Statistics and Probability Commons^™

Time Series Models For Predicting Application Gpu Utilization And Power Draw Based On Trace Data, Dorothy Xiaoshuang Parry

Electrical & Computer Engineering Theses & Dissertations

This work explores collecting performance metrics and leveraging various statistical and machine learning time series predictive models on a memory-intensive application, Inception v3. Trace data collected using nvidia-smi measured GPU utilization and power draw for two runs of Inception3. Experimental results from the statistical and machine learning-based time series predictive algorithms showed that the predictions from statistical-based models were unable to capture the complex changes in the trace data. The Probabilistic TNN model provided the best results for the power draw trace, according to the test evaluation metrics. For the GPU utilization trace, the RNN models produced the most accurate …

Reinforcement Learning: Applying Low Discrepancy Action Selection To Deep Deterministic Policy Gradient, Aleksandr Svishchev

Electronic Theses and Dissertations

Reinforcement learning (RL) is a subfield of machine learning concerned with agents learning to behave optimally by interacting with an environment. One of the most important topics in RL is how the agent should explore, that is, how to choose actions in order to rate their impact on long-term reward. For example, a simple baseline strategy might be uniformly random action selection. This thesis investigates the heuristic idea that agents will learn faster if they explore by factoring the environment’s state into their decision and intentionally choose actions which are as different as possible from what they have previously observed. …

Statistical And Machine Learning Approaches To Describe Factors Affecting Preweaning Mortality Of Piglets, Md Towfiqur Rahman, Tami M. Brown-Brandl, Gary A. Rohrer, Sudhendu R. Sharma, Vamsi Manthena, Yeyin Shi

Department of Biological Systems Engineering: Papers and Publications

High preweaning mortality (PWM) rates for piglets are a significant concern for the worldwide pork industries, causing economic loss and well-being issues. This study focused on identifying the factors affecting PWM, overlays, and predicting PWM using historical production data with statistical and machine learning models. Data were collected from 1,982 litters from the United States Meat Animal Research Center, Nebraska, over the years 2016 to 2021. Sows were housed in a farrowing building with three rooms, each with 20 farrowing crates, and taken care of by well-trained animal caretakers. A generalized linear model was used to analyze the various sow, …

Data-Optimized Spatial Field Predictions For Robotic Adaptive Sampling: A Gaussian Process Approach, Zachary Nathan

Computer Science Senior Theses

We introduce a framework that combines Gaussian Process models, robotic sensor measurements, and sampling data to predict spatial fields. In this context, a spatial field refers to the distribution of a variable throughout a specific area, such as temperature or pH variations over the surface of a lake. Whereas existing methods tend to analyze only the particular field(s) of interest, our approach optimizes predictions through the effective use of all available data. We validated our framework on several datasets, showing that errors can decline by up to two-thirds through the inclusion of additional colocated measurements. In support of adaptive sampling, …

Identifying Key Activity Indicators In Rats' Neuronal Data Using Lasso Regularized Logistic Regression, Avery Woods

Honors Theses

This thesis aims to identify timestamps of rats’ neuronal activity that best determine behavior using a machine learning model. Neuronal data is a complex and high-dimensional dataset, and identifying the most informative features is crucial for understanding the underlying neuronal processes. The Lasso regularization technique is employed to select the most relevant features of the data to the model’s prediction. The results of this study provide insights into the key activity indicators that are associated with specific behaviors or cognitive processes in rats, as well as the effect that stress can have on neuronal activity and behavior. Ultimately, it was …

Intra-Hour Solar Forecasting Using Cloud Dynamics Features Extracted From Ground-Based Infrared Sky Images, Guillermo Terrén-Serrano

Electrical and Computer Engineering ETDs

Due to the increasing use of photovoltaic systems, power grids are vulnerable to the projection of shadows from moving clouds. An intra-hour solar forecast provides power grids with the capability of automatically controlling the dispatch of energy, reducing the additional cost for a guaranteed, reliable supply of energy (i.e., energy storage). This dissertation introduces a novel sky imager consisting of a long-wave radiometric infrared camera and a visible light camera with a fisheye lens. The imager is mounted on a solar tracker to maintain the Sun in the center of the images throughout the day, reducing the scattering effect produced …

Characterizing Convolutional Neural Network Early-Learning And Accelerating Non-Adaptive, First-Order Methods With Localized Lagrangian Restricted Memory Level Bundling, Benjamin O. Morris

Theses and Dissertations

This dissertation studies the underlying optimization problem encountered during the early-learning stages of convolutional neural networks and introduces a training algorithm competitive with existing state-of-the-art methods. First, a Design of Experiments method is introduced to systematically measure empirical second-order Lipschitz upper bound and region size estimates for local regions of convolutional neural network loss surfaces experienced during the early-learning stages. This method demonstrates that architecture choices can significantly impact the local loss surfaces traversed during training. Next, a Design of Experiments method is used to study the effects convolutional neural network architecture hyperparameters have on different optimization routines' abilities to …

Novel Statistical Modeling Methods For Traffic Video Analysis, Hang Shi

Dissertations

Video analysis is an active and rapidly expanding research area in computer vision and artificial intelligence due to its broad applications in modern society. Many methods have been proposed to analyze the videos, but many challenging factors remain untackled. In this dissertation, four statistical modeling methods are proposed to address some challenging traffic video analysis problems under adverse illumination and weather conditions.

