Physical Sciences and Mathematics Commons^™

Mathematical Models Yield Insights Into Cnns: Applications In Natural Image Restoration And Population Genetics, Ryan Cecil

Electronic Theses and Dissertations

Due to a rise in computational power, machine learning (ML) methods have become the state-of-the-art in a variety of fields. Known to be black-box approaches, however, these methods are oftentimes not well understood. In this work, we utilize our understanding of model-based approaches to derive insights into Convolutional Neural Networks (CNNs). In the field of Natural Image Restoration, we focus on the image denoising problem. Recent work have demonstrated the potential of mathematically motivated CNN architectures that learn both `geometric' and nonlinear higher order features and corresponding regularizers. We extend this work by showing that not only can geometric features …

Dimension And Ramsey Results In Partially Ordered Sets., Sida Wan

Electronic Theses and Dissertations

In this dissertation, there are two major parts. One is the dimension results on different classes of partially ordered sets. We developed new tools and theorems to solve the bounds on interval orders using different number of lengths. We also discussed the dimension of interval orders that have a representation with interval lengths in a certain range. We further discussed the interval dimension and semi dimension for posets. In the second part, we discussed several related results on the Ramsey theory of grids, the results involve the application of Product Ramsey Theorem and Partition Ramsey Theorem

Cross-Validation For Autoregressive Models., Christina Han

Electronic Theses and Dissertations

There are no set rules for choosing the lag order for autoregressive (AR) time series models. Currently, the most common methods employ AIC or BIC. However, AIC has been proven to be inconsistent and BIC is inefficient. Racine proposed an estimator based on Shao's work which he hypothesized would also be consistent, but left the proof as an open problem. We will show his claim does not follow immediately from Shao. However, Shao offered another consistent method for cross validation of linear models called APCV, and we will show that AR models satisfy Shao's conditions. Thus, APCV is a consistent …

A New Sir Model With Mobility., Ciana Applegate

Electronic Theses and Dissertations

In this paper, a mobility-based SIR model is built to understand the spread of the pandemic. A traditional SIR model used in epidemiology describes the transition of particles among states, such as susceptible, infected, and recovered states. However, the traditional model has no movement of particles. There are many variations of SIR models when it comes to the factor of mobility, the majority of studies use mobility intensity or population density as a measure of mobility. In this paper, a new dynamical SIR model, including the spatial motion of three-type particles, is constructed and the long-time behavior of the first …

An Application Of Differential Mathematical Modeling Techniques To Study The Ongoing Rabies Epizootic In China, Christopher Turner

Electronic Theses and Dissertations

Rabies remains a global public health issue with a wide variety of neurological symptoms such as confusion, slight paralysis, hypersalivation, and hydrophobia. Rabies is usually fatal once symptoms appear. Many species are reservoirs for rabies, such as foxes, racoons, and wild dogs, which in turn can transmit the disease to humans, leading to complex transmission chains. There is a long latent period of rabies, between 1 to 3 months after infection, which further complicates control efforts. Mathematical modeling is a valuable tool in the study of infectious disease outbreaks and there have been many models applied to rabies outbreaks. However, …

Development And Evaluation Of Modeling Approaches For Extrusion-Based Additive Manufacturing Of Thermoplastics, Christopher C. Bock

Electronic Theses and Dissertations

This work focuses on evaluating different modeling approaches and model parameters for thermoplastic AM, with the goal of informing more efficient and effective modeling approaches. First, different modeling approaches were tested and compared to experiments. From this it was found that all three of the modeling approaches provide comparable results and provide similar results to experiments. Then one of the modeling approaches was tested on large scale geometries, and it was found that the model results matched experiments closely. Then the effect of different material properties was evaluated, this was done by performing a fractional factorial design of experiments where …

Efficient Numerical Optimization For Parallel Dynamic Optimal Power Flow Simulation Using Network Geometry, Rylee Sundermann

Electronic Theses and Dissertations

In this work, we present a parallel method for accelerating the multi-period dynamic optimal power flow (DOPF). Our approach involves a distributed-memory parallelization of DOPF time-steps, use of a newly developed parallel primal-dual interior point method, and an iterative Krylov subspace linear solver with a block-Jacobi preconditioning scheme. The parallel primal-dual interior point method has been implemented and distributed in the open-source PETSc library and is currently available. We present the formulation of the DOPF problem, the developed primal dual interior point method solver, the parallel implementation, and results on various multi-core machines. We demonstrate the effectiveness our proposed block-Jacobi …

Reinforcement Learning: Low Discrepancy Action Selection For Continuous States And Actions, Jedidiah Lindborg

Electronic Theses and Dissertations

In reinforcement learning the process of selecting an action during the exploration or exploitation stage is difficult to optimize. The purpose of this thesis is to create an action selection process for an agent by employing a low discrepancy action selection (LDAS) method. This should allow the agent to quickly determine the utility of its actions by prioritizing actions that are dissimilar to ones that it has already picked. In this way the learning process should be faster for the agent and result in more optimal policies.

Eeg Signals Classification Using Lstm-Based Models And Majority Logic, James A. Orgeron

Electronic Theses and Dissertations

The study of elecroencephalograms (EEGs) has gained enormous interest in the last decade with the increase of computational power and availability of EEG signals collected from various human activities or produced during medical tests. The applicability of analyzing EEG signals ranges from helping impaired people communicate or move (using appropriate medical equipment) to understanding people's feelings and detecting diseases.

We proposed new methodology and models for analyzing and classifying EEG signals collected from individuals observing visual stimuli. Our models rely on powerful Long-Short Term Memory (LSTM) Neural Network models, which are currently the state of the art models for performing …