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 Fri, 15 May 2026 07:36:41 +0000 https://escholarship.org/uc/item/45t7f7xx This thesis focuses on the design, fabrication, and characterization of gold microelectrode array (MEA) devices for applications in electrochemical aptamer-based (EAB) sensing. EAB sensors leverage target-induced conformational changes in surface-bound aptamers to transduce molecular recognition phenomena into measurable electrochemical signals, enabling real-time, reversible, and reagentless detection of target molecules in complex biological environment. Such sensors are highly necessary for pharmacokinetic measurements, where continuous monitoring of target concentrations is required.Motivated by the need to extend EAB sensing beyond single-point measurements toward spatially resolved target monitoring, this work develops planar, multi-site gold microelectrode arrays fabricated using standard microfabrication techniques. The MEAs were designed to support dense packing of sensing sites while maintaining well-defined electrode geometry and compatibility with aptamer functionalization.... https://escholarship.org/uc/item/45t7f7xx Mon, 4 May 2026 00:00:00 +0000 Chatterjee, Oindrila https://escholarship.org/uc/item/0g4425k7 Understanding how humans learn, monitor, and adapt their behavior across multiple timescales is a central challenge at the intersection of neuroscience and reinforcement learning. Human behavior spans fine-grained motor corrections, structured action execution, and higher-level strategic decision-making under uncertainty. This thesis develops computational frameworks that connect reinforcement learning theory with behavioral and neural evidence across these scales.First, we investigate how humans monitor and correct actions during continuous motor learning. We propose that the reinforcement learning state–action value function (Q) provides a principled signal for value-based error monitoring during movement execution. Using a spatial–temporal navigation task with continuous feedback, we compute frame-level value trajectories and quantify their curvature to capture within-trial correction dynamics. Neural analyses reveal that these value-based signals correlate with activity in... https://escholarship.org/uc/item/0g4425k7 Mon, 4 May 2026 00:00:00 +0000 Yang, Dengxian https://escholarship.org/uc/item/9w50n1j6 The aim of this research is to expand approaches to Robert Schumann’s Gesänge der Frühe (Songs of Dawn), op. 133, through a performer-centered lens informed by concepts of lateness and disability. Schumann’s late works have long been marginalized due to assumptions that his mental illness affected his compositions from the last two years of his life. These assumptions have been challenged, particularly following the release of additional primary sources in the 1980s, prompting renewed scholarly and performative attention to his late music. This study aligns with that trend while emphasizing a performative perspective, examining Gesänge der Frühe, op. 133, not merely as a material object, but as a living work that continues to be reinterpreted. Focusing on musical features such as introspection, fragmentation, tonal ambiguity, and structural looseness, the research demonstrates how Schumann balances historical influence, literary engagement, and personal expression. Drawing... https://escholarship.org/uc/item/9w50n1j6 Fri, 1 May 2026 00:00:00 +0000 Nobumori, Chika https://escholarship.org/uc/item/9t5166r6 This dissertation examines how formations of stigmatized, rejected, or alternative knowledge arise in relation to established social authorities—elite and popular—and circulate through society. Ideas dismissed as “deviant”—hereafter referred to as heterodox—recur across cultures and historical periods, repeatedly forming communities organized around shared dissent against dominant institutions of knowledge and power. In the contemporary global media environment, such counter-discourses have become especially visible as they spread rapidly across digital networks—from QAnon, anti-vaccination activism, and Flat Earth cosmology to movements such as MAGA, antifa, Anonymous, and global “color revolutions.” Although these milieus often differ profoundly in political orientation, epistemology, and goals, they participate in shared, recognizable practices rooted in challenging established authority. More broadly, counter-discourses emerge across all domains of culture, including religion,... https://escholarship.org/uc/item/9t5166r6 Fri, 1 May 2026 00:00:00 +0000 Whitesides, Kevin https://escholarship.org/uc/item/9nx503b4 Multi-task learning (MTL) improves model generalization by leveraging shared information across related tasks. A central challenge is designing architectures that balance shared representations with task-specific flexibility while remaining scalable as the number of tasks grows. This dissertation proposes two families of deep multi-task learning models for time series forecasting, organized