Matching Items (172)
Filtering by
- All Subjects: Machine Learning
- Member of: Barrett, The Honors College Thesis/Creative Project Collection
Leveraging Machine Learning and Wireless Sensing for Robot Localization - Location Variance Analysis
Description
Object localization is used to determine the location of a device, an important aspect of applications ranging from autonomous driving to augmented reality. Commonly-used localization techniques include global positioning systems (GPS), simultaneous localization and mapping (SLAM), and positional tracking, but all of these methodologies have drawbacks, especially in high traffic indoor or urban environments. Using recent improvements in the field of machine learning, this project proposes a new method of localization using networks with several wireless transceivers and implemented without heavy computational loads or high costs. This project aims to build a proof-of-concept prototype and demonstrate that the proposed technique is feasible and accurate.
Modern communication networks heavily depend upon an estimate of the communication channel, which represents the distortions that a transmitted signal takes as it moves towards a receiver. A channel can become quite complicated due to signal reflections, delays, and other undesirable effects and, as a result, varies significantly with each different location. This localization system seeks to take advantage of this distinctness by feeding channel information into a machine learning algorithm, which will be trained to associate channels with their respective locations. A device in need of localization would then only need to calculate a channel estimate and pose it to this algorithm to obtain its location.
As an additional step, the effect of location noise is investigated in this report. Once the localization system described above demonstrates promising results, the team demonstrates that the system is robust to noise on its location labels. In doing so, the team demonstrates that this system could be implemented in a continued learning environment, in which some user agents report their estimated (noisy) location over a wireless communication network, such that the model can be implemented in an environment without extensive data collection prior to release.
Modern communication networks heavily depend upon an estimate of the communication channel, which represents the distortions that a transmitted signal takes as it moves towards a receiver. A channel can become quite complicated due to signal reflections, delays, and other undesirable effects and, as a result, varies significantly with each different location. This localization system seeks to take advantage of this distinctness by feeding channel information into a machine learning algorithm, which will be trained to associate channels with their respective locations. A device in need of localization would then only need to calculate a channel estimate and pose it to this algorithm to obtain its location.
As an additional step, the effect of location noise is investigated in this report. Once the localization system described above demonstrates promising results, the team demonstrates that the system is robust to noise on its location labels. In doing so, the team demonstrates that this system could be implemented in a continued learning environment, in which some user agents report their estimated (noisy) location over a wireless communication network, such that the model can be implemented in an environment without extensive data collection prior to release.
ContributorsChang, Roger (Co-author) / Kann, Trevor (Co-author) / Alkhateeb, Ahmed (Thesis director) / Bliss, Daniel (Committee member) / Electrical Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2020-05
Description
Many factors are at play within the genome of an organism, contributing to much of the diversity and variation across the tree of life. While the genome is generally encoded by four nucleotides, A, C, T, and G, this code can be expanded. One particular mechanism that we examine in this thesis is modification of bases—more specifically, methylation of Adenine (m6A) within the GATC motif of Escherichia coli. These methylated adenines are especially important in a process called methyl-directed mismatch repair (MMR), a pathway responsible for repairing errors in the DNA sequence produced by replication. In this pathway, methylated adenines identify the parent strand and direct the repair proteins to correct the erroneous base in the daughter strand. While the primary role of methylated adenines at GATC sites is to direct the MMR pathway, this methylation has also been found to affect other processes, such as gene expression, the activity of transposable elements, and the timing of DNA replication. However, in the absence of MMR, the ability of these other processes to maintain adenine methylation and its targets is unknown.
To determine if the disruption of the MMR pathway results in the reduced conservation of methylated adenines as well as an increased tolerance for mutations that result in the loss or gain of new GATC sites, we surveyed individual clones isolated from experimentally evolving wild-type and MMR-deficient (mutL- ;conferring an 150x increase in mutation rate) populations of E. coli with whole-genome sequencing. Initial analysis revealed a lack of mutations affecting methylation sites (GATC tetranucleotides) in wild-type clones. However, the inherent low mutation rates conferred by the wild-type background render this result inconclusive, due to a lack of statistical power, and reveal a need for a more direct measure of changes in methylation status. Thus as a first step to comparative methylomics, we benchmarked four different methylation-calling pipelines on three biological replicates of the wildtype progenitor strain for our evolved populations.
While it is understood that these methylated sites play a role in the MMR pathway, it is not fully understood the full extent of their effect on the genome. Thus the goal of this thesis was to better understand the forces which maintain the genome, specifically concerning m6A within the GATC motif.
