http://www.rssboard.org/rss-specification 720 https://escholarship.org/uc/ucsb_etd/rss Recent eScholarship items from UC Santa Barbara Electronic Theses and Dissertations Sun, 30 Aug 2026 01:39:38 +0000 https://escholarship.org/uc/item/82k864kp Financial markets produce vast amounts of multimodal data, including textual sources, such as financial news, earnings conference calls, and financial reports, coupled with structured numerical data, such as market prices and accounting statements. This combination of rich multimodal inputs and the natural availability of market-based supervision presents a compelling opportunity for the development of machine learning and natural language processing methods. However, financial data exhibits several unique challenges. Financial documents are long, complex, and domain-specific. The data has strong temporal structure and information must be interpreted in the context of prior related events. In addition, textual and numerical modalities must be jointly modeled to capture their unique patterns and interactions. While existing language models are remarkably capable, they remain limited in their ability to address these challenges, as they struggle with long-range temporal dependencies,... https://escholarship.org/uc/item/82k864kp Thu, 27 Aug 2026 00:00:00 +0000 Koval, Ross https://escholarship.org/uc/item/4zh085qd The underlying theme that characterizes my research arc has been that of phenomenological modeling of dynamical systems and mathematical abstraction of networks. By abstracting complicated networks (such as neural networks trained to learn dynamic system behavior), we can build much lower parameter and geometrically interpretable models that capture the same properties of the more expensive and harder to understand deep neural networks. Additionally, an abstracted vision of biochemical networks allows us to discover new mechanisms within cellular transcriptional networks to better understand how bacteria can adapt to and perform computation in challenging and changing environments. In Chapter 2 we analyze the learned dictionaries from the deep network based Koopman system identification algorithm, deepDMD. We explore the theoretical basis for the strong performance of this algorithm. We do so by discovering dictionaries to satisfy a property of subspace approximation, which... https://escholarship.org/uc/item/4zh085qd Thu, 27 Aug 2026 00:00:00 +0000 Johnson, Charles Addison https://escholarship.org/uc/item/5sj15084 This study examines how partner perceptions, more specifically, expectation of partner's emotion influence guilt-related emotional responses in interpersonal contexts. Using interactive transgression tasks across two studies (N=121 and N=185), we demonstrate that individuals' emotional responses closely track their expectations of partners' emotions, with this pattern persisting across American and Chinese samples. Inter-subject representational similarity analysis revealed associations between general social trait judgments style and context-specific emotion expectations. Cultural differences emerged in avoidance behaviors, with American participants showing increased avoidance when anticipating partner anger, while Chinese participants maintained low avoidance regardless of expected anger. These findings highlight the interpersonal nature of self-conscious emotions and the importance of studying emotions within authentic social contexts. https://escholarship.org/uc/item/5sj15084 Mon, 24 Aug 2026 00:00:00 +0000 Zhao, Shangcheng https://escholarship.org/uc/item/26m6b4n0 Ectomycorrhizal association is a prevalent symbiosis that plays a crucial role in supporting temperate and boreal forest ecosystems through the provisioning of fungal foraged nutrients for plant-derived carbon. The stability of this ecological relationship can be viewed as heavily dependent on environmental constraints and resource dynamics. To test how associations with distinct fungal communities impact the performance of Pinus muricata seedlings growing under controlled greenhouse conditions, we inoculated seedlings with ectomycorrhizal fungi using field-collected soil from the Jug Handle State Reserve Ecological Staircase in Mendocino, California. We collected field soils from two terraces which have been previously described to have distinct fungal communities, and established four inoculum treatments: one from Terrace 2 (T2), a relatively resource-replete site with tall trees and low carbon-demand fungi, one from Terrace 3 (T3), a comparatively resource-deplete site with... https://escholarship.org/uc/item/26m6b4n0 Mon, 24 Aug 2026 00:00:00 +0000 Sedano, Jazmin https://escholarship.org/uc/item/15d574jp High-dimensional dynamical systems arise in many scientific and engineering applications, yet scalable probabilistic methods for estimating latent processes and forecasting future states with reliable uncertainty quantification remain limited. Latent factor models with Gaussian process (GP) priors on the latent factors offer a flexible probabilistic framework that naturally captures spatiotemporal dependence, quantifies uncertainty, and performs well in data-limited settings. However, existing approaches face two fundamental challenges. First, high-dimensional dynamical systems require estimating a large number of parameters, while GP priors on latent factors incur cubic computational complexity, making scalable inference prohibitive. Second, forecasting nonlinear dynamics with uncertainty quantification remains challenging when the governing equations are unknown.This dissertation addresses both obstacles through three contributions. First, in Chapter 2, we introduce a latent... https://escholarship.org/uc/item/15d574jp Mon, 24 Aug 2026 00:00:00 +0000 Lin, Yizi https://escholarship.org/uc/item/9td4j1wv Despite rapid advances in computing and multimedia, there is a digital divide surrounding our sense of touch: Existing technologies cannot facilitate the sharing of haptic experiences in the ways that are possible with other media. The challenges in realizing such technologies arise from the hand’s biomechanical complexity and extensive sensory capabilities, as well as from the intrinsically active nature of touch interactions. This dissertation builds upon prior research on haptic perception and hand function to investigate opportunities for new haptic technologies and methods that can facilitate the sharing of haptic experiences. After two chapters of introduction and background information, the third chapter addresses the need for the digital transcription of interpersonal touch communication through tactile sign languages (TSLs), used by people with multiple sensory impairments. It presents a wrist-worn smart bracelet that senses touch-evoked skin oscillations, extracts... https://escholarship.org/uc/item/9td4j1wv Sat, 22 Aug 2026 00:00:00 +0000 Dinulescu, Stejara Iulia https://escholarship.org/uc/item/9tb951hx Environmental shocks, such as marine heatwaves and harmful algal blooms, greatly damage fishing communities. Fishers have limited access to financial tools to ameliorate income loss due to diminished fishery productivity. Index insurance is a promising new tool to help protect fishers during environmental downturns. However, designing effective index insurance contracts for fisheries presents distinct and unresolved challenges. Unique fishery characteristics such as vulnerability to overexploitation, complex environmental to harvest relationships, and interaction with management have stymied deployment. This dissertation addresses each of the limitations in three chapters to enable effective and sustainable use of index insurance in fisheries.Chapter one models fisher behavior changes induced through insurance moral hazard. My results show standard fishery production models in unconstrained settings lead insurance to incentivize increased harvest. Insurance will not be viable... https://escholarship.org/uc/item/9tb951hx Sat, 22 Aug 2026 00:00:00 +0000 Grimes, Nathaniel https://escholarship.org/uc/item/9rz7x2pz Machine learning promises to help networks monitor, diagnose, secure, and optimize themselves under conditions that increasingly exceed the reach of static rules and manual operation. Yet this promise in academia is limited by a credibility crisis: many network ML systems report excellent benchmark performance while remaining opaque, brittle under deployment shifts, and dependent on artifacts of the data-collection or evaluation pipeline. This dissertation argues that credible machine learning for networking requires treating diagnosis, data collection, and model design as a connected problem rather than as isolated stages.The dissertation studies this problem through three complementary methods. First, it develops diagnostic techniques for scrutinizing what network ML models learn. Trustee extracts high-fidelity, low-complexity decision-tree explanations and trust reports from black-box models, and the Intrinsic Evaluation Framework probes network foundation models through embedding... https://escholarship.org/uc/item/9rz7x2pz Sat, 22 Aug 2026 00:00:00 +0000 Beltiukov, Roman https://escholarship.org/uc/item/9q165317 The earliest known occurrence of eukaryotic biomineralization is the 812–717 Ma apatitic scale microfossils (ASMs) from the Fifteenmile Group in the Yukon Territory, Canada. ASMs are considered analogous to the body scales produced by extant unicellular eukaryotes (protists) and exhibit a broad range of body shapes with distinct structural features, including lattices with hexagonally-packed pores, centrally projecting barbs, and folded rims bearing radial ribs. Various phylogenetic affinities of ASMs to extant clades of scale-bearing protists have been proposed, predominantly on the basis of shape, and, in some cases, mineralogy. We used scanning, focused ion beam, and transmission electron microscopy with energy dispersive X-ray spectroscopy to observe the internal structures of three scale specimens to better resolve fine-scale morphological details that may clarify their phylogenetic placement. Our results indicate a previously unrecognized bilayer construction of the scales,... https://escholarship.org/uc/item/9q165317 Sat, 22 Aug 2026 00:00:00 +0000 Felder, Emily https://escholarship.org/uc/item/9ph5z3nq Across organizations and occupations, emerging technologies are reshaping expert work in ways that the literature has struggled to make sense of holistically. Research on expertise has developed along three largely separate tracks, treating expertise as knowledge held by individuals, as performance enacted for audiences, or as recognition conferred by others, and each has produced its own account of how technology disrupts expertise based on these ontological divides. However, these are only partial accounts of a larger phenomenon that is difficult to capture in one study. This dissertation attempts to do just that. I argue that knowing, performing, and recognizing are not competing definitions of expertise but component, interwoven constitutive processes within a unified ecological framework of expert work, and that technology, once it enters expert work, becomes endogenous to each of them and to the recursive loops that bind them together. I develop this argument through three... https://escholarship.org/uc/item/9ph5z3nq Sat, 22 Aug 2026 00:00:00 +0000 Shen, Roni https://escholarship.org/uc/item/9p3309rp Vast evidence supports the claim that 85% of matter in the universe is “dark,” interacting only gravitationally. Its influence is observed across length and time scales, from the rotation of galaxies to the cosmic microwave background. Despite this, the composition of dark matter stands as one of the most significant open questions in modern physics. One of the most well-motivated candidates is a class of particles called Weakly Interacting Massive Particles (WIMPs). The LUX-ZEPLIN (LZ) dark matter experiment is a dual-phase xenon time projection chamber designed to search for WIMPs, which has produced world-leading exclusion limits on the WIMP-nucleon scattering cross-section.This thesis presents involvement across multiple facets of LZ: Outer Detector (OD) characterization, statistical inference, and the proposed HydroX upgrade. First, the identification of intrinsic radioactivity in the OD provides an in-situ, continuous calibration source, which is used to determine detector... https://escholarship.org/uc/item/9p3309rp Sat, 22 Aug 2026 00:00:00 +0000 Trask, Makayla https://escholarship.org/uc/item/9md321rk Mezcal, an artisanal agave spirit produced primarily in Oaxaca, Mexico, has undergone rapid commercialization over the past two decades, expanding from a locally consumed cultural product to a globally exported commodity. This boom has generated significant economic activity, yet the rural Oaxacan communities whose traditions and labor sustain mezcal production frequently do not share in the value the industry generates. This thesis examines that disjuncture, asking how global demand is reshaping mezcal production and cultural heritage in Oaxaca, whether mezcal can function as a tool for economic development for rural producers, and what differing organizational responses to inequality in the industry reveal about the possibilities and limits of development intervention.The research is based on five months of fieldwork in Oaxaca, conducted through semi-formal and informal interviews with mezcal producers across three comparative field sites - Santiago Matatlán, Santa Catarina... https://escholarship.org/uc/item/9md321rk Sat, 22 Aug 2026 00:00:00 +0000 Stollenwerk Cavallaro, Marinella Stollenwerk https://escholarship.org/uc/item/9mc7f6kj Immigration and health scholars tend to define resilience as the process of successful adaptation to adversity and risk, often through internally directed goals. This study