Matching Items (15,933)
Filtering by
- Creators: Arizona State University
- Creators: Brahms, Johannes, 1833-1897
Description
Recently, Generative Adversarial Networks (GANs) have been applied to the problem of Cold-Start Recommendation, but the training performance of these models is hampered by the extreme sparsity in warm user purchase behavior. This thesis introduces a novel representation for user-vectors by combining user demographics and user preferences, making the model a hybrid system which uses Collaborative Filtering and Content Based Recommendation. This system models user purchase behavior using weighted user-product preferences (explicit feedback) rather than binary user-product interactions (implicit feedback). Using this a novel sparse adversarial model, Sparse ReguLarized Generative Adversarial Network (SRLGAN), is developed for Cold-Start Recommendation. SRLGAN leverages the sparse user-purchase behavior which ensures training stability and avoids over-fitting on warm users. The performance of SRLGAN is evaluated on two popular datasets and demonstrate state-of-the-art results.
ContributorsShah, Aksheshkumar Ajaykumar (Author) / Venkateswara, Hemanth (Thesis advisor) / Berman, Spring (Thesis advisor) / Ladani, Leila J (Committee member) / Arizona State University (Publisher)
Created2022
Description
Violence has been characterized as a force for both political change and maintenance of the status quo in human societies. The present study examines how outbreaks of violent events led to a legacy of prolonged warfare among neighboring communities and shaped the formation of new political institutions during the late prehispanic era in the North-Central Andes. Drawing on data collected through archaeological excavation, osteological analysis of human remains, and radiocarbon dating, this work reconstructs life and death histories of 287 individuals recovered from nine archaeological sites to investigate diachronic patterns in physical violence. The observed individuals inhabited settlements located within the high-altitude, mountainous terrain of the Callejón de Huaylas, a region that has received little attention from bioarchaeologists, and the majority lived during the Late Intermediate Period (1000-1450 CE). Archaeological research has indicated local livelihoods changed significantly around 1000 CE. In the wake of Wari state disintegration and an increasingly arid climate, communities faced a series of social, political, and economic transformations. Less is known about how these shifts affected embodied practices of violence in the region. This study documents a stark change in who experienced head injuries during the Late Intermediate Period, as compared to data from preceding eras. Individuals of all ages exhibited both antemortem and perimortem trauma throughout the four and a half centuries. Results reveal people experienced novel forms of physical violence beginning in the mid-1200s—not only did more individuals sustain head injuries, including juveniles, but the inflicted trauma was more lethal and severe at this time. These trauma patterns persisted for generations, continuing through Inka conquest around 1450 CE. The frequency and type of observed cranial trauma are consistent with warfare documented ethnographically among some small-scale societies, suggesting an association between violence and political autonomy. Beyond identifying cultural transformations in victim identities and intergroup dynamics, this research contributes to a growing body of work across the Americas investigating mounting evidence of social strife and conflict from the 13th to 15th centuries. Finally, it sheds light on intergenerational consequences of violent actions by centering individual experiences within contexts of long-term historical trajectories.
ContributorsSharp, Emily Anne (Author) / Buikstra, Jane E. (Thesis advisor) / Knudson, Kelly J. (Committee member) / Stojanowski, Christopher M. (Committee member) / Arizona State University (Publisher)
Created2022
Description
How to teach a machine to understand natural language? This question is a long-standing challenge in Artificial Intelligence. Several tasks are designed to measure the progress of this challenge. Question Answering is one such task that evaluates a machine's ability to understand natural language, where it reads a passage of text or an image and answers comprehension questions. In recent years, the development of transformer-based language models and large-scale human-annotated datasets has led to remarkable progress in the field of question answering. However, several disadvantages of fully supervised question answering systems have been observed. Such as generalizing to unseen out-of-distribution domains, linguistic style differences in questions, and adversarial samples. This thesis proposes implicitly supervised question answering systems trained using knowledge acquisition from external knowledge sources and new learning methods that provide inductive biases to learn question answering. In particular, the following research projects are discussed: (1) Knowledge Acquisition methods: these include semantic and abductive information retrieval for seeking missing knowledge, a method to represent unstructured text corpora as a knowledge graph, and constructing a knowledge base for implicit commonsense reasoning. (2) Learning methods: these include Knowledge Triplet Learning, a method over knowledge graphs; Test-Time Learning, a method to generalize to an unseen out-of-distribution context; WeaQA, a method to learn visual question answering using image captions without strong supervision; WeaSel, weakly supervised method for relative spatial reasoning; and a new paradigm for unsupervised natural language inference. These methods potentially provide a new research direction to overcome the pitfalls of direct supervision.