To determine if the disruption of the MMR pathway results in the reduced conservation of methylated adenines as well as an increased tolerance for mutations that result in the loss or gain of new GATC sites, we surveyed individual clones isolated from experimentally evolving wild-type and MMR-deficient (mutL- ;conferring an 150x increase in mutation rate) populations of E. coli with whole-genome sequencing. Initial analysis revealed a lack of mutations affecting methylation sites (GATC tetranucleotides) in wild-type clones. However, the inherent low mutation rates conferred by the wild-type background render this result inconclusive, due to a lack of statistical power, and reveal a need for a more direct measure of changes in methylation status. Thus as a first step to comparative methylomics, we benchmarked four different methylation-calling pipelines on three biological replicates of the wildtype progenitor strain for our evolved populations.
While it is understood that these methylated sites play a role in the MMR pathway, it is not fully understood the full extent of their effect on the genome. Thus the goal of this thesis was to better understand the forces which maintain the genome, specifically concerning m6A within the GATC motif.
ContributorsBoyer, Gwyneth (Author) / Lynch, Michael (Thesis director) / Behringer, Megan (Committee member) / Geiler-Samerotte, Kerry (Committee member) / School of Life Sciences (Contributor) / Department of Psychology (Contributor) / Barrett, The Honors College (Contributor)
Created2020-05
Description
At present, the vast majority of human subjects with neurological disease are still diagnosed through in-person assessments and qualitative analysis of patient data. In this paper, we propose to use Topological Data Analysis (TDA) together with machine learning tools to automate the process of Parkinson’s disease classification and severity assessment. An automated, stable, and accurate method to evaluate Parkinson’s would be significant in streamlining diagnoses of patients and providing families more time for corrective measures. We propose a methodology which incorporates TDA into analyzing Parkinson’s disease postural shifts data through the representation of persistence images. Studying the topology of a system has proven to be invariant to small changes in data and has been shown to perform well in discrimination tasks. The contributions of the paper are twofold. We propose a method to 1) classify healthy patients from those afflicted by disease and 2) diagnose the severity of disease. We explore the use of the proposed method in an application involving a Parkinson’s disease dataset comprised of healthy-elderly, healthy-young and Parkinson’s disease patients.
ContributorsRahman, Farhan Nadir (Co-author) / Nawar, Afra (Co-author) / Turaga, Pavan (Thesis director) / Krishnamurthi, Narayanan (Committee member) / Electrical Engineering Program (Contributor) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2020-05
Description
In this project, the use of deep neural networks for the process of selecting actions to execute within an environment to achieve a goal is explored. Scenarios like this are common in crafting based games such as Terraria or Minecraft. Goals in these environments have recursive sub-goal dependencies which form a dependency tree. An agent operating within these environments have access to low amounts of data about the environment before interacting with it, so it is crucial that this agent is able to effectively utilize a tree of dependencies and its environmental surroundings to make judgements about which sub-goals are most efficient to pursue at any point in time. A successful agent aims to minimizes cost when completing a given goal. A deep neural network in combination with Q-learning techniques was employed to act as the agent in this environment. This agent consistently performed better than agents using alternate models (models that used dependency tree heuristics or human-like approaches to make sub-goal oriented choices), with an average performance advantage of 33.86% (with a standard deviation of 14.69%) over the best alternate agent. This shows that machine learning techniques can be consistently employed to make goal-oriented choices within an environment with recursive sub-goal dependencies and low amounts of pre-known information.
ContributorsKoleber, Derek (Author) / Acuna, Ruben (Thesis director) / Bansal, Ajay (Committee member) / W.P. Carey School of Business (Contributor) / Software Engineering (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
Description
This thesis dives into the world of artificial intelligence by exploring the functionality of a single layer artificial neural network through a simple housing price classification example while simultaneously considering its impact from a data management perspective on both the software and hardware level. To begin this study, the universally accepted model of an artificial neuron is broken down into its key components and then analyzed for functionality by relating back to its biological counterpart. The role of a neuron is then described in the context of a neural network, with equal emphasis placed on how it individually undergoes training and then for an entire network. Using the technique of supervised learning, the neural network is trained with three main factors for housing price classification, including its total number of rooms, bathrooms, and square footage. Once trained with most of the generated data set, it is tested for accuracy by introducing the remainder of the data-set and observing how closely its computed output for each set of inputs compares to the target value. From a programming perspective, the artificial neuron is implemented in C so that it would be more closely tied to the operating system and therefore make the collected profiler data more precise during the program's execution. The program is designed to break down each stage of the neuron's training process into distinct functions. In addition to utilizing more functional code, the struct data type is used as the underlying data structure for this project to not only represent the neuron but for implementing the neuron's training and test data. Once fully trained, the neuron's test results are then graphed to visually depict how well the neuron learned from its sample training set. Finally, the profiler data is analyzed to describe how the program operated from a data management perspective on the software and hardware level.
