we conducted two ecologically valid empirical studies with a preexisting VR training application and evaluated the effects of delayed interaction cues, in addition to comparing the purposes of interaction cues for learning and retention. Our results indicated that immediate interaction cues …

Discovering Vulnerabilities And Designing Trustworthy Defenses In Iot Systems And Devices, Bryan Pearson

Electronic Theses and Dissertations, 2020-

Internet of Things (IoT) dominates many functions in the modern world, from sensing and reporting temperature, humidity, and air quality, to controlling and automating homes, commercial buildings, and equipment. However, IoT systems have received scrutiny in recent years due to countless security incidents, which can have physical and even deadly consequences. This research provides a comprehensive assessment of the security of IoT systems and devices, including low-cost microcontroller (MCU) based sensors, cloud services, and Building Automation Systems (BAS). We begin by exploring the current landscape of vulnerabilities and defenses in modern IoT applications. We show that many security needs can …

Towards A Holistic And Comparative Analysis Of The Free Content Web: Security, Privacy, And Performance, Abdulrahman Alabduljabbar

Electronic Theses and Dissertations, 2020-

Free content websites that provide free books, music, games, movies, etc., have existed on the Internet for many years. While it is a common belief that such websites might be different from premium websites providing the same content types in terms of their security, a rigorous analysis that supports this belief is lacking from the literature. In particular, it is unclear if those websites are as safe as their premium counterparts. In this dissertation, we set out to investigate the similarities and differences between free content and premium websites, including their risk profiles. Moreover, we analyze and quantify through measurements …

Low-Resource Machine Learning Techniques For The Analysis Of Online Social Media Textual Data, Toktam Amanzadeh Oghaz

Electronic Theses and Dissertations, 2020-

Low-resource and label-efficient machine learning methods can be described as the family of statistical and machine learning techniques that can achieve high performance without needing a substantial amount of labeled data. These methods include both unsupervised learning techniques, such as LDA, and supervised methods, such as active learning, each providing different benefits. Thus, this dissertation is devoted to the design and analysis of unsupervised and supervised techniques to provide solutions for the following problems: Unsupervised narrative summary extraction for social media content, Social media text classification with Active Learning (AL), Investigating restrictions and benefits of using Curriculum Learning (CL) for …

Torward Real-World Cross-View Image Geo-Localization, Sijie Zhu

Electronic Theses and Dissertations, 2020-

Cross-view image geo-localization aims to determine the locations of street-view query images by searching in a GPS-tagged reference image database from aerial view. One fundamental challenge is the dramatic view-point/domain difference between the street-view query images and aerial-view reference images. Recent works have made great progress on bridging the domain gap with advanced deep learning techniques and geometric prior knowledge, i.e. the query is aligned at the center of one aerial-view reference image (spatial alignment) and the orientation relationship between the two views is known (orientation alignment). However, such prior knowledge of the geometry correspondence of the two views is …

Advanced Deep Learning Methodologies For Deepfake Detection, Aminollah Khormali

Electronic Theses and Dissertations, 2020-

The recent advances in the field of Artificial Intelligence (AI), particularly Generative Adversarial Networks (GANs) and an abundance of training samples along with robust computational resources have significantly propelled the field of AI-generated fake information in all kinds, e.g., deepfakes. Deepfakes are among the most sinister types of misinformation, posing large-scale and severe security and privacy risks targeting critical governmental institutions and ordinary people across the world. The fact that deepfakes are AI-generated digital content and not actual events captured by a camera implies that they still can be detected using advanced AI models. Although the deepfake detection task has …

Exploring The Privacy Dimension Of Wearables Through Machine Learning-Enabled Inference, Ulku Meteriz Yildiran