around LSTM-based and Transformer-based architectures.In the first part, we develop two LSTM-based frameworks. RM-LSTM (Regularized Multi-task LSTM) introduces cross-task regularization via soft parameter sharing, encouraging task-specific parameters to remain similar across related tasks. EnM-LSTM (Ensemble-based Multi-task LSTM) replaces explicit penalization with a self-attention-based similarity ensemble that dynamically borrows strength across tasks at prediction time. To handle large numbers of tasks, we further propose a task clustering strategy that reduces pairwise attention complexity... https://escholarship.org/uc/item/9nx503b4 Fri, 1 May 2026 00:00:00 +0000 Xiao, Lihao https://escholarship.org/uc/item/9mb7q3xd Large Vision-Language Models (LVLMs) demonstrate significant progress in multimodal understanding and reasoning, yet object hallucination remains a critical challenge. While existing research focuses on mitigating language priors or high-level statistical biases, they often overlook the internal complexities of the visual encoding process. We identify that visual statistical bias, arising from the inherent Bag-of-Patches behavior of Vision Encoders under weak structural supervision, acts as a contributing factor of object hallucinations. Under this bias, models prioritize local texture features within individual patches over holistic geometric structures. This tendency may induce spurious visual confidence and result in hallucinations. To address this, we introduce a training-free algorithm called Structure-Disrupted Contrastive Decoding (SDCD), which performs contrastive calibration of the output distribution by introducing a shuffled structure-disrupted view. By penalizing tokens... https://escholarship.org/uc/item/9mb7q3xd Fri, 1 May 2026 00:00:00 +0000 Xia, Yuxuan https://escholarship.org/uc/item/92m9c0wj In Chapter 1, we provide the necessary background on rare-event simulation, large-deviation theory, importance sampling, and stochastic calculus. We also review point processes and numerical methods for simulating stochastic processes, and introduce an efficient importance sampling approach for rare-event simulation based on the exponential change of measure (ECM). These tools form the theoretical and computational foundation for the development and analysis of variance-reduction techniques for rare-event estimation throughout the thesis. Finally, we introduce linear Hawkes processes and discuss their generalization to affine jump-diffusion models, which serve as the primary stochastic frameworks within which the proposed rare-event simulation and importance sampling methodologies are applied and evaluated.In Chapter 2, we develop exact importance sampling schemes for (linear) Hawkes processes to efficiently compute their tail probabilities. We show that the classical approach... https://escholarship.org/uc/item/92m9c0wj Fri, 1 May 2026 00:00:00 +0000 Feng, Liang https://escholarship.org/uc/item/8z67t530 Members of society routinely manage trouble in speaking while conversing with one another; occasionally, they make multiple tries in succession at resolving the trouble(s)-at-hand in the course of constructing turns at talk. In this thesis, I use conversation-analytic methods to examine a collection of fragments of naturally-occurring interactions in which speakers are observed to successively adjust their verbal production of (part of) an ongoing turn. I describe a range of practices that speakers may employ in making such successive tries at delivering turn elements and an array of next actions that recipients of an ongoing try-series may produce in different placements. On the basis of these speakers’ turn-constructional practices and recipients’ responsive action designs, I propose that members orient to three tries as a normative maximum for constructing elements within the same turn and managing the trouble(s) that arises in the process. Lastly, I conclude by considering... https://escholarship.org/uc/item/8z67t530 Fri, 1 May 2026 00:00:00 +0000 Zheng, Marat https://escholarship.org/uc/item/8kv1k2pc GPGPU has been widely adopted by developers and super scalers for training and deployment of AI models. Memory and cache subsystems in GPGPU face unique challenges due to the nature of the SIMT architecture and workloads that are increasingly bounded by memory bandwidth. Existing open source GPGPU projects often use a simplified cache system and/or implementation that is not representative of real-world products and use cases. We integrate a high performance, out-of-order data cache into the L1 cache system of a RISC-V GPGPU, and observe up to 7.28% average IPC improvement on high performance computing kernels. We demonstrate an open-source high performance cache design can achieve competitive performance in a variety of processor architectures under the ecosystem of RISC-V, while being optimized for power and area. https://escholarship.org/uc/item/8kv1k2pc Fri, 1 May 2026 00:00:00 +0000 Zhou, Shaoqian