33760 research outputs found
Sort by
Black Lives Matter and Black Millennial Meaning-Making
No full textThis dissertation is a study of Black millennial meaning-making in relation to the Black Lives Matter (BLM) Movement. As a case, this research supports an investigation of the social psychological processes through which oppressed racial groups interpret the character and actions of racial justice movements and the social contexts in which they mobilize. I use Symbolic Interactionism and critical theories of race and racism to address deficits in the way traditional social movement theories explain the relationship between social structure, interpretive processes, and mobilization—particularly for race-based movements. The study uses data from interviews with 36 Black millennials, conducted across an approximately one-year span starting in early 2019. Although these data provide for analysis of a wide range of topics (e.g. gender dynamics, social and news media impacts, participation, tactics, and collective identity), the dissertation focuses on three areas of meaning construction about the movement that became clear through preliminary analysis. First, I explore interpretations of the purpose and goals of BLM, extending the framing perspective to capture the external framing processes that occur in response to movements. The second portion of the analysis explicates the powerful role of collective memory of past iterations of Black activism in the U.S. in evaluations of BLM. Finally, examining definitions and projections of success for BLM, the third empirical chapter interrogates the relationship between political opportunity structures and prospectus for social change through activism. Across all three chapters, I specifically point to the ways that Black millennial meaning-making about BLM in the U.S. context was decidedly dependent on sociostructural understandings of racism. Thus, I argue that theories of social movements must explicitly incorporate historically constituted systems of racial domination.
FROM INSTANCE-SPECIFIC TRAINING TO GENERALIZABLE SOLVERS: ADVANCING MACHINE LEARNING FOR SCIENTIFIC COMPUTING
No full textNeural networks have played a crucial role in scientific computing by providing data-driven solutions to complex mathematical and physical problems. However, traditional neural network solvers remain fundamentally limited in accuracy and lack interpretability, as they operate as black-box models with no explicit mathematical structure. To address these challenges, this thesis explores the Finite Expression Method (FEX), a symbolic regression approach designed to enhance interpretability and accuracy in scientific computing. FEX leverages deep reinforcement learning to discover interpretable mathematical expressions, offering a principled approach to solving high-dimensional partial differential equations (PDEs) and uncovering governing equations from experimental data. Despite these advantages, both neural network solvers and FEX still require retraining for each new equation or change in initial and boundary conditions, limiting their scalability and adaptability.
To overcome these fundamental constraints, scientific computing is now transitioning to the second stage, characterized by foundation models inspired by large language models. These models are pretrained on diverse scientific data and employ in-context learning to generalize across a wide range of problems without requiring instance-specific retraining. In this thesis, we introduce FMint, a foundation model designed for the fast and accurate simulation of dynamical systems. FMint builds upon a decoder-only transformer architecture and functions as an error corrector for coarse simulations, significantly improving accuracy while maintaining computational efficiency. By learning from a broad set of dynamical system trajectories, FMint generalizes well to out-of-distribution dynamics, demonstrating superior performance compared to traditional neural network solvers.
This shift from single-instance solvers to foundation models marks a major transformation in scientific computing. FMint exemplifies how pretrained models can serve as a foundation for more advanced PDE solvers, demonstrating the potential of leveraging large-scale pretraining and in-context learning for scientific applications. Its success highlights a broader direction for future research, where foundation models trained on diverse physical systems can enable more generalizable, efficient, and interpretable solutions across a wide range of computational science and engineering problems
Molecular Layer Deposition of a Polymeric Crown Ether
No full textCrown ethers (CEs) are macrocyclic molecules known for their ion-selective complexation and have been widely used in solution-phase systems for sensing, separation, and ion transport. Despite their well-characterized coordination chemistry and thermal stability, their integration into vapor-phase thin film fabrication has not been previously demonstrated. In this work, we report the first molecular layer deposition of a crown ether–containing polymer using 4,10-diaza-15-crown-5 (15AC5) and malonyl chloride (MC) as precursors. In-situ spectroscopic ellipsometry confirms layer-by-layer growth behavior, while X-ray photoelectron spectroscopy (XPS) offers strong evidence that the crown ether ring is preserved due to the presence of both C-O-C oxygen in the 15AC5 ring and C=O oxygen in the MC. Comparison between measurement and Ab initio simulations of the valence band spectra support the observed bonding structure and further validate the retention of molecular integrity during deposition. These findings establish a new route for incorporating functional macrocycles into conformal, nanoscale coatings.This work was initially supported by the U.S. Department of Energy Office of Science, Materials Chemistry Program, DE-SC0017620, “Molecular Design and Vapor Phase Synthesis of Hybrid Ion-Conducting Materials Based on Crown Ethers. Continued in manuscript writing and discussions were supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Award Number DE-SC002107
Reweighting Methods for Elucidation of Conformational Ensembles of Proteins
No full textThis dissertation explores the application and advancement of reweighting methods in computational biology, focusing on increasing their ability to provide reliable solutions for understanding the conformational states of biological systems. We begin by introducing reweighting methods, discussing their different categories, available implementations, and the various types of input data they use. We review applications of these methods in biological systems, showing their use in understanding complex molecular systems and we discuss the limitations and challenges faced by reweighting methods, setting the foundation for the methodology proposed in this work.The second chapter outlines the methodology we developed to increase the reliability of reweighting methods. An important contribution of this work is the incorporation of multiple experimental data to improve the robustness of the solutions. This methodology is applied to the ubiquitin dimer system, where we used two reweighting methods, SES and BME, in combination with a variety of experimental data: PCS, RDC, PRE, Diamagnetic RDC, and SANS. The combination of different data types increases the accuracy of the results, with each dataset deconvoluting specific information of the system. We fully discuss the limitations of each type of experimental data and the importance of their proper integration.