Electronic Theses and Dissertations, 2020-

Today's hyper-connected consumers demand convenient ways to tune into information without switching between devices, which led the industry leaders to the wearables. Wearables such as smartwatches, fitness trackers, and augmented reality (AR) glasses can be comfortably worn on the body. In addition, they offer limitless features, including activity tracking, authentication, navigation, and entertainment. Wearables that provide digestible information stimulate even higher consumer demand. However, to keep up with the ever-growing user expectations, developers keep adding new features and interaction methods to augment the use cases without considering their privacy impacts. In this dissertation, we explore the privacy dimension of wearables …

Distance Perception Through Head-Mounted Displays, Sina Masnadi

Electronic Theses and Dissertations, 2020-

It has been shown in numerous research studies that people tend to underestimate distances while wearing head-mounted displays (HMDs). We investigated various possible factors affecting the perception of distance is HMDs through multiple studies. Many contributing factors has been identified by researchers in the past decades, however, further investigation is required to provide a better understanding of this problem. In order to find a baseline for distance underestimation, we performed a study to compare the distance perception in real world versus a fake headset versus a see-through HMD. Users underestimated distances while wearing the fake headset or the see-through HMD. …

Methods For Defending Neural Networks Against Adversarial Attacks, Sharvil Shah

Electronic Theses and Dissertations, 2020-

Convolutional Neural Networks (CNNs) have been at the frontier of the revolution within the field of computer vision. Since the advent of AlexNet in 2012, neural networks with CNN architectures have surpassed human-level capabilities for many cognitive tasks. As the neural networks are integrated in many safety critical applications such as autonomous vehicles, it is critical that they are robust and resilient to errors. Unfortunately, it has recently been observed that deep neural network models are susceptible to adversarial perturbations which are imperceptible to human vision. In this thesis, we propose a solution to defend neural networks against white box …

Static Analysis Of The Build System To Accelerate Continuous Testing Of Highly Configurable Software, Necip Fazil Yildiran

Electronic Theses and Dissertations, 2020-

Continuous testing is widely used for facilitating fast and reliable software delivery. However, build-time configurability makes such testing harder for configurable software. As configurable software forms the basis of much of our computing infrastructure, there is even more need for better continuous testing for configurable software. In this dissertation, our goal is to improve the quality of configurable software. To this end, we tackle two, previously unsolved problems. The build system of configurable software is one of the biggest reasons why testing configurable software is hard. Therefore, in our solutions, we deal with the build system by using a comprehensive …

Towards Leveraging Sparse Infrared Datasets For Multiple View Synthesis, Few Shot Learning And Background Invariant Recognition, Maliha Arif

Electronic Theses and Dissertations, 2020-

This dissertation presents a study of various machine learning techniques for recognizing vehicular objects in infrared images. State of the art methods for computer vision have not been widely explored for this part of the electromagnetic spectrum (EM). Challenges that arise due to the dearth of infrared training images, terrain clutter, and thermal phenomenology have not been fully addressed. Infrared dataset collection and annotation is both difficult and expensive. What if there is a way we can generate infrared images and diminish the need for collecting data out in the field? Our first research study encompasses an encoder-decoder model that …

Algorithms For The Detection Of Resolved And Unresolved Targets In The Infrared Bands, Bruce Mcintosh

Electronic Theses and Dissertations, 2020-

This dissertation proposes algorithms for the detection of both resolved and unresolved targets in the infrared bands. Recent breakthroughs in deep learning have spurred major advancements in computer vision, but most of the attention and progress has been focused on RGB imagery from the visual band. The infrared bands such as Long Wave Infrared (LWIR), Medium Wave Infrared (MWIR), Short Wave Infrared (SWIR) and Near Infrared (NIR) each respond differently to physical phenomena, providing information that can be used to better understand the environment. The first task addressed is that of detecting vehicles in heavy clutter in MWIR imagery. A …

Applications For Machine Learning On Readily Available Data From Virtual Reality Training Experiences, Alec Moore

