• Aquatic and Riparian Connectivity in Arid Landscapes

      Sharma, Akanksha (The University of Arizona., 2018)
      Aquatic and riparian ecosystems are of critical importance in arid environments, supporting a diverse suite of resident and migratory species over different life stages. Ecological connectivity is an important property in the functioning of these ecosystems, and a significant subject of interest for researchers,scientists,resource managers, practitioners and other stakeholders. Furthermore, a variety of perceptions exists on aquatic and riparian connectivity among stakeholders, and connectivity of these ecosystems in arid landscapes is a relatively unexplored subject. I focused on these issues in the US portion of the Madrean Archipelago by combining qualitative methods to capture the diversity of perspectives among experts and quantitative spatial analysis to capture the variety of factors that influence aquatic and riparian connectivity. I synthesized the resultant expert perspectives into a Connectivity Component-Dimension Framework that deconstructs aquatic and riparian connectivity into connectivity components and their dimensions. Using GIS and regression analysis, I applied this framework to a case study of the threatened Chiricahua leopard frog (Rana chiricahuensis) in the Cienega Creek basin in Arizona and created connectivity indices for this focal species. Some factors that emerged significant in this case study included elevation, fire hazard potential, and density of leopard frog sightings. This connectivity framework and the related indices provide customizable options for stakeholders to assess aquatic and riparian connectivity multidimensionally using readily available data. These tools can be used by stakeholders for exploratory analysis, assessment and visualization of aquatic and riparian connectivity, in arid landscapes, and beyond.
    • Passive Strategies to Improve Energy Efficiency in Existing and Pursuing Leed® Certified Buildings in Arid Regions

      Simmerman, Cecilia (The University of Arizona., 2018)
      Energy efficiency in buildings is vital for the environment and sustainability. Edifices are responsible for significant energy consumption and carbon dioxide emissions. “LEED® provides a framework to create healthy, highly efficient and cost-saving green buildings” (10). This framework that LEED® developed and the variety of paths to achieve points for certification make it very easy to bypass the energy category and produce underachieving buildings regarding energy efficiency. I think to create sustainable structures it is essential to employ passive strategies, and this study will illustrate that some LEED® Certify Building rely more on active systems rather than passive systems. This research will also demonstrate through energy simulation that passive strategies minimized external loads due to climate and are very effective in a hot arid climate. These strategies are sustainable reduce energy consumption are cost effective and without risk of mechanical or user failure. Because of investigation, a check list was developed to aid designers create more efficient structure using passive strategies.
    • Experimental and Flight Investigation of the Laminar Separation Bubble on an Oscillating X-56A Wing Section Near Stall

      Agate, Mark (The University of Arizona., 2018)
      An investigation of laminar separation bubble behavior on an oscillating X-56A wing section has been performed experimentally at Reynolds number 200,000. Wind tunnel results along with Implicit Large Eddy Simulations (CFD) quantify the behavior of the laminar separation bubble. The oscillation parameters were selected based on a scaled flight vehicle at the University of Arizona. Wind tunnel results were validated against theory using static angle of attack sweeps and an unsteady case at an angle of attack of $\alpha = 10$ degrees. The static results show excellent agreement between the experimental data, Thin Airfoil Theory, a computational vortex lattice method (XFLR5), and CFD results in both pressure coefficient and lift coefficient. The unsteady validation case of $\alpha=10$ degrees (nondimensional plunging frequency of $k = \frac{\pi f c}{U_{\infty}}=0.7$, where $f$ is the dimensional plunging frequency, $c$ is the wing section chord, and $U_{\infty}$ is the free-stream velocity, and nondimensional plunging amplitude $h = \frac{amplitude}{chord} = 3.2\%$) also showed agreement for comparison between the experiment, Theodorsen's theory (analytical solution to plunging wing sections), and CFD results. Pressure coefficient behaved similarly between the experiment and CFD with the laminar separation bubble changing pressures at similar times in the cycle. The lift coefficient was found to oscillate sinusoidally, achieving higher lift than the static case with no moment stall. Near static stall angle of attack ($\alpha=12$ degrees, where stall $\alpha=12.25$ degrees), Theodorsen's theory is no longer applicable. Oscillation parameters were $k=0.7$ and $h=4.8\%$ and effective angles of attack reached nearly $16$ degrees. The airfoil continued to produce lift past static stall at the consequence of a moment stall. Pressure measurements indicate that the laminar separation bubble is shed from the leading edge which was confirmed through 2D particle image velocimetry. The shedding behavior was modeled differently in the CFD simulation with a lack of free-stream turbulence. However, pressure coefficient and lift coefficient are in excellent agreement for over $75\%$ of the oscillation cycle. It is shown that the experimental setup is valid and the increased aerodynamic efficiency comes at the consequence of a moment stall for the high angle of attack case ($\alpha=12$ degrees). Additionally, free-flight tests have been completed including maiden flights of the 1/3 scale X-56A vehicle built at The University of Arizona. The flight vehicle is the motivation for the wind tunnel parameters. Flight instruments have been verified against previously collected data including pressure sensors, wing accelerometers (to track the motion), and a stand-alone constant temperature anemometry (CTA) system to measure free-stream turbulence. The instrumentation was flown on a stable platform to compare to historical data (1/5 scale Ximango) and is performing nearly 10 times as fast (data collection frequency) of the expected phenomenon occurring with the laminar separation bubble shedding on the 1/3 X-56A vehicle. This will need to be analyzed in future work as the laminar separation bubble is sensitive to free-stream turbulence conditions.
    • Differential Impacts of Passive versus Active Irrigation on Semiarid Urban Forests

