• 

  Cavitation and cavitation damage

  Rogers, W. L.; Wang, Shih-cheng, 1938- (The University of Arizona., 1965)
• 

  Stress-Induced Dynamics of FOXO Transcription Factors

  Paek, Andrew; Lasick, Kathleen Anne; El-Kareh, Ardith; Schroeder, Joyce; Weinert, Ted; Yao, Guang (The University of Arizona., 2023)

  The FOXO family of transcription factors have several important roles in multicellular organisms: they are required for proper development of different tissues, maintain homeostasis in response to diverse cellular stresses, function as tumor suppressors, and have an evolutionarily conserved role in prolonging lifespan. Consistent with their role in diverse cellular processes, FOXO proteins are activated by several different stimuli, leading to the promotion of many different downstream programs often with opposing outcomes. How FOXO protein activation can lead to stimulus-dependent transcriptional outcomes is not known, though several mechanisms have been proposed. Possible mechanisms include differences in FOXO post-translational modifications, binding partners, and the dynamics of FOXO activation. Here, I will describe the current evidence in the literature supporting these mechanisms, and our investigation into the role of dynamic patterns of activation of the FOXO transcription factors. Specifically, I set out to determine whether FOXO responds to different stresses with different temporal patterns of activation. I have shown that FOXO responds to oxidative stress in a sustained, bimodal pattern, while it responds to serum starvation in a stochastic pulsatile pattern. I also found that in MCF7 breast cancer cells, both patterns are controlled by the activity of the primary negative regulator of FOXO, Akt.
• 

  Monolithic Three-Mirror Anastigmat Telescope

  Kim, Daewook; Han, Yuqiao; Choi, Heejoo; Kupinski, Meredith (The University of Arizona., 2023)

  This thesis presents the design of a near-infrared (near-IR) three-mirror anastigmat telescope with a focal length of 38,675.1 ??, an entrance pupil diameter of 1,933.75 ??, and a 0.12° full field-of-view (FoV). This diffraction-limited telescope has a RMS spot radius of 36.6 μm (similar to the radius of Airy disk) to satisfy the design resolution requirements of the IR telescope. Designed specifically for near-IR sensing, the system benchmarked a widely known module of an IR detector, called the Near-Infrared Camera (NIRCam) detector, with ten Teledyne HgCdTe H2RG detectors. The NIRCam detector was assumed in this TMA design study, similar to the James Webb Space Telescope case. This three-mirror telescope design contains only two substrates, one consisting of two mirror surfaces with different radii of curvature. Consequently, the structure of this robust telescope design is straightforward and it is therefore easy to manufacture and align. The starting point for the optical design of this telescope was the Vera C. Rubin Observatory (a.k.a. Large Synoptic Survey Telescope), a three-mirror anastigmat system combined with more refractive lenses and a color filters. Importantly, the specifications of these three mirrors were changed to achieve the goal of aberration balancing. To determine the tolerances of the optical design, the M2 mirror substrate was misaligned with respect to the M1/M3 monolithic mirror, and the realistic maximum error was evaluated. As a result, in the range of 36.6 ?? root mean square (RMS) spot radius for the three degrees of freedom (DoF), the calculated tolerances were 0.0072° for X-rotation, 43 µ? for Y-translation, and 7.6 µ? for Z-translation. This result confirmed that the required machining accuracy to build the proposed telescope is within the advanced optical fabrication (e.g., diamond turning) and modern integration capabilities.
• 

  Efficacy of Chlorine and Peracetic Acid to Reduce Shiga Toxin-Producing Escherichia Coli and Impact of Simultaneous Nitrogen-Based Fertilizer Use on Microbial Die-Off in Preharvest Agricultural Water

  Rock, Channah; Scott, Zoe; Cooper, Kerry; Gerba, Charles (The University of Arizona., 2023)