First, a new foreground detection method is presented to detect the foreground objects in videos. A novel Global Foreground Modeling (GFM) method, which estimates a global probability density function for the foreground and applies the Bayes decision rule …

Applying Deep Learning To The Ice Cream Vendor Problem: An Extension Of The Newsvendor Problem, Gaffar Solihu

Electronic Theses and Dissertations

The Newsvendor problem is a classical supply chain problem used to develop strategies for inventory optimization. The goal of the newsvendor problem is to predict the optimal order quantity of a product to meet an uncertain demand in the future, given that the demand distribution itself is known. The Ice Cream Vendor Problem extends the classical newsvendor problem to an uncertain demand with unknown distribution, albeit a distribution that is known to depend on exogenous features. The goal is thus to estimate the order quantity that minimizes the total cost when demand does not follow any known statistical distribution. The …

Classification Of Chess Games: An Exploration Of Classifiers For Anomaly Detection In Chess, Masudul Hoque

All Graduate Theses, Dissertations, and Other Capstone Projects

Chess is a strategy board game with its inception dating back to the 15th century. The Covid-19 pandemic has led to a chess boom online with 95,853,038 chess games being played during January, 2021 on lichess.com. Along with the chess boom, instances of cheating have also become more rampant. Classifications have been used for anomaly detection in different fields and thus it is a natural idea to develop classifiers to detect cheating in chess. However, there are no specific examples of this, and it is difficult to obtain data where cheating has occurred. So, in this paper, we develop 4 …

Multimodal Data Fusion And Attack Detection In Recommender Systems, Mehmet Aktukmak

USF Tampa Graduate Theses and Dissertations

The commercial platforms that use recommender systems can collect relevant information to produce useful recommendations to the platform users. However, these sources usually contain missing values, imbalanced and heterogeneous data, and noisy observations. Such characteristics render the process of exploiting the information nontrivial, as one should carefully address them during the data fusion process. In addition to the degenerative characteristics, some entries can be fake, i.e., they can be the outcomes of malicious intents to manipulate the system. These entries should be eliminated before incorporation to any recommendation task. Detecting such malicious attacks quickly and accurately and then mitigating them …

Analysis Of Gameplay Strategies In Hearthstone: A Data Science Approach, Connor W. Watson

Theses

In recent years, games have been a popular test bed for AI research, and the presence of Collectible Card Games (CCGs) in that space is still increasing. One such CCG for both competitive/casual play and AI research is Hearthstone, a two-player adversarial game where players seeks to implement one of several gameplay strategies to defeat their opponent and decrease all of their Health points to zero. Although some open source simulators exist, some of their methodologies for simulated agents create opponents with a relatively low skill level. Using evolutionary algorithms, this thesis seeks to evolve agents with a higher skill …

Semi-Supervised Regression With Generative Adversarial Networks Using Minimal Labeled Data, Greg Olmschenk

Dissertations, Theses, and Capstone Projects

This work studies the generalization of semi-supervised generative adversarial networks (GANs) to regression tasks. A novel feature layer contrasting optimization function, in conjunction with a feature matching optimization, allows the adversarial network to learn from unannotated data and thereby reduce the number of labels required to train a predictive network. An analysis of simulated training conditions is performed to explore the capabilities and limitations of the method. In concert with the semi-supervised regression GANs, an improved label topology and upsampling technique for multi-target regression tasks are shown to reduce data requirements. Improvements are demonstrated on a wide variety of vision …

Field Drilling Data Cleaning And Preparation For Data Analytics Applications, Daniel Cardoso Braga

LSU Master's Theses

Throughout the history of oil well drilling, service providers have been continuously striving to improve performance and reduce total drilling costs to operating companies. Despite constant improvement in tools, products, and processes, data science has not played a large part in oil well drilling. With the implementation of data science in the energy sector, companies have come to see significant value in efficiently processing the massive amounts of data produced by the multitude of internet of thing (IOT) sensors at the rig. The scope of this project is to combine academia and industry experience to analyze data from 13 different …

Statistical Machine Learning Methods For Mining Spatial And Temporal Data, Fei Tan

Dissertations

Spatial and temporal dependencies are ubiquitous properties of data in numerous domains. The popularity of spatial and temporal data mining has thus grown with the increasing prevalence of massive data. The presence of spatial and temporal attributes not only provides complementary useful perspectives, but also poses new challenges to the representation and integration into the learning procedure. In this dissertation, the involved spatial and temporal dependencies are explored with three genres: sample-wise, feature-wise, and target-wise. A family of novel methodologies is developed accordingly for the dependency representation in respective scenarios.