Electronic Theses and Dissertations, 2020-

The purpose of the research presented in this dissertation is to improve virtual reality (VR) training systems by enhancing their understanding of users. While the field of intelligent tutoring systems (ITS) has seen value in this approach, much research into making use of biometrics to improve user understanding and subsequently training, relies on specialized hardware. Through the presented research, I show that with machine learning (ML), the VR system itself can serve as that specialized hardware for VR training systems. I begin by discussing my explorations into using an ecologically valid, specialized training simulation as a testbed to predict knowledge …

Addressing Human-Centered Artificial Intelligence: Fair Data Generation And Classification And Analyzing Algorithmic Curation In Social Media, Amirarsalan Rajabi

Electronic Theses and Dissertations, 2020-

With the growing impact of artificial intelligence, the topic of fairness in AI has received increasing attention. Artificial intelligence is observed to have caused unanticipated negative consequences. In this dissertation, we address two critical aspects regarding human-centered artificial intelligence (HCAI), a new paradigm for developing artificial intelligence that is ethical, fair, and helps to improve the human condition. In the first part of this dissertation, we investigate the effect that AI curation of contents by social media platforms has on an online discussions, by studying a polarized discussion in the Twitter network. We then develop a network communication model that …

Methodology Of Augmented Reality Chinese Language Articulatory Pronunciation Practice: Game And Study Design, Daria Sinyagovskaya

Electronic Theses and Dissertations, 2020-

Learning a language can be hard. Learning a language that contains tones to convey meaning is even harder. This dissertation presents a novel methodology for creating a language practice using augmented reality that has never been used before. The design of a new app in AR and non-AR versions can evaluate the same practice methodology. This methodology was applied to new software and was examined in regard to the importance of this software. Although the study results are inconclusive, progress has been made in answering research questions on the effectiveness of AR versus non-AR and the reliability of peer assessment. …

A Human-Centered Approach To Improving Adolescent Online Sexual Risk Detection Algorithms, Afsaneh Razi

Electronic Theses and Dissertations, 2020-

Computational risk detection has the potential to protect especially vulnerable populations from online victimization. Conducting a comprehensive literature review on computational approaches for online sexual risk detection led to the identification that the majority of this work has focused on identifying sexual predators after-the-fact. Also, many studies rely on public datasets and third-party annotators to establish ground truth and train their algorithms, which do not accurately represent young social media users and their perspectives to prevent victimization. To address these gaps, this dissertation integrated human-centered approaches to both creating representative datasets and developing sexual risk detection machine learning models to …

Examining Cooperative System Responses Against Grid Integrity Attacks, Alexander D. Parady

Honors Undergraduate Theses

Smart grid technologies are integral to society’s transition to sustainable energy sources, but they do not come without a cost. As the energy sector shifts away from a century’s reliance on fossil fuels and centralized generation, technology that actively monitors and controls every aspect of the power infrastructure has been widely adopted, resulting in a plethora of new vulnerabilities that have already wreaked havoc on critical infrastructure. Integrity attacks that feedback false data through industrial control systems, which result in possible catastrophic overcorrections and ensuing failures, have plagued grid infrastructure over the past several years. This threat is now at …

Comparative Evaluation Of Assemblers For Metagenomic Data Analysis, Matheus Pavini Franco Ferreira

Honors Undergraduate Theses

Metagenomics is a cultivation-independent approach for obtaining the genomic composition of microbial communities. Microbial communities are ubiquitous in nature. Microbes which are associated with the human body play important roles in human health and disease. These roles span from protecting us against infections from other bacteria, to being the causes of these diseases. A deeper understanding of these communities and how they function inside our bodies allows for advancements in treatments and preventions for these diseases. Recent developments in metagenomics have been driven by the emergence of Next-Generation Sequencing technologies and Third-Generation Sequencing technologies that have enabled cost-effective DNA sequencing …