https://escholarship.org/uc/item/8f35q20h Functional verification through simulation is one of the most time-consuming phases of the digital design cycle. Graphics processing units offer a natural avenue for acceleration, since nodes within a single topological level of a synthesized netlist can be evaluated independently and in parallel. Prior GPU-based approaches operate at the granularity of individual logic gates, inheriting the full complexity of the original netlist. This work presents an alternative: we first subject the input design to logical synthesis using Yosys, which restructures the logic into a smaller number of multi-input lookup tables (LUTs) of up to eight inputs, and then simulate the simplified netlist on the GPU. Across a suite of 13 benchmark designs ranging from a 2-gate half adder to a 14,000-gate FIR filter, synthesis reduces the number of logical components by 66–98% for non-trivial designs. When simulating for 10,000 cycles on an NVIDIA RTX A4000, LUT-level simulation achieves GPU kernel speedups... https://escholarship.org/uc/item/8f35q20h Fri, 1 May 2026 00:00:00 +0000 Dhulipala, Krishna https://escholarship.org/uc/item/86d216fd Modeling how radio signals propagate through cities is a prerequisite for nearly everything in modern wireless network design — coverage planning, beamforming, and the deployment of next-generation systems all depend on it. Physics-based ray tracing can do this accurately, but at significant computational cost, particularly in large outdoor environments where every reflection, diffraction, and scattering interaction has to be resolved across detailed 3D geometry. Learning-based methods offer a way to sidestep this bottleneck: train on ray tracing data once, then predict propagation behavior at a fraction of the cost. This thesis examines two such methods — Geo2SigMap and RF-3DGS — within a common outdoor simulation environment built on BostonTwin, a city-scale digital twin of Boston, and the Sionna ray-tracing framework. Ray-tracing simulations on this environment serve as ground truth for both training and evaluation. The two methods represent distinct modeling philosophies.... https://escholarship.org/uc/item/86d216fd Fri, 1 May 2026 00:00:00 +0000 Ozarslan, Emir Saltuk https://escholarship.org/uc/item/8568h090 This dissertation consists of three essays. The first chapter studies how the spread of transboundary biological threats stimulates innovation. Developed countries often possess the capacity to innovate technologies that mitigate global biological threats, such as vaccines for infectious diseases, but innovate little when not directly exposed. This chapter develops a spatial dynamic game to study endogenous innovation incentives as biological threats expand into developed countries. In the model, all countries can control threats locally, while only a few can innovate. I decompose how two externalities—threat diffusion and technology spillovers—and their interaction shape strategic innovation incentives. Using evidence of dengue-transmitting mosquitoes expanding into the U.S., I estimate that endogenous U.S. vaccine innovation could reduce dengue cases in the Americas by 54% relative to a no-innovation scenario and assess the resulting global welfare implications. Finally, I show... https://escholarship.org/uc/item/8568h090 Fri, 1 May 2026 00:00:00 +0000 Tsao, Shu-Chen https://escholarship.org/uc/item/7w58h3tv As artificial intelligence (AI) systems are increasingly embedded in organizational decision-making, understanding how power dynamics shape human–AI collaboration has become a critical theoretical and practical concern. This dissertation extends Dyadic Power Theory (DPT) to human–AI contexts, testing whether power dynamics observed in human relationships operate similarly when one partner is artificial. Using a 3 × 2 experimental design (N = 202), participants collaborated with AI chatbots on a travel-planning task while power positioning (high/equal/low) and anthropomorphism (high/low) were manipulated. The study hypothesized that equal-power conditions would yield the highest satisfaction and control attempts, with anthropomorphism and cognitive load moderating these effects.Although none of the hypotheses were supported as predicted, the findings reveal theoretically meaningful extensions of DPT. Satisfaction followed the asymmetrical L-shaped “hook” pattern observed in recent... https://escholarship.org/uc/item/7w58h3tv Fri, 1 May 2026 00:00:00 +0000 Wilkenfeld, Jennifer Nan https://escholarship.org/uc/item/7sc312xw Factors driving fire outcomes change across scale and context, making it difficult to identify particularly hazardous plants (community scale) and conditions (landscape scale). This work traces the evolution of fire hazard across scales, focusing on the regional context of Santa Barbara, California. At the plant scale, laboratory combustion was used to assess differences in wild-land urban interface plant species’ flammability and Mixed Effects modeling was used to