differs in that it centers immigrant communities’ conceptualizations of resilience to understand its differing uses across legal statuses. Based on 36 interviews with primarily Latinx immigrant workers, I engage in qualitative and narrative analysis to explore how resilience is understood across legal barriers during the COVID-19 pandemic. My analysis frames respondent’s narrative employment of resilience as a dialogic resource for making meaning which reproduce or resist neoliberal discourses. I found that immigrant communities’ conception of resilience during the COVID-19 pandemic was complex; it served as an empowering resource for those relatively privileged and legally secure while those without legal status were weighed down by the need to explain their experiences within this framework, sometimes contesting... https://escholarship.org/uc/item/9mc7f6kj Sat, 22 Aug 2026 00:00:00 +0000 Perez, Timothy https://escholarship.org/uc/item/9gz4587r Polyurethane (PU) is the sixth most produced plastic globally, with 21 million metric tons produced in 2021. Its properties can be tuned by varying the structure and ratio of the polyol and isocyanate monomers used during synthesis, enabling broad use across industrial and consumer applications. As a highly crosslinked polymer with high chemical and thermal stability, conventional mechanical recycling of PU typically yields low-value products. As a result, about 80% of PU is landfilled or incinerated. Chemical recycling offers a more sustainable end-of-life pathway by recovering high-value starting materials that can be reused in PU production.Among the available chemical recycling routes, acidolysis and glycolysis show promise due to their relatively mild reaction conditions and high-purity product streams. However, limited kinetic and mechanistic understanding has hindered their broader implementation. In this work, I combine computational chemistry and kinetic modeling to identify... https://escholarship.org/uc/item/9gz4587r Sat, 22 Aug 2026 00:00:00 +0000 Richardson, Kelsey https://escholarship.org/uc/item/9gx8x29x Fire safety in the built environment presents a paradox: advances in prevention, detection, building standards, and emergency response have reduced many fire impacts, yet structure fires continue to produce a disproportionate share of deaths, injuries, and property losses. In fire-prone coastal communities, this problem is intensified by the overlap of urban development and wildland-urban interface exposure, where fires can move from vegetation into neighborhoods and, once structures are involved, from building to building. Santa Barbara County illustrates this coupled risk context. Its coastal communities sit between the Santa Ynez Mountains and the Pacific Ocean, where dense development, complex road networks, seasonal winds, steep terrain, and wildfire exposure converge. In such settings, fire preparedness depends not only on preventing ignition or deploying firefighters, but also on whether the infrastructure that supports suppression is spatially configured to function where... https://escholarship.org/uc/item/9gx8x29x Sat, 22 Aug 2026 00:00:00 +0000 Heffernan, Vanessa https://escholarship.org/uc/item/9dw9r52b The SLn spider from is a diagrammatic category that describes the representation theory of the quantum group Uq(sln). Cobwebs, first defined in , are another type of diagram that encode the combinatorial data of sln. I define a new set of cobweb relations, including the pop-switch relation from [12], and prove that these relations compute the Ocneanu cells in type An. With this modified set of cobweb relations, I demonstrate how cobwebs provide a diagrammatic way to compute the Ocneanu cells.The main result in implies that any closed web in the SLn spider can be reduced to a scalar multiple of the empty diagram, by repeatedly applying spider relations. Computational evidence strongly suggests that any closed web simplifies to a symmetric unimodal polynomial over the Laurent ring Z[q, q−1 ]. I make significant progress towards a proof of this conjecture by defining a collection of relations in the SLn spider that preserve the symmetric and unimodal properties. Each relation must... https://escholarship.org/uc/item/9dw9r52b Sat, 22 Aug 2026 00:00:00 +0000 Mantell, Sarah https://escholarship.org/uc/item/96j8g610 What are the neural and behavioral consequences of severing the cerebral hemispheres in humans? Split-brain patients are individuals whose corpus callosum is surgically severed to treat intractable epilepsy. Their unique condition has attracted neuroscientists, psychologists, and philosophers, raising fundamental questions about the nature of human consciousness, selfhood, and agency. In this thesis, I use behavioral and functional neuroimaging approaches, and develop new experimental tools, to examine what remains integrated versus disconnected after hemispheric severance. In Chapter 1, I review different conceptualizations of “unity” in split-brain patients. In Chapter 2, I investigate the behavioral effects of hemispheric severance using both bedside and computerized testing. First, I develop a bedside assessment battery that can be used clinically to characterize disconnection effects across various sensory-motor and cognitive domains, and a novel “disconnection score” that... https://escholarship.org/uc/item/96j8g610 Sat, 22 Aug 2026 00:00:00 +0000 Bekir, Selin https://escholarship.org/uc/item/9523z99t This qualitative study examines how Korean American undergraduate students interpret and experience food within campus and local food environments at a public research university on the West Coast, referred to by the pseudonym Uri University. Drawing on semi-structured individual interviews, “gimbap chat” group conversation sessions, dining hall observations, systemic analysis of dining hall menus, and contextual institutional interviews, I explore the emotions, memories, and cultural meanings Korean American students attach to their on-campus dining experiences and examine how those meanings shape their interpretations of campus food environments and their sense of belonging. Guided by the Culturally Engaging Campus Environments (CECE) model and Suda, a Korean cultural conversational practice theorized as a decolonizing research methodology, I find that campus dining systems organized around White, Western, and standardized food norms produce experiences of cultural exclusion,... https://escholarship.org/uc/item/9523z99t Sat, 22 Aug 2026 00:00:00 +0000 Hong, Damhee https://escholarship.org/uc/item/8zg9f7zd This dissertation examines how changes in fossil fuel activity reshape labor markets, youth outcomes, and local public finances in the United States. Chapter 1 uses matched employer-employee data on 1.35 million workers separated from fossil fuel extraction (1999-2019) to estimate the impacts of local fossil fuel labor demand shocks. Employment probabilities fall sharply after exposure, and earnings drift down for seven years with only partial recovery by year ten. Workers who remain in fossil fuels, disproportionately men in sector-specific roles, lose nearly twice as much as sector switchers, consistent with limited occupational mobility. Among non-switchers, losses are largest in markets with high employer concentration, indicating monopsony depresses reemployment wages. Geographic movers fare worse than stayers, reflecting negative selection and moves into metropolitan niches where fossil fuel or low-skill services remain highly concentrated. Chapter 2 links coal... https://escholarship.org/uc/item/8zg9f7zd Sat, 22 Aug 2026 00:00:00 +0000 Hyun, Minwoo https://escholarship.org/uc/item/8xs1886f Under what conditions do anti-gender campaigns succeed in undermining transgender rights? This dissertation addresses this question by investigating how structural country-level conditions and advocacy strategies shape the outcomes of anti-transgender mobilization across countries. I argue that two structural conditions, the level of liberal democracy and the institutionalization of organized religion, are central to determining whether anti-gender campaigns succeed in reversing or constraining transgender rights, and I identify five strategies through which anti-gender actors pursue this goal: transnational coalition building, diplomacy, contestation of liberal human rights, alliances with right-wing populism, and scapegoating of sexual and gender minorities.I use mixed methods to test these arguments in two parts. First, I conduct a cross-national panel analysis covering 162 countries across three four-year periods between 2011 and 2022, estimating pooled OLS models with period... https://escholarship.org/uc/item/8xs1886f Sat, 22 Aug 2026 00:00:00 +0000 Sacramento Moreira, Vitoria https://escholarship.org/uc/item/8qj3h0q0 I. The Suzuki-Miyaura coupling reaction is a cornerstone of modern organic chemistry and, in the decades since its first disclosure, significant strides have been made to enhance its efficiency and environmental sustainability. However, the directly preceding reaction, the Miyaura borylation, routinely involves the use of toxic and wasteful solvents, harsh reaction conditions, and high loadings of transition metal catalysts. Reported is a new ’near-neat’ method for synthesizing ethylpinacol boronic esters (BEpins) that generates these stable boron reagents in concentrated EtOAc, with moderate reaction temperatures, and with highly efficient palladium catalyst loadings. Additionally, despite their enhanced stability, the BEpins are also shown to be effectively implemented in a sustainable 1-pot, on-water borylation/Suzuki-Miyaura coupling protocol to generate a diverse library of bi(hetero)aryl substrates, including several notable pharmacologically relevant examples.II. A new,... https://escholarship.org/uc/item/8qj3h0q0 Sat, 22 Aug 2026 00:00:00 +0000 Nelson, Chandler Blaine https://escholarship.org/uc/item/8mn5d98h Herbivorous fishes provide critical ecological roles on coral reefs. However, coral reefs are undergoing local and global stressors that can alter the dominant benthic organisms on reefs, often leading to reefs with abundant algae. Given the importance of herbivorous fishes in mediating coral-algal competition, it is essential to re-evaluate their functional roles in the age of the Anthropocene. This dissertation addresses three important knowledge gaps. 1) Studies increasingly show that herbivore foraging behavior varies along axes beyond food choice, underscoring the need to consider additional dimensions of foraging such as bite rates and space use to understand patterns of herbivory. While these important niche axes have been identified, there is a deficit of research evaluating across all of these ecological dimensions for a diverse suite of herbivorous fishes. 2) Many coral reefs across the globe are experiencing a loss of coral and increase in algae in response to global... https://escholarship.org/uc/item/8mn5d98h Sat, 22 Aug 2026 00:00:00 +0000 Molina, Nury Elizabeth https://escholarship.org/uc/item/8k31h6nn Cilia are microscopic hair-like structures which transport fluid in biological systems. In mucociliary clearance, cilia pump a multilayered film of fluid to remove contaminants from the lungs. At points deep in the lungs (where the mucus layer becomes thin), the cilia interact near the free surface, an effect which has been largely unexplored in an experimental set-up. We first develop an analytic model for the flow response to cilium motion confined between a free-slip surface and a no-slip boundary. Next, we design an experimental set-up which can approximate the motion of biologic cilia using a rotating magnet and a polymer seeded with iron particles. Through particle image velocimetry (PIV), we show deviation from theory in the response of the fluid, specifically that under certain conditions the flow can completely reverse from the expected direction. Lastly, we show there is a transient response occurring which would violate the conditions in a purely Stokes flow and we... https://escholarship.org/uc/item/8k31h6nn Sat, 22 Aug 2026 00:00:00 +0000 Abell, Ryan Michael https://escholarship.org/uc/item/8jr107fw The diverse structure and function of biological systems often depend on the dynamic conformations and assemblies of biomolecules. In particular, electrostatic interactions arising from ionizable amino acid residue side chains play a central role in regulating biomolecular structure in vivo through changes in pH or post-translational modifications. Developing methods that can actively trigger and control assembly is therefore essential to understand protein function and to gain insight into the biophysical mechanisms that govern folding and assembly pathways.This thesis presents an electrochemical approach for precisely modulating electrostatic interactions in biomolecules by neutralizing protonated amino acid residues, thereby enabling the study of protein structural evolution driven by changes in Coulombic interactions and intermolecular assembly. Electrochemistry was combined with in situ spectroscopies (e.g., dynamic light scattering, circular dichroism, ellipsometry, and... https://escholarship.org/uc/item/8jr107fw Sat, 22 Aug 2026 00:00:00 +0000 Lin, Yin-Chen https://escholarship.org/uc/item/8gc109rc Post-fire debris flows (PFDFs) are a cascading hazard resulting from the interaction between intense rainfall and susceptible terrain altered by wildfire. These events have proven devastating in the past, damaging infrastructure and communities, particularly in California, where fire season in the warm months is closely followed by rainfall during the cool season. The expansion of wildfire occurrence, along with pronounced variability in terrain and climate across California, highlights the need to better understand the rainfall conditions associated with