The third chapter shifts focus to the application of reweighting methods using only SANS data, which is often the most accessible experimental data for many systems. This section details how SANS data can reliably be used to predict the pH dependence of conformations in the ubiquitin dimer and explores the potential for studying larger systems, such as trimer and tetramer ubiquitin. Additionally, we address the critical issue of ensemble generation, proposing methods to assess and ensure that the initial ensemble adequately covers the conformational space.
Finally, we suggest future directions for reweighting methods, including their integration with molecular dynamics, emphasizing the potential for these approaches to transform the study of complex biological systems
Individual-, County-, And Structural-Level Factors In Prenatal Care Utilization And Birth Outcomes
No full textThis research explores Black/White differences in prenatal care utilization, the contribution of these differences to racial disparities in birth outcomes, and the effect of community context on these outcomes. Black women initiate and continue prenatal care at rates that are lower than those for White women, which may contribute to their high relative risk of adverse birth outcomes, including infant mortality. The decision to use prenatal care might seem like an easy one for most pregnant women. However, pregnant Black women face significant financial, geographic, and psychosocial barriers to initiating and continuing care. Although there is a wide body of literature examining a variety of barriers to prenatal care utilization and the racial differences in utilization between Black and White women, this has not systematically examined the variation in utilization that exists between Black women of different socioeconomic statuses (SES). This research employs an intersectional framework to examine the effect of individual- and community-level socioeconomic factors on Black women’s prenatal care utilization patterns. This is done using the most recently released version of the Center for Disease Control’s (CDC) Linked Birth/Death dataset linked with the 5-year file of the American Community Survey (ACS) dataset. Black women are not a monolith, which is why it is important to examine within group variation to better understand the factors that push and pull Black women into and out of prenatal care, and the associated birth outcomes
TOWARDS HYBRID QUANTUM NETWORKING: INTERFACING ION TRAPS WITH NEUTRAL ATOM SYSTEMS
No full textBuilding large-scale modular quantum computers and quantum networks require scalable high fidelity, high efficiency, and long lifetime quantum memories [1]. Quantum memories are proposed to increase photon-mediated matter-qubit entanglement rates by synchronizing photon interference between network nodes [2]. Hybrid quantum networkingleverages trapped ions’ high fidelity operations and neutral-atoms’ single photon manipulation for increased entanglement rates over single-species quantum networks [3, 4, 5, 6, 7, 8]. Here, we aim to demonstrate flying-qubit photon storage in a neutral-atom system using frequency-converted photons entangled with a trapped barium ion. The quantum information encoded in the flying qubit’s polarization states is reversibly mapped to a multiplexed dual-rail encoding scheme during storage. This work helps enable long-distance quantum networking by synthesizing hybrid components in entanglement distribution [9]
A SPECTRAL ANALYSIS OF SPEECH PRE- AND POST- GLOSSECTOMY
No full textThis study aims to use spectral analysis to explain the changes in speech caused by a glossectomy. A single patient is observed using video recordings across 30 years at four stages: before tongue cancer, after a partial glossectomy with a radial forearm free flap, after additional surgeries including removal of the flap, and without the flap in the cold, a known area of speech difficulty for the patient. Formant frequencies and consonant spectra were analyzed to quantify the changes in speech production. Results show the greatest changes in formants occur with front vowels, indicating difficulty making a constriction with the tongue tip; however, overall change in vowel formants is still minimal. Significant spectral differences were observed in the production of the sibilant fricatives /s/ and /S/, with lower spectral peaks and reduced spectral distinctiveness between the sibilant fricatives across all post-glossectomy stages, but most prominently in the cold. Other consonants are less affected, indicating the disproportionate impact a glossectomy has on speech that requires finer control of the tongue-tip. These results underscore the value of surgical techniques that preserve tongue tip mobility where possible and have further implications in post-glossectomy targeted speech therapy
TARGETING CELLULAR METABOLISM BY LEVERAGING THE PHOTOSENSITIZER VERTEPORFIN TO OVERCOME CANCER MULTIDRUG RESISTANCE
No full textOvercoming multidrug resistance (MDR) remains one of the major challenges to successful treatment outcomes for cancer patients. P-glycoprotein (P-gp) is an ATP-binding cassette (ABC) drug transporter that effectively translocates substrates (chemotherapeutics in this context) from cancer cells, thus reducing their efficacy over time. Chemoresistance is responsible for approximately 90% of treatment failure in the clinic, demonstrating a need for effective P-gp inhibition strategies to overcome MDR.