  Several foodborne disease outbreaks in the United States have been linked to the consumption of various types of leafy greens in which irrigation water was suspected as the potential source of contamination. To reduce potential produce contamination from agricultural water, the U.S. Food and Drug Administration (FDA) has proposed regulations/metrics which would require growers to assess their agricultural water systems. In some cases, this would mean monitoring their water quality and taking corrective action, by way of antimicrobial treatments, when agricultural waters are deemed as a “reasonably likely foreseeable hazard”. Additionally, the Arizona and California Leafy Greens Marketing Agreements (AZ/CA LGMA) require growers utilizing surface water for overhead irrigation to treat their water within 21 days of harvest to meet acceptable risk indicators; generic Escherichia coli (E. coli) (non-detect per 100mL) and Total Coliform bacteria (<99 MPN/100mL). For many growers, this will be the first time that water quality data may necessitate them to use an antimicrobial treatment before irrigation can be applied safely. Additionally, growers are faced with a myriad of options related to antimicrobial water treatment with very little guidance on the most appropriate treatment option for their ranch, or the requirements needed for successful implementation. With limited guidance, water treatment decisions are likely to be unsuccessful and expend both excess time and money while seeing little to no reduction in potential pathogen loading in an agricultural water source and thus little to no reduction in microbiological risk. To provide guidance on antimicrobial agricultural water treatment options available to industry, the efficacy of two antimicrobial treatments Peroxyacetic Acid (PAA) and Calcium Hypochlorite (Cl) were tested, in triplicate. Tests were executed for various rates of each antimicrobial product (sanitizer), 6 & 8 PPM for PAA and 2 & 4 PPM for Cl. For each sanitizer at each PPM, tests were conducted at temperatures 12°C and 32°C. To evaluate sanitization efficacy, the team measured the reduction of a 109 CFU/mL cocktail of Shiga toxin-producing E. coli (STEC) strains (ATCC MP-9 and 43895) in four water sources from across the southwestern United States (Yuma and Maricopa, AZ, Uvalde and Edinburgh, TX). Four different water sources were used to gauge if water quality impacted sanitization efficacy. The experimental design was based on an EPA/FDA protocol to assess the efficacy of an antimicrobial product to reduce foodborne bacteria in pre-harvest agricultural water (https://www.fda.gov/media/140640/download) . This protocol dictates that STEC cocktail be added to agricultural water then equilibrated at either temperature (12°C or 32°C); post equilibration, each sanitizer, for each concentration, is injected into the mixed solution. The appointed contact time (1 or 5 minutes) is given and then the solution is neutralized and evaluated. To further growers’ comprehension of best management practices for successful antimicrobial treatment application, the impact of two nitrogen-based fertilizers (UAN32 and CAN17) on the efficacy of Sodium Hypochlorite 6% (chlorine) and PAA against naturally occurring coliforms was also evaluated. The first study provides evidence that chlorine meets EPA’s required 3-log reduction of pathogens in order to receive label approval. At a one-minute contact time, the chlorine treatment resulted in log reduction values (LRVs) ranging from 3.24 to 6.15 regardless of temperature, dose/PPM, or water source. PAA however did not perform as well with LRVs ranged from 0.0 to 1.10 with higher reduction occurring at the higher temperature and dose of PAA. When the contact time of PAA treatment was increased to five minutes, LRVs increased and ranges from 1.5 to 5.4 were observed; the efficacy of the sanitizer increased with increased solution temperature. Furthermore, the addition of nitrogen-based fertilizer to the water source in tandem with treatment application significantly affected the antimicrobial capabilities of chlorine. For chlorine, when applied unaccompanied an average log reduction of 3 logs was seen. However, LRVs decreased on average by 1.34 logs when fertilizer was introduced: with the greatest reduction in efficacy resulting in a nearly 2-log decrease. Contrarily, combined application of PAA and either fertilizer showed little to no interaction with a 0.4 log increase in disinfection efficacy when UAN32 was used. Results indicate that a prolonged contact time may be needed to meet regulations when PAA is used as an antimicrobial treatment. As well, growers must be cautious when applying fertilizer conjointly with antimicrobial treatment to their agricultural waters to ensure compliance with new proposed food safety metrics.
• 

  Computational Design Optimization and Reliability Assessment of Thermal Systems

  Missoum, Samy; Pidaparthi, Bharath; Li, Peiwen; Chan, Cholik; Ditzler, Gregory (The University of Arizona., 2023)

  Transition to renewable energy solutions, while managing surging energy demand, calls for novel thermal designs. These designs, especially with the advent of additive manufacturing, are becoming increasingly complex and computationally driven. Optimization and reliability analysis of such complex designs typically require several evaluations of the quantities of interest. This is usually accomplished by querying computational models often involving expensive numerical methods like Finite Element Analysis and Computational Fluid Dynamics. To alleviate the computational cost of these design routines, surrogate models can be employed in place of the original model, as they are cheaper to evaluate. In addition, several models, both computational and experimental, are often available to describe a system of interest. These models have varying evaluation costs and fidelities. In general, an expensive high-fidelity model describes the system with the accuracy required for the task at hand, while lower-fidelity models are less accurate but computationally cheaper. In such situations, multi-fidelity procedures can combine information from different levels of fidelity to accelerate the optimization and reliability routines. In this work, these two concepts (i.e., surrogate modeling and multi-fidelity) are employed for optimization and reliability analysis of concentrated solar receiver tubes and heat exchangers.
• 

  Structural and Diffusion MRI to Study the Effects of Hypertension in Rat Brain Macrostructure and Microstructure

  Trouard, Theodore; Wiskoski, Haley Elizabeth; Hutchinson, Elizabeth; Chen, Nan-kuei (The University of Arizona., 2023)