draw connections between plant traits and components of plant scale flammability (ignitibility, combustibility, sustainability) (Prior et. al. 2018). Live fuel moisture was found to be the strongest driver of ignitibility while sample mass was found to be the strongest driver of combustibility and sustainability. At the landscape scale, Generalized Linear Modeling was used to investigate how weather, human factors, terrain, and fuels affected the probability that an ignition would become a wildfire. The interaction... https://escholarship.org/uc/item/7sc312xw Fri, 1 May 2026 00:00:00 +0000 Fauss, Kristina https://escholarship.org/uc/item/7rf9j3m8 Visual scene understanding is a central problem in computer vision, underpinning applications ranging from autonomous navigation to human-computer interaction. While prior Scene Graph Generation (SGG) methods have shown promise on static images, they remain fundamentally limited when applied to dynamic, real-world video: they generalize poorly to unseen object–relationship compositions, fail to enforce temporal consistency, and rely on coarse representations that preclude user interaction. This dissertation advances the state of Dynamic Scene Graph Generation (DSG) by addressing these limitations along three complementary dimensions: compositional generalization, temporal coherence, and interactive, fine-grained scene understanding.First, to mitigate dataset bias and improve generalization, we introduce the Decoupled Dynamic Scene-Graph (DDS) network, which explicitly separates object and relationship representation learning. This architectural decoupling promotes disentangled... https://escholarship.org/uc/item/7rf9j3m8 Fri, 1 May 2026 00:00:00 +0000 Ruschel dos Santos, Raphael https://escholarship.org/uc/item/7gx040ms It is becoming increasingly evident that one of the ways that cells interpret and encode information into multiple cell fates is by multiplexing information through pathway dynamics. Recently, WNT signaling was shown to propagate through the 2D gastruloid in a wave, challenging the traditional stable signaling gradient model. Here we identify the features of WNT signaling dynamics that are essential for cell fate decisions and patterning of the gastruloid. We combine 2D-micropatterning, cellular optogenetic control and live cell reporters to dissect the essential signaling requirements for germ layer patterning. By eliminating endogenous WNT signaling and re-introducing it with light, we define the timing and duration of signaling that determines fate and germ layer positioning in the gastruloid. Employing automated high throughput delivery of light patterns with our “LITOS” device, we determine the minimum amount of WNT necessary to recover mesoderm patterning in the gastruloid... https://escholarship.org/uc/item/7gx040ms Fri, 1 May 2026 00:00:00 +0000 Baxter, Naomi https://escholarship.org/uc/item/6hs88228 Although misinformation is not a new phenomenon, misinformation spreads at unprecedented scale and speed in the digital age due to the boundless and decentralized nature of the internet. As human fact-checkers struggle to keep pace with the volume of misinformation online, automated fact-checking powered by Artificial Intelligence (AI) has emerged as a promising tool. Yet, little is known about how audiences evaluate the credibility of AI-based versus human-based fact-checking and to what extent they differ in the effectiveness in reducing misinformation. Meanwhile, despite the implicit assumption that transparency can increase credibility perceptions and improve corrective effectiveness in fact-checking, little research has been conducted to test this assumption empirically.Therefore, this dissertation investigated how fact-check source type, machine heuristics (both positive and negative), misinformation topic, and transparency of fact-checking methods jointly influence credibility... https://escholarship.org/uc/item/6hs88228 Fri, 1 May 2026 00:00:00 +0000 Qi, Li https://escholarship.org/uc/item/66t1z0r5 Chern insulators are quantum states characterized by broken time-reversal symmetry and topologically nontrivial electronic bands that support dissipationless edge transport without an external magnetic field. Twisted transition metal dichalcogenide (TMD) moiré superlattices provide a highly tunable platform for realizing interaction-driven integer and fractional Chern states in narrow topological bands. Although transport measurements provide sharp global signatures of these phases, local magnetic imaging is required to assess the role of structural disorder and to directly measure the thermodynamic topological gap.In this thesis, I describe the development of nanoSQUID-on-tip (nSOT) magnetometry into a robust and quantitative probe of correlated topological states. An improved fabrication protocol significantly enhances device yield, reproducibility, and magnetic field sensitivity, enabling stable high-resolution magnetic imaging in scanning configurations.Using this