      Luketich, Anthony (The University of Arizona., 2018)
      Trees provide benefits to the urban environment and irrigation is common to support these ecosystem services. In dryland communities where water resources are limited, collection and retention of stormwater runoff is used to passively irrigate the urban forest. However, the effects of passive irrigation versus regular, controlled moisture inputs, or active irrigation, is largely unquantified. We monitored the ecohydrology of urban mesquite trees (Prosopis spp) under these contrasting irrigation regimes in semiarid Tucson, AZ. Measurements included soil moisture, sap flow, canopy greenness, and leaf-area index. We expected both irrigation types to provide additional deep (>20 cm) soil moisture compared to natural conditions, and that trees would depend on this deep soil moisture for transpiration and phenological activity. Results show that active irrigation supported higher soil moisture, sap flow, and greenness during the dry conditions of spring. Following summer rain, greenness was higher under passive irrigation, despite sustained elevated soil moisture under active irrigation. Deep soil moisture had only slightly stronger controls over mesquite productivity than shallow moisture, and these relationships were stronger in the spring, rather than summer months. Finally, passive irrigation generally failed to provide additional deep soil moisture, though treatments in closer proximity to impervious surfaces did provide wetter soil conditions. This research aims to contribute empirical observations of green infrastructure performance and improved understanding of urban forest function for watershed management and planning.
    • Rationality and Resentment in the Egyptian Critique of Orientalism: The Example of Anouar Abdel-Malek and Ḥasan Ḥanafī

      Allosh, Islam (The University of Arizona., 2018)
      Fifteen years before Edward Said published his seminal book Orientalism, Anouar Abdel-Malek (1924-2012), an Egyptian alumnus of the Sorbonne and a Sociologist in the French National Center for Scientific Research (CNRS), had published his contentious article entitled “Orientalism in Crisis” in 1963. The essay placed Abdel-Malek as the first Arab thinker to critique Orientalism in a European language. In 1991, Ḥasan Ḥanafī (b. 1935), an Egyptian philosopher and Sorbonne graduate, published Introduction to the Science of Occidentalism. He presents the book as the first serious formation of an Eastern science capable of challenging Eurocentrism and countering Western Orientalism. The present study implements Partha Chatterjee’s (b. 1947) model of the three moments in the development of the Nationalist thought in India on Anouar Abdel-Malek and Ḥasan Ḥanafī in the context of restructuring the power relations between the East and the West. Chatterjee argued that nationalist thought in the colonial world, while seeking to liberate itself from the imperialist influence, remained a prisoner of the post-Enlightenment Western thought. It will be argued that Ḥasan Ḥanafī, who fits in the third moment, the moment of arrival, could not escape the Orientalist mode of knowledge. It will also be argued that Anouar Abdel-Malek, who fits in the third moment as well, has successfully managed to overcome the nationalist dilemma suggested by Chatterjee. The moment of arrival represents a fully developed ideology that embraces the different components of a nation.
    • Mean Flow Structure of Swept Impinging Oblique Shock Boundary Layer Interactions