  Hypertension (HTN) is associated with an increased risk of cardiovascular disease (CVD) and cognitive decline in aging humans, with onset occurring around middle age, and responsible for roughly 7 million deaths worldwide, annually. Prior research has also shown that mid-life HTN is associated with negative effects on brain structure and function in late life. Therefore, it is important to study the symptoms of HTN on the central nervous system as the disease progresses with age, and specifically how this may affect neurological anatomy, development, and function. Animal models are an integral tool in preclinical, translational research of the human body, facilitating greater understanding, treatment, and prevention of diseases such as HTN. The Fischer-344 Cyp1a1-Ren2 transgenic xenobiotic-inducible rat model is an appreciable strain in studies of HTN due to the fact that the induction of increased blood pressure, as induced via the administration of dietary molecule Indole-3-Carbinol (I3C), is reversible, controllable, and dose-dependent in magnitude. The purpose of this study was to investigate the longitudinal effects of induced HTN in macrostructural and microstructural neuroanatomy of F344 Cyp1a1-Ren2 transgenic rats through the use of noninvasive diffusion-weighted MRI (dMRI) and imaging analyses. Results of this study show that even in the face of sustained increases in blood pressure and end-organ damage in the heart and kidney, a majority of the brain remained unaffected in terms of volume and microstructural characteristics. This indicates the presence of an intrinsic, protective mechanism of the brain in this model, forestalling the onset of detrimental effects of HTN on brain structure and function.
• 

  Estimation of Diffractive Surface Profile using Phase Retrieval Techniques

  Schwiegerling, James T.; Ryu, Jieun; Milster, Thomas D; Sasian, Jose M. (The University of Arizona., 2023)

  An intraocular lens (IOL) is an artificial lens that is inserted into the eye as part of a treatment for cataract or myopia. Among the many types of IOLs, multifocal IOLs with diffractive optics design have been demonstrated to provide superior vision for both distance and near vision after surgery. In this dissertation, methods for estimating the diffractive surface profile of a multifocal diffractive IOL are investigated. Traditionally, several types of instruments have been proposed to determine the phase profile, such as conventional interferometers and Shack-Hartmann sensor. Holography-based setups have also been widely used for surface profile measurement. However, the proposed conventional methods require additional optical components or a reference beam, increasing system complexity and cost of the system. To avoid the limitations of the conventional system stated above, phase retrieval technique is implemented to estimate the diffractive surface profile of an IOL. The phase retrieval technique is the process of recovering the complex-valued function given the magnitude of its Fourier transform. It is natural to investigate the phase of an object, as optical imaging devices only measure the intensity of light and cannot measure the associated phase directly. This dissertation examines several phase retrieval algorithms. The multiplane phase retrieval algorithm described by Gerchberg is implemented for embodiment of the methods. To acquire multiple diffraction patterns, some techniques were used, such as displacing the imaging sensor to record intensities at different planes, modulating phase in Fourier domain using spatial light modulator (SLM) to record a sequence of intensities with different image planes, and modulating phase in pupil plane using SLM to record diffraction patterns at an image plane with phase diversity. Employing the considered phase retrieval schemes, the phase profile of diffractive multifocal intraocular lens was estimated. This dissertation begins with background section, which describes the human eye and basic principles of multifocal intraocular lenses. Chapter 1 covers basic optical structure of the human eye, age-related eye conditions, and fundamentals of intraocular lenses. Chapter 2 motives and describes several basic phase retrieval algorithms and limitations in finding the phase profile of the diffractive surface of an intraocular lens, followed by computational simulations. Chapter 3 deals with phase retrieval technique employed in combination with the multi-plane phase retrieval method, as well as the selected SLM-based phase retrieval technique suitable for finding phase profile of the diffractive intraocular lens. Chapter 4 describes the experimental setup employed for validating the phase retrieval technique to measure the wavefront profile of the intraocular lenses. The calibration procedure implemented in the experiment is discussed. Chapter 5 concludes the dissertation.
• 

  3 Essays in Natural Disaster Adaptation

  Bakkensen, Laura; Blair, Logan; Baldwin, Elizabeth; Smith, Craig (The University of Arizona., 2023)

  Natural disasters present a considerable threat to the welfare of society. This is especially true for growing coastal populations facing stronger and more frequent hurricanes. In response, special building codes in the US are now available through the International Code Council (ICC) that mandate technology like reinforced roof-to-wall connections in new coastal homes ex-ante, and ex-post aid programs through the Federal Emergency Management Agency (FEMA) have grown tremendously. However, despite their critical roles, ICC codes have only been fully adopted by one state, Florida, while others debate their costs and market alternatives. There also remain salient concern over the equitable distribution of FEMA aid for which the literature struggles to untangle. This dissertation asks whether hazard-resistant building code institutions hold legitimizing public values outside of the market, in particular the ability to sidestep damage associated with behavioral bias, internalize physical externalities, and attenuate long run recovery for low-income households. In addition, I argue that our current understanding of equity in ex-post aid is limited by geographic and program aggregation, and ask whether FEMA’s Blue Roof Program, which installs temporary roofs following a hurricane, contribute to class and race-based disparities at the program-property level. Employing a novel property-level dataset on damages over time derived from remotely sensed images of Bay County Florida following Hurricane Michael, I find strong evidence that hazard-resistant codes reduced roof damage through improved preparedness, spill over beyond the individual to protect nearby homes from debris, and attenuated long recoveries---especially in low-income areas. This implies that places exposed to similar disaster risk, but resist modern code institutions, may miss welfare gains and a considerable opportunity to protect their most vulnerable populations. Second, I find that FEMA’s Blue Roof Program prioritizes homes with larger roofs, likely due to associated private contract incentives. However, market alternatives appear to perform worse along these dimensions, and neither arrangement discriminated along indicators of class or race such as home value and percent non-white.
• 