Traditional approaches for inhibiting ATP-dependent P-gp-mediated drug efflux using as small molecule inhibitors and antibodies have had limited clinical success in part due to low selectivity causing systemic toxicity. Photochemical approaches emerge as an alternative inhibition strategy to overcome P-gp-mediated MDR. Previous research has shown that photoactivation of the photosensitizer, verteporfin (VP), can directly cause P-gp aggregation due to cross-linking of the protein. Photoactivation of VP resulting in indirect P-gp inhibition, such as by depleting ATP levels, has been understudied. In this context, this thesis aims to understand the mechanisms through which ATP-depletion is an indirect target for P-gp inhibition using the photosensitizer verteporfin (VP) with and without light irradiation.
This thesis addresses this goal by (1) evaluating the effects of photodynamic priming (PDP) using VP on ATP depletion and its subsequent impact on P-gp function, (2) investigating how different formulations of VP influence its retention, subcellular localization, and ability to engage, evade, and exploit P-gp to enhance drug retention, and (3) examining the light-independent effects of VP on ATP depletion and sustained P-gp inhibition to improve chemotherapy efficacy in drug-resistant cancer cells
Looking Up: Reading for Settler Colonialism in Contemporary Asian American Literature
No full textThis dissertation examines how contemporary Asian American literature engages settler colonialism not through direct representation of Indigenous characters or moments of Asian-Indigenous encounter, but through formal and narrative strategies that illuminate the structural logics of settler colonialism. While much of Asian American literary scholarship explores settler colonialism through explicit references to or representations of such interactions, this project instead asks we might read for its presence in the absence of such representations. I argue that literary form offers a crucial site to understand how Asian American literature engages the entangled logics of settler colonialism and racism. To this end, I read four contemporary Asian American novels for their formal and narrative strategies: the compromised status of truth and memory in the confessional form in The Sympathizer by Viet Thanh Nguyen in Chapter One; the narrative cartography produced by a polyvocal crowd of narrators in Karen Tei Yamashita’s Tropic of Orange in Chapter Two; and the questionable unity of first-person plural narrators in Chang-rae Lee’s On Such a Full Sea and Vauhini Vara’s The Immortal King Rao in Chapter Three. Taking temporality as its organizing heuristic, the dissertation interrogates the dominant narrative arc through with Asian American history is typically plotted: a linear progression from migrant exclusion, to multicultural inclusion, to model minority assimilation. I read these novels as they relate to constructions of linear time—of past, to present, to future—through critical Asian American, Native, Black, and queer temporal interventions that disrupt such normative teleologies, and foreground the entangled histories of migration, racialization, and settler violence. By doing so, these I hope this project helps us expand the dominant formal and temporal frames and reading practices through which Asian American literature and subjectivity has been understood, and situate it within a broader critique of U.S. settler colonialism
AUGMENTATION OF A HIGH-FIDELITY SOLVER TO RESOLVE UNSTEADY FLOW PHENOMENON IN SUPERSONIC RETROPROPULSION FLOWFIELDS
No full textA scale-resolving solver is augmented to incorporate adaptive mesh refinement to reduce the computational expense of application flowfields by over an order of magnitude. This new capability, after a validation effort, was successfully deployed on supersonic retropropulsion (SRP) test cases to provide a large, complete dataset of quantities of interest for the EDL and computational fluid dynamics communities.
Augmentations to the in-house CRoCCo codebase are made to enable high-fidelity simulations of complex applications such as SRP. Adaptive mesh refinement is implemented on generalized curvilinear coordinates to enable wall-resolved simulations with careful considerations to preserve the bandwidth-resolving efficiencies of the underlying numerical methods. The newly developed CRoCCo-AMR codebase is validated through direct numerical simulation (DNS) and wall-resolved large eddy simulation (WRLES) of shock wave and turbulent boundary layer interactions (STBLI) at supersonic conditions. This validation exercise demonstrates a 40% for DNS and 27% for WRLES reduction in grid points. AMR validation shows strong agreement for the mean flow variables, turbulent intensities, and spectral quantities of existing datasets. This dissertation provides in-depth discussion and defense of decisions, implementation, and validation of scale-resolving simulation capabilities in the pursuit of deploying such methods on SRP flowfields to improve the state-of-the-art simulation capabilities for human-class exploration of Mars.
High-fidelity simulations of nozzle flows relevant to supersonic retropropulsion environments are conducted using CRoCCo-AMR. The computational approach is discussed in detail. Initial results are presented for a single-nozzle condition with comparison to available experimental data. Two additional single-nozzle retropropulsion configurations are simulated with reductions in grid size of an order of magnitude. Strong agreement is demonstrated in key regions of the flowfield. Mean and statistical analyses are performed to characterize unsteady, turbulence quantities present in the canonical retropropulsion environment. This study provides new and important insight into the behavior of the vortex shedding present within the SRP flowfields. A detailed accounting of unsteadiness and frequency content is provided in a novel analysis coupling the effects of off-body wake flow to fluctuations in wall quantities