ContributorsRichards, Nicholas Giovanni (Author) / Miller, Phillip (Thesis director) / Meuth, Ryan (Committee member) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
Description
Breast cancer is the leading cause of cancer-related deaths of women in the united states. Traditionally, Breast cancer is predominantly treated by a combination of surgery, chemotherapy, and radiation therapy. However, due to the significant negative side effects associated with these traditional treatments, there has been substantial efforts to develop alternative therapies to treat cancer. One such alternative therapy is a peptide-based therapeutic cancer vaccine. Therapeutic cancer vaccines enhance an individual's immune response to a specific tumor. They are capable of doing this through artificial activation of tumor specific CTLs (Cytotoxic T Lymphocytes). However, in order to artificially activate tumor specific CTLs, a patient must be treated with immunogenic epitopes derived from their specific cancer type. We have identified that the tumor associated antigen, TPD52, is an ideal target for a therapeutic cancer vaccine. This designation was due to the overexpression of TPD52 in a variety of different cancer types. In order to start the development of a therapeutic cancer vaccine for TPD52-related cancers, we have devised a two-step strategy. First, we plan to create a list of potential TPD52 epitopes by using epitope binding and processing prediction tools. Second, we plan to attempt to experimentally identify MHC class I TPD52 epitopes in vitro. We identified 942 potential 9 and 10 amino acid epitopes for the HLAs A1, A2, A3, A11, A24, B07, B27, B35, B44. These epitopes were predicted by using a combination of 3 binding prediction tools and 2 processing prediction tools. From these 942 potential epitopes, we selected the top 50 epitopes ranked by a combination of binding and processing scores. Due to the promiscuity of some predicted epitopes for multiple HLAs, we ordered 38 synthetic epitopes from the list of the top 50 epitope. We also performed a frequency analysis of the TPD52 protein sequence and identified 3 high volume regions of high epitope production. After the epitope predictions were completed, we proceeded to attempt to experimentally detected presented TPD52 epitopes. First, we successful transduced parental K562 cells with TPD52. After transduction, we started the optimization process for the immunoprecipitation protocol. The optimization of the immunoprecipitation protocol proved to be more difficult than originally believed and was the main reason that we were unable to progress past the transduction of the parental cells. However, we believe that we have identified the issues and will be able to complete the experiment in the coming months.
ContributorsWilson, Eric Andrew (Author) / Anderson, Karen (Thesis director) / Borges, Chad (Committee member) / School of Molecular Sciences (Contributor) / Barrett, The Honors College (Contributor)
Created2016-05
Description
Background: Noninvasive MRI methods that can accurately detect subtle brain changes are highly desirable when studying disease-modifying interventions. Texture analysis is a novel imaging technique which utilizes the extraction of a large number of image features with high specificity and predictive power. In this investigation, we use texture analysis to assess and classify age-related changes in the right and left hippocampal regions, the areas known to show some of the earliest change in Alzheimer's disease (AD). Apolipoprotein E (APOE)'s e4 allele confers an increased risk for AD, so studying differences in APOE e4 carriers may help to ascertain subtle brain changes before there has been an obvious change in behavior. We examined texture analysis measures that predict age-related changes, which reflect atrophy in a group of cognitively normal individuals. We hypothesized that the APOE e4 carriers would exhibit significant age-related differences in texture features compared to non-carriers, so that the predictive texture features hold promise for early assessment of AD. Methods: 120 normal adults between the ages of 32 and 90 were recruited for this neuroimaging study from a larger parent study at Mayo Clinic Arizona studying longitudinal cognitive functioning (Caselli et al., 2009). As part of the parent study, the participants were genotyped for APOE genetic polymorphisms and received comprehensive cognitive testing every two years, on average. Neuroimaging was done at Barrow Neurological Institute and a 3D T1-weighted magnetic resonance image was obtained during scanning that allowed for subsequent texture analysis processing. Voxel-based features of the appearance, structure, and arrangement of these regions of interest were extracted utilizing the Mayo Clinic Python Texture Analysis Pipeline (pyTAP). Algorithms applied in feature extraction included Grey-Level Co-Occurrence Matrix (GLCM), Gabor Filter Banks (GFB), Local Binary Patterns (LBP), Discrete Orthogonal Stockwell Transform (DOST), and Laplacian-of-Gaussian Histograms (LoGH). Principal component (PC) analysis was used to reduce the dimensionality of the algorithmically selected features to 13 PCs. A stepwise forward regression model was used to determine the effect of APOE status (APOE e4 carriers vs. noncarriers), and the texture feature principal