Matching Items (13,634)
Filtering by
- Creators: Arizona State University
- Creators: Ravel, Maurice, 1875-1937
Description
Engineered nanomaterials (ENMs) are added to numerous consumer products to enhance their effectiveness, whether it be for environmental remediation, mechanical properties, or as dietary supplements. Uses of ENMs include adding to enhance products, carbon for strength or dielectric properties, silver for antimicrobial properties, zinc oxide for UV sun-blocking properties, titanium dioxide for photocatalysis, or silica for desiccant properties. However, concerns arise from ENM functional properties that can impact the environment and a lack of regulation regarding ENMs leads to potential public exposure to ENMs and results in ill-informed public or manufacturer perceptions of ENMs. My dissertation evaluates the environmental, human health, and societal impacts of using ENMs, with a focus on ionic silver and nanosilver, in consumer and industrial products. Reproducible experiments served as functional assays to assess ENM distributions among various environmental matrices. Functional assay results were visualized using radar plots and aid in a framework to estimate likely ENM disposition in the environment. To assess beneficial uses of ENMs, bromide ion removal from drinking waters to limit disinfection by-product formation was studied. Silver-enabled graphene oxide materials were capable of removing bromide from water, and exhibited less competition from background solutes (e.g. natural organic matter) when compared against solely ionic silver addition to water for bromide removal. To assess complex interactions of ENMs with the microbiome, batch experiments were performed using fecal samples spiked with ionic silver or commercial dietary silver nanoparticles. Dietary nanosilver and ionic silver exposures to the fecal microbiome for 24 hours reduce short chain fatty acid (SCFA) production and changes the relative abundance of the microbiota. To understand the social perceptions of ENMS, statistically rigorous surveys were conducted to assess related perceptions related to the use of ENMs in drinking water treatment devices the general public and, separately, industrial manufacturers. These stakeholders are influenced by costs and efficiency of the technologies, consumer concerns of the safety of technologies, and environmental health and safety of the technologies. This dissertation represents novel research that took an interdisciplinary approach, spanning from wet-lab engineering bench scale testing to social science survey assessments to better understand the environmental, human health, and societal impacts of using ENMs such as nanosilver and ionic silver in industrial processes and consumer products.
ContributorsKidd, Justin (Author) / Westerhoff, Paul (Thesis advisor) / Krajmalnik-Brown, Rosa (Committee member) / Perreault, Francois (Committee member) / Maynard, Andrew (Committee member) / Arizona State University (Publisher)
Created2020
Description
There have been multiple attempts of coupling neural networks with external memory components for sequence learning problems. Such architectures have demonstrated success in algorithmic, sequence transduction, question-answering and reinforcement learning tasks. Most notable of these attempts is the Neural Turing Machine (NTM), which is an implementation of the Turing Machine with a neural network controller that interacts with a continuous memory. Although the architecture is Turing complete and hence, universally computational, it has seen limited success with complex real-world tasks.
In this thesis, I introduce an extension of the Neural Turing Machine, the Neural Harvard Machine, that implements a fully differentiable Harvard Machine framework with a feed-forward neural network controller. Unlike the NTM, it has two different memories - a read-only program memory and a read-write data memory. A sufficiently complex task is divided into smaller, simpler sub-tasks and the program memory stores parameters of pre-trained networks trained on these sub-tasks. The controller reads inputs from an input-tape, uses the data memory to store valuable signals and writes correct symbols to an output tape. The output symbols are a function of the outputs of each sub-network and the state of the data memory. Hence, the controller learns to load the weights of the appropriate program network to generate output symbols.
A wide range of experiments demonstrate that the Harvard Machine framework learns faster and performs better than the NTM and RNNs like LSTM, as the complexity of tasks increases.
