Digital Commons Network^™

A Deep Machine Learning Approach For Predicting Freeway Work Zone Delay Using Big Data, Abdullah Shabarek

Dissertations

The introduction of deep learning and big data analytics may significantly elevate the performance of traffic speed prediction. Work zones become one of the most critical factors causing congestion impact, which reduces the mobility as well as traffic safety. A comprehensive literature review on existing work zone delay prediction models (i.e., parametric, simulation and non-parametric models) is conducted in this research. The research shows the limitations of each model. Moreover, most previous modeling approaches did not consider user delay for connected freeways when predicting traffic speed under work zone conditions. This research proposes Deep Artificial Neural Network (Deep ANN) and …

Human Activity Recognition Using Wearable Sensors: A Deep Learning Approach, Jialun Xue

Theses

In the past decades, Human Activity Recognition (HAR) grabbed considerable research attentions from a wide range of pattern recognition and human–computer interaction researchers due to its prominent applications such as smart home health care. The wealth of information requires efficient classification and analysis methods. Deep learning represents a promising technique for large-scale data analytics. There are various ways of using different sensors for human activity recognition in a smartly controlled environment. Among them, physical human activity recognition through wearable sensors provides valuable information about an individual’s degree of functional ability and lifestyle. There is abundant research that works upon real …

Leveraging The Inductive Bias Of Large Language Models For Abstract Textual Reasoning, Christopher Michael Rytting

Theses and Dissertations

Large natural language models (such as GPT-2 or T5) demonstrate impressive abilities across a range of general NLP tasks. Here, we show that the knowledge embedded in such models provides a useful inductive bias, not just on traditional NLP tasks, but also in the nontraditional task of training a symbolic reasoning engine. We observe that these engines learn quickly and generalize in a natural way that reflects human intuition. For example, training such a system to model block-stacking might naturally generalize to stacking other types of objects because of structure in the real world that has been partially captured by …

Forecasting Bitcoin Prices Using N-Beats Deep Learning Architecture, Alikhan Bulatov

Student Theses

The use of computationally intensive systems that employ machine learning algorithms is increasingly common in the field of finance. New state of the art deep learning architectures for time series forecasting are being developed each year making them more accurate than ever. This study evaluates the predictive power of the N-BEATS deep learning architecture trained on Bitcoin daily, hourly, and up-to-the-minute data in comparison with other popular time series forecasting methods such as LSTM and ARIMA. Prediction errors are measured with Mean Average Percentage Error (MAPE), and Root Mean Squared Error (RMSE). The results suggest that the developed N-BEATS model …

Methods For Generative Adversarial Output Enhancement, Michael B. Brodie

Theses and Dissertations

Generative Adversarial Networks (GAN) learn to synthesize novel samples for a given data distribution. While GANs can train on diverse data of various modalities, the most successful use cases to date apply GANs to computer vision tasks. Despite significant advances in training algorithms and network architectures, GANs still struggle to consistently generate high-quality outputs after training. We present a series of papers that improve GAN output inference qualitatively and quantitatively. The first chapter, Alpha Model Domination, addresses a related subfield of Multiple Choice Learning, which -- like GANs -- aims to generate diverse sets of outputs. The next chapter, CoachGAN, …

Metarec: Meta-Learning Meets Recommendation Systems, James Le

Theses

Artificial neural networks (ANNs) have recently received increasing attention as powerful modeling tools to improve the performance of recommendation systems. Meta-learning, on the other hand, is a paradigm that has re-surged in popularity within the broader machine learning community over the past several years. In this thesis, we will explore the intersection of these two domains and work on developing methods for integrating meta-learning to design more accurate and flexible recommendation systems.