platform,... https://escholarship.org/uc/item/66t1z0r5 Fri, 1 May 2026 00:00:00 +0000 Redekop, Evgeny https://escholarship.org/uc/item/65f7q0th The integration of deep learning into physical modeling faces significant hurdles when applied to non-Euclidean domains, such as curved surfaces and manifolds. Standard neural architectures, typically predicated on regular grids or flat Euclidean spaces, often struggle to capture intrinsic geometric properties and handle the irregularity of unstructured data. To address these limitations, this dissertation introduces Geometric Neural Operators (GNPs), a novel deep learning framework designed to learn operators on manifolds directly from scattered, noisy point cloud data. The contributions of this work are threefold. First, we establish the capabilities of GNPs by demonstrating their effectiveness in estimating fundamental geometric quantities, such as metric tensors and curvatures, and in learning solution maps for geometric Partial Differential Equations (PDEs). We showcase their use as efficient surrogate models in a Bayesian inverse problem for shape identification.... https://escholarship.org/uc/item/65f7q0th Fri, 1 May 2026 00:00:00 +0000 Quackenbush, Blaine Raymond https://escholarship.org/uc/item/60m9h8hb Program analysis research has produced a rich set of techniques for discovering, exercising, and demonstrating software bugs. These methods, however, often struggle to scale to the modern, complex, and stateful applications that underpin critical infrastructure. In this thesis, we demonstrate how academic program analysis techniques can be tailored and strategically applied to improve the analysis of real-world targets.First, our research on symbolic execution is typically restricted to small or highly constrained programs, limiting its applicability to large, stateful kernel components. Specifically, first we demonstrate how symbolic execution, which traditionally is plagued by path-explosion issues, can nonetheless be applied to Windows Kernel Drivers to effectively discover vulnerabilities at scale across thousands of drivers.Secondly, we present our research on concolic execution, which has been traditionally regarded as useful in hybrid fuzzing alongside a greybox-mutational... https://escholarship.org/uc/item/60m9h8hb Fri, 1 May 2026 00:00:00 +0000 Dresel, Lukas Patrick https://escholarship.org/uc/item/5sn7z5sd Phagocytic cells play a critical role in both the immune response to pathogens, killing and eliminating unhealthy cells, and the repair and revascularization of damaged tissues following infection or sterile physical injury. Phagocytes are also responsible for maintaining tissue homeostasis, clearing dead cells at the end of their normal lifespan, and stimulating stem cells to replace them during normal turnover. Underlying these two opposing roles is the remarkable ability of phagocytes to detect the health of a tissue and reversibly switch between a pro-inflammatory (called M1) and pro-regenerative (M2 phenotype). Defects in polarization are associated with a number of human diseases: the inability to switch between pro-inflammatory M1 to pro-regenerative M2 following infection can lead to chronic inflammation and autoimmunity, and the ability of a tumor to switch M1 cells to an M2 state promote immune suppression and tumor cell proliferation. Despite the clinical importance... https://escholarship.org/uc/item/5sn7z5sd Fri, 1 May 2026 00:00:00 +0000 Singh, Shambhavi https://escholarship.org/uc/item/58g3d4zn Line-of-sight (LoS) multiple-input multiple-output (MIMO) communication at millimeter-wave and terahertz frequencies is a promising approach for achieving extremely high data rates by combining large bandwidths with spatial multiplexing. In principle, such systems can approach data rates competitive with optical links while using compact antenna form factors. In practice, however, conventional LoS MIMO designs are fragile: they rely on precise transceiver alignment, antenna geometries carefully matched to a specific link range, and bulky hardware to sustain spatial degrees of freedom. These requirements significantly limit flexibility and hinder large-scale deployment, particularly in urban environments.Our objective is to enhance the flexibility and efficiency of LoS MIMO communication systems by developing architectures that remain reliable under realistic deployment constraints. We focus on two complementary operating regimes. In settings where compact form factors and opportunistic... https://escholarship.org/uc/item/58g3d4zn Fri, 1 May 2026 00:00:00 +0000 Giridhar, Lalitha https://escholarship.org/uc/item/5409j7j4 Support for outdoor learning is gaining momentum worldwide for its social-emotional, physical, and cognitive benefits. In parallel, green schoolyard initiatives that increase tree canopy and naturalize school landscapes are becoming widespread across the globe. Given the potential benefits of outdoor learning and green schoolyards, it would be expected to see increased