this hazard under both current and future climates. Consequently, this dissertation integrates multiple datasets and methods to characterize rainfall and atmospheric conditions associated with storms triggering post-fire debris flows, providing new insight into the rainfall environments, precipitation-system structure, predictive controls, and future changes in PFDF hazard across California. First, an inventory of PFDF is compiled... https://escholarship.org/uc/item/8gc109rc Sat, 22 Aug 2026 00:00:00 +0000 Castillo Guerra, Silvana https://escholarship.org/uc/item/8g66n5c0 This thesis develops stochastic many-body methodologies for the electronic structure description of chemically relevant systems with strong electron correlation and nonequilibrium dynamics. Many phenomena of interest in chemistry and materials science, including excited-state processes, charge transfer at catalytic interfaces, and plasmon-mediated photocatalysis, require an accurate treatment of interacting electrons beyond mean-field theory. However, direct solutions of the many-body Schrodinger equation remain computationally intractable for realistic systems due to the exponential growth of the configurational Hilbert space with system size. This challenge motivates the development of reduced representations and embedding strategies capable of retaining essential correlation effects at lower computational cost. The work presented here focuses on stochastic matrix compression and downfolding techniques for many-body electronic structure theory. In particular, new stochastic... https://escholarship.org/uc/item/8g66n5c0 Sat, 22 Aug 2026 00:00:00 +0000 Canestraight, Annabelle Louise https://escholarship.org/uc/item/8g3405nf Cell boundary detection and segmentation enable the quantitative analysis of cellular organization, morphology, and dynamics, which are fundamental to understanding biological processes and disease mechanisms. Accurate detection and segmentation from time-lapse microscopy videos are essential for providing longitudinal information of cell boundaries. Popular unsupervised cell segmentation tools, such as ImageJ, require manual tuning of parameters and rely on restrictive assumptions about the shape of the objects. While recently developed supervised segmentation tools based on convolutional neural networks enhance accuracy, they depend on high-quality labeled images for training, making them unsuitable for segmenting new types of objects not in existing databases. To address these limitations, we developed an unsupervised cell segmentation algorithm from noisy microscopy images using fast Gaussian processes without the need for parameter tuning or restrictive assumptions... https://escholarship.org/uc/item/8g3405nf Sat, 22 Aug 2026 00:00:00 +0000 King, Blythe https://escholarship.org/uc/item/88z9w5fj This dissertation seeks to unsettle the entrenched boundaries of the human, offering a crip (disabled and queer) intervention into the parameters of the humanist subject. My research engages with recent affective and cognitive turns in literary and cultural studies, turning to neurodivergence as a mode of critique. Through a neuroqueer (queer and neurodivergent) lens, my work critiques contemporary understandings of the human and the humane that are rooted in intertwined histories that pathologize those who fall outside of relational norms. Such pathologization often plays out through psychiatric and social narratives of neurological difference. In opposition to accounts of neurodivergence that foreclose agency, I propose that neurodivergent creators deploy aesthetic, rhetorical, and affective strategies that disrupt, reshape, and queer dominant discourses. In doing so, they exercise more expansive notions of agency that move beyond the parameters of humanism to embrace neuroabjection... https://escholarship.org/uc/item/88z9w5fj Sat, 22 Aug 2026 00:00:00 +0000 Killeen, Shanna https://escholarship.org/uc/item/80j8n51z This thesis presents three languages for reasoning about timed, concurrent, or reactive, systems. The field of reactive systems is vast, covering technologies from distributed systems to automata, from industrial process to electronic circuits. The core characteristics include sensitivity to features such as timing, communication between components, mutual access to shared data, and intentional non-termination. Research in reactive systems is traditionally associated with model checking and other forms of post-hoc verification. While this traditional approach provides a large degree of automation, covering a wide array of verification tasks, it also has several drawbacks. These drawbacks include scalability issues, absence of equational reasoning, and a lack of modularity between models and property encodings. Additionally, the field of reactive systems is relatively understudied from the perspective of programming language design and theory.We approach reactive systems through... https://escholarship.org/uc/item/80j8n51z Sat, 22 Aug 2026 00:00:00 +0000 Kringen, Harlan https://escholarship.org/uc/item/7zv8h7ff Photorealistic image synthesis has long been a central goal of computer graphics. Classical rendering pursues this goal by simulating light transport on an explicit scene description: it delivers precise artistic control and consistency, but falls short of ultimate photorealism and depends on costly assets and expensive light transport. Generative image and video models pursue the same goal differently. They learn the distribution of real photographs and videos and synthesize new samples directly, achieving high photo-realism, yet they lack precise control and consistency.In this dissertation, we combine the strengths of both approaches through generative rendering, using G-buffers – per-pixel scene quantities such as albedo, normals, roughness, metallicity, depth, and lighting – as a structured control interface between a rendering pipeline and a learned generative model. G-buffers are structured enough to carry precise artistic control, yet compact enough to leave a generative... https://escholarship.org/uc/item/7zv8h7ff Sat, 22 Aug 2026 00:00:00 +0000 Zeng, Zheng https://escholarship.org/uc/item/7x048410 Optimal control of nonlinear dynamical systems is often approached through reduced-order models, gradient-based optimization, or stochastic search methods. While computationally tractable, these approaches implicitly restrict where optimal solutions may reside in state space. This thesis studies coupled nonlinear oscillators and shows that these assumptions can fail simultaneously. By solving the full four-dimensional Hamilton-Jacobi-Bellman equation on a GPU-accelerated grid, we identify a critical coupling strength beyond which the globally optimal control leaves the limit-cycle manifold entirely. This regime is missed by phase reduction, gradient methods, and stochastic search alike, establishing the transition as a multi-method failure boundary rather than a limitation of any one approach.The thesis makes two main contributions. First, it establishes off-manifold optimal control as a fundamentally full-state phenomenon and identifies a mechanism, termed amplitude collapse,... https://escholarship.org/uc/item/7x048410 Sat, 22 Aug 2026 00:00:00 +0000 Rajabi, Faranak https://escholarship.org/uc/item/7qx9n418 Family involvement and engagement at school are each associated with student outcomes; however, few studies have examined cultural considerations at school associated with family involvement at school. Likewise, school climate is also associated with student outcomes, however, cultural considerations have largely been absent in the school climate research. No existing tools measure culturally appropriate methods to understand family engagement. This study examines the School Cultural Congruity Scale (SCCS) which measures cultural alignment between home and schools to promote family engagement. Considering cultural and contextual factors, this tool will help to provide schools with information by identifying areas for further improving family engagement, school climate, and ultimately enhancing student outcomes. Participants were from 5 countries (Australia, China, Georgia, Latvia, and USA). Study 1 employs Exploratory Factor Analysis to further understand the SCCS factor structure,... https://escholarship.org/uc/item/7qx9n418 Sat, 22 Aug 2026 00:00:00 +0000 Kaur, Lakhvir https://escholarship.org/uc/item/7md8k4r3 In this thesis, I analyze Indigenous peoples' interactions with contemporary Colombian political and social spheres regarding their health, autonomy, practices, and systems, specifically in relation to the passing of SISPI, the Self-Determined and Intercultural Indigenous Health System. After years of efforts involving local stakeholders, Indigenous leaders, and health rights activists, Decree Law 0480 was passed in April 2025, meant to strengthen Indigenous peoples' own health systems and mechanisms for coordination with the national health system. Although widely celebrated across Indigenous communities in Colombia, SISPI was passed under a left-wing government more inclined toward protecting Indigenous autonomy and rights, raising urgent questions about funding, implementation, and continuity ahead of the May 2026 elections. This thesis asks: can national health systems genuinely support and coexist with Indigenous health systems and practices? And what do Wayuu women's own... https://escholarship.org/uc/item/7md8k4r3 Sat, 22 Aug 2026 00:00:00 +0000 Gonzalez, Paula https://escholarship.org/uc/item/7gr1d8j3 As computing has evolved, general-purpose processors have been expected to support an increasingly diverse set of applications, from latency-sensitive single-threaded programs to throughput-oriented and memory-intensive workloads. These workloads can differ substantially in their demand for instruction-level parallelism (ILP), memory capacity, and execution resources. A fixed monolithic core must therefore be designed around one compromise point in a much larger design space. Scaling such cores also increases the cost and complexity of tightly coupled microarchitectural structures. Together, workload diversity and inflexible hardware architectures limit both performance and utilization.Prior work has addressed this mismatch through configurable and composable architectures that allow compute resources to be reorganized around workload demand. By grouping or partitioning execution resources, these architectures aim to improve resource utilization and provide a more flexible alternative... https://escholarship.org/uc/item/7gr1d8j3 Sat, 22 Aug 2026 00:00:00 +0000 Bui, Minh https://escholarship.org/uc/item/7bx5g17w Eye movements are an important part of human vision. We constantly make them, directing our foveal processing to different locations. Research has shown eye movements to optimize accuracy in various tasks, such as search and navigation, implying that human eye movements are an active process, constantly seeking information that could aid decision-making. But what happens when there is no task to guide eye movements (free viewing task)? Is there any purpose to those eye movements? A prevalent idea is that during free viewing, humans make eye movements to low-level visually salient regions. However, decades of research have shown that humans fixate on people in scenes, on objects, on gaze, and on text, and, more recently, on local regions judged to be meaningful. The goal of this thesis is to understand which perceptual tasks humans engage in during free viewing and to assess whether free-viewing eye movements reflect the goal of optimizing task accuracy (optimal eye movements).... https://escholarship.org/uc/item/7bx5g17w Sat, 22 Aug 2026 00:00:00 +0000 Murlidaran, Shravan https://escholarship.org/uc/item/7bw5x81d Flavoproteins comprise an estimated 1-3% of all eukaryotic and prokaryotic proteins, functioning as light-, oxygen-, and voltage-sensors, photosensitizers, catalysts, and electron-transfer reagents. These proteins contain a yellow redox-active flavin prosthetic group in the form of flavin mononucleotide (FMN) or flavin adenine dinucleotide (FAD), derived from riboflavin (vitamin B2). Flavins contain an aromatic isoalloxazine ring system whose redox properties are highly sensitive to the characteristics of the surrounding protein medium. Thus, they are highly versatile, as their reduction potentials can be finely tuned to match those of various substrates determined by the specific protein in which they are bound.Depending on the specific thermodynamic arrangement of a flavin’s two one-electron reduction potentials, they can be coaxed to participate in either one- or two-electron transfer reactions. While the mechanisms that flavoproteins use to modulate the reduction potentials... https://escholarship.org/uc/item/7bw5x81d Sat, 22 Aug 2026 00:00:00 +0000 Jones, Benjamin Joseph https://escholarship.org/uc/item/79j5j9fs Women in early modern Japan (1603-1868) faced strict social and governmental regulations on their behavior and mobility. However, retirement— marked by the rearrangement of household duties following the arrival of a daughter-in-law— granted these women a unique space for self- discovery. In my dissertation I argue that, due to a shift in neo-Confucianist morals codes which came into effect as people entered retirement, retired women experienced greater freedoms compared to earlier stages in their lives. I further argue that this newfound freedom manifested in travel, religious expression, and, for literate women, literary endeavors, allowing retired women to fashion new identities and femininities.Societal norms, rooted in neo-Confucianist moral codes, largely confined women to domestic roles earlier in their lives. However, my research reveals that while prescriptive literature for women instructed them on how to behave as