ContributorsBanerjee, Pratyay (Author) / Baral, Chitta (Thesis advisor) / Yang, Yezhou (Committee member) / Blanco, Eduardo (Committee member) / Li, Baoxin (Committee member) / Arizona State University (Publisher)
Created2022
Description
With increasing costs of higher education, community college pathways to baccalaureate transfer degrees are attractive student alternatives to starting at a traditional four-year institution. However, disparate student outcomes, particularly for underserved student populations, continue to be a concern when considering equitable four-year degree completion rates. Previous literature demonstrates that student satisfaction and student informational capital play key roles in the success of community college transfer students to persist to four-year institutions and attain their educational and career goals. The role of academic advising in the transfer context provides a uniquely collaborative opportunity to address factors of success and student outcomes. Via this mixed methods action research study, I utilized archival student enrollment data, a student survey, and student and advisor interviews to examine an academic advising model that I created in partnership between Cochise Community College and the University of Arizona (i.e., the Colaborativo Advising for Transfer Success Model, or CATS Advising Model), whereby I assigned a singular academic advisor (i.e., a CATS advisor) a student caseload across the two institutions in a deliberate effort to facilitate successful transfer. I used a combined framework of the Model of Student Departure, Transfer Student Capital, and Appreciative Inquiry to inform the advising intervention. I found that students who received the advising intervention were significantly more likely to a) be satisfied with their transfer advising experience, b) perceive increased transfer knowledge (capital), and c) retain through transfer and university enrollment, in comparison to their peers who received advising via a more traditional transfer advising model. Importantly, the students experiencing the advising intervention were also able to articulate their appreciation and recognition of the impact of their relationship with the CATS advisors on their transfer success.
ContributorsWieland, Sarah (Author) / Beardsley, Audrey (Thesis advisor) / Smith, Stephanie (Committee member) / Urquídez, Kasandra (Committee member) / Arizona State University (Publisher)
Created2022
Description
This dissertation explores the effect of school competition on the human capital accumulation of students. Policies that expand the scope for school choice have become increasingly popular largely due to the belief that this will create incentives for low-performing, incumbent schools to improve academic outcomes. However, there is a general lack of empirical support for these positive academic spillover effects in most contexts. In the first chapter, I demonstrate that if schools respond to competition through channels not typically considered in standard arguments in favor of school choice, it means that these policies may lead to negative, unintended consequences for academic achievement. I find that increasing the number of schools serving a given market can have a negative effect on test scores through creating incentives for schools to increase the provision of non-academic services that do not contribute to academic preparation, and through the creation of excess costs in the public school system. I use an empirical strategy designed to address strategic location decisions by new entrants as well as student selection across schools to show that entry of a new charter middle school during a recent large-scale charter expansion in North Carolina decreased average traditional public middle school test scores across a school district.
The second chapter considers the extent to which policymakers have tools available to them that can improve the ability of competition to generate the increases in test scores at incumbent schools that they have prioritized. I show that the efficacy of school choice can be improved by providing short-term, partial reimbursements to public school districts for increases in charter school enrollment by resident pupils. I also demonstrate that these effects occur not only due to the direct increase in district revenue associated with reimbursements, but also because the presence of this aid reduces the incentives of school administrators to compete for students through non-academic channels. The empirical strategy that I use to generate these results leverages plausibly exogenous cutoffs for aid eligibility induced by a unique policy in the state of New York.
ContributorsTobin, Zachary Benjamin (Author) / Aucejo, Esteban (Thesis advisor) / Silverman, Daniel (Committee member) / Murphy, Alvin (Committee member) / Arizona State University (Publisher)
Created2022
Description
Mucosal membranes represent a major site of pathogen transmission and cancer development. Enhancing T cell migration to mucosal surfaces could improve immune-based therapies for these diseases, yielding better clinical outcomes. All-trans-retinoic acid (ATRA) is a biologically active form of vitamin A that has been shown to increase T cell migration to mucosal sites, however its therapeutic use is limited by its toxicity potential and unstable nature. ATRA-related compounds with lower toxicity and higher stability were assessed for their ability to induce similar immune migration effects as ATRA, using in vitro and in vivo model systems. Chapter 2 summarizes the first project, in which synthetic, ATRA-like compounds called rexinoids were used to modulate T cell expression of mucosal homing proteins chemokine receptor 9 (CCR9) and integrin alpha 4 beta 7 (α4β7), and alter their physical migration in vitro. Several rexinoids independently mimicked the activity of ATRA to enhance protein expression and migration, while others worked synergistically with subtoxic doses of ATRA to produce similar results. Furthermore, rexinoid administration in vivo was well-tolerated by animal models, a finding not seen with ATRA.