utilization of green schoolyards for outdoor learning. Yet, this is not a widespread practice at K-12 public schools in California. This case study of a K-8 school district situated on the Central Coast of California investigated a successful model of outdoor learning in schoolyards.A conceptual framework informed by systems theory of organizations was used to examine how a school district successfully provided outdoor learning in the schoolyard from a holistic perspective. Through interviews, observations, publicly available documents, and student work, this study investigated the factors that led to the successful... https://escholarship.org/uc/item/5409j7j4 Fri, 1 May 2026 00:00:00 +0000 Azzam, Devon https://escholarship.org/uc/item/52x8z92p In Chapter 2 a robust method is described to synthesize degradable copolymers under aqueous miniemulsion conditions using α-lipoic acid as a cheap and scalable building block. Simple formulations of α-lipoic acid (up to 10 mol%), n-butyl acrylate, surfactant, and co-stabilizer generate stable micelles in water with particle sizes &lt;200 nm. The ready availability of these starting materials facilitated performing polymerization reactions at large scales (4 L), yielding 600 g of poly(n-butyl acrylate–stat–α-lipoic acid) latexes that degrade under reducing conditions (250 kg mol–1 → 8 kg mol–1). Substitution of α-lipoic acid with ethyl lipoate further improves the solubility of dithiolane in n-butyl acrylate, resulting in copolymers that degrade to even lower molecular weights after polymerization and reduction. In summary, this simple and scalable strategy provides access to large quantities of degradable copolymers and particles using cheap and commercially available starting... https://escholarship.org/uc/item/52x8z92p Fri, 1 May 2026 00:00:00 +0000 Morris, Parker Triplett https://escholarship.org/uc/item/4xz3171b Ferruccio Busoni’s piano transcriptions occupy an unusual position in the piano repertoire. Some of them, most notably his version of Bach’s Chaconne, are among the most frequently performed works associated with his name. In contrast, many others remain rarely played and often treated as secondary to his so-called “original” compositions. This divide reflects a broader uncertainty about how Busoni’s use of pre-existing music should be understood and, consequently, how these works ought to be interpreted in performance.This document approaches Busoni’s transcriptions from the perspective of the pianist. Rather than treating them simply as arrangements of earlier works, it examines them as products of a distinctive compositional and aesthetic vision. Drawing on Busoni’s own theoretical writings and editorial comments, the study traces three different ways in which he transforms existing music: works that aim to recreate the sound world of another instrument on the piano, works... https://escholarship.org/uc/item/4xz3171b Fri, 1 May 2026 00:00:00 +0000 Pascucci, Alvise https://escholarship.org/uc/item/4rs6124v Trichodesmium are a notoriously enigmatic and critically important diazotrophic cyanobacteria that inhabit the tropical and subtropical surface oligotrophic ocean. In addition to their unpredictable blooming patterns in the ocean, they are unpredictable in laboratory cultures as well. Recently, a few labs have been able to bring more environmentally relevant species and strains into culture which has made the research presented here possible. In this study we measured Trichodesmium lipids to understand how they vary with species and morphology. Lipids are important molecules and are used for a variety of processes including energy storage, membrane structure, and chemical signaling. There are very few studies on the lipid composition of Trichodesmium and the primary study on their lipids, in 1967, found unusual results – that a medium chain fatty acid dominated the lipid pool. Following up on that study, we analyzed the lipids of Trichodesmium blooms in the Gulf of Mexico... https://escholarship.org/uc/item/4rs6124v Fri, 1 May 2026 00:00:00 +0000 Gosselin, Kelsey https://escholarship.org/uc/item/4p06v9qj This study advocates for the inclusion of isotopic analyses in standard tissue sampling protocols and emphasizes the importance of retaining tissue samples, particularly in cases where law enforcement anticipates challenges in identification. While isotopic studies have long been established in geochemistry, their integration into forensic anthropology is relatively recent. In the discipline, isotopic analyses are still sparingly employed and are typically reserved for the identification of decedents in unresolved or inactive cases. Many isotopic reference values are readily accessible through online databases, and can provide insights into the geographic origin, dietary habits, migration patterns, and environmental exposures of unidentified individuals. This dissertation explores the historical development of forensic anthropology, provides a comprehensive overview of isotopic