      Doehrmann, Adam (The University of Arizona., 2018)
      An experimental investigation has been conducted to assess the e↵ect of sweep on the mean flow structure of impinging oblique Shock/Boundary Layer Interactions (SBLIs), specifically focused surface flow visualization and mean wall pressures. Four shock generators are utilized with x-y plane deflection of ✓ = 12.5!, and x-z plane sweep angles of 15.0!, 22.5!, 30.0!, and 40.0!. The swept oblique shocks impinge upon the naturally turbulent Mach 2.3 boundary layer along the tunnel floor (Re✓ ⇡ 5000). The resultant SBLIs all exhibit significant separation, with a structure that grows in the spanwise direction. Surface flow visualization shows a quasi-infinite region of separation that is limited by corner e↵ects at the root and tip of the interaction. The rise in mean pressure near separation scales locally with cylindrical similarity suggesting the three-dimensional separation along the span obeys Free Interaction Concept. Local reattachment behavior is only mildly dependent upon span. Convention from literature states that when the flow features, such as separation and reattachment lines are parallel, the interaction scales cylindricalyly. Conversely, when these flow features diverge from each other, the interaction scales conically. Divergence of separation and reattachment lines indicated that the global shock structure scales cylindrically for shock generator sweep angles less than 22.5! and conically above this angle. Another wind tunnel configuration suggests that the incoming boundary layer can influence this behavior. Similar trends to compression ramp observations (Settles and Teng, 1984) are seen for the asymptotic behavior of the inception length near the root of the SBLI. This suggests a cylindrical/conical boundary similar to that found from the divergence of separation and reattachment lines. The root behavior was further investigated using a delta shock generator producing an inviscid shock similar to the shock generator with an x-z plane sweep angle of 22.5!. Surface flow visualization shows good agreement between the two shock generators at the separation line. The pressure at separation also appears to align between the two, but the delta span, which is limited by tunnel size, is not sufficient to generate a quasi-infinite region.
    • Lossless Image Compression using Reversible Integer Wavelet Transforms and Convolutional Neural Networks

      Ahanonu, Eze (The University of Arizona., 2018)
      Image compression is an area of data compression which looks to exploit various redundancies that exist within images to reduce storage and transmission requirements. In information critical applications such as professional photography, medical diagnostics, and remote sensing, lossless image compression may be used to ensure the original data can be restored at a later time. In this work, a lossless compression framework is proposed which incorporates Convolutional Neural Networks (CNNs) to predict wavelet detail coefficients from coefficients within neighboring subbands. The main premise of the proposed framework is that information which can be recovered at the decoder via CNN prediction can be excluded from the compressed codestream, resulting in reduced file sizes. An end-to-end encoder and decoder is implemented to test the validity of the proposed, model and compression performance is compared with current state of the art methods.
    • Gender Differences in Achievement Emotions: A Control-Value Theory Approach

      Di, Shuxin (The University of Arizona., 2018)
      The current study examines whether there are gender differences in general academic contexts within three achievement emotions: prospective outcome emotions, retrospective outcome emotions and activity emotions. I combined Pekrun’s control-value theory with the Achievement Emotion Questionnaire (AEQ) to assess participants’ achievement emotions. Discriminant function analysis revealed statistically and practically significant gender differences in prospective outcome emotions and activity emotions, but not in retrospective outcome emotions. Moreover, females scored higher on three achievement emotions: prospective outcome emotions, retrospective outcome emotions, and activity emotions than males in this study. The current study filled in the gap of prior studies which have not explored gender differences in three achievement emotions: prospective outcome emotions, retrospective outcome emotion and activity emotions, in general domains. Future studies could replicate the current study and explore if other factors would influence the impact of gender on achievement emotions, for example, culture and age. Additionally, researchers could try to apply achievement emotions to improve students’ academic performance.
    • Spinning Records: How Hip-Hop is Used in the Tucson Community