daughters, wives, daughters-in-law, and mothers, it... https://escholarship.org/uc/item/79j5j9fs Sat, 22 Aug 2026 00:00:00 +0000 Trumble, Erin Kelly https://escholarship.org/uc/item/6xm39997 Geometric inequalities are one of the main ways in which curvature controls the analysis and geometry of a Riemannian manifold. Sobolev-type inequalities compare a function with its derivatives and are closely connected to volume growth, heat flow, and concentration phenomena. Isoperimetric inequalities compare the volume of a region with the area of its boundary and give a geometric measure of how efficiently a space encloses volume. In comparison geometry, such inequalities are often understood by comparing a manifold or submanifold with a model space, such as Euclidean space, a sphere, or a space form of constant curvature. This thesis studies two problems in this direction: logarithmic Sobolev inequalities for submanifolds and comparison theorems for isoperimetric profile functions. In both cases, the goal is to extend classical comparison principles under weaker or more flexible curvature assumptions. The results are obtained in joint work with Jihye Lee.The first part of... https://escholarship.org/uc/item/6xm39997 Sat, 22 Aug 2026 00:00:00 +0000 Ricci, Fabio https://escholarship.org/uc/item/6sc2k4z5 This thesis studies regime-switching polynomial diffusion models and their statistical inference, and proposes a novel approach that incorporates forward-looking information via the Onsager–Machlup (OM) functional.Classical likelihood-based inference relies on pointwise observations and does not incorporate forward-looking information. In this thesis, we propose a path wise inference approach based on the OM functional, which provides a continuous-time analogue of likelihood for diffusion trajectories and enables most-probable trajectory estimation. As a result, the proposed framework allows for more timely detection of regime changes and yields more accurate reconstruction of the underlying diffusion process.Building on this idea, we extend the Topological Hidden Markov Model (THMM) framework to regime-switching polynomial diffusion models. This formulation allows inference to remain computationally simple while preserving the use of standard Hidden Markov Model algorithms, including... https://escholarship.org/uc/item/6sc2k4z5 Sat, 22 Aug 2026 00:00:00 +0000 Xu, Minxing https://escholarship.org/uc/item/6r744391 The most extreme phenomena in the universe present the most intriguing opportunities to test our understanding of physics in environments no laboratory on Earth can simulate. In this dissertation, I present a series of observational and theoretical studies of extreme astrophysical systems, spanning infant supernovae and horizon-scale black hole physics. First, I investigate the earliest phases of core-collapse supernovae in both optical and radio wavelengths, describing and applying a framework for the robust application of shock-cooling models across the hydrogen-rich envelope mass spectrum. Next, I consider theoretical properties of the black hole shadow in general relativity, and extend these to perform angular momentum inference using modern very-long-baseline interferometers in dynamically evolving horizon-scale systems. These results directly contribute to the second Event Horizon Telescope image of a black hole (the first image of Sagittarius A*) and help explore the science... https://escholarship.org/uc/item/6r744391 Sat, 22 Aug 2026 00:00:00 +0000 Farah, Joseph Rachid https://escholarship.org/uc/item/6p5865v4 Among the mechanisms of cellular regulation, DNA supercoiling is crucial in modulating gene expression; whether DNA is underwound (negative supercoiling) or overwound (positive supercoiling) directly affects the accessibility of polymerases to initiate transcription. As transcription occurs, the opening of the transcription bubble and the subsequent movement of RNAP have an immense influence on the state of local supercoiling. Recent studies have shown how the natural propagation of negative and positive supercoils via RNAP movement leads to significant changes in gene expression relative to the nearby genetic syntax (convergent, divergent, and tandem gene orientations). While this discovery has offered great insight into the importance of genetic syntax on gene expression, there is much to know about the extent of these different orientations on cell fate, especially among pathogenic bacteria. Here, we explore how cell fate of bacteria is determined by modeling virulence-associated... https://escholarship.org/uc/item/6p5865v4 Sat, 22 Aug 2026 00:00:00 +0000 Nguyen, Annie https://escholarship.org/uc/item/6j04r7n3 Technology mapping translates hardware designs into the concrete primitives provided by target platforms such as simulators, FPGAs, and ASICs. In current flows, this process is often brittle, syntax-driven, and heavily phase-ordered: designers duplicate implementations across technologies, while tools depend on narrowly structured inputs and carefully tuned pass sequences. This thesis argues that technology mapping should instead be approached in a principled way: target technologies should be exposed through explicit semantic abstractions, and compilation should reason about equivalence across representations rather than committing early to a single syntactic form.The first part of the thesis develops this argument for memories through Memo, a language and compilation framework built around fungible memories. Fungible memories expose the semantic structure needed for automated mapping while hiding target-specific implementation details, allowing designers to specify a memory... https://escholarship.org/uc/item/6j04r7n3 Sat, 22 Aug 2026 00:00:00 +0000 Kong, Sijie https://escholarship.org/uc/item/64b21215 This thesis presents a measurement of the cross section for the production of two W bosons and one Z boson. The measurement is performed using data collected by the CMS experiment at the CERN LHC at center-of-mass energies √s = 13 and 13.6 TeV, corresponding to an integrated luminosity of 200 fb−1. Events with four charged leptons (electrons or muons) are selected for this analysis. Both nonresonant W W Z production and ZH production, with the Higgs boson decaying into two W bosons, are reported. For the first time, the two processes are measured separately in a simultaneous fit. Combining the two modes, signal strengths relative to the standard model (SM) predictions of 0.75+0.34 −0.29 and 1.74+0.71 −0.60 are measured for √s = 13 and 13.6 TeV, respectively. The observed (expected) significance for the triboson signal is 3.8 (2.5) standard deviations for √s = 13.6 TeV, thus providing the first evidence for triboson production at this center-of-mass energy. Combining the two modes... https://escholarship.org/uc/item/64b21215 Sat, 22 Aug 2026 00:00:00 +0000 Downham, Keegan https://escholarship.org/uc/item/62q393bn This dissertation develops a state-space modeling framework for analyzing stratospheric ozone concentration profiles over altitude and time. Ozone profiles exhibit smooth vertical structure and temporal variation, including seasonal patterns whose strength and timing can differ across altitude. To capture these features, the proposed model represents the vertical profile using cubic B-spline basis functions and models seasonal temporal variation through harmonic state components. The seasonal effects are allowed to vary smoothly across altitude through B-spline loading functions in the observation equation, which keeps the state dimension relatively low while maintaining flexibility. Model parameters are estimated by maximum likelihood using the Kalman Filter likelihood, and latent states are estimated through Kalman filtering and smoothing. An innovation-based bootstrap is used to quantify parameter uncertainty. Identifiability issues in the seasonal component are also... https://escholarship.org/uc/item/62q393bn Sat, 22 Aug 2026 00:00:00 +0000 Xiong, Moya https://escholarship.org/uc/item/61g9d5sc Nonequilibrium quantum dynamics, such as those in irradiated quantum matter, continue to hold surprises despite decades of significant theoretical and experimental advances. This thesis reports a series of results on the nonequilibrium control and engineering of quantum phases using strontium condensates in a strongly driven incommensurate optical lattice and dipole traps. These results include coherently tunable localization phase transitions, interlaced phase diagrams, and Floquet engineering of novel multifractal phases, all controlled by light of variable polarization, even in 1D. We will connect the implications of our findings to the mapping between our 1D quasicrystal and the 2D irradiated Harper model. Lastly, the parallel to a strongly-driven condensate and an irradiated gas allows us to observe strong-field stabilization in a strongly out-of-equilibrium condensate. https://escholarship.org/uc/item/61g9d5sc Sat, 22 Aug 2026 00:00:00 +0000 Bai, Yifei https://escholarship.org/uc/item/6086r90d Based on fourteen months of ethnographic fieldwork in Ainu Mosir (Hokkaido, Japan), this dissertation examines the ways in which Indigenous Ainu communities develop strategies to flourish in the 21st century. Through river-based activism, contemporary Ainu communities envision and enact futures of flourishing and abundance that are rooted in ancestral relationships to native lands and waters and the more-than-human beings who inhabit them. In so doing, they challenge Japan’s settler colonial project of Ainu dispossession and elimination, insisting upon a future for the Ainu as an Indigenous people. Rivers connect future visions to ancestral ways inherited from the past and remain an important medium through which kin-based, territorial, and cosmological rights and responsibilities are expressed and carried into the future. In their struggles for Indigenous rights today, Ainu activists continue to focus on rivers as the central locus of Ainu sovereignty and the origin from which... https://escholarship.org/uc/item/6086r90d Sat, 22 Aug 2026 00:00:00 +0000 Ioannides, Michael John https://escholarship.org/uc/item/5jt649q5 Global climate change is reshaping the Arctic with long-term impacts to both the environment and Indigenous cultures, but to enact sustainable policies addressing these changes a better understanding of stakeholder perceptions is needed than presently exists. By investigating core issues in the Arctic affected by environmental change including intangible cultural heritage, last chance tourism, and portrayals of the region in film, this dissertation challenges prevailing narratives while offering key recommendations for sustainable policy. Three major research questions were developed: (1) How do stakeholders perceive global climate change impacts in the North Slope Borough and the interconnections with intangible cultural heritage and life ways?; (2) How have films depicted Indigenous peoples, and especially women, living in the Arctic?; and (3) How have stakeholder experiences and perspectives of Arctic tourism developed in the region since the establishment of the North Slope... https://escholarship.org/uc/item/5jt649q5 Sat, 22 Aug 2026 00:00:00 +0000 Grafton, Daniel Jeremy https://escholarship.org/uc/item/5jm552xp My dissertation contains three chapters examining how changes in educational and labor market opportunities for women in India affect household formation, marriage, fertility, and outcomes for the next generation. The first chapter “Educational Expansion and Assortative Matching: Evidence from the Marriage Market in India” studies how school construction reshaped educational attainment and marital outcomes over the long run in India. Using staggered school construction across districts, I estimate the effects of increased school access on women’s education, marriage timing, and educational assortative matching between spouses. I find that greater access to schooling increased women’s years of education, delayed marriage, reduced child marriage, and narrowed spousal age gaps. Beyond individual outcomes, educational expansion also strengthened assortative matching, particularly among less educated groups, suggesting greater stratification at the lower end of the education... https://escholarship.org/uc/item/5jm552xp Sat, 22 Aug 2026 00:00:00 +0000 Arora, Akanksha https://escholarship.org/uc/item/5j29m2zw Why do some rebel groups attempt to create their own currencies and monetary institutions during armed conflict, before gaining formal statehood? This dissertation confronts a striking puzzle: since 1945, at least 38 rebel groups have engaged in governing money, building central banks, issuing currencies, or manipulating the state's monetary system, despite operating under economically destructive conditions of war. This phenomenon is puzzling on several dimensions. Rebels who govern money divert scarce resources away from fighting, assume that civilians will accept a monetary system backed by an actor without meaningful assets, and compete against the existing monetary system of the state. Scholarship on central banking and currency typically characterizes sovereign states and international organizations as the sole actors capable of conducting monetary affairs and views monetary institutions as tools for credible commitment and access to credit. Separately, scholarship on rebel... https://escholarship.org/uc/item/5j29m2zw Sat, 22 Aug 2026 00:00:00 +0000 Vester, Annjulie https://escholarship.org/uc/item/57x5g3rm South Korea's one-person households have doubled in two decades—a demographic transformation that unfolded over 30–50 years in other developed countries. Young adults aged 20–39 now comprise over one-third of individuals living alone. Drawing on 15 years of nationally representative panel data from the Korea Welfare Panel