First, dependencies among discrete, continuous and repeated observations are studied …

A Data-Driven Approach For Modeling Agents, Hamdi Kavak

Computational Modeling & Simulation Engineering Theses & Dissertations

Agents are commonly created on a set of simple rules driven by theories, hypotheses, and assumptions. Such modeling premise has limited use of real-world data and is challenged when modeling real-world systems due to the lack of empirical grounding. Simultaneously, the last decade has witnessed the production and availability of large-scale data from various sensors that carry behavioral signals. These data sources have the potential to change the way we create agent-based models; from simple rules to driven by data. Despite this opportunity, the literature has neglected to offer a modeling approach to generate granular agent behaviors from data, creating …

A Statistical Analysis And Machine Learning Of Genomic Data, Jongyun Jung

All Graduate Theses, Dissertations, and Other Capstone Projects

Machine learning enables a computer to learn a relationship between two assumingly related types of information. One type of information could thus be used to predict any lack of informaion in the other using the learned relationship. During the last decades, it has become cheaper to collect biological information, which has resulted in increasingly large amounts of data. Biological information such as DNA is currently analyzed by a variety of tools. Although machine learning has already been used in various projects, a flexible tool for analyzing generic biological challenges has not yet been made. The recent advancements in the DNA …

Regression Tree Construction For Reinforcement Learning Problems With A General Action Space, Anthony S. Bush Jr

Electronic Theses and Dissertations

Part of the implementation of Reinforcement Learning is constructing a regression of values against states and actions and using that regression model to optimize over actions for a given state. One such common regression technique is that of a decision tree; or in the case of continuous input, a regression tree. In such a case, we fix the states and optimize over actions; however, standard regression trees do not easily optimize over a subset of the input variables\cite{Card1993}. The technique we propose in this thesis is a hybrid of regression trees and kernel regression. First, a regression tree splits over …

Data Patterns Discovery Using Unsupervised Learning, Rachel A. Lewis

Electronic Theses and Dissertations

Self-care activities classification poses significant challenges in identifying children’s unique functional abilities and needs within the exceptional children healthcare system. The accuracy of diagnosing a child's self-care problem, such as toileting or dressing, is highly influenced by an occupational therapists’ experience and time constraints. Thus, there is a need for objective means to detect and predict in advance the self-care problems of children with physical and motor disabilities. We use clustering to discover interesting information from self-care problems, perform automatic classification of binary data, and discover outliers. The advantages are twofold: the advancement of knowledge on identifying self-care problems in …

Cognitive Virtual Admissions Counselor, Kumar Raja Guvindan Raju, Cory Adams, Raghuram Srinivas

SMU Data Science Review

Abstract. In this paper, we present a cognitive virtual admissions counselor for the Master of Science in Data Science program at Southern Methodist University. The virtual admissions counselor is a system capable of providing potential students accurate information at the time that they want to know it. After the evaluation of multiple technologies, Amazon’s LEX was selected to serve as the core technology for the virtual counselor chatbot. Student surveys were leveraged to collect and generate training data to deploy the natural language capability. The cognitive virtual admissions counselor platform is currently capable of providing an end-to-end conversational dialog to …

Old English Character Recognition Using Neural Networks, Sattajit Sutradhar

Electronic Theses and Dissertations

Character recognition has been capturing the interest of researchers since the beginning of the twentieth century. While the Optical Character Recognition for printed material is very robust and widespread nowadays, the recognition of handwritten materials lags behind. In our digital era more and more historical, handwritten documents are digitized and made available to the general public. However, these digital copies of handwritten materials lack the automatic content recognition feature of their printed materials counterparts. We are proposing a practical, accurate, and computationally efficient method for Old English character recognition from manuscript images. Our method relies on a modern machine learning …

Deep Energy-Based Models For Structured Prediction, David Belanger

Doctoral Dissertations

We introduce structured prediction energy networks (SPENs), a flexible frame- work for structured prediction. A deep architecture is used to define an energy func- tion over candidate outputs and predictions are produced by gradient-based energy minimization. This deep energy captures dependencies between labels that would lead to intractable graphical models, and allows us to automatically discover discrim- inative features of the structured output. Furthermore, practitioners can explore a wide variety of energy function architectures without having to hand-design predic- tion and learning methods for each model. This is because all of our prediction and learning methods interact with the energy …