  Managing Uncertainty in Collaborative Governance: Multi-Method Evidence

  Baldwin, Elizabeth; Emerson, Kirk; Ahn, Minwoo; Smith, Craig R.; Jo, Suyeon (The University of Arizona., 2023)

  Scholars have recognized the importance of uncertainty as institutional context in collaborative water management, but the relationship between uncertainty and collaborative performance is mixed. On the one hand, increased uncertainty will positively enhance the performance of collaborative governance through new ties and innovations. On the other hand, certain types of uncertainties are negatively associated with collaborative performance. To understand the puzzle of uncertainty and collaborative performance, I take the problem of groundwater management for theory development and empirical testing. Aquifers are being depleted faster than they can recharge, leaving municipalities, irrigators, and ranchers vulnerable to ever-reducing water availability over time, but the management of groundwater problems is wicked because it involves complex social, ecological, scientific, administrative, and political issues. The effectiveness of collaborative governance depends in large part on the way in which stakeholders perceive, interpret, and use uncertain information.This dissertation fills the theoretical and empirical gap by using multi method research design. The first research question is: What are the nature and characteristics of uncertainty in collaborative governance? This question is addressed based on the in-depth case study of Upper San Pedro Watershed Partnership in Arizona, U.S. Based on the various sources of empirical data, including 22 in-depth interviews, policy reports, and local news articles, conceptual typology and theoretical propositions are proposed to develop theories of collaborative governance under uncertainties. Results suggest that scientific and managerial uncertainty are significant and tend to have negative effects on the performance of groundwater management, but the relationship between uncertainty and collaborative performance can be positively or negatively moderated by the quality of relation management including integrative leadership and cohesion building between participants. Results also suggest that levels and sources of uncertainty tend to change as collaboration evolves and thus the relationship between uncertainty and performance may shift over time. Having recognized that understanding scientific uncertainty is important in groundwater management based on the case study, this dissertation asks two questions: How and to what extent does scientific uncertainty affect collaborative performance? Do collaborative management tools have an impact on different types of collaborative outcomes, particularly under the condition of scientific uncertainty? This dissertation modified a groundwater game experiment where groups of 4-5 participants play a crop choice game for multiple rounds as resource users (Meinzen-Dick et al. 2016). The goal of this game for each participant is to grow as many profitable crops as possible under conditions where all users share groundwater resources with limited ability to recharge. But unlike the original game, where participants had full information about recharge rate, two treatments are introduced about scientific uncertainty in water recharge: uncertainty operationalized as a range of values (Treatment 1) and uncertainty operationalized as competing hydrological models (Treatment 2). Using quantitative and qualitative game experimental data from 30 groups, results suggest that more uncertain information tends to reduce individual earnings and thus increase shared resources. A range of uncertain information has a more significant impact on resource behavior than competing information. Finally, post-experimental analysis shows that diverse collaboration strategies tend to reduce distributional inequity among game participants. This dissertation contributes to the literature of collaborative governance and collective action by explicitly theorizing and modelling the relationship between uncertainty, collaboration process, and performance.
• 

  Advancing Neural Networks Towards Realistic Settings Using Few-Shot

  Ditzler, Gregory; Hess, Samuel Thomas; Akoglu, Ali; Tandon, Ravi (The University of Arizona., 2022)

  Neural networks have shown remarkable performance across many tasks, including classification, object detection, and image segmentation. Advances in high-performance computing have enabled neural networks to train on extremely large datasets that have resulted in superior performance, often outperforming humans in many tasks. In fact, conventional supervised learning neural networks trained with large volumes of labeled data can produce highly accurate models to classify images, videos, and audio signals. Despite the success of neural networks, their deployment and evaluation are limited to the classes and experiences observed during training. The success of neural networks, however, poses a serious challenge if large labeled datasets are not available to train. Thus, these models are not expected to achieve the same success if there are only a few labeled samples per class. To address this weakness of sample size, an area of research is rapidly evolving known as few-shot learning. Specifically, few-shot learning classifies unlabeled data from novel classes with only one or "a few'' labeled exemplary samples. Unfortunately, few-shot learning comes with its challenges, including reduced classification accuracy with respect to supervised counterparts, requirements on the overall size of the training data, classifier explainability, and evaluation assumptions that can quickly break down with many real-world applications. It is against this background that in this thesis, we present five contributions that expand few-shot performance, explainability, and applicability to new novel tasks. Specifically, our contributions are: (1) A novel few-shot network that improves the classification accuracy over prior models by learning to weight features conditioned on the samples. Conventional techniques perform a one-way comparison of an unlabeled query to a labeled support set; however, the soft weight network allows for two-way cross-comparisons of both query-to-support and support-to-query, which is shown to improve the performance of a few-shot model. (2) A new application and novel few-shot network, namely OrderNet, that can accurately learn an ordering of data given a small labeled dataset. Through pairwise subsampling and episodic training, OrderNet was shown to significantly reduce the amount of training data required to achieve regression accuracy. (3) A new approach for eXplainable Artifical Intelligence (XAI), namely ProtoShotXAI, that uses a few-shot architecture to explain black-box neural networks and is the first approach that is directly applicable to the explanation of few-shot neural networks. (4) A novel similarity metric for a few-shot network that achieves state-of-the-art performance on inductive few-shot tasks. The metric is motivated by the fast approximation of exponentially distributed features in the final layer of a trained few-shot classifier, and maximum log-likelihood estimation. State-of-the-art 1-shot transductive performance is also achieved on imbalanced data using a simple iterative approach with our similarity metric. (5) A novel framework for online detection and classification using few-shot classifiers. In contrast to related work, our lifelong learning framework assumes a continuous data stream of unlabeled and imbalanced data. Additionally, our approach continuously refines classes as new data becomes available while considering computational storage constraints. We demonstrate the capabilities of our proposed approach on benchmark data streams and achieve competitive detection performance and state-of-the-art online classification accuracy.
• 