      Barbre, Joshua (The University of Arizona., 2018)
      The phenomenon of hip-hop began as a local musical practice in New York in the 1970s and from that local practice developed into a formally recognized musical genre, and furthermore, into a viable and distinct culture – a way of life – in its own right. Hip-hop has expanded its formerly narrowly-defined demographics and indoctrinated a broad cultural diversity of contributing artists to become a truly global musical and cultural phenomenon. Hip-hop culture is signaled, enacted, and expressed fundamentally through rapping, deejaying, graffiti, and dance. It was designed to accommodate and support dual identities for its practitioners, through both an acquired identity of affiliation within hip-hop and an identity of affiliation within the locale in which they develop and operate. The members of the Tucson, AZ hip-hop community, the subjects of this study, claim that what defines Tucson’s hip-hop is not how it sounds, but how it is used within the local hip-hop community as well as within the greater Tucson community. This study examines the relationships, symmetrical and asymmetrical, that exist between hip-hop and Tucson. Furthermore, I demonstrate how hip-hop deejays in Tucson serve a pivotal role in connecting the local hip-hop community to the greater Tucson community. Their idea of a hip-hop identity is fluid; therefore, through adaptable performance practice, they can achieve different aims at different times to satisfy different target audiences. Ultimately, what is most vital to their success, and by extension that of the hip-hop community in Tucson, in general, are their multipronged efforts to establish and maintain a strong sense of community through education, collaboration, and support of their fellow artists.
    • Determination of Stress in Humans Using Data Fusion of Off-The-Shelf Wearable Sensors Data for Electrocardiogram and Galvanic Skin Response

      Jeroh, Odafe (The University of Arizona., 2018)
      Stress detection helps individuals understand their stress levels and advises them when to take a break from activities causing stress. Physical activities and environmental influences can affect a person’s stress levels. People with professions as first responders, pilots, and working parents with newborns are examples of people exposed to a large amount of stress. Acquisition and proper analysis of physiological data is helpful in managing stress. In this paper, the results from two sensors, electrocardiogram (ECG) and galvanic skin response (GSR) measurements, are fused to analyze stress in individuals; these sensors are noninvasive and wearable. Data from these sensors are collected simultaneously over a period of 25 minutes from 25 people which are undergoing a simulated stressor. Support Vector Machine (SVM) and Multilayer Perceptron (MLP) are used as the classifiers while Linear Discriminant Analysis (LDA) is used as the stress detection algorithm. The stress detection accuracy achieved varies with individuals and ranges from 87% to 95%. This approach of measuring stress is very suitable for real-time applications and can be used by practically anybody who wants to improve their performance.
    • A Study of Adversarial Attacks Against an LSTM Language Model and the Impact of Normalization in SNN

      Liang, Zhengzhong (The University of Arizona., 2018)
      Artificial Neural Networks (ANNs) have been used to many application-driven fields and have been shown to be quite successful, however, some aspects of ANNs are not well understood. One such area is learning an ANN in the presence of an adversary.In such a context, it is assumed that the attacker can manipulate the training (also referred to causative attack or poison) or testing data (also referred to exploratory attack) to disrupt its normal functionality. In turn, the defender aims at reducing the impact of such attacks as much as possible. The first part of this thesis focuses on causative attacks against an Long Short-Term Memory (LSTM) neural network in a language model. In causative attacks, it is assumed that the attacker can only change the training text in the language model. We study the behavior of the LSTM language model under different causative attacks and propose several simple measures that can reduce the impact of the attacks. Our results show that the poisoning ratio, the poisoning position and the generation of poisoned text can all influence the performance the LSTM language model. Furthermore, we show that proper use of dropout and gradient clipping can reduce the impact of poisoning the training data to some extent. We also contribute to understanding how to globally learn a Spiking Neural Network (SNN). SNNs are a type of ANN; however SNNs are much more biologically realistic than other ANNs. SNNs have not been widely adopted because of several critical issues of SNNs that are not well studied. One such effect is the training of SNNs and the encoding/decoding of signals in SNNs. In the second part of this thesis, we build an SNN based image classifier to study the encoding/decoding of signals and compare several learning rules for training an SNN. Results reveal that (i) classical STDP learning windows generally obtain the best performance using different decoding schemes; (ii) first-spike decoding has worse accuracy than count decoding classifier does when no normalization rules are applied, although first-spike decoding classifier consumes much less time than count decoding classifier; (iii) the performance of first-spike decoding classifier can be largely enhanced with proper use of normalization rules.
    • Electroless Deposition of Cobalt by Coordination and Reduction of Palladium Catalyst by Surface Amines Groups