Study (2005–2019; 48,058 person-years of observation), this study examines associations between living alone and health outcomes—self-rated health and depressive symptoms—among young adults. Using random-effects panel models and logistic regression, the analysis shows that living alone is associated with elevated risks of poor physical health (33% higher odds) and depression (23% higher odds). These associations are partially attenuated by social insurance contributions, indicating the protective role of employment-based social protection. However, women living alone show elevated physical health risks that persist even when controlling for relationship satisfaction... https://escholarship.org/uc/item/57x5g3rm Sat, 22 Aug 2026 00:00:00 +0000 Kim, Bomi https://escholarship.org/uc/item/55g2t586 A positive school climate is foundational to creating a space where students are able to succeed and progress in their academic performance, well-being, and overall development. This dissertation used Phenomenological Variant of Ecological Systems Theory (PVEST) and intersectionality as conceptual guides to examine perceptions of safety and belonging in a statewide sample (N = 8,224) of Black middle and high school students in California. The broader project encompassed two studies. The first study employed a latent class analysis to identify patterns in 7th, 9th, and 11th grade students’ perceptions of school climate. This study analyzed the response patterns and identified the number of classes enumerated to understand how students perceived school climate by grade. The second study conducted a covariate analysis to explore how perceptions of school climate varied by gender and parental education level. Study 2 found significant differences at different grade levels and the... https://escholarship.org/uc/item/55g2t586 Sat, 22 Aug 2026 00:00:00 +0000 Holland, Shemiyah https://escholarship.org/uc/item/55g1753d Modern software systems are growing in complexity at a pace that surpasses our ability to secure them. Vulnerabilities hide in millions of lines of code in large open source software repositories, in the intricate business logic of financial protocols, and in the silent behavioral shifts introduced by software updates. At the heart of the problem is a deceptively simple question: how do we recognize when software is behaving in a way that it should not? Answering this question requires a reference for what "normal" looks like, and this dissertation argues that such a reference already exists, embedded in the history of software itself. The development process of a codebase, the execution traces of a deployed protocol, and the version history of an upgradeable contract all encode rich information about expected behavior. By mining these retrospective artifacts intelligently, we can build practical, automated tools that identify security threats across a wide range of software domains,... https://escholarship.org/uc/item/55g1753d Sat, 22 Aug 2026 00:00:00 +0000 Meng, Dongyu https://escholarship.org/uc/item/50c923qg Nonparametric regression has been well established as a flexible and applicable approach to regression. Many nonparametric regression techniques, however, break down in the presence of correlated noise, leading to fits that are far too noisy. In this thesis we develop an estimation technique for nonparametric regression that can be extended to cases where the noise is correlated. In the first chapter, we review relevant prerequisites from nonparametric regression, Gaussian Processes, and the spectral domain. We also include an overview of relevant literature. In the second chapter, we propose our estimation procedure, which is rooted in an Empirical Bayes framework. Hyperparameters are estimated in the spectral domain using a Generalized Linear Model, which allows our technique to be extended to problems where noises are correlated. We also demonstrate the effectiveness of the estimation procedure on several non-simulated datasets, also demonstrating that our estimation procedure... https://escholarship.org/uc/item/50c923qg Sat, 22 Aug 2026 00:00:00 +0000 Marzban, Ethan P https://escholarship.org/uc/item/4v1469cp This dissertation project centers on the persistent use of torture by liberal democratic states and the institutional and ideological logics that perpetuate it. While it is commonly assumed that torture perpetrated by liberal democracies represents a momentary, uncharacteristic lapse of morality amidst an otherwise unshakable commitment to the rule of law, this research seeks to illuminate consistent, deeply entrenched ties between liberal ideology, democratic statehood, and torture. This dissertation approaches the issue through an analysis of torture perpetrated in the twenty-first century by the United States and Israel. These two countries present themselves as exemplars of liberal morality, while also maintaining particularly extensive and entrenched practices of torture. The dialectic of these seemingly contradictory characteristics illuminates the extent to which liberal democratic governance accommodates and rationalizes torture. Close analysis of torture survivor testimonies,... https://escholarship.org/uc/item/4v1469cp Sat, 22 Aug 2026 00:00:00 +0000 Rosner, Em https://escholarship.org/uc/item/4rz33506 Advancements in deep learning optimization have sparked significant interest in the geometry of high-dimensional loss landscapes. To make these landscapes interpretable, researchers heavily rely on low-dimensional projection techniques, such as random directions, filter normalization, and trajectory-based PCA. However, these visualizations are typically validated through heuristic visual inspection. A key open question remains: can we define a principled, task-agnostic criterion to quantify how well a chosen projection captures the underlying loss landscape? This thesis introduces the disagreement metric, a theoretically grounded criterion that evaluates projection fidelity by measuring the unexplained fraction of projected gradient energy. Geometrically, this metric quantifies the divergence between the true landscape gradients and the gradients evaluated over the projected subspace. We demonstrate theoretically that in the local quadratic region near a minimum, minimizing... https://escholarship.org/uc/item/4rz33506 Sat, 22 Aug 2026 00:00:00 +0000 Zhang, Justin https://escholarship.org/uc/item/4qt2m15p This dissertation examines how environmental challenges and governance shape economic behavior, resource use, and inequality in developing countries. Across three chapters, I combine high-frequency panel data, geospatial analysis, satellite imagery, and quasi-experimental methods. The first chapter, estimates the short-run labor supply impacts of piped water shortages in Mexico City. Urban water shortages are an increasingly urgent challenge in large cities, driven by population growth, climate change, and aging infrastructure. In Mexico City—where households lack alternative sources—disruptions to piped supply force residents to wait at home for public deliveries or purchase water from private trucks and bottled suppliers. These shocks affect both time and income, creating theoretically ambiguous effects on labor supply. I combine high-frequency labor panel data with a novel proxy that interacts rainfall-driven variation in upstream reservoir storage with each neighborhood’s... https://escholarship.org/uc/item/4qt2m15p Sat, 22 Aug 2026 00:00:00 +0000 Rico Straffon, Sara Jimena https://escholarship.org/uc/item/4h04v7hv The historical study of Roman pregnancy is typically focused on either the High Empire (the first to third centuries CE) or the Byzantine period (the sixth century CE onwards). Moreover, these studies reify a very normative progression for pregnancy that goes uninterrupted from conception to childbirth. This dissertation explores the cultural history of pregnancy in the late antique period between the third and sixth centuries CE to trace the combined influences of Greek medical and early Christian thought on reproductive thought. I widen the scope of typical reproductive narratives to show that late antique authors and families also possessed a range of scripts for navigating infertility, abortion, and miscarriage. Furthermore, I trace the rising challenge to pregnancy and birth as the standard for women’s health. In comparing medical handbooks, legislation, homilies, theological letters and polemics, and hagiographies, I demonstrate that late antique Christian gynecological... https://escholarship.org/uc/item/4h04v7hv Sat, 22 Aug 2026 00:00:00 +0000 Nguyen, Misa https://escholarship.org/uc/item/4g91c9wm Graphs provide a natural abstraction for representing relationships and connectivity in real-world systems, including social networks, web pages, road networks, and recommendation systems. As a result, many modern applications rely on graph algorithms, making graph processing an active area of research across algorithms, software systems, and hardware architectures. However, graph algorithms are difficult to execute efficiently on modern processors, especially when processing large-scale input graphs. Much of their execution time is spent on traversing edges and accessing vertex data through pointer chasing and dependent memory accesses. These access patterns expose irregular memory-access behavior that stresses the cache and TLB hierarchies, leading to data-cache misses, TLB misses, and costly page table walks (PTWs). This thesis focuses on the address-translation behavior of graph workloads. We observe that PTWs are not uniformly distributed across the page footprint. Instead,... https://escholarship.org/uc/item/4g91c9wm Sat, 22 Aug 2026 00:00:00 +0000 He, Yuchen https://escholarship.org/uc/item/4fh2469p MKIDs are a class of superconducting detector with high sensitivity, straightforward scalability, and inherent energy resolution. The applications for Optical MKIDs include astronomy, dark matter detection, photonic quantum computing, and biological process imaging. To motivate greater use in these applications, improvements to MKID yield and uniformity at the array scale, detection efficiency, dark counts, and resolving power are required. This dissertation details a variety of detector designs towards these ends. First, the performance of several styles of phonon blocking architectures is discussed: superconducting bilayers, polymer suspended inductors, and air-suspended inductors. Next, the characterization of the dark count rate of a large-format MKID array is presented, with some investigation of the sources of this background. Finally, a large-format array is presented which made improvements to the yield, resonator quality factor, and quantum efficiency of the previous... https://escholarship.org/uc/item/4fh2469p Sat, 22 Aug 2026 00:00:00 +0000 Clay, Will Hawkins https://escholarship.org/uc/item/49x4p852 All microbes live in community, and have the potential to engage in both positive and negative interactions with each other. The complexity of naturally occurring, environmental communities, furthermore, makes the dissection of these interactions difficult and hard to interpret. Despite these interactions being difficult to detect, leveraging a multifaceted approach that employs microscale methods can aid in the decryption of these cryptic interactions. My interest in exploring microbial community interactions drove me to the pink berries of the Sippewissett salt marsh, microbial aggregates comprised of two main sulfur cycling symbionts: a purple sulfur bacteria (PB-PSB1) and a sulfate reducing bacteria (PB-SRB1). This naturally occurring community is a tractable, low-diversity system that provides a model microbial system to enable the study of bacterial metabolic interactions. As pink berry bacteria live in tightly associated, reproducible communities across space and time,... https://escholarship.org/uc/item/49x4p852 Sat, 22 Aug 2026 00:00:00 +0000 Castillo Villaseñor, Braulio Saúl https://escholarship.org/uc/item/49c0z1dh We will show that $SL(4,\mathbb{R})$ and $SU(3,1)$ contain strongly dense hyperbolic 3-manifold subgroups. Then we will establish the existence of Schottky groups in $PSL(2,\mathbb{C})$ where no pair of non-commuting elements have translation axes that lie on a common plane. https://escholarship.org/uc/item/49c0z1dh Sat, 22 Aug 2026 00:00:00 +0000 Lee, Ricky https://escholarship.org/uc/item/3zq0r1s3 This thesis examines how external dependence shaped growth and development in Nigeria. The argument is that the challenges of economic growth, structural transformation, and household welfare are connected by a common structural feature, which is the dependence on a narrow export base for foreign exchange together with persistent reliant on importation of production inputs. This combination repeatedly exposed the economy to external pressures and influenced macroeconomic performance, policy responses, and the available opportunities for households across different regions and sectors.This thesis analyzes the question from historical, household, and macroeconomic perspectives. First, it reviews the evolution of Nigeria’s development trajectory, focusing on the how oil dependence, foreign exchange availability, and structural transformation are connected. Second, it analyses trends in labor allocation, gendered participation in economic activities, and differences in household welfare... https://escholarship.org/uc/item/3zq0r1s3 Sat, 22 Aug 2026 00:00:00 +0000 Ogundipe, Michael Oluwasegun https://escholarship.org/uc/item/3rd3v3bp Human-induced landscape change and climate change are reshaping ecosystems globally, yet their interactive effects remain in need of further understanding. Arthropod vectors, particularly ticks and mosquitoes, depend on both biotic and abiotic environmental conditions throughout their life