components on age (as a continuous variable). After identification of 5 significant predictors of age in the model, the individual feature coefficients of those principal components were examined to determine which features contributed most significantly to the prediction of an aging brain. Results: 70 texture features were extracted for the two regions of interest in each participant's scan. The texture features were coded as 70 initial components andwere rotated to generate 13 principal components (PC) that contributed 75% of the variance in the dataset by scree plot analysis. The forward stepwise regression model used in this exploratory study significantly predicted age, accounting for approximately 40% of the variance in the data. The regression model revealed 5 significant regressors (2 right PC's, APOE status, and 2 left PC by APOE interactions). Finally, the specific texture features that contributed to each significant PCs were identified. Conclusion: Analysis of image texture features resulted in a statistical model that was able to detect subtle changes in brain integrity associated with age in a group of participants who are cognitively normal, but have an increased risk of developing AD based on the presence of the APOE e4 phenotype. This is an important finding, given that detecting subtle changes in regions vulnerable to the effects of AD in patients could allow certain texture features to serve as noninvasive, sensitive biomarkers predictive of AD. Even with only a small number of patients, the ability for us to determine sensitive imaging biomarkers could facilitate great improvement in speed of detection and effectiveness of AD interventions..
ContributorsSilva, Annelise Michelle (Author) / Baxter, Leslie (Thesis director) / McBeath, Michael (Committee member) / Presson, Clark (Committee member) / School of Life Sciences (Contributor) / Department of Psychology (Contributor) / Barrett, The Honors College (Contributor)
Created2016-05
Description
Mammary gland development in humans during puberty involves the enlargement of breast tissue, but this is not true in non-human primates. To identify potential causes of this difference, I examined variation in substitution rates across genes related to mammary development. Genes undergoing purifying selection show slower-than-average substitution rates, while genes undergoing positive selection show faster rates. These may be related to the difference between humans and other primates. Three genes were found to be accelerated were FOXF1, IGFBP5, and ATP2B2, but only the latter one was found in humans and it seems unlikely that it would be related to the differences between mammary gland development at puberty between humans and non-human primates.
ContributorsArroyo, Diana (Author) / Cartwright, Reed (Thesis director) / Wilson Sayres, Melissa (Committee member) / Schwartz, Rachel (Committee member) / School of Life Sciences (Contributor) / Barrett, The Honors College (Contributor)
Created2016-05
Description
Epilepsy affects numerous people around the world and is characterized by recurring seizures, prompting the ability to predict them so precautionary measures may be employed. One promising algorithm extracts spatiotemporal correlation based features from intracranial electroencephalography signals for use with support vector machines. The robustness of this methodology is tested through a sensitivity analysis. Doing so also provides insight about how to construct more effective feature vectors.
ContributorsMa, Owen (Author) / Bliss, Daniel (Thesis director) / Berisha, Visar (Committee member) / Barrett, The Honors College (Contributor) / Electrical Engineering Program (Contributor)
Created2015-05
Description
microRNAs (miRNAs) are short ~22nt non-coding RNAs that regulate gene output at the post-transcriptional level. Via targeting of degenerate elements primarily in 3'untranslated regions (3'UTR) of mRNAs, miRNAs can target thousands of varying genes and suppress their protein translation. The precise mechanistic function and bio- logical role of miRNAs is not fully understood and yet it is a major contributor to a pleth- ora of diseases, including neurological disorders, muscular disorders, and cancer. Cer- tain model organisms are valuable in understanding the function of miRNA and there- fore fully understanding the biological significance of miRNA targeting. Here I report a mechanistic analysis of miRNA targeting in C. elegans, and a bioinformatic approach to aid in further investigation of miRNA targeted sequences. A few of the biologically significant mechanisms discussed in this thesis include alternative polyadenylation, RNA binding proteins, components of the miRNA recognition machinery, miRNA secondary structures, and their polymorphisms. This thesis also discusses a novel bioinformatic approach to studying miRNA biology, including computational miRNA target prediction software, and sequence complementarity. This thesis allows a better understanding of miRNA biology and presents an ideal strategy for approaching future research in miRNA targeting.
ContributorsWeigele, Dustin Keith (Author) / Mangone, Marco (Thesis director) / Katchman, Benjamin (Committee member) / Barrett, The Honors College (Contributor) / Department of Chemistry and Biochemistry (Contributor) / School of Life Sciences (Contributor)
Created2014-12