      Ng, Amy Sum-Yee (The University of Arizona., 2018)
      Solution-deposited palladium on amine-terminated self-assembled monolayers (SAMs) is a well-characterized catalyst and adhesion layer combination for electroless metallization of dielectric films. A reducing agent is typically added to the deposition bath or a sensitizer such as tin is co-deposited producing relatively thick Pd layers. Thinner Pd deposits would enable barrier seed layers for filling < 10 nm wide gaps in patterned dielectric films with metal. In this work, we eliminated the reducing agent from the deposition bath and worked at pH < 2 to deposit monomeric Pd(2+) species and show that the amine groups terminating the SAM reduce Pd(2+) to Pd(0). The amount of Pd deposited depended on the coverage of the two types of amines on the SAM. The adsorption of PdCl42- ions in solution on protonated amine groups (–NH3+) is well known. Our data suggest that the nonprotonated amine groups (–NH2), which coexist with –NH3+, chemically reduce the Pd(2+) ion to Pd metal by oxidizing to the amine radical cation (–NH2•+). Pd bonds to and covers the –NH2•+ groups in the process depositing around a monolayer of Pd from solution on the SAM-covered silicon oxide surface. The Pd layer served as a catalyst for solution deposition of cobalt films on the surface using a reducing agent under oxygen-free conditions. The cobalt deposited initially as islands that grew together into a closed film with good adhesion.
    • Postnatal Microcephaly in MASA/CRASH Syndrome: L1 Cellular Adhesion Molecule Loss Leads to Smaller, Fewer, and Directionally Challenged Neurites

      Gottman, Moriah (The University of Arizona., 2018)
      Postnatal microcephaly in MASA/CRASH (mental retardation, aphasia, shuffling gate, adducted thumbs) syndrome is caused by compromised neurite outgrowth, arborization, and directionality. Neurite outgrowth is lost due to a lack of L1-cellular adhesion molecule (L1CAM, or L1) and ankyrins B and G interaction. Neurite arborization is compromised due to a loss of L1CAM and ezrin-radixin-moesin (ERM) protein interaction. L1CAM and neuropilin-1(NP1) interaction leads to a loss of neurite directionality. How these neurite growth changes affect neurogenesis throughout the brain, and lead to microcephaly (MiC), has not yet been modeled. Future drug discovery projects aiming to halt the progression of postnatal MiC in MASA syndrome could narrow the scope of focus if modeling software parameters were adjusted to reflect phenotypic changes to neurites lacking L1 demonstrated in the literature.
    • Informatics for Trauma-Mediated Psychiatric Illness

      Gilchrist, Collin Andrew (The University of Arizona., 2018)
      Traumatic events such as direct or indirect exposure to serious injury or situations inciting strong fear response have been shown to result in mental disorders that are complex to diagnose, prognosticate, and treat. Neuropsychiatric sequelae following head injuries, including post concussive syndrome (PCS) and post-traumatic stress disorder (PTSD) are particularly challenging because of overlapping symptoms and the profound nature of the injury itself. This work aims to elucidate the similarities and differences between PCS and PTSD, two common sequelae following traumatic brain injury, using an emerging framework (Research Domain Criteria – RDoC) for understanding mental health disorders, and better characterize the mechanisms leading to psychiatric illness as a result of neurotrauma. The application of RDoC is demonstrated using analyses of clinical data from the Alzheimer's Disease Neuroinitiative. This analysis involves identification of subtle symptom variants for differentiating sequelae when clinical prognosis is uncertain based self-reported symptoms. In addition to specific self-reported symptoms, features in positron emission tomography and diffusion tensor imaging are also identified. These results are further supplemented by identification of brain structures associated with self-reported symptoms using imaging data. Overall, this project demonstrates the application of RDoC to better characterize PCS and PTSD, which would potentially allow for more effective treatment and management of these disorders.
    • Demographics of Riparian Lizards in the Chiricahua Mountains in Relation to Water Availability and Emerging Aquatic Insects as a Potential Food Source