cycles, making them uniquely sensitive indicators of concurrent anthropogenic pressures and an informative lens for studying ecosystem responses to multiple interacting stressors. My dissertation investigates how the combined and interacting effects of climate variability and land-use alteration shape vector populations and vector-borne disease risk across three distinct systems. In my first study, a replicated exclosure experiment along a topoclimatic gradient revealed that both large-herbivore identity and climate mediate tick population dynamics. My second study examined mosquito density across alpine lakes in the Sierra Nevada, finding that seasonal ice-off timing and human recreational... https://escholarship.org/uc/item/3rd3v3bp Sat, 22 Aug 2026 00:00:00 +0000 Copeland, Stephanie Joyce https://escholarship.org/uc/item/3pm7s4ds The mechanical response of SiC/SiC ceramic matrix composites is governed by the properties and spatial arrangement of fibers, coatings, and matrix. Under idealized conditions, the single fiber composite model accurately predicts tensile behavior. In practice, however, real composites deviate from these idealizations in two important ways. First, incomplete matrix densification––particularly in composites fabricated by chemical vapor infiltration––produces discontinuous matrix regions and non-uniform load transfer. These features introduce extrinsic geometric effects that influence the measured composite response. Second, microstructural non-uniformities—including fiber-fiber contacts, coating discontinuities, and variations in fiber spacing—promote local load sharing. This leads to spatially correlated fiber fracture and reductions in composite strength. Together with the inherently brittle and stochastic nature of the constituents, these factors complicate the relationship between... https://escholarship.org/uc/item/3pm7s4ds Sat, 22 Aug 2026 00:00:00 +0000 Han, Nicholas https://escholarship.org/uc/item/3j73c106 The South American hydroclimate is driven by the transport of heat, moisture, and momentum from the deep tropics into the subtropics. This moisture advection is primarily facilitated by northerly low-level jets (LLJs) channeled along the eastern slopes of the Andes, which serve as the primary catalyst for the continent's most extreme weather, specifically, Mesoscale Convective Systems (MCSs). However, predicting the variability of this coupled LLJ-MCS relationship across historical and future climates remains a critical challenge due to complex, multiscale atmospheric interactions. This dissertation investigates the spatiotemporal variability of northerly LLJs and their multiscale controls on Mesoscale Convective Systems (MCSs), spanning from historical oceanic teleconnections to high-resolution projections of future climate change. Chapter 1 investigates the synoptic-scale variability of distinct LLJ branches (Central, Northern, and Andes) driven by coupled oceanic teleconnections.... https://escholarship.org/uc/item/3j73c106 Sat, 22 Aug 2026 00:00:00 +0000 Mu, Ye https://escholarship.org/uc/item/3gs8x5s6 Plastic trash, microplastics, and per- and polyfluoroalkyl substances (PFAS) are co-occurring, ubiquitous, and persistent urban pollutants with demonstrated adverse impacts on human and ecosystem health. Though their sources may differ, urban runoff is a major transport pathway for all three of these pollutants into surface waters. Given emerging regulatory and legislative responses at federal and state levels, accurately estimating these pollutants’ loads with measures of variance is critical to monitoring, determining risk, informing policy design, and ensuring compliance with remediation requirements. However, field campaigns to directly measure plastic trash, microplastics, and PFAS are time- and cost-intensive, resulting in data limitations. In the case of plastic trash, current single-pass surveying methods are known to underestimate environmental loads, and lack measures of effort and variance, preventing cross-study comparisons. Surface waters receiving plastic trash,... https://escholarship.org/uc/item/3gs8x5s6 Sat, 22 Aug 2026 00:00:00 +0000 Reynolds, Rebecca https://escholarship.org/uc/item/3cb055v1 The utilization of tectonically offset geomorphic markers to quantify differential motion between adjacent crustal blocks provides constraints on the kinematics of active fault systems and their associated hazards. These geological data are foundational inputs for probabilistic seismic hazard analyses (PSHAs), which inform regional building codes that are designed to reduce loss of life during earthquake events.The Santa Cruz Island Fault (SCIF) is an ~83 km-long segment of the larger ~200 km-long Western Transverse Range (WTR) fault system in Southern California. The SCIF’s proximity and structural connection to the Los Angeles basin have implications for earthquake hazard models that impact ~20 million people. Current PSHA’s for the entire fault rely on geologic data based on the westernmost ~3 km segment. This study analyzes ~13 km of the terrestrial segment of the SCIF using field mapping, 1-m resolution LiDAR, radiocarbon and luminescence dating. I measured 75 lateral and... https://escholarship.org/uc/item/3cb055v1 Sat, 22 Aug 2026 00:00:00 +0000 Love, Joshua S https://escholarship.org/uc/item/3c817361 This dissertation studies decision-making under uncertainty through the lenses of stochastic control, stochastic differential games, and mean-field games. The primary focus is on developing theoretical analyses, computational methods, and modeling applications for complex multi-agent and partially observed systems. The first part studies adversarial interactions in stochastic control. We develop a strategic misdirection model that integrates hypothesis testing statistics into control frameworks, yielding principled methods for optimally perturbing trajectories to deceive an adversary. This model is then extended to a Stackelberg game formulation for designing counter-deception strategies. Concurrently, we study likelihood-based stealthy attacks, leading to optimal attack designs that balance effectiveness and detectability under deterministic and adaptive information structures. Numerical experiments demonstrate the advantages of the proposed designs over heuristic alternatives... https://escholarship.org/uc/item/3c817361 Sat, 22 Aug 2026 00:00:00 +0000 Zhou, Haosheng https://escholarship.org/uc/item/3979b8z7 Recurrent neural networks (RNNs) are widely used both as practical architectures for machine learning tasks involving sequential data and as computational models that let neuroscientists study recurrent circuits in the brain, yet their training dynamics are only partially understood. A key reason is dimensionality: a small RNN can have thousands of parameters that interact nonlinearly through the recurrent dynamics, making the loss landscape difficult to analyse and the role of any single design choice difficult to isolate. This thesis takes a different approach. Rather than studying full-rank networks directly, we analyse a minimal model—the rank-1 RNN with connectivity W = mn⊤ trained on a one-bit flip-flop task. The training dynamics get reduced to a two-dimensional phase plane over effective connectivity parameters, allowing the network’s behaviour to be both derived mathematically and tested empirically.Using a model reduction we show that the rank-1 RNN is characterised... https://escholarship.org/uc/item/3979b8z7 Sat, 22 Aug 2026 00:00:00 +0000 Ou, Tzu-Ming https://escholarship.org/uc/item/3064r35g Large language models (LLMs) have demonstrated strong performance across a wide range of tasks, yet these gains come at the cost of substantial memory consumption and inference latency. This dissertation investigates methods for improving the training and inference efficiency of LLMs across three parts. The first part of this dissertation presents several novel methods for improving zeroth-order (ZO) optimization for LLM fine-tuning. To mitigate the high estimation variance inherent in ZO methods, we propose reducing the number of trainable parameters. Specifically, we introduce LoRETTA and AdaZeta, which achieve competitive performance compared to first-order and ZO fine-tuning methods while using an ultra-low parameter budget. In addition, we propose SharpZO, a method that significantly advances ZO fine-tuning for vision-language models by smoothing the loss landscape with a novel sharpness-aware strategy. The second part focuses on accelerating LLM inference through improved... https://escholarship.org/uc/item/3064r35g Sat, 22 Aug 2026 00:00:00 +0000 YANG, YIFAN https://escholarship.org/uc/item/2z9636sq With the acceptance of Black art songs and spirituals in the Western Classical canon, published repertoire has become more readily available. However, today, Black repertoire is still neglected. There is an underlying fear of singing and performing Black art songs and spirituals. As it pertains to vocal technique, the technique of Black art music is equivalent to that of the classical technique. So, where does the fear manifest? Many non-Black singers fear the Black repertoire because of the social histories that surround the music; the fear of making a caricature out of the music; the fear of offending. It is first important to know and understand the history behind and the style within the repertoire, just as all classical musicians are challenged to learn that of Western Classical music. When we research and study repertoire as performers, we gain deeper understanding. We learn more about the composers and their experiences, including the histories surrounding them and their... https://escholarship.org/uc/item/2z9636sq Sat, 22 Aug 2026 00:00:00 +0000 Johnson Jr, Lorenzo Dewayne https://escholarship.org/uc/item/2t99f1qw This dissertation investigates the macroeconomic and welfare implications of climate change and climate policies, with particular emphasis on firm dynamics and market frictions. The three chapters approach these questions using a combination of quantitative macroeconomic modeling, theoretical analysis, and econometric evidence from micro- and macro-level data. The first chapter studies optimal carbon taxation when capital is misallocated across firms. Using data on U.S. public firms, I document that firm-level revenue-based emissions intensity is negatively correlated with the return to capital, indicating that capital is misallocated toward dirty firms. Embedding this pattern into a misallocation framework, I demonstrate a reallocation mechanism: a carbon tax can reduce resource misallocation and boost aggregate productivity by reallocating capital from dirty to clean firms. I develop a dynamic, general-equilibrium climate-economy model with heterogeneous firms and tangibility-dependent... https://escholarship.org/uc/item/2t99f1qw Sat, 22 Aug 2026 00:00:00 +0000 Gao, Yang https://escholarship.org/uc/item/2rt5q6rx This dissertation examines how institutional design, resource allocation, and formal rules shape behavior in political and social systems. Across three chapters, I combine quasi-experimental empirical methods with original panel datasets to evaluate outcomes in local public finance, electoral competition, and professional sports.The first chapter investigates the asymmetric effects of intergovernmental transfers on local public finances and political dynamics in Paraguay. Exploiting a flawed census that generated plausibly exogenous variation in transfer allocations, I find that municipalities receiving unexpected increases sharply reduce their own-source tax effort, while those facing cuts expand local tax collection. Beyond this fiscal substitution, these exogenous transfer shocks distort local accountability, reducing voter turnout and shifting mayoral ambitions away from local reelection toward central government positions. The second chapter examines the electoral returns... https://escholarship.org/uc/item/2rt5q6rx Sat, 22 Aug 2026 00:00:00 +0000 Abbate Granada, Enzo Camilo https://escholarship.org/uc/item/2rs0d272 The social clock shapes our lives by prescribing the “right” ages for milestones like starting college, getting married, or having children. The present research examines cultural differences in the social clock’s rigidity and how intersubjective perceptions of the social clock implicate subjective perception and social judgment. In Chapter 1, I investigate how and why the rigidity of the social clock may vary across cultures. In three studies (two preregistered), participants from China and the U.S. were asked to provide the earliest and the latest ages they think appropriate for engaging in several life events. The social clock’s rigidity was operationalized as the width of the time windows for these life events. I found notable cultural differences: the social clock was more rigid in China than in the U.S., and filial piety beliefs are likely explanation for these differences. I further assessed anticipated negative self-conscious emotions of deviation from the social clocks.... https://escholarship.org/uc/item/2rs0d272 Sat, 22 Aug 2026 00:00:00 +0000 Zang, Lu https://escholarship.org/uc/item/2p38g7v7 This dissertation examines how social environments and early-life shocks shape cognitive, socioemotional, and behavioral outcomes in developing countries. Across three chapters, I combine experimental data, panel data, survey data, and quasi-experimental methods. The first chapter studies the causal impact of friendships on children's outcomes in elementary schools in Ecuador. While the literature on peer effects typically focuses on broad measures of peer composition, such as average ability or class characteristics, friendships represent a more direct and behaviorally relevant form of interaction. Furthermore, little is known about whether friendships help or hurt learning and