In the present work, we propose a meta-learning framework for the design of collaborative filtering methods in recommendation systems, drawing from ideas, models, and solutions from modern approaches …

Automated Intelligent Cueing Device To Improve Ambient Gait Behaviors For Patients With Parkinson's Disease, Nader Naghavi

Doctoral Dissertations

Freezing of gait (FoG) is a common motor dysfunction in individuals with Parkinson’s disease (PD). FoG impairs walking and is associated with increased fall risk. Although pharmacological treatments have shown promise during ON-medication periods, FoG remains difficult to treat during medication OFF state and in advanced stages of the disease. External cueing therapy in the forms of visual, auditory, and vibrotactile, has been effective in treating gait deviations. Intelligent (or on-demand) cueing devices are novel systems that analyze gait patterns in real-time and activate cues only at moments when specific gait alterations are detected. In this study we developed methods …

Nature-Inspired Topology Optimization Of Recurrent Neural Networks, Abdelrahman A. Elsaid

Theses

Hand-crafting effective and efficient structures for recurrent neural networks (RNNs) is a difficult, expensive, and time-consuming process. To address this challenge, this work presents three nature-inspired (NI) algorithms for neural architecture search (NAS), introducing the subfield of nature-inspired neural architecture search (NI-NAS). These algorithms, based on ant colony optimization (ACO), progress from memory cell structure optimization, to bounded discrete-space architecture optimization, and finally to unbounded continuous-space architecture optimization. These methods were applied to real-world data sets representing challenging engineering problems, such as data from a coal-fired power plant, wind-turbine power generators, and aircraft flight data recorder (FDR) data.

Initial work …

Multiphoton Microscopy And Deep Learning Neural Networks For The Automated Quantification Of In Vivo, Label-Free Optical Biomarkers Of Skin Wound Healing, Jake D. Jones

Graduate Theses and Dissertations

Non-healing ulcerative wounds that occur frequently in diseases such as diabetes are challenging to diagnose and treat due to numerous possible etiologies and the variable efficacy of wound care products. With advanced age, skin wound healing is often delayed, leaving elderly patients at high risk for developing these chronic injuries. As it is challenging to discriminate age-related delays from disease-related chronicity, there is a critical need to develop new quantitative biomarkers that are sensitive to wound status. Multiphoton microscopy (MPM) techniques are well-suited for 3D imaging of epithelia and are capable of non-invasively detecting metabolic cofactors (NADH and FAD) without …

Acquisition, Processing, And Analysis Of Video, Audio And Meteorological Data In Multi-Sensor Electronic Beehive Monitoring, Sarbajit Mukherjee

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

In recent years, a widespread decline has been seen in honey bee population and this is widely attributed to colony collapse disorder. Hence, it is of utmost importance that a system is designed to gather relevant information. This will allow for a deeper understanding of the possible reasons behind the above phenomenon to aid in the design of suitable countermeasures.

Electronic Beehive Monitoring is one such way of gathering critical information regarding a colony’s health and behavior without invasive beehive inspections. In this dissertation, we have presented an electronic beehive monitoring system called BeePi that can be placed on top …

Lightweight Deep Learning For Botnet Ddos Detection On Iot Access Networks, Eric A. Mccullough

MSU Graduate Theses

With the proliferation of the Internet of Things (IoT), computer networks have rapidly expanded in size. While Internet of Things Devices (IoTDs) benefit many aspects of life, these devices also introduce security risks in the form of vulnerabilities which give hackers billions of promising new targets. For example, botnets have exploited the security flaws common with IoTDs to gain unauthorized control of hundreds of thousands of hosts, which they then utilize to carry out massively disruptive distributed denial of service (DDoS) attacks. Traditional DDoS defense mechanisms rely on detecting attacks at their target and deploying mitigation strategies toward the attacker …

Unsupervised Structural Graph Node Representation Learning, Mikel Joaristi

Boise State University Theses and Dissertations

Unsupervised Graph Representation Learning methods learn a numerical representation of the nodes in a graph. The generated representations encode meaningful information about the nodes' properties, making them a powerful tool for tasks in many areas of study, such as social sciences, biology or communication networks. These methods are particularly interesting because they facilitate the direct use of standard Machine Learning models on graphs. Graph representation learning methods can be divided into two main categories depending on the information they encode, methods preserving the nodes connectivity information, and methods preserving nodes' structural information. Connectivity-based methods focus on encoding relationships between nodes, …