      McGee, Earyn Nycole (The University of Arizona., 2018)
      Severe drought driven by climate change and water use by humans are causing formerly perennial streams to flow intermittently, presenting an unprecedented level of disturbance. The loss of emerging aquatic insects as potential prey items could negatively impact riparian and terrestrial species, including lizards. Because lizards play important roles in riparian food webs (e.g. predators, nutrient cycling), it is crucial to understand the cascading effects of stream drying on lizard communities. We hypothesized that perennial streams provide aquatic subsidies to riparian lizards, reducing competition and opening niches. We predicted that lizard abundances would be greater, and that individuals within a species would grow larger and faster, along perennial streams compared to ephemeral streams. We studied three paired 100-meter perennial and ephemeral reaches with similar microhabitat but differing water availability in the Chiricahua Mountains of southeastern Arizona. We measured individual growth rates during a 2-month mark-recapture study of Yarrow’s spiny lizards (Sceloporus jarrovii), striped plateau lizards (Sceloporus virgatus), and ornate tree lizards (Urosaurus ornatus). We used emergence traps to quantify the availability of aquatic prey. Aquatic insects were collected in high abundances, suggesting a potential food source for lizards along perennial streams that may be unavailable along ephemeral streams. When considering mass at first capture, we found that S. jarrovii were larger at perennial versus ephemeral reaches. However, this pattern did not hold true for S. virgatus. Additionally, we failed to detect differences in abundances between paired perennial and ephemeral reaches for either S. jarrovii or S. virgatus. Low sample sizes prevented us from performing any statistical analyses for Urosaurus ornatus and on the mark-recapture data for S. jarrovii and S. virgatus. Although more research is needed to confirm these results, they indicate that emerging aquatic insects may be an important resource to riparian lizard species in arid environments. Future research should quantify trophic links between lizards and potential aquatic subsidies.
    • A Systematic Review of Recruitment and Retention Strategies Used in Dietary Randomized Controlled Interventions in Cancer Survivors

      Lavelle, Sarah Arline (The University of Arizona., 2018)
      Interpretation of results of dietary intervention trials in cancer survivors may be limited by insufficient recruitment or retention of study participants. This systematic review describes recruitment strategies, accrual of participants, and attrition (withdrawal) rates for dietary interventions conducted in breast, prostate, and colorectal cancer survivors. PubMed, CINAHL, Cochrane Central Register of Controlled Trials, Embase, PsychINFO, and Web of Science databases were searched. Eligible studies included national and international dietary randomized controlled trials (RCTs), with at least 12 weeks of intervention and 6 months of follow-up. Trials were required to include a CONSORT (CONsolidated Standards of Reporting Trials) diagram. Twenty cohorts were included: breast cancer (BC) survivors (n=11), prostate cancer (PC) (n=3), colorectal cancer (CRC) (n=1), and combined (n=5). Primary recruitment methods included health care providers (n=13) or cancer registries (n=9). Of studies that set a priori sample sizes, 12 met accrual targets and five did not. Attrition rates averaged 18.6% at 6 months, 16.3% at 12-13 months, and 20.3% at 2 years. Among completed studies (n=18), seven trials met a priori retention targets, three trials did not, one assessed feasibility, and seven trials did not provide a clearly defined retention goal. There were few trials in PC and CRC survivors. Missing CONSORT diagrams reduced the eligible studies. The majority of studies met recruitment goals (n=12). Overall, attrition rates averaged approximately 17.4%. Improved understanding of effective recruitment and retention strategies requires more diligent reporting. Qualitative research may allow for more systematic and detailed evaluation of challenges that contribute to insufficient recruitment and retention of cancer survivors in dietary intervention trials. Registration can be found at PROSPERO ref: CRD42018070396.
    • Deep Neural Networks for Modeling Sequential Prediction Tasks with Applications in Brain Tumors