fundamentals and fractionation, and evaluates the role of radiocarbon dating for chronological age estimation... https://escholarship.org/uc/item/4p06v9qj Fri, 1 May 2026 00:00:00 +0000 Lucero, Brittany-Nicole Margaret https://escholarship.org/uc/item/4hv7h8sf Scholars and commentators have felt uneasy about describing the spectrum of Japanese politics as being Left Wing versus Right Wing, Liberal versus Conservative, or Hawk versus Dove. This thesis proposes a heuristic, Dogmatism versus Pragmatism, as an alternate description of variation in Japanese political behavior. It briefly assesses the applicability of standard spectra to the political history and present-day politics of Japan. It examines an indigenous variation of standard spectra used to frame the pre-1945 Japanese political struggles. It offers a new spectrum of dogmatism versus pragmatism as a better heuristic for Japanese political variety. To counteract criticism that pragmatism, both the word and philosophical method, are both distinctly American and thus inappropriate for Japan, this thesis presents evidence that pragmatism, both indigenous and introduced, is firmly rooted in in Japanese political behavior. https://escholarship.org/uc/item/4hv7h8sf Fri, 1 May 2026 00:00:00 +0000 Cucek, Michael Thomas https://escholarship.org/uc/item/454541dn The growing adoption of 48-V distribution architectures in data centers and high-performance computing systems has intensified the need for compact and efficient intermediate-bus converters capable of delivering regulated low-voltage outputs. Conventional inductor-based converters suffer from limited power density due to magnetic component size, while switched-capacitor converters offer improved integration but are constrained by fixed conversion ratios and limited regulation capability.This thesis presents a reconfigurable hybrid inductor-assisted Dickson DC–DC converter targeting conversion from a nominal 48-V bus to a regulated 12-V output. The proposed architecture fundamentally relocates the inductor current path from the output side to the input side of the converter, significantly reducing magnetic-component current stress. For a given output power, the inductor current is reduced by up to 80%, corresponding to an approximately 96% reduction in copper loss for a fixed ESR,... https://escholarship.org/uc/item/454541dn Fri, 1 May 2026 00:00:00 +0000 Fan, Shibo https://escholarship.org/uc/item/44q908wb The III-Nitrides remain a rich material family for study and new applications. Emerging ultra-wide bandgap (UWBG) semiconductors, such as high Al-content AlGaN, are attractive due to their large direct bandgaps, resulting high critical electric field, and large saturation velocity for electrons. While this materials system has been the subject of intense research and commercial success for several decades, there are many challenges still associated with making practical devices from UWBG III-nitrides, including lattice defects, low mobilities and conductivities, and difficulty forming contacts. In this work, I seek to improve the understanding of these materials when grown by ammonia molecular beam epitaxy (NH3-MBE), first by establishing suitable homoepitaxial AlN buffer layers, followed by characterizing and later avoiding plastic relaxation in graded alloys, and then by integrating these materials into polarization-doped field effect transistors for next-generation RF power... https://escholarship.org/uc/item/44q908wb Fri, 1 May 2026 00:00:00 +0000 Wissel-Garcia, Ashley Elizabeth https://escholarship.org/uc/item/3c63z5dk Feedback control determines how systems respond to measured states to achieve desired objectives. The linear quadratic regulator (LQR), developed in the 1960s, became one of the most widely applied control methods due to its elegant solution. The robust control extension for bounded disturbances, the disturbance attenuation regulator (DAR), formulated by American and Soviet researchers in the 1960s, has remained unsolved for six decades.This dissertation provides a complete state feedback solution to the DAR for linear systems, addressing signal bound disturbances where cumulative magnitude is constrained over the horizon, and stage bound disturbances where magnitudes are independently constrained at each time step. We also provide a complete feedback solution to the constrained LQR with signal and stage bound control constraints. The optimal controls to DAR and constrained LQR are nonlinear in the state and require solving a tractable convex optimization; the control is then... https://escholarship.org/uc/item/3c63z5dk Fri, 1 May 2026 00:00:00 +0000 Mannini, Davide https://escholarship.org/uc/item/389958wj Advances in satellite data collection and imagery have made large amounts of data available for inferring and predicting Earth systems. To use this data effectively, statistical methods that address measurement bias, account for differences between models and reality, and provide reliable long-term predictions are