Classification With Large Sparse Datasets: Convergence Analysis And Scalable Algorithms, Xiang Li

Electronic Thesis and Dissertation Repository

Large and sparse datasets, such as user ratings over a large collection of items, are common in the big data era. Many applications need to classify the users or items based on the high-dimensional and sparse data vectors, e.g., to predict the profitability of a product or the age group of a user, etc. Linear classifiers are popular choices for classifying such datasets because of their efficiency. In order to classify the large sparse data more effectively, the following important questions need to be answered.

1. Sparse data and convergence behavior. How different properties of a dataset, such as …

Towards Deeper Understanding In Neuroimaging, Rex Devon Hjelm

Computer Science ETDs

Neuroimaging is a growing domain of research, with advances in machine learning having tremendous potential to expand understanding in neuroscience and improve public health. Deep neural networks have recently and rapidly achieved historic success in numerous domains, and as a consequence have completely redefined the landscape of automated learners, giving promise of significant advances in numerous domains of research. Despite recent advances and advantages over traditional machine learning methods, deep neural networks have yet to have permeated significantly into neuroscience studies, particularly as a tool for discovery. This dissertation presents well-established and novel tools for unsupervised learning which aid in …

Scaling Mcmc Inference And Belief Propagation To Large, Dense Graphical Models, Sameer Singh

Doctoral Dissertations

With the physical constraints of semiconductor-based electronics becoming increasingly limiting in the past decade, single-core CPUs have given way to multi-core and distributed computing platforms. At the same time, access to large data collections is progressively becoming commonplace due to the lowering cost of storage and bandwidth. Traditional machine learning paradigms that have been designed to operate sequentially on single processor architectures seem destined to become obsolete in this world of multi-core, multi-node systems and massive data sets. Inference for graphical models is one such example for which most existing algorithms are sequential in nature and are difficult to scale …

Incorporating Boltzmann Machine Priors For Semantic Labeling In Images And Videos, Andrew Kae

Doctoral Dissertations

Semantic labeling is the task of assigning category labels to regions in an image. For example, a scene may consist of regions corresponding to categories such as sky, water, and ground, or parts of a face such as eyes, nose, and mouth. Semantic labeling is an important mid-level vision task for grouping and organizing image regions into coherent parts. Labeling these regions allows us to better understand the scene itself as well as properties of the objects in the scene, such as their parts, location, and interaction within the scene. Typical approaches for this task include the conditional random field …

Convergence Of A Reinforcement Learning Algorithm In Continuous Domains, Stephen Carden

All Dissertations

In the field of Reinforcement Learning, Markov Decision Processes with a finite number of states and actions have been well studied, and there exist algorithms capable of producing a sequence of policies which converge to an optimal policy with probability one. Convergence guarantees for problems with continuous states also exist. Until recently, no online algorithm for continuous states and continuous actions has been proven to produce optimal policies. This Dissertation contains the results of research into reinforcement learning algorithms for problems in which both the state and action spaces are continuous. The problems to be solved are introduced formally as …

Online Multi-Stage Deep Architectures For Feature Extraction And Object Recognition, Derek Christopher Rose

Doctoral Dissertations

Multi-stage visual architectures have recently found success in achieving high classification accuracies over image datasets with large variations in pose, lighting, and scale. Inspired by techniques currently at the forefront of deep learning, such architectures are typically composed of one or more layers of preprocessing, feature encoding, and pooling to extract features from raw images. Training these components traditionally relies on large sets of patches that are extracted from a potentially large image dataset. In this context, high-dimensional feature space representations are often helpful for obtaining the best classification performances and providing a higher degree of invariance to object transformations. …

Mobile Computing: Challenges And Opportunities For Autonomy And Feedback, Ole J. Mengshoel, Bob Iannucci, Abe Ishihara

Ole J Mengshoel

Machine Learning Approaches For Determining Effective Seeds For K -Means Algorithm, Kaveephong Lertwachara

Doctoral Dissertations

In this study, I investigate and conduct an experiment on two-stage clustering procedures, hybrid models in simulated environments where conditions such as collinearity problems and cluster structures are controlled, and in real-life problems where conditions are not controlled. The first hybrid model (NK) is an integration between a neural network (NN) and the k-means algorithm (KM) where NN screens seeds and passes them to KM. The second hybrid (GK) uses a genetic algorithm (GA) instead of the neural network. Both NN and GA used in this study are in their simplest-possible forms.

In the simulated data sets, I investigate two …