      Guo, Jiashu (The University of Arizona., 2018)
      Gliomas are malignant brain tumors that are associated with high neurological morbidity and poor outcomes. Patients diagnosed with low-grade gliomas are typically followed by a sequence of measurements of the tumor size as they visit their medical physician. To optimize the timing of therapy of patients, the effective methodologies are needed to predict the future behavior of gliomas by modeling the mechanisms that mediate the characteristic features of gliomas. Machine learning is deployed in a range of computing tasks where designing and programming explicit algorithms with good performance is difficult or infeasible, example applications include email filtering, optical character recognition (OCR), and computer vision. In the broad sweep of machine learning’s current worldly ambitions, healthcare applications seem to top the list for funding and press in the last three years. In this thesis, we demonstrate the promise of Long Short-Term Memory Neural Networks (LSTMs), and a 1-Dimensional Convolutional Neural Network (1-D CNN) to address two important clinical questions in low-grade gliomas: 1) classification and prediction of future behavior; and 2) early detection of dedifferentiation to a higher grade or more aggressive growth. The model of brain tumor growth was recently modeled to be a complex system of partial differential equations (PDEs). This system of PDE provides a significant amount of information about the grade of the tumor that cannot be easily measured from the tumor mass. For example, the PDEs have constants that are related to motility, angiogenesis and the mitotic rate; however, a single measurement of the brain tumor mass cannot easily be connected to these constants. This thesis presents a machine learning-based approach to estimate these parameters using a sequence of tumor mass measurements with an LSTM and a 1-D CNN network to solve the inverse problem of PDE parameter estimation. Experimental results show that through working with different architectures, for both of two models, accuracy increases as a function of the number of tumor measurements. This result shows that the LSTM and 1-D CNN provide better and better approximations to these PDE parameters as more measurements are made available (e.g., a tumor mass measurement every month). Moreover, the perplexity can be used for LSTMs to effectively detect a change in the tumor grade. These findings demonstrate the potential that machine learning and neural networks can provide the medical community for solving the inverse problem of PDE parameter estimation.
    • Experimental Investigation of Boundary Layer Separation

      Balthazar, Michael Aeko (The University of Arizona., 2018)
      A study of boundary layer separation in wind tunnel, water tunnel, and free-flight conditions is reported on within these pages. Methods and results for building a sub-scale wing glove, installing active flow control, and testing in flight are presented. Water tunnel dye flow visualizations and particle image velocimetry of laminar separation bubbles are also presented.
    • Failure Prediction of Structures Subjected to Random Vibrations

      Ballesteros, Luis Enrique (The University of Arizona., 2018)
      The purpose of this thesis is to study the failure prediction of structures subjected to random vibrations. First, the most relevant and fundamental theory in random vibrations is presented. A known method to efficiently calculate RMS von Mises stress in a random vibration environment is re-derived. Next, two practical case studies are presented: a cantilever beam and a rocket launcher payload subjected to random vibrations are analyzed using ANSYS. From this analysis, the expected frequency of the response and the root mean square (RMS) response for the displacement and stress are calculated. The statistical quantities obtained from the random vibration analysis, such as the expected frequency and RMS stress response, are used to determine the probability of the first-passage failure as well as the fatigue life.
    • Analytical Study on the Stability of Collector Members in Composite Steel Structures

      Lizarraga, Daniel (The University of Arizona., 2018)
      Seismic collectors transfer seismic forces to the primary vertical elements of the Seismic Force Resisting System (SFRS). In steel structures, collectors are typically elements in the floor framing system specially designed to carry these forces primarily though axial thrust. Steel collector members are designed as beam-columns under the compressive collector action. The controlling limit state in design may be strong axis beam-column behavior, weak axis buckling, torsional buckling or constrained-axis flange torsional buckling depending on the bracing condition providing by the floor slab or roof deck system attached to the collector. This paper presents an analytical study examining the stability behavior of steel seismic collectors under different structural configurations, including roof deck and composite slab. The orientation of the deck and slab is alternated between the parallel and deck perpendicular configurations, relative to the collector. The sizing of collector members, collector connections, and gravity system members are based on the design of a prototype structure in order to evaluate collectors under realistic design conditions. The compression strength of collector elements under these conditions is compared to design values per the American Institute of Steel Construction (AISC). Design considerations are provided for the stability design of steel seismic collectors.