  Chiefs, Elections, and Violence: Mobilization and Demobilization of African Voters

  Braithwaite, Jessica M.; Chen, Xiran; Braithwaite, Alex; Cyr, Jennifer; Schuler, Paul; Turnbull, Megan (The University of Arizona., 2023)

  Why do some electoral districts experience more pre-election violence than the others in national elections? This dissertation examines how a particular type of local actor – African chiefs – affects pre-election violence locally. I make two arguments regarding the conditions under which chiefs are capable of deterring pre-election violence targeting their communities, and under which chiefs and their subjects are motivated to participate in pre-election violence. In my dissertation I first argue that centralization of kin groups in precolonial era enhances chiefs’ capability of voter coordination in contemporary time, and in turn reduces the risk of pre-election violence. Using survey- and event-based data from both existing and original datasets, I find a negative relationship between precolonial centralization of kin groups and pre-election violence. Further results of two-stage least squares regressions confirm the internal validity of this relationship. These findings apply to cases where indirect rule was adopted and customary land tenure preserved under colonial government, such as much of the Anglophone West Africa, because precolonial institutions have been better preserved in such cases. My second argument concerns how kin-group-based chieftaincy disputes drive royal families to fight one another during the election periods. Having local aspirants in the challenger families – who seek to change the status quo in chieftaincy disputes – increases the risk that chieftaincy disputes escalate into violent conflicts during the elections. Local aspirants, politicians who have dual identities as political party and royal family members, have particular interests in causing political parties to interfere with chieftaincy disputes. As the outcomes of chieftaincy disputes become associated with the outcomes of national elections, disputing royal families have strong motivation to fight each other during elections. I adopt a most-similar case design based on qualitative data collected through field research in Ghana, and inductively develop a theory of politicization of chieftaincy disputes. The findings of this dissertation demonstrate the complex functions of chiefs and their institutional foundations in African elections. In particular, the institutions of kin group structure local actors’ interests in such a way that they could be motivated to support and undermine democratic processes at the same time. These findings contribute new arguments and evidence to the debate about the relationships between traditional and democratic institutions. In addition, they also highlight the heterogenous colonial legacies between the Anglophone countries in Africa. Precolonial institutions are in general better preserved in Anglophone countries in West Africa then in other countries. Lastly, the findings also have policy implications. Chiefs can become valuable local non-state actors that join forces with international and national actors in pre-election violence prevention. It is also necessary to develop legal and policy instruments that separate politicians from traditional affairs.
• 

  Team Approach to Prescription of Naloxone with Opioid Therapy

  Prettyman, Allen; Soto, Anne-Jeanette; Sowicz, Timothy; Gray, Bianca (The University of Arizona., 2023)

  Purpose: This quality improvement (QI) project utilized an educational session to increase the prescription of naloxone with opioid therapy and to introduce a team approach that encompasses the Primary Care Provider (PCP), Clinical Pharmacist, Registered Nurse (RN), and Licensed Practical Nurse (LPN). Background: Although naloxone is an opioid antagonist and can rapidly reverse opioid overdose, there continues to be hesitation among primary care providers to prescribe naloxone with opioid therapy. Potential barriers include lack of proper training, fear of offending patients, and logistical prescribing difficulties attributed to the lack of naloxone co-prescription. Method: An educational session discussing the benefits of naloxone, the myths associated with naloxone prescription, and the proposed change in workflow regarding the prescription of naloxone with opioid therapy will be presented. Results: Two weeks after the educational session, there was a 23% increase in naloxone prescriptions with opioid therapy at the Sierra Vista CBOC. Conclusions: Despite federal recommendations and organizational resources, there continued to be hesitation in prescribing naloxone with opioid therapy due to barriers and stigma related to the medication. Educating providers and healthcare professionals about the importance of naloxone have proven that using a team approach is effective in advocating naloxone with opioid therapy.
• 

  Gender Identity and Expression in Drum Corps Performance: An Ethnographic Study of Performers' Experiences

  Gubner, Jennie; Frew, Johnathan; Alexander, Kathryn; Mora, Amalia (The University of Arizona., 2023)