Trace: A Differentiable Approach To Line-Level Stroke Recovery For Offline Handwritten Text, Taylor Neil Archibald

Theses and Dissertations

Stroke order and velocity are helpful features in the fields of signature verification, handwriting recognition, and handwriting synthesis. Recovering these features from offline handwritten text is a challenging and well-studied problem. We propose a new model called TRACE (Trajectory Recovery by an Adaptively-trained Convolutional Encoder). TRACE is a differentiable approach using a convolutional recurrent neural network (CRNN) to infer temporal stroke information from long lines of offline handwritten text with many characters. TRACE is perhaps the first system to be trained end-to-end on entire lines of text of arbitrary width and does not require the use of dynamic exemplars. Moreover, …

Traffic Time Headway Prediction And Analysis: A Deep Learning Approach, Saumik Sakib Bin Masud

Theses and Dissertations

In the modern world of Intelligent Transportation System (ITS), time headway is a key traffic flow parameter affecting ITS operations and planning. Defined as “the time difference between any two successive vehicles when they cross a given point”, time headway is used in various traffic and transportation engineering research domains, such as capacity analysis, safety studies, car-following, and lane-changing behavior modeling, and level of service evaluation describing stochastic features of traffic flow. Advanced travel and headway information can also help road users avoid traffic congestion through dynamic route planning, for instance. Hence, it is crucial to accurately model headway distribution …

Deep Learning-Based, Passive Fault Tolerant Control Facilitated By A Taxonomy Of Cyber-Attack Effects, Dean C. Wardell

Theses and Dissertations

In the interest of improving the resilience of cyber-physical control systems to better operate in the presence of various cyber-attacks and/or faults, this dissertation presents a novel controller design based on deep-learning networks. This research lays out a controller design that does not rely on fault or cyber-attack detection. Being passive, the controller’s routine operating process is to take in data from the various components of the physical system, holistically assess the state of the physical system using deep-learning networks and decide the subsequent round of commands from the controller. This use of deep-learning methods in passive fault tolerant control …

New Methods For Deep Learning Based Real-Valued Inter-Residue Distance Prediction, Jacob Barger

Theses

Background: Much of the recent success in protein structure prediction has been a result of accurate protein contact prediction--a binary classification problem. Dozens of methods, built from various types of machine learning and deep learning algorithms, have been published over the last two decades for predicting contacts. Recently, many groups, including Google DeepMind, have demonstrated that reformulating the problem as a multi-class classification problem is a more promising direction to pursue. As an alternative approach, we recently proposed real-valued distance predictions, formulating the problem as a regression problem. The nuances of protein 3D structures make this formulation appropriate, allowing predictions …

Deep Learning Application On American Sign Language Database For Video-Based Gesture Recognition, Muhammad Murtaza Saleem

Student Theses

ASL speaking individuals always bring a companion as a translator [1]. This creates barriers for those who wish to take part in activities alone. Online translators exist however, they are limited to the individual characters instead of the gestures which group characters in a meaningful way, and connectivity is not always accessible. Thus, this research tackles the limitations of existing technologies and presents a model, implemented in MATLAB 2020b, to be used for predicting and classifying American sign language gestures/characters. The proposed method looks further into current neural networks and how they can be utilized against our transformed World Largest …

Organ Segmentation Of Pediatric Computed Tomography (Ct) With Generative Adversarial Networks, Chi Nok Enoch Kan

Master's Theses (2009 -)

Accurately segmenting organs in abdominal computed tomography (CT) is crucial for many clinical applications such as organ-specific dose estimation. With the recent emergence of deep learning techniques for computer vision, many powerful frameworks are proposed for organ segmentation in abdominal CT images. A major problem with these state-of-the-art methods is that they depend on large amounts of training data to achieve high segmentation accuracy. Pediatric abdominal CTs are particularly hard to obtain since these children are much more sensitive to ionizing radiation than adults. It is extremely challenging to train automatic segmentation algorithms on pediatric CT volumes. To address these …