In this thesis, I introduce an extension of the Neural Turing Machine, the Neural Harvard Machine, that implements a fully differentiable Harvard Machine framework with a feed-forward neural network controller. Unlike the NTM, it has two different memories - a read-only program memory and a read-write data memory. A sufficiently complex task is divided into smaller, simpler sub-tasks and the program memory stores parameters of pre-trained networks trained on these sub-tasks. The controller reads inputs from an input-tape, uses the data memory to store valuable signals and writes correct symbols to an output tape. The output symbols are a function of the outputs of each sub-network and the state of the data memory. Hence, the controller learns to load the weights of the appropriate program network to generate output symbols.
A wide range of experiments demonstrate that the Harvard Machine framework learns faster and performs better than the NTM and RNNs like LSTM, as the complexity of tasks increases.
ContributorsBhatt, Manthan Bharat (Author) / Ben Amor, Hani (Thesis advisor) / Zhang, Yu (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2020
Description
In this paper, I study many-to-one matching markets in a dynamic framework with the
following features: Matching is irreversible, participants exogenously join the market
over time, each agent is restricted by a quota, and agents are perfectly patient. A
form of strategic behavior in such markets emerges: The side with many slots can
manipulate the subsequent matching market in their favor via earlier matchings. In
such a setting, a natural question arises: Is it possible to analyze a dynamic many-to-one
matching market as if it were either a static many-to-one or a dynamic one-to-one
market? First, I provide sufficient conditions under which the answer is yes. Second,
I show that if these conditions are not met, then the early matchings are "inferior"
to the subsequent matchings. Lastly, I extend the model to allow agents on one side
to endogenously decide when to join the market. Using this extension, I provide
a rationale for the small amount of unraveling observed in the United States (US)
medical residency matching market compared to the US college-admissions system.
Micro Finance Institutions (MFIs) are designed to improve the welfare of the poor.
Group lending with joint liability is the standard contract used by these institutions.
Such a contract performs two roles: it affects the composition of the groups that form,
and determines the properties of risk-sharing among their members. Even though the
literature suggests that groups consist of members with similar characteristics, there
is evidence also of groups with heterogeneous agents. The underlying reason is that
the literature lacked the risk-sharing behavior of the agents within a group. This
paper develops a model of group lending where agents form groups, obtain capital
from the MFI, and share risks among themselves. First, I show that joint liability
introduces inefficiency for risk-averse agents. Moreover, the composition of the groups
is not always homogeneous once risk-sharing is on the table.
following features: Matching is irreversible, participants exogenously join the market
over time, each agent is restricted by a quota, and agents are perfectly patient. A
form of strategic behavior in such markets emerges: The side with many slots can
manipulate the subsequent matching market in their favor via earlier matchings. In
such a setting, a natural question arises: Is it possible to analyze a dynamic many-to-one
matching market as if it were either a static many-to-one or a dynamic one-to-one
market? First, I provide sufficient conditions under which the answer is yes. Second,
I show that if these conditions are not met, then the early matchings are "inferior"
to the subsequent matchings. Lastly, I extend the model to allow agents on one side
to endogenously decide when to join the market. Using this extension, I provide
a rationale for the small amount of unraveling observed in the United States (US)
medical residency matching market compared to the US college-admissions system.
Micro Finance Institutions (MFIs) are designed to improve the welfare of the poor.
Group lending with joint liability is the standard contract used by these institutions.
Such a contract performs two roles: it affects the composition of the groups that form,
and determines the properties of risk-sharing among their members. Even though the
literature suggests that groups consist of members with similar characteristics, there
is evidence also of groups with heterogeneous agents. The underlying reason is that
the literature lacked the risk-sharing behavior of the agents within a group. This
paper develops a model of group lending where agents form groups, obtain capital
from the MFI, and share risks among themselves. First, I show that joint liability
introduces inefficiency for risk-averse agents. Moreover, the composition of the groups
is not always homogeneous once risk-sharing is on the table.