  This project seeks to understand how gender is constructed, performed, and experienced through participation in drum corps. Drum corps is a North American youth performing arts practice in which musicians and dancers perform themed marching band shows throughout the United States and are judged competitively against one another. The data used in this project comes from four weeks of ethnographic fieldwork with the Blue Knights drum corps during the summer of 2022. Personal accounts from the corps’ members and staff, fieldwork photographs, and the author’s observations are analyzed in this project to explore performers’ experiences with gender in drum corps.
• 

  Connecting Immigrants to Mental Health Providers and Community Resources

  Prettyman, Allen; Velez, Juan C.; Kiser, Lisa; Kading, Valerie (The University of Arizona., 2023)

  Purpose: This project was designed to determine whether connecting immigrants with mental health practitioners and community resources could be beneficial for increasing awareness of such resources within the immigrant community. Background: Immigrants represent a large percentage of the population in the United States (US), accounting for over 44.8 million individuals. However, the immigrant population is disproportionally affected by mental health disorders, with 30.8% and 30.65% suffering from depression and post-traumatic disorder, respectively. Immigrants often lack access to mental healthcare due to frequent barriers, including undocumented status, lack of health insurance, low income, communication barriers, and coping behaviors, among other obstacles. Methods: The project adopted a quantitative, descriptive approach to determine the impact of connecting immigrants at a local clinic in Tucson, Arizona, with mental health providers and mental health community resources. Self-reported data were collected from 16 immigrants using a self-developed survey. The data was then analyzed using descriptive statistics. Results: The survey findings demonstrated an increased awareness of food banks providing groceries and an understanding of shelter organizations and self-help groups designed to address needs within the immigrant community. Linking immigrants to mental health providers increased their knowledge of available resources. The Wilcoxon Signed-Ranks Test yielded P<0.05, demonstrating a significant improvement in awareness of mental health resources. Conclusion: Connecting immigrants to mental health providers and community resources significantly improve immigrants’ awareness of mental health resources and strengthens a collaborative care approach compared to current practice
• 

  Mitigating Data Scarcity for Neural Language Models

  Surdeanu, Mihai; Van, Hoang Nguyen Hung; Surdeanu, Mihai; Bethard, Steven; Hahn-Powell, Gus; Levine, Joshua (The University of Arizona., 2023)

  In recent years, pretrained neural language models (PNLMs) have taken the field of natural language processing by storm, achieving new benchmarks and state-of-theart performances. These models often rely heavily on annotated data, which may not always be available. Data scarcity are commonly found in specialized domains, such as medical, or in low-resource languages that are underexplored by AI research. In this dissertation, we focus on mitigating data scarcity using data augmentation and neural ensemble learning techniques for neural language models. In both research directions, we implement neural network algorithms and evaluate their impact on assisting neural language models in downstream NLP tasks. Specifically, for data augmentation, we explore two techniques: 1) creating positive training data by moving an answer span around its original context and 2) using text simplification techniques to introduce a variety of writing styles to the original training data. Our results indicate that these simple and effective solutions improve the performance of neural language models considerably in low-resource NLP domains and tasks. For neural ensemble learning, we use a multi-label neural classifier to select the best prediction outcome from a variety of individual pretrained neural language models trained for a low-resource medical text simplification task.
• 

  Provider Education Regarding Hospice and Palliative Care: Improving Patient Outcomes

  Martin Plank, Lorraine; Spall, Laura Elizabeth; Godfrey, Timian; Daly, Patricia (The University of Arizona., 2023)

  PurposeThe purpose of this quality improvement initiative was to provide hospice and palliative care education to healthcare providers to increase their willingness and intent to refer patients to them. Background Hospice and palliative care have been used for many years to improve patient quality of life by decreasing pain, anxiety, and overall symptom burden for patients suffering from terminal illnesses (American Cancer Society [ACS], 2019; Center to Advance Palliative Care, 2022). While these services have well established benefits for patients and their families, they are often not utilized until very late into a patient’s terminal diagnosis (Devi, 2011). There are many potential reasons for a lack of or late referrals to hospice and palliative care, but a healthcare provider knowledge deficit may play a major role. Methods An online educational PowerPoint session (OEPS) on hospice and palliative care was administered to the healthcare providers at the Bear River Valley Hospital, along with a follow up post-then-pretest survey. The healthcare providers were asked to answer questions about their knowledge of, attitude towards and perception of intent to refer to, and experiences with hospice and palliative care both before and after viewing the OEPS. The healthcare providers were asked to provide two Likert scale scores of 1 to 5 for each question, one being associated with before the education and one being associated with after. Results Eleven healthcare providers participated in the OEPS and follow up survey. There was an overall reported increase in Likert scale score for each question following administration of the OEPS. The greatest increase in score was noted for the questions regarding knowledge of and intent to refer patients to hospice and palliative care. Conclusions Hospice and palliative care education for healthcare providers has the potential to increase their intent to refer patients to these important services. This may translate to an increase in actual patient use of hospice and palliative care, potentially improving their outcomes.
• 

  Ultrafast Photoelectron Spectroscopy of Electron Dynamics in Atoms and Molecules

  Sandhu, Arvinder; Plunkett, Alexander C.; LeRoy, Brian; Schaibley, John R.; Hassan, Mohammed; Golubev, Nikolay (The University of Arizona., 2022)