Translations To Support Loop Invariant Generation In Jml, Kohei Koja

Electronic Theses and Dissertations, 2020-

Software is used in many critical systems in the real world such as autonomous cars and medical devices. Such software must be reliable to protect the general public. One standard way to make reliable software is to use Hoare-style verification techniques. However, for Hoare-style verification of loop correctness, loop invariants are necessary but are difficult for people to write themselves. Since Java is one of the most popular programming languages in the world, it is useful to have a tool to generate loop invariants for Java programs. OpenJML is a widely used program verification tool for Java. However, it does …

Reverse Engineering Of Adversarial Samples By Leveraging Patterns Left By The Attacker, Rahul Ambati

Electronic Theses and Dissertations, 2020-

Intrinsic susceptibility of deep learning to adversarial examples has led to a plethora of attack techniques with a common broad objective of fooling deep models. However, we find slight compositional differences between the algorithms achieving this objective. These differences leave traces that provide important clues for attacker profiling in real-life scenarios. Inspired by this, we introduce a novel problem of 'Reverse Engineering of aDversarial attacks' (RED). Given an adversarial example, the objective of RED is to identify the attack used to generate it. Under this perspective, we can systematically group existing attacks into different families, leading to the sub-problem of …

Multivariate Cognitive Walkthrough Of Qubitvr: An Educational Quantum Computing, Virtual Reality Application, Pauline Johnson

Electronic Theses and Dissertations, 2020-

Quantum computing is a promising field but understanding how it works can be challenging for a beginner. There are also not many educational resources to visualize and learn about quantum computing. To advance knowledge in this area, we have created QubitVR, which employs a Bloch sphere representation of a qubit, and supports trajectory visualizations and state equations in a virtual reality (VR) setting. We also conducted a multivariate cognitive walkthrough with subject matter experts (SMEs) on QubitVR to assess the effectiveness of trajectory visualizations and state equations in learning about quantum gates. The results were that trajectory visualizations aided users …

Towards Automated Data Mining: Reinforcement Intelligence For Self-Optimizing Feature Engineering, Kunpeng Liu

Electronic Theses and Dissertations, 2020-

Feature engineering is one of the most important components in data mining and machine learning. One of the key thrusts in data mining is to answer: How should a low-dimensional geometry structure be extracted and reconstructed from high-dimensional data? To solve this issue, researchers proposed feature selection, PCA, sparsity regularization, factorization, embedding, and deep learning. However, existing techniques are limited in achieving full automation, globally optimal, and explainable explicitness. Can I address the automation, optimal, and explainability challenges in data geometry reconstruction? A low-dimensional data geometry structure is crucial for SciML methods (e.g., GP models), and the accuracy of these …

Visual Question Answering: Exploring Trade-Offs Between Task Accuracy And Explainability, Aisha Urooj

Electronic Theses and Dissertations, 2020-

Given visual input and a natural language question about it, the visual question answering (VQA) task is to answer the question correctly. To improve a system's reliability and trustworthiness, it is imperative that it links the text (question and answer) to specific visual regions. This dissertation first explores the VQA task in a multi-modal setting where questions are based on video as well as subtitles. An algorithm is introduced to process each modality and their features are fused to solve the task. Additionally, to understand the model's emphasis on visual data, this study collects a diagnostic set of questions which …

How Adolescents In The Child Welfare System Seek Support For Their Sexual Risk Experiences Online, Taylor L. Moraguez

Honors Undergraduate Theses

Youth in the foster care system experience unique and challenging situations online, such as higher risks of inappropriate messaging (e.g., sexting) and unwanted solicitations from strangers. As a vulnerable group of adolescents, foster youth often use online platforms as a resource to express themselves and seek support over their sexual experiences online. This thesis analyzes how foster youth seek support online for their sexual risk experiences, including sexual abuse, sexting, and sexuality. To understand how adolescents (ages 13-17) in the child welfare system seek support for these experiences, we conducted a thematic analysis of 541 individual posts made by 121 …