ContributorsAltinok, Ahmet (Author) / Chade, Hector (Thesis advisor) / Manelli, Alejandro (Committee member) / Friedenberg, Amanda (Committee member) / Kovrijnykh, Natalia (Committee member) / Arizona State University (Publisher)
Created2020
Description
This work is concerned with the study and numerical solution of large reaction diffusion systems with applications to the simulation of degradation effects in solar cells. A discussion of the basics of solar cells including the function of solar cells, the degradation of energy efficiency that happens over time, defects that affect solar cell efficiency and specific defects that can be modeled with a reaction diffusion system are included. Also included is a simple model equation of a solar cell. The basics of stoichiometry theory, how it applies to kinetic reaction systems, and some conservation properties are introduced. A model that considers the migration of defects in addition to the reaction processes is considered. A discussion of asymptotics and how it relates to the numerical simulation of the lifetime of solar cells is included. A reduced solution is considered and a presentation of a numerical comparison of the reduced solution with the full solution on a simple test problem is included. Operator splitting techniques are introduced and discussed. Asymptotically preserving schemes combine asymptotics and operator splitting to use reasonable time steps. A presentation of a realistic example of this study applied to solar cells follows.
ContributorsShapiro, Bruce G. (Author) / Ringhofer, Christian (Thesis advisor) / Gardner, Carl L (Committee member) / Jackiewicz, Zdzislaw (Committee member) / Platte, Rodrigo B (Committee member) / Vasileska, Dragica (Committee member) / Arizona State University (Publisher)
Created2020
Description
In the last decade deep learning based models have revolutionized machine learning and computer vision applications. However, these models are data-hungry and training them is a time-consuming process. In addition, when deep neural networks are updated to augment their prediction space with new data, they run into the problem of catastrophic forgetting, where the model forgets previously learned knowledge as it overfits to the newly available data. Incremental learning algorithms enable deep neural networks to prevent catastrophic forgetting by retaining knowledge of previously observed data while also learning from newly available data.
This thesis presents three models for incremental learning; (i) Design of an algorithm for generative incremental learning using a pre-trained deep neural network classifier; (ii) Development of a hashing based clustering algorithm for efficient incremental learning; (iii) Design of a student-teacher coupled neural network to distill knowledge for incremental learning. The proposed algorithms were evaluated using popular vision datasets for classification tasks. The thesis concludes with a discussion about the feasibility of using these techniques to transfer information between networks and also for incremental learning applications.
This thesis presents three models for incremental learning; (i) Design of an algorithm for generative incremental learning using a pre-trained deep neural network classifier; (ii) Development of a hashing based clustering algorithm for efficient incremental learning; (iii) Design of a student-teacher coupled neural network to distill knowledge for incremental learning. The proposed algorithms were evaluated using popular vision datasets for classification tasks. The thesis concludes with a discussion about the feasibility of using these techniques to transfer information between networks and also for incremental learning applications.
ContributorsPatil, Rishabh (Author) / Venkateswara, Hemanth (Thesis advisor) / Panchanathan, Sethuraman (Thesis advisor) / McDaniel, Troy (Committee member) / Arizona State University (Publisher)
Created2020
Description
The National Center for Educational Statistics (2018) reported that only 59% of first time college students will retain from their first to second year. The institutional effects of retention are wide ranging and nationwide colleges and universities are seeking effective methods of improving the retention of first year students. Isaak, Graves, & Mayers (2007) identified both emotional intelligence and resilience as important factors contributing to student retention. According to Daniel Goleman (1995), emotional intelligence is integral to success in life, and a significant relationship has been found with grades and successful acclimation to the college environment (Ciarrochi, Deane, & Anderson, 2002; Liff, 2003; and Pekrun, 2006). This study explored the impact of an emotional intelligence (EI) intervention within a First Year Experience course on students’ emotional intelligence, resilience, and academic success. Forty four students at a small, private, liberal arts institution in the southeastern United States participated in the EI intervention and were measured for EI and resilience utilizing the EQ-i 2.0 and the 5x5RS measures as pre and posttests. Based on the results of this study, the EI intervention may have positive implications on EI, resilience and academic success. Institutions and researchers should continue to explore EI as a mechanism to improve resilience and academic success among first year students.