well-being. Friendships can provide emotional support and foster motivation, but they may also distract from academic activities or reinforce negative behaviors. We use a unique randomized experiment in which students are randomly assigned to classrooms in every grade from Kindergarten to 6th grade,... https://escholarship.org/uc/item/2p38g7v7 Sat, 22 Aug 2026 00:00:00 +0000 Fuertes-Segura, Nicolas https://escholarship.org/uc/item/2jr1k9fd Graduate education is a critical period for socialization and professional identity development, particularly for students aspiring to become professors. However, our understanding of discipline-specific pathways into the professoriate and the development of professional identities within those pathways remains limited. Mathematics carries distinctive cultures, values, and discourses that shape how graduate students come to understand themselves as professionals. To better understand how disciplinary and institutional contexts influence this socialization process, I qualitatively examined the narratives of mathematics graduate students at a research-intensive university as they navigated their training and developed identities as prospective faculty. This three-paper dissertation explores an overarching question from multiple perspectives: What does it mean to become a mathematics professor? Across the papers, I draw on methodologies such as case study design and narrative inquiry,... https://escholarship.org/uc/item/2jr1k9fd Sat, 22 Aug 2026 00:00:00 +0000 Olarte, Timothy Royce https://escholarship.org/uc/item/2jb5d9mr Resolutions of singularities are a key idea in birational algebraic geometry. Resolutions are important as they allow one to find, for a singular scheme, a non-singular scheme in the same birational class. Resolutions have been studied extensively in classical algebraic geometry, but are not as well understood in the derived category setting. In this thesis we extend recent results for computing crepant categorical resolutions and their kernels in the A1 and A2 cases to the arbitrary An case. https://escholarship.org/uc/item/2jb5d9mr Sat, 22 Aug 2026 00:00:00 +0000 Pion, Joel Ethan https://escholarship.org/uc/item/2fk6046g At any given moment, we are presented with far more information in the environment than we can consciously perceive. To combat this limitation, the attentional system prioritizes specific stimuli at the expense of others; however, failures in prioritization can lead to increased distraction and missed goal-related opportunities. Existing research on this topic is dominated by spatial attention tasks, even though attention is also constrained in the temporal domain. My dissertation research is focused on better understanding the spatiotemporal dynamics of attentional priority by exploring how the prioritization process unfolds across features commonly favored in the selection process and as a function of task relevance. In Chapters 2 and 3, I use both behavioral and neural measures to reveal the rapid allocation and delayed disengagement of attention toward stimuli associated with previous experience, facilitating performance when task-relevant and delaying the consolidation of... https://escholarship.org/uc/item/2fk6046g Sat, 22 Aug 2026 00:00:00 +0000 Chak, Carly https://escholarship.org/uc/item/2cq383hw Zero-knowledge (ZK) proof systems let one party convince another that a computation was performed correctly while revealing nothing about its secret inputs, and they have rapidly become foundational infrastructure for privacy-preserving and blockchain applications. In nearly all such systems, the computation to be proven must first be expressed as an arithmetic circuit: a system of polynomial constraints over a large finite field. Writing these circuits by hand is both error-prone, and performance-critical, since the size and structure of a circuit directly determine the cost of proving and/or verification. Existing domain-specific languages ease the programming burden but address neither problem fully: they give developers no way to specify and verify what a circuit is meant to compute, and their compilers often leave substantial efficiency unexploited.This dissertation argues that both the correctness and the efficiency of ZK circuits can be addressed systematically through... https://escholarship.org/uc/item/2cq383hw Sat, 22 Aug 2026 00:00:00 +0000 Liu, Junrui https://escholarship.org/uc/item/2bp0f1s5 For many decades, Roberto Rossellini has been widely celebrated in studies of film history as the preeminent filmmaker responsible for inaugurating the influential genre known as Italian Neorealism. However, far fewer viewers and scholars are familiar with the fact that Rossellini began his filmic career as a director of Fascist wartime propaganda films, produced in close collaboration with Mussolini and the regime. This raises an important question: How is it that the director of a series of Fascist propaganda films could go on to develop a film genre– Neorealism– which has long been identified by many critics as intrinsically anti-Fascist? This study seeks to interrogate this apparent contradiction by probing the aesthetics and, more importantly, the ethics of these Fascist films before comparing them to Rossellini’s later contributions to the Neorealist canon. Ultimately, it finds that Rossellini’s early entanglement with Fascist cinema cannot be so easily overlooked, if only... https://escholarship.org/uc/item/2bp0f1s5 Sat, 22 Aug 2026 00:00:00 +0000 Shaw, Vernon B https://escholarship.org/uc/item/1vg6d8xj This dissertation consists of three chapters. The first chapter examines the health impacts of urban heat and its implications for adaptation. Cities tend to be warmer than their suburban counterparts, a phenomenon known as the heat island effect. This effect varies significantly over time and across different neighborhoods. Using high-resolution, satellite-derived surface temperature data and emergency medical service records from Northern California, this chapter employs fixed-effects Poisson regressions to analyze these variations. The findings show that neighborhood-level extreme heat has significant adverse health effects. However, urban heat—a highly localized and predominantly anthropogenic component of overall temperature—is less harmful than temperature shifts on a larger geographic scale. This is likely because people can better adapt to urban heat. These results highlight the importance of considering both overall temperature exposure and adaptation potential in urban... https://escholarship.org/uc/item/1vg6d8xj Sat, 22 Aug 2026 00:00:00 +0000 Bian, Luorao https://escholarship.org/uc/item/1hm962vw “Deeper learning” is a confusing term. Depending on who you ask, it could mean processes that happen in the brain, processes that happen in the classroom, or various learning outcomes. Despite a lack of consensus on what the term “deeper learning” means, scholars and teacher educators agree that deeper learning and teaching for deeper learning are both important for ensuring that students are able to apply their knowledge (NRC, 2018). This means we must also understand how to teach K12 teachers to teach their students for deeper learning in teacher education programs. This qualitative research study focuses on deeper learning in teacher education: focusing both on teacher educators and preservice math and science teachers. It is well understood that how teachers conceive of teaching influences their practice (Kang & Wallace, 2004; Butcher et al., 2022). We also know that teachers develop mental models for teaching (Greca & Moreira, 2000; Jones et al., 2011; Leinhardt... https://escholarship.org/uc/item/1hm962vw Sat, 22 Aug 2026 00:00:00 +0000 Bennett, Matthew David https://escholarship.org/uc/item/1d01p67m Interacting particle systems are complex systems whose many components interact and co-evolve over time, with individual interactions giving rise to a wide variety of collective behaviors such as clustering, alignment, swarming, and synchronization. Such systems are ubiquitous, from the orbits of planets and the flocking of birds and schooling of fish to the aggregation of cells, the consensus of opinions, and the synchronization of oscillators over networks. Understanding how local interaction rules produce these global, collective behaviors is one of the most fundamental problems across disciplines, and differential equation models have been central to studying it. Yet the governing laws are often unknown, the systems are high-dimensional, and the many dynamical couplings among components are difficult to disentangle, motivating data-driven approaches that learn these laws directly from trajectory data. To be scientifically useful, such a model must do more than fit the observed... https://escholarship.org/uc/item/1d01p67m Sat, 22 Aug 2026 00:00:00 +0000 Kulick, Charles https://escholarship.org/uc/item/1bq1w68p Procrastination is a pervasive self-regulation failure characterized by the voluntary delay of intended action despite anticipated negative consequences. Although prior research has identified multiple contributing factors—including emotion regulation, temporal discounting, and individual differences—less is known about the mechanisms that operate at the moment of task initiation and early engagement, as well as how repeated experiences during task engagement contribute to more enduring patterns of procrastination. The present dissertation addresses this gap by examining how task structure and reinforcement processes influence state procrastination and by testing whether repeated reinforcement of effort can increase engagement with effortful tasks over time.Across three studies, this dissertation investigates interventions targeting distinct phases of procrastination. Study 1 tested a brief micro-intervention based on the Temporal Decision Model (Zhang & Feng, 2020), designed... https://escholarship.org/uc/item/1bq1w68p Sat, 22 Aug 2026 00:00:00 +0000 Garg, Anusha https://escholarship.org/uc/item/16p212h7 DNA Methyltransferase 3A (DNMT3A) is an epigenetic enzyme known to drive gene silencing, genomic imprinting, and X-chromosome inactivation in a variety of organisms, including humans. By catalyzing the de novo methylation of DNA at CpG sites, DNMT3A establishes the methylation pattern imperative for transcriptional regulation and cellular differentiation during embryonic development. DNMT3A achieves this through interactions with other cellular partners (ncRNA, nucleosomes, p53, and EZH2) to precisely methylate the human genome. Dysregulation of DNMT3A, whether through overexpression, mutation, or loss of regulatory control, leads to aberrant hypermethylation and contributes to various forms of cancer development. For instance, DNMT3A mutations are found in 20-30% of all acute myeloid leukemia (AML) cases, making DNMT3A the most mutated gene in AML patients. Current FDA-approved mechanism-based inhibitors targeting DNMT3A (Azacitidine and Decitabine) are nucleoside analogs that... https://escholarship.org/uc/item/16p212h7 Sat, 22 Aug 2026 00:00:00 +0000 Schmelmer, Ben Manfred Felix https://escholarship.org/uc/item/14g941m4 Motivated by the desingularization of certain orbifolds and the construction of geometric quantization, we consider a vector bundle E n+k → Mn over a closed manifold M, and introduce a regularization of analytic torsion on E by replacing the Hodge Laplacian in the torsion zeta function with a Witten Laplacian. We also investigate the relationship of the primary and secondary invariants between the total space and the base manifold. By passing to the adiabatic limit, we show that the index of E coincides with the index of M. For analytic torsion, with the additional assumption that k is even and n is odd, we show that the Quillen metric associated with the analytic torsion of E is independent of the metric chosen, and equals the usual Quillen metric of M.  https://escholarship.org/uc/item/14g941m4 Sat, 22 Aug 2026 00:00:00 +0000 Liu, Debin https://escholarship.org/uc/item/13v8z3t1 Text has become a central interface in modern vision-language models. In visual generative models, text provides an intuitive mechanism for controlling image and video synthesis. In multimodal large language models, however, text can also become a shortcut that causes models to underutilize visual information. This dissertation studies these two roles of text: as a controller for visual generation, and as a crutch that reveals limitations in multimodal understanding. The first part of this dissertation investigates text-based controllability in visual generation. I study how natural-language instructions can guide image editing, image generation, and video synthesis across different forms of visual control. For image editing, I develop a multimodal large language model agent that translates user requests into executable editing operations in professional software. Since ground-truth edits are difficult to define, I further introduce a generalist reward model that evaluates... https://escholarship.org/uc/item/13v8z3t1 Sat, 22 Aug 2026 00:00:00 +0000 Wu, Qiucheng https://escholarship.org/uc/item/13q122pm When Brazil joined the BRICS bloc in 2009, the project was framed as a vehicle for South-South cooperation that could alter the terms of its integration into the global economy: new trade partners, development finance free from IMF conditionality, and instruments for reducing dependence on the U.S. dollar. This thesis asks whether BRICS delivered that transformation, inverting the question the existing literature has asked to examine what the bloc produced for one of its members whose structural position most needed to change. Combining historical-structural analysis, quantitative data, and interviews with officials inside the institutions examined, and drawing on dependency theory, International