Deep Learning For Quantitative Susceptibility Mapping Reconstruction, Juan Liu

Dissertations (1934 -)

Quantitative susceptibility mapping (QSM) is a magnetic resonance imaging (MRI) technique that estimates tissue magnetic susceptibility from Larmor frequency offset measurements. The generation of QSM requires solving ill-posed background field removal (BFR) and field-to-source inversion problems. Incorrect BFR often introduces erroneous local field outputs and subsequently affects susceptibility quantification accuracy. Inaccurate field-to-source inversion often causes large susceptibility estimation errors that appear as streaking artifacts in the QSM, especially in massive hemorrhagic regions. Because current QSM techniques struggle to generate reliable QSM, the clinical translation of QSM is greatly hindered. Recently, deep learning (DL) has achieved state-of-the-art performance in many computer …

Semantic-Driven Unsupervised Image-To-Image Translation For Distinct Image Domains, Wesley Ackerman

Theses and Dissertations

We expand the scope of image-to-image translation to include more distinct image domains, where the image sets have analogous structures, but may not share object types between them. Semantic-Driven Unsupervised Image-to-Image Translation for Distinct Image Domains (SUNIT) is built to more successfully translate images in this setting, where content from one domain is not found in the other. Our method trains an image translation model by learning encodings for semantic segmentations of images. These segmentations are translated between image domains to learn meaningful mappings between the structures in the two domains. The translated segmentations are then used as the basis …

A Semi-Automated Approach To Medical Image Segmentation Using Conditional Random Field Inference, Yu-Chi Hu

Dissertations, Theses, and Capstone Projects

Medical image segmentation plays a crucial role in delivering effective patient care in various diagnostic and treatment modalities. Manual delineation of target volumes and all critical structures is a very tedious and highly time-consuming process and introduce uncertainties of treatment outcomes of patients. Fully automatic methods holds great promise for reducing cost and time, while at the same time improving accuracy and eliminating expert variability, yet there are still great challenges. Legally and ethically, human oversight must be integrated with ”smart tools” favoring a semi-automatic technique which can leverage the best aspects of both human and computer.

In this work …

Physics-Constrained Hyperspectral Data Exploitation Across Diverse Atmospheric Scenarios, Nicholas M. Westing

Theses and Dissertations

Hyperspectral target detection promises new operational advantages, with increasing instrument spectral resolution and robust material discrimination. Resolving surface materials requires a fast and accurate accounting of atmospheric effects to increase detection accuracy while minimizing false alarms. This dissertation investigates deep learning methods constrained by the processes governing radiative transfer to efficiently perform atmospheric compensation on data collected by long-wave infrared (LWIR) hyperspectral sensors. These compensation methods depend on generative modeling techniques and permutation invariant neural network architectures to predict LWIR spectral radiometric quantities. The compensation algorithms developed in this work were examined from the perspective of target detection performance using …

Joint 1d And 2d Neural Networks For Automatic Modulation Recognition, Luis M. Rosario Morel

Theses and Dissertations

The digital communication and radar community has recently manifested more interest in using data-driven approaches for tasks such as modulation recognition, channel estimation and distortion correction. In this research we seek to apply an object detector for parameter estimation to perform waveform separation in the time and frequency domain prior to classification. This enables the full automation of detecting and classifying simultaneously occurring waveforms. We leverage a lD ResNet implemented by O'Shea et al. in [1] and the YOLO v3 object detector designed by Redmon et al. in [2]. We conducted an in depth study of the performance of these …

Computational Methods For Predicting Protein-Protein Interactions And Binding Sites, Yiwei Li

Electronic Thesis and Dissertation Repository

Proteins are essential to organisms and participate in virtually every process within cells. Quite often, they keep the cells functioning by interacting with other proteins. This process is called protein-protein interaction (PPI). The bonding amino acid residues during the process of protein-protein interactions are called PPI binding sites. Identifying PPIs and PPI binding sites are fundamental problems in system biology.