ContributorsDavis, Alexander M (Author) / Wylie, Ruth (Thesis advisor) / Correa, Kevin (Committee member) / Duncan, Tisha (Committee member) / Arizona State University (Publisher)
Created2020
Description
The presence of huge amounts of waste heat and the constant demand for electric energy makes this an appreciable research topic, yet at present there is no commercially viable technology to harness the inherent energy resource provided by the temperature differential between the inside and outside of buildings. In a newly developed technology, electricity is generated from the temperature gradient between building walls through a Seebeck effect. A 3D-printed triply periodic minimal surface (TPMS) structure is sandwiched in copper electrodes with copper (I) sulphate (Cu2SO4) electrolyte to mimic a thermogalvanic cell. Previous studies mainly concentrated on mechanical properties and the electric power generation ability of these structures; however, the goal of this study is to estimate the thermal resistance of the 3D-printed TPMS experimentally. This investigation elucidates their thermal resistances which in turn helps to appreciate the power output associated in the thermogalvanic structure. Schwarz P, Gyroid, IWP, and Split P geometries were considered for the experiment with electrolyte in the thermogalvanic brick. Among these TPMS structures, Split P was found more thermally resistive than the others with a thermal resistance of 0.012 m2 K W-1. The thermal resistances of Schwarz D and Gyroid structures were also assessed experimentally without electrolyte and the results are compared to numerical predictions in a previous Mater's thesis.
ContributorsDasinor, Emmanuel (Author) / Phelan, Patrick (Thesis advisor) / Milcarek, Ryan (Committee member) / Bhate, Dhruv (Committee member) / Arizona State University (Publisher)
Created2020
Description
Subject Pronoun Expression (SPE) has been extensively studied in monolingual and bilingual varieties of Spanish using the variationist framework. The goal of these studies has been to examine the linguistic and extra-linguistic factors that condition the expression and the omission of personal subject pronouns. Nonetheless, to date, there is no study of SPE in the Spanish of Equatorial Guinea, the only African country where it is an official language, and the single country where Spanish is exclusively a second language (L2). This dissertation fills this gap in the literature by accounting for SPE in Equatoguinean Spanish.
The research questions guiding this study concern the rates of Subject Pronoun Expression, its conditioning factors, and universal accounts of L2 acquisition, in particular, the Interface Hypothesis (IH). The study had 30 participants from Malabo, who took part in sociolinguistic interviews. These interviews were transcribed and analyzed using the mixed effects software Rbrul. Along the lines of the literature reviewed, the linguistic factor groups studied were grammatical person and number, reference, reflexivity, verb type, and ambiguity. By the same token, the extra linguistic factors analyzed were age, sex, education, native language (L1), and speaker as a random factor.
The results indicate that the Equatoguinean variety of Spanish has one of the lowest pronoun rates (19.1%), a finding that goes against the predictions of the IH. With regard to the linguistic factor groups that condition Subject Pronoun Expression, Equatoguinean Spanish shows an unorthodox ranking: grammatical person and number, ambiguity, verb class, and reference. Interestingly, the low ranking of reference gives support to the IH, which argues that L2 speakers have problems with constraints like the switch of the reference in subjects because it integrates discourse and pragmatic interfaces. The only significant extra-linguistic factor was education, whereas speakers’ L1 exerted no effect on SPE. Individual speaker was a significant random factor group, indicating that variation is great even in speakers with comparable education.
In sum, this study of a unique speech community provides new information on SPE of L2 Spanish. It also contributes to the fields of language contact, language variation, and second language acquisition.
The research questions guiding this study concern the rates of Subject Pronoun Expression, its conditioning factors, and universal accounts of L2 acquisition, in particular, the Interface Hypothesis (IH). The study had 30 participants from Malabo, who took part in sociolinguistic interviews. These interviews were transcribed and analyzed using the mixed effects software Rbrul. Along the lines of the literature reviewed, the linguistic factor groups studied were grammatical person and number, reference, reflexivity, verb type, and ambiguity. By the same token, the extra linguistic factors analyzed were age, sex, education, native language (L1), and speaker as a random factor.