Chapter 3 focuses on the second project, where plasmids containing ATRA-synthesizing proteins were assessed for their in vivo ability to act as mucosal vaccine adjuvants and enhance T cell migration to mucosal sites during DNA vaccination. Though increased mucosal migration was seen with use of the adjuvant plasmids, these findings were not determined to be significant. Immune-mediated protection following viral challenge was also not determined to be significant in animal models receiving both vaccine and adjuvant plasmids.
The data shows that several novel rexinoids may possess enhanced clinical utility compared to ATRA, lending support for their use in immunotherapeutic approaches towards mucosal maladies. While the potential mucosal vaccine adjuvants did not show great significance in enhancing T cell migration or viral protection, further optimization of the model system may produce better results. This work helps advance knowledge of immune cell trafficking to afflicted mucosal regions. It can be used as a basis for understanding migration to other body areas, as well as for the development of better immune-based treatments.
ContributorsManhas, Kavita Rani (Author) / Blattman, Joseph (Thesis advisor) / Marshall, Pamela (Committee member) / Lake, Douglas (Committee member) / Ugarova, Tatiana (Committee member) / Arizona State University (Publisher)
Created2022
Description
With the rapid development of both hardware and software, mobile devices with their advantages in mobility, interactivity, and privacy have enabled various applications, including social networking, mixed reality, entertainment, authentication, and etc.In diverse forms such as smartphones, glasses, and watches, the number of mobile devices is expected to increase by 1 billion per year in the future.
These devices not only generate and exchange small data such as GPS data, but also large data including videos and point clouds.
Such massive visual data presents many challenges for processing on mobile devices.
First, continuously capturing and processing high resolution visual data is energy-intensive, which can drain the battery of a mobile device very quickly.
Second, data offloading for edge or cloud computing is helpful, but users are afraid that their privacy can be exposed to malicious developers.
Third, interactivity and user experience is degraded if mobile devices cannot process large scale visual data in real-time such as off-device high precision point clouds. To deal with these challenges, this work presents three solutions towards fine-grained control of visual data in mobile systems, revolving around two core ideas, enabling resolution-based tradeoffs and adopting split-process to protect visual data.In particular, this work introduces: (1) Banner media framework to remove resolution reconfiguration latency in the operating system for enabling seamless dynamic resolution-based tradeoffs;
(2) LesnCap split-process application development framework to protect user's visual privacy against malicious data collection in cloud-based Augmented Reality (AR) applications by isolating the visual processing in a distinct process;
(3) A novel voxel grid schema to enable adaptive sampling at the edge device that can sample point clouds flexibly for interactive 3D vision use cases across mobile devices and mobile networks. The evaluation in several mobile environments demonstrates that, by controlling visual data at a fine granularity, energy efficiency can be improved by 49% switching between resolutions, visual privacy can be protected through split-process with negligible overhead, and point clouds can be delivered at a high throughput meeting various requirements.Thus, this work can enable more continuous mobile vision applications for the future of a new reality.