Experimental methods for solving these two problems are slow and expensive. Therefore, great efforts are being made towards increasing the performance of computational methods.

We present DELPHI, a deep learning based program for PPI site prediction and SPRINT, an algorithmic …

Material Evaluation And Structural Monitoring Of Early-Age Masonry Structures, Kyle Dunphy

Electronic Thesis and Dissertation Repository

During the initial construction period, “early-age” masonry walls are susceptible to lateral loads induced by wind or earthquake, which may result in damages or catastrophic failures. To mitigate such consequences at construction sites, temporary bracings are adopted to provide lateral support to masonry walls until they are matured enough to serve as the inherent lateral system of the structure. However, current temporary bracing guidelines provide oversimplified design due to the lack of available information on the material properties of early-age masonry. Moreover, there are no existing techniques for monitoring masonry walls to detect cracks due to construction activities. …

Discrimination Of Leucine And Isoleucine In De Novo Peptide Sequencing Using Deep Neural Networks, Bingran Shen

Electronic Thesis and Dissertation Repository

De novo peptide sequencing from tandem MS data is a key technology in proteomics for understanding the structure of proteins, especially for first seen sequences. Although this technique has advanced rapidly in recent years and become more effective, one crucial problem remained unsolved. Due to the isomerism of leucine and isoleucine, they are practically indistinguishable in de novo sequencing using traditional tandem MS data. Some experimental attempts have been made to resolve this ambiguity such as EThCD fragmentation process. In this study, we took a data focused approach rather than only looking for characteristic satellite ions produced by the EThCD …

Deep Learning To Predict Ocean Seabed Type And Source Parameters, David Franklin Van Komen

Theses and Dissertations

In the ocean, light from the surface dissipates quickly leaving sound the only way to see at a distance. Different sediment types on the ocean floor and water properties like salinity, temperature, and ocean depth all change how sound travels across long distances. Hard sediment types, such as sand and bedrock, are highly reflective while softer sediment types, such as mud, are more absorptive and change the received sound upon arrival. Unfortunately, the vast majority of the ocean floor is not mapped and the expenses involved in creating such a map are far too great. Traditional signal processing methods in …

Biomedical Data Classification With Improvised Deep Learning Architectures, Heta Desai

Computer Science Dissertations

With the rise of very powerful hardware and evolution of deep learning architectures, healthcare data analysis and its applications have been drastically transformed. These transformations mainly aim to aid a healthcare personnel with diagnosis and prognosis of a disease or abnormality at any given point of healthcare routine workflow. For instance, many of the cancer metastases detection depends on pathological tissue procedures and pathologist reviews. The reports of severity classification vary amongst different pathologist, which then leads to different treatment options for a patient. This labor-intensive work can lead to errors or mistreatments resulting in high cost of healthcare. With …

Advances In Deep Learning Through Gradient Amplification And Applications, Sunitha Basodi

Computer Science Dissertations

Deep neural networks currently play a prominent role in solving problems across a wide variety of disciplines. Improving performance of deep learning models and reducing their training times are some of the ongoing challenges. Increasing the depth of the networks improves performance but suffers from the problem of vanishing gradients and increased training times. In this research, we design methods to address these challenges in deep neural networks and demonstrate deep learning applications in several domains. We propose a gradient amplification based approach to train deep neural networks, which improves their training and testing accuraries, addresses vanishing gradients, as well …

Point Completion Networks And Segmentation Of 3d Mesh, Naga Durga Harish Kanamarlapudi

Theses

Deep learning has made many advancements in fields such as computer vision, natural language processing and speech processing. In autonomous driving, deep learning has made great improvements pertaining to the tasks of lane detection, steering estimation, throttle control, depth estimation, 2D and 3D object detection, object segmentation and object tracking. Understanding the 3D world is necessary for safe end-to-end self-driving. 3D point clouds provide rich 3D information, but processing point clouds is difficult since point clouds are irregular and unordered. Neural point processing methods like GraphCNN and PointNet operate on individual points for accurate classification and segmentation results. Occlusion of …