The results indicate that the Equatoguinean variety of Spanish has one of the lowest pronoun rates (19.1%), a finding that goes against the predictions of the IH. With regard to the linguistic factor groups that condition Subject Pronoun Expression, Equatoguinean Spanish shows an unorthodox ranking: grammatical person and number, ambiguity, verb class, and reference. Interestingly, the low ranking of reference gives support to the IH, which argues that L2 speakers have problems with constraints like the switch of the reference in subjects because it integrates discourse and pragmatic interfaces. The only significant extra-linguistic factor was education, whereas speakers’ L1 exerted no effect on SPE. Individual speaker was a significant random factor group, indicating that variation is great even in speakers with comparable education.
In sum, this study of a unique speech community provides new information on SPE of L2 Spanish. It also contributes to the fields of language contact, language variation, and second language acquisition.
ContributorsPADILLA, LILLIE VIVIAN KARLE (Author) / Cerron-Palomino, Alvaro (Thesis advisor) / Lafford, Barbara (Committee member) / Beas, Omar (Committee member) / Otabela, Joseph-Désiré (Committee member) / Arizona State University (Publisher)
Created2020
Description
Hyperspectral unmixing is an important remote sensing task with applications including material identification and analysis. Characteristic spectral features make many pure materials identifiable from their visible-to-infrared spectra, but quantifying their presence within a mixture is a challenging task due to nonlinearities and factors of variation. In this thesis, physics-based approaches are incorporated into an end-to-end spectral unmixing algorithm via differentiable programming. First, sparse regularization and constraints are implemented by adding differentiable penalty terms to a cost function to avoid unrealistic predictions. Secondly, a physics-based dispersion model is introduced to simulate realistic spectral variation, and an efficient method to fit the parameters is presented. Then, this dispersion model is utilized as a generative model within an analysis-by-synthesis spectral unmixing algorithm. Further, a technique for inverse rendering using a convolutional neural network to predict parameters of the generative model is introduced to enhance performance and speed when training data are available. Results achieve state-of-the-art on both infrared and visible-to-near-infrared (VNIR) datasets as compared to baselines, and show promise for the synergy between physics-based models and deep learning in hyperspectral unmixing in the future.
ContributorsJaniczek, John (Author) / Jayasuriya, Suren (Thesis advisor) / Dasarathy, Gautam (Thesis advisor) / Christensen, Phil (Committee member) / Arizona State University (Publisher)
Created2020
Description
People who participate in correctional treatment programming are viewed as making positive steps towards their reentry into society. However, this is often assessed through a simple “yes” or “no” response to whether they are currently participating without much emphasis on how, why, or to what degree that participation is meaningful for reentry preparedness. The present study aims to a) identify to what extent there is variation in the degree to which women participate in programming and are prepared for reentry, b) identify the characteristics that distinguish highly-involved programmers from less involved programmers, c) identify the characteristics that distinguish women who are highly-prepared for reentry from women who are less prepared, and d) assess whether levels of involvement in programming relates to levels of reentry preparedness. The sample comes from interviewer-proctored surveys of 200 incarcerated women in Arizona. Two indices were created: one for the primary independent variable of program involvement and one for the dependent variable of reentry preparedness. Logistic and multivariate regressions were run to determine the indices’ relatedness to each other and the characteristic variables. The two indices did not have a statistically significant relationship with each other. However, variation across them is found. This indicates that programmers may not be a homogenous group and that they may engage with programming to varying degrees based on a multitude of indicators.
ContributorsRodriguez, Bianca (Author) / Wright, Kevin (Thesis advisor) / Young, Jacob (Committee member) / Telep, Cody (Committee member) / Arizona State University (Publisher)
Created2020