  This dissertation presents experiments studying electron dynamics in atoms and molecules using the tools of ultrafast photoelectron spectroscopy. The first experiment studies extreme-ultraviolet excitation and relaxation dynamics in molecular oxygen and identifies a previously undiscovered excitation and dissociation pathway, filling a hole in the spectroscopic library of oxygen that had existed for almost twenty years. This is in fact a multielectron excitation and would therefore be expected to have a very low excitation cross-section. Fresh analysis of the Fano approach to autoionizing states revealed that this can be explained as a sort of backwards autoionization– “excitation” initially proceeds to the continuum and population is transferred back to the discrete state via the same coupling mechanism as autoionization. The second and third experiments stem serendipitously from the same few sets of exploratory data sets taken in argon when testing a new optical technique. In one an argon atom is excited into a superposition of autoionizing states and this wavepacket is probed with a delayed IR pulse. This has the unintended but beneficial effect of inducing Raman transitions among the wavepacket constituent states and allowing study of its dynamics with unprecedented time and energy resolution simultaneously. Finally, a similar Raman process is used to probe a superposition of bound states, this time leading to rich and unanticipated angular structure in the electron emission. Categorically, the method used across all these experiments is known as “time-resolved photoelectron spectroscopy” and consists of an initial excitation, or “pump,” of an atom or molecule followed by a time-delayed “probe” inducing some measurable effect. Most simply, the probe photoionizes the system, thereby producing a photoelectron with measurable momentum and energy, but more subtle probes can be designed where, for example, it effects the rate of autoionization. The experimental apparatus including the laser system and the velocity-map imaging spectrometer are presented and discussed.
• 

  The Role of Social Identities, Motivational Beliefs and Environmental Support in College Aspirations

  Marx, Ronald W.; Friesen, Rebecca; Cheng, Katherine C.; Cimetta, Adriana D. (The University of Arizona., 2023)

  The benefits of higher education for society are numerous: addressing many pressing social and environmental challenges requires a diverse workforce with advanced degrees; democracies thrive with knowledgeable and caring voters; and a productive economy needs a well-trained workforce. Consequently, increasing the number and diversity of students attending college is a forefront concern. However, gender and racial/ethnic differences in college enrollment and completion persist. Given the importance of higher education for individuals and society, understanding the factors underlying these differences is critical. Students’ expectations of success and how they perceive the value of education affect their post-secondary plans. Social identities and the sociocultural context influence the development of these expectations and values. In particular, an individual’s identified race/ethnicity, affiliated gender, motivational beliefs and perceived environmental support can enhance or detract from the formation of college aspirations. However, few studies have examined how these factors interact to affect intention to complete college. With a diverse sample of 4803 high school students from the 2018 Programme for International Student Assessment United States data, I used logistic regression and structural equation modeling to examine gender and racial/ethnic differences in the extent to which 1) motivational beliefs, including self-efficacy, perceived usefulness of education, mastery motivation, and sense of purpose predict college aspirations; 2) environmental support (i.e. parent support, sense of belonging, and bullying), mediated by motivational beliefs, predict college aspirations. Results show that each motivational belief significantly predicts college aspirations for students overall, but differences emerge when examining individual gender and racial/ethnic groups. High perceived utility of education and mastery motivation significantly increased college aspirations for Black and White students; self-efficacy and mastery motivation for Hispanic students; and high mastery motivation only for Asian American students. In addition, motivational beliefs fully mediate environmental support for Black, Hispanic, and White students, but not Asian American students. The structural equation model did not significantly explain the variance in college aspirations (R2 = 0.28, p = .31) for Asian American students, but it did significantly explain 36% (R2 = 0.36, p < .001) for Black students, 29% (R2 = 0.29, p < .001) for Hispanic students, and 34% (R2 = 0.34, p < .001) for White students. The study highlights the value of increasing high school students’ motivational beliefs, particularly mastery motivation, but it also indicates that harnessing existing motivation beliefs and environmental supports promises to foster college aspirations, ultimately benefitting both individuals and society.
• 

  Host Galaxies of Extragalactic Transients

  Zabludoff, Ann; Qin, Yujing; Smith, Nathan; Sand, David; Behroozi, Peter; Merchant, Nirav (The University of Arizona., 2022)