ContributorsHu, Jinhan (Author) / LiKamWa, Robert (Thesis advisor) / Wu, Carole-Jean (Committee member) / Doupe, Adam (Committee member) / Jayasuriya, Suren (Committee member) / Arizona State University (Publisher)
Created2022
Description
Modern life is full of challenging optimization problems that we unknowingly attempt to solve. For instance, a common dilemma often encountered is the decision of picking a parking spot while trying to minimize both the distance to the goal destination and time spent searching for parking; one strategy is to drive as close as possible to the goal destination but risk a penalty cost if no parking spaces can be found. Optimization problems of this class all have underlying time-varying processes that can be altered by a decision/input to minimize some cost. Such optimization problems are commonly solved by a class of methods called Dynamic Programming (DP) that breaks down a complex optimization problem into a simpler family of sub-problems. In the 1950s Richard Bellman introduced a class of DP methods that broke down Multi-Stage Optimization Problems (MSOP) into a nested sequence of ``tail problems”. Bellman showed that for any MSOP with a cost function that satisfies a condition called additive separability, the solution to the tail problem of the MSOP initialized at time-stage k>0 can be used to solve the tail problem initialized at time-stage k-1. Therefore, by recursively solving each tail problem of the MSOP, a solution to the original MSOP can be found. This dissertation extends Bellman`s theory to a broader class of MSOPs involving non-additively separable costs by introducing a new state augmentation solution method and generalizing the Bellman Equation. This dissertation also considers the analogous continuous-time counterpart to discrete-time MSOPs, called Optimal Control Problems (OCPs). OCPs can be solved by solving a nonlinear Partial Differential Equation (PDE) called the Hamilton-Jacobi-Bellman (HJB) PDE. Unfortunately, it is rarely possible to obtain an analytical solution to the HJB PDE. This dissertation proposes a method for approximately solving the HJB PDE based on Sum-Of-Squares (SOS) programming. This SOS algorithm can be used to synthesize controllers, hence solving the OCP, and also compute outer bounds of reachable sets of dynamical systems. This methodology is then extended to infinite time horizons, by proposing SOS algorithms that yield Lyapunov functions that can approximate regions of attraction and attractor sets of nonlinear dynamical systems arbitrarily well.
ContributorsJones, Morgan (Author) / Peet, Matthew M (Thesis advisor) / Nedich, Angelia (Committee member) / Kawski, Matthias (Committee member) / Mignolet, Marc (Committee member) / Berman, Spring (Committee member) / Arizona State University (Publisher)
Created2021
Description
Many options for mouthpieces and reeds are available to clarinetists, which makes selecting the right equipment complicated. In addition to personal research, potential influences in this process included teachers, professors, and private instructors. To provide some clarity in the current trends impacting the selection of new equipment, a survey of clarinetists was conducted. All participants were asked what equipment they were currently using and to specify what elements of the purchase were most important. Aspects such as price, instructor influence, personal research, conferences, brand loyalty, new releases, and social media were ranked by level of importance. Additionally, questions were asked of participants who taught clarinet about what they recommended to their students at various skill levels. The opinions of clarinetists along with the analysis of the data confirmed which mouthpieces and reeds were being suggested for different skill levels. The results were analyzed by type of career with strong trends in the participants current equipment and their suggested equipment.
ContributorsDruesedow, Elizabeth Jane (Author) / Spring, Robert (Thesis advisor) / Gardner, Joshua (Thesis advisor) / Knowles, Kristina (Committee member) / Caslor, Jason (Committee member) / Arizona State University (Publisher)
Created2021
Description
This work comprises a cumulative effort to provide analysis of proteins relevant to understanding and treating human disease. This dissertation focuses on two main protein complexes: the structure of the Chimp adenovirus Y25 capsid assembly, as used in the SARS-CoV-2 vaccine, Vaxzveria, and the Dbl family RhoGEF (guanosine exchange factor) Syx and its associated small G protein, RhoA. The course of research was influenced heavily by the onset of the Covid-19 pandemic and associated lockdown, which pushed anyone with the means to do meaningful research to shift priorities towards addressing the greatest public health crisis since the 1918 flu pandemic.
Analysis of the Syx-RhoA complex for the purposes of structurally guided drug design was initially the focus of heavy optimization efforts to overcome the numerous challenges associated with expression, purification, and handling of this protein. By analyzing E. Coli derived protein new important knowledge was gained about this protein’s biophysical characteristics which contribute to its behavior and may inform drug design efforts. Expression in SF9 insect cells resulted in promising conditions for production of homogeneous and monodispersed protein. Homology modeling and molecular dynamics simulation of this protein support hypotheses about its interactions with both RhoA as well as regions of the cytoplasmic leaflet of the cell membrane.
Structural characterization of ChAdOx1, the adenoviral vector used in the AstraZeneca Covid-19 vaccine, Vaxzveria resulted in the highest resolution adenovirus structure ever solved (3.07Å). Subsequent biochemical analysis and computational simulations of PF4 with the ChAdOx1 capsid reveal interactions with important implications for vaccine induced thrombocytic throbocytopenia syndrome, a disorder observed in approximately 0.000024% of patients who receive Vaxzveria.
ContributorsBoyd, Ryan J (Author) / Fromme, Petra (Thesis advisor) / Chiu, Po-Lin (Committee member) / Liu, Wei (Committee member) / Arizona State University (Publisher)
Created2021