Financial Subordination, and the currency hierarchy framework, it argues that BRICS did not break Brazil's pattern of subordinated integration but modernized it. Across trade, finance, and money, Brazil gained new creditors, new buyers, and a seat at a new multilateral table, but not... https://escholarship.org/uc/item/13q122pm Sat, 22 Aug 2026 00:00:00 +0000 Silva Coelho, Larissa https://escholarship.org/uc/item/12q6c7tw The rapid scaling of artificial intelligence computation has exposed electrical interconnects as a critical bottleneck in modern data centers, driving the industry toward silicon photonics and high-speed wavelength division multiplexing links. As networking standards have evolved to demand >100G per lane, the architectural paradigm has shifted from massively parallel, moderate-speed microring arrays to fewer, high-baud-rate channels. Despite this aggressive scaling, existing literature remains heavily focused on isolated device-level demonstrations, lacking systematic parametric studies that connect fundamental microring geometry to system-level Bit Error Rate performance. To address this gap, this thesis presents a comprehensive, multi-physics numerical modeling framework for high-speed Silicon Microring Modulators (MRM) and Microring Resonators (MRR) on a 160 nm standard platform. Utilizing an end-to-end simulation flow (MODE, FDTD, HEAT, CHARGE, and INTERCONNECT), the fundamental... https://escholarship.org/uc/item/12q6c7tw Sat, 22 Aug 2026 00:00:00 +0000 Kim, Aaron https://escholarship.org/uc/item/1101j8ct Modern quantum platforms naturally realize many-body states shaped by measurement, decoherence, and dissipation. The appropriate object is often not a wave function or ground state, but a density matrix. This thesis asks how phases and phase transitions can be defined for such mixed states, especially when universal structure appears only through nonlinear, replicated, or doubled diagnostics. The thesis develops this perspective using nonlinear diagnostics, doubled-Hilbert- space representations, statistical-mechanics mappings, and generalized symmetry. It shows that weak measurement can suppress higher-form symmetry breaking and drive mixed-state transitions; maps fidelity diagnostics of decohered Chern insulators to effective conformal-field-theory partition functions; and maps Rényi diagnostics of noisy stabilizer and CSS codes to dual statistical-mechanics models connected to decodability. It also shows that measurement and Lindbladian evolution can prepare density matrices... https://escholarship.org/uc/item/1101j8ct Sat, 22 Aug 2026 00:00:00 +0000 Su, Kaixiang https://escholarship.org/uc/item/0wf0c4nf As transitional kindergarten (TK) expands, there is an increasing need to identify reliable measures of children’s school readiness and to understand how teachers support early learning across diverse classroom contexts. The present dissertation examined how teacher training relates to both teachers’ perceptions of children’s school readiness and the instructional practices implemented in TK classrooms. In Study 1, multilevel modeling was used to examine whether teachers’ training backgrounds were associated with their perceptions of children’s school readiness at TK entry. Participants included 50 TK teachers and 829 students from a school district in Central Coastal California. Teachers completed the Prekindergarten School Entry Profile (PSEP) during the first month of the school year. Results indicated that teacher training background was not a significant predictor of school readiness ratings. However, student gender was significant, with girls receiving higher readiness ratings... https://escholarship.org/uc/item/0wf0c4nf Sat, 22 Aug 2026 00:00:00 +0000 Perez, Emanuel Jose https://escholarship.org/uc/item/0v80p317 Microelectronics underpin virtually every aspect of modern technology, and continued advances in their performance and miniaturization depend critically on the ability to fabricate increasingly smaller features on semiconductor devices. The continued scaling of these devices has placed increasing demands on photoresist materials, exposing fundamental limitations in the chemically amplified resists (CARs) that have long underpinned semiconductor lithography. These issues primarily stem from chemical heterogeneity imparted by conventional polymer synthesis and formulation which, along with insufficient absorption of extreme ultraviolet (EUV) radiation, contributes to degraded patterning resolution and process reliability at the nanoscale. Sequence-defined polymers are an emerging class of materials that offers a possible solution to these problems. By precisely controlling chain length, monomer sequence, and sidechain identity, these materials are uniquely suited to isolate and... https://escholarship.org/uc/item/0v80p317 Sat, 22 Aug 2026 00:00:00 +0000 Adams, Cameron Patrick https://escholarship.org/uc/item/0td9v6s0 The demand for high-performance amplifiers in wireless communication systems continues to grow as radio-frequency front ends require improved sensitivity, gain, and noise performance. Low-noise amplifiers (LNAs) with high gain and low noise figure are therefore critical for enhancing overall receiver performance. To improve gain at the target operating frequency, an inductive-neutralization technique can be employed to mitigate the effect of parasitic feedback capacitance. This thesis presents the design and characterization of an inductively neutralized low-noise amplifier implemented in a 130nm complementary NPN–PNP SiGe BiCMOS process. The proposed architecture uses a complementary NPN–PNP amplifier with inductive neutralization to investigate the cancellation of the base–collector overlap capacitance, Cbc. The fabricated amplifier achieves an average measured small-signal gain of approximately 9dB at 15.5GHz while dissipating 9.12mW of DC power. A minimum noise figure of 4.92dB... https://escholarship.org/uc/item/0td9v6s0 Sat, 22 Aug 2026 00:00:00 +0000 Liu, Lian https://escholarship.org/uc/item/0rz3j4gk This dissertation asks how policy components influence perceptions of fairness and identifies the policy bundles that Americans are more likely to support. The research develops and validates a unique multidimensional measurement of fairness, advancing the study of public opinion methodologically and pushes the understanding of immigration attitudes beyond conventional debates. Chapter 1 outlines the theory and measurement approach and Chapter 2 uses an experimental design to identify the policy bundles Americans consider fair and support. Using an additional conjoint experiment, Chapter 3 tests whether the findings hold when explicit economic costs are included in the design. Using two original and nationally representative samples of more than 5,500 U.S. respondents from April 2025 and April 2026, I utilize pre-registered experiments to test how state-level immigrant integration policy features affect perceptions of fairness and policy support. I find that state policies... https://escholarship.org/uc/item/0rz3j4gk Sat, 22 Aug 2026 00:00:00 +0000 Trummel, Taylor https://escholarship.org/uc/item/0r1882jr Machine learning systems increasingly make decisions with long-term consequences for individuals and society. Designing trustworthy machine learning systems therefore requires going beyond static accuracy and one-shot fairness criteria. Decisions may alter future populations through feedback, individuals may improve and reapply, uncertainty may call for set-valued predictions rather than point estimates, and large-scale deployment may be constrained by the efficiency of modern generative models. This thesis explores these challenges through contributions spanning fairness under dynamics, counterfactual fairness under predictive uncertainty, and efficient inference serving for large language models. We first study fairness in dynamic decision-making settings such as lending, hiring, and admissions, where rejected individuals may invest effort to improve and reapply. We introduce Equal Improvability, a fairness notion requiring that rejected individuals from different groups... https://escholarship.org/uc/item/0r1882jr Sat, 22 Aug 2026 00:00:00 +0000 Guldogan, Ozgur https://escholarship.org/uc/item/0ph9508b Although women account for over half of the world’s population, there has been little work devoted to studying women’s health and cognition, especially from an endocrine perspective. This is particularly striking when changes in ovarian hormones (such as during menopause) are assumed to play a role in the risk of developing severe cognitive impairments, such as Alzheimer’s Disease. Navigation ability is one of the first cognitive abilities to deteriorate in early Alzheimer’s, women account for 2/3rds of all late-onset cases, and the risk for developing Alzheimer’s significantly increases after menopause. As such, it is important to understand how ovarian hormones are associated with the ability to navigate, as this can be a crucial marker for detecting early, more preventable stages of Alzheimer’s.Animal research has previously demonstrated positive effects of estradiol on navigation, especially on the ability to take novel, more efficient shortcuts (i.e., a place-based strategy)... https://escholarship.org/uc/item/0ph9508b Sat, 22 Aug 2026 00:00:00 +0000 Topete, Alexis https://escholarship.org/uc/item/0nx992bz Thermopower measurements have been used since the early 1980s to study quantum Hall systems because they probe aspects of transport that are difficult to access from electrical resistance alone. Diffusion thermopower is related, through the Mott formula, to the energy derivative of the conductivity at the Fermi energy. It is therefore sensitive to the electronic structure, disorder, and scattering mechanisms near the Fermi level. Under appropriate conditions, thermopower is proportional to entropy per carrier or quasiparticle, making it a useful probe of correlated and topological quantum Hall states.This dissertation investigates magneto-thermoelectric transport in the quantum Hall regime in cadmium arsenide (Cd3As2) thin films. Bulk Cd3As2 is a three-dimensional Dirac semimetal, while quantum confinement in thin films can open a gap and, over a certain thickness range, realize a two-dimensional topological insulator. The combination of high electronic mobility and low thermal... https://escholarship.org/uc/item/0nx992bz Sat, 22 Aug 2026 00:00:00 +0000 Huynh, William https://escholarship.org/uc/item/09h8w20s Ammonia-oxidizing organisms are key contributors to the nitrogen cycle, yet little is known about their genomic adaptations to saline or alkaline conditions. This study combines physiological characterization and comparative genomics of ammonia-oxidizing bacteria (AOB) with a metagenomics to better understand potential adaptations to salt and high pH environments. Halotolerant AOB possess numerous adaptations that help regulate the osmotic stress caused by varying salt and pH conditions including osmolyte biosynthesis pathways and an inventory of transporters that helps regulate the electrochemical gradient imposed. Genomic analysis of Nitrosomonas sp. ANs5, an extremely alkalitolerant species within the Nitrosomonas lineage, identified multiple genomic features predicted to facilitate osmotic and pH homeostasis, including biosynthesis of compatible solutes and the presence of a V-type ATPase, unique within the Nitrosomonas. Ammonia-oxidizing enrichment cultures were established... https://escholarship.org/uc/item/09h8w20s Sat, 22 Aug 2026 00:00:00 +0000 Valera, Jose Rafael https://escholarship.org/uc/item/06n0q66g Prevailing depictions of retirement conceptualize it as the end of a person’s working life. However, with increasing longevity, aging populations, and shifting social policies in industrialized nations, millions of older adults are working beyond the years typically associated with retirement. In formal organizational employment, a career model that positions retirement as the complete end of work remains prevalent, resulting in limited options for older workers. At the same time, new employment arrangements are emerging, spurred by digital technologies, including online labor platforms. Thus, a growing number of older workers are retiring from full-time organizational employment and continuing to work through platforms. On the surface, platforms promote autonomy, flexibility, and opportunity. However, they introduce new challenges through algorithmic management, limited worker protections, and a lack of direct interpersonal interaction. Therefore, this dissertation examines whether... https://escholarship.org/uc/item/06n0q66g Sat, 22 Aug 2026 00:00:00 +0000 Parker, Sienna Helena Sinense https://escholarship.org/uc/item/05n034w9 The Koopman operator, which describes a dynamical system via a linear representation that can be approximately learned from data, allows application of linear control techniques to nonlinear systems. One may consider these aspects of system representation and controller design separately, but successful application will ultimately require a fusion of both. This thesis studies contributions to both aspects. The predictive accuracy of a Koopman-based representation is heavily dependent on the chosen set of lifting functions, termed observables. Identifying an appropriate lifting is a challenging domain-specific problem. We focus our study on a class of systems that are "entity-based systems" that we introduce and formally define. The states of such systems may be subdivided into a variable number of components, which we refer to as "entities". In the Koopman operator framework, we introduce a type of observable that describes a notion of the density of these entities. We discuss... https://escholarship.org/uc/item/05n034w9 Sat, 22 Aug 2026 00:00:00 +0000 Blischke, Madeline