  Host galaxy properties of extragalactic transients provide valuable insights into the nature of their progenitors. This dissertation focuses on the automated identification of transient host galaxies, the systematic difference in host properties and rate dependencies across supernova subtypes, and the demographics and cosmic environments of supernovae without host galaxies detected.Chapter 1 presents a brief overview of transient-host connections. Chapter 2 describes a novel machine learning-based algorithm to identify transient host galaxies with high accuracy (above 97%). Based on this algorithm, we develop software pipelines to identify transient host galaxies, collect their measured properties across multiple sky surveys, and assemble the largest-ever value-added database of host galaxies. Chapter 3 compares host galaxy physical properties (stellar mass, star formation rate, metallicity) and photometric properties (absolute magnitude, rest-frame color) across supernova subtypes and characterizes how supernova rates depend on galaxy properties. We find subtle but statistically significant differences in host properties across several core-collapse supernova (CC SN) subtypes. Contrary to common belief, CC SN rates are not proportional to their host star formation rates -- either a fraction of long-lived progenitors or a metallicity-dependent supernova production efficiency better interpret the observed host properties. SN Ia subtypes feature heterogeneous host properties attributable to the dramatic contrast of progenitor ages. Chapter 4 presents the analysis of a supernova sample without hosts detected. There are more interacting and superluminous (especially hydrogen-deficient) subtypes among hostless supernovae than in a typical transient sample. Wide-field galaxy surveys put rigorous limits on their host luminosities -- the faintest ones are close to the typical luminosity of dwarf spheroidal galaxies or even globular clusters. The lack of spatial association with galaxy clusters and the excess of slowly-declining SN Ia imply a population of star-forming dwarf hosts in the field. Finally, Chapter 5 summarizes the main results and provides a future outlook.
• 

  Autonomous Vehicle Security Framework (AVSF)

  Hariri, Salim; Satam, Shalaka; Akoglu, Ali; Salehi, Soheil (The University of Arizona., 2022)

  The rapid growth of the Internet and Cloud Computing has started a data revolution, with data analytics playing a critical role in all areas, including transportation. Vehicles are increasingly relying on integration of data from an array of sensors (over different communication networks) to perform semi/fully autonomous operations. These vehicles use sensors like onboard cameras, LiDARs, Global Positioning System (GPS) deployed as Engine Control Units (ECUs) communicating over Controller Area Network (CAN) bus. As vehicles become increasingly autonomous, they will increasingly rely on V2X communications, including Vehicle to Vehicle communications, and Vehicle to Infrastructure communications, for information sharing and performing predictive analysis for a faster, safer, and more comfortable ride. However, this increased vehicle automation relying on sensors and communications has exposed them to cyberattacks. Researchers have successfully demonstrated attacks on the latest Jeep Cherokee, Tesla, and Lexus models, by disabling critical safety features in the vehicle, creating fatal accident-causing scenarios. Vehicle manufacturers rely (unsuccessfully) on encryption, obfuscation techniques, and individual intrusion detection systems to detect and prevent cyber-attacks. These approaches fail to secure the vehicles (as showcased in attacks on Tesla and Lexus), as these vehicular attacks are highly complex, where the attacker exploits multiple systems simultaneously. Such complex attacks are very hard to detect, and counter (once detected) as countermeasures taken to prevent the attack has the potential to disable a critical autonomous vehicle subsystem. Thus, there is a need for a security framework that takes into account a holistic perspective of vehicle operations before taking mitigating actions. This dissertation presents an Autonomous Vehicle Security Framework (AVSF) that uses a combination of highly granular anomaly behavior analysis based intrusion detection systems to detect cyberattacks on various vehicle subsystems. The proposed AVSF performs decision fusion of individual threats to obtain a global vehicle impact and risk analysis before taking mitigation actions to stop the cyberattack. To showcase the efficacy of the proposed AVSF, I present a threat model analyzing the potential threats (and attack vectors) targeting autonomous vehicles and present three anomaly-based intrusion detection systems to detect attacks on different components of autonomous vehicles. For the threat modeling process, I present an autonomous vehicle architecture that splits the complex vehicle into four layers: End Devices layers, Intra Vehicle Communications layer, Inter-Vehicle Communications layer, and applications layer. The presented threat model analyzes and rates each threat on every layer in the autonomous vehicle architecture using the DREAD framework. This dissertation presents an Intrusion Detection System to detect attacks on vehicle sensors and actuators (End Devices). This sensor intrusion detection system (IDS) models the sensor's normal behavior using Discrete Wavelet Transform (DWT). The DWT uses Biorthogonal, Daubechie, Coiflets, Discrete Meyer, Reverse Biorthogonal, and Symlets wavelets to model the spatial and temporal features of the sensor. Experimental analysis shows the IDS can detect attacks like Denial-of-Service attacks, Impersonation Attacks, Random signal attacks, and Replay attacks with One-Class SVM, Local Outlier Factor, and Elliptical Envelope. The One-Class SVM performed the best compared to the results of other machine learning techniques. This dissertation presents a second IDS to detect attacks on the Bluetooth Protocol (Inter-Vehicle Networks). This Bluetooth IDS uses an n-gram based approach to create a behavior model for characterizing the normal behavior of the AV using the Bluetooth protocol by monitoring the protocol's state transitions. Attacks on the Bluetooth network are detected using machine learning algorithms like Decision Trees, AdaBoostM1, SVM, Naïve Bayes, Ripper, and Bagging algorithm, with precision up to 99.6% and recall up to 99.6%. Lastly, I present an IDS to detect attacks on CAN-BUS (Intra Vehicle networks). Attacks on the CAN bus are detected using novelty detection algorithms like SVM, Isolation Forest, Local Outlier Factor, and Elliptic Envelope. I also use two class classifiers like SVC, Multilayer perceptron, Decision tree, Random Forest and AdaBoost to classify the data with precision up to 100% and recall up to 100%.

View more