needed. In this thesis, I present new Gaussian process (GP) methods to address these challenges in two main applications: calibrating geophysical models and forecasting long-term environmental changes. Chapter 2 introduces a two-phase machine learning approach for long-term forecasting of vegetation indices from satellite data. First, we construct the generalized parallel partial Gaussian process (G-PPGP) to model the relationship between climate variables and vegetation. Next, we predict key climate variables one year in advance and use these forecasts as inputs to the G-PPGP. When applied to the Four Corners region in the Southwestern United States, this method... https://escholarship.org/uc/item/389958wj Fri, 1 May 2026 00:00:00 +0000 McPhillips, Erika Lois https://escholarship.org/uc/item/2zs929j9 Geometric inequalities play a central role in understanding the interaction between curvature, topology, and analysis on Riemannian manifolds. Classical results in comparison geometry show that pointwise curvature bounds, such as non-negative Ricci curvature, lead to sharp inequalities including Sobolev, logarithmic Sobolev, and isoperimetric inequalities. In this dissertation, we investigate how such inequalities remain valid under weaker curvature assumptions.The first part of this work establishes logarithmic Sobolev inequalities for submanifolds in complete Riemannian manifolds with asymptotically non-negative intermediate Ricci curvature. By combining refined Jacobi field estimates, Riccati-type differential inequalities, and an ABP method adapted to submanifolds, we obtain inequalities that capture the geometry at infinity and extend known results under non-negative Ricci curvature.The second part focuses on Willmore-type inequalities for hypersurfaces in complete non-compact... https://escholarship.org/uc/item/2zs929j9 Fri, 1 May 2026 00:00:00 +0000 Lee, Jihye https://escholarship.org/uc/item/2sp954gp This study examines how traditional Chinese musical elements are integrated into contemporary Western cello repertoire, focusing on Bright Sheng’s Seven Tunes Heard in China and Tan Dun’s Crouching Tiger, Hidden Dragon Cello Concerto. The purpose is to understand how cultural musical languages, especially those rooted in folk tunes, oral traditions, modal systems, and distinct timbral ideals, can be translated onto a Western classical instrument. For cellists trained in Western conservatory traditions, these works present technical and interpretative challenges, such as performance practice, stylistic understanding, and the expanding role of the cello in culturally blended compositions. The method of the study combines close score analysis. Selected movements from Sheng’s suite and the first movement of Dun’s concerto were examined for their use of modal structures, timbral imitation, extended techniques, and culturally informed phrasing. The two scores were also compared with... https://escholarship.org/uc/item/2sp954gp Fri, 1 May 2026 00:00:00 +0000 Guo, Qiele https://escholarship.org/uc/item/2sg6k1ff Visual autoregressive models (VAR) using next-scale prediction represent a scalable frontier for visual content generation. However, their practical deployment is severely hindered by a critical bottleneck: the massive memory footprint of the quadratically growing Key-Value (KV) cache, which can rapidly expand to exceed the size of the model weights themselves. This presents a unique optimization challenge and impedes deployment on memory-constrained devices. To address this bottleneck, we introduce VAR-Q, the first quantization framework specifically designed for the KV cache in visual autoregressive models. VAR-Q pushes the state of the art, enabling aggressive 3-bit quantization on large-scale Infinity-8B model without any degradation in generation quality. The code is publicly available at https://github.com/lujiaji/VAR-Q. 基于下一尺度预测（next-scale prediction）的视觉自回归模型（Visual Autoregressive Models, VAR）代表了视觉内容生成领域中一种具有良好扩展潜力的发展方向。然而，这类模型在实际部署中面临一个关键瓶颈，即随序列长度呈平方级增长的键值缓存（Key-Value... https://escholarship.org/uc/item/2sg6k1ff Fri, 1 May 2026 00:00:00 +0000 Lu, Jiaji https://escholarship.org/uc/item/2pb1606h This thesis explores the pervasive issue of religious othering in Pakistan, focusing particularly on religious minorities in Punjab. It delves into how religious identity and politics intersect, leading to systemic discrimination, violence, and marginalization. The study aims to elucidate the historical, social, and political contexts that have shaped the experiences of religious minorities, particularly Christians and Ahmadis, in this region. The research is grounded in a thorough analysis of Pakistan’s post-colonial history, emphasizing the role of successive governments and political actors in fostering an environment of religious exclusion. The project argues that the state's policies and actions have often sanctioned, and even encouraged, violence and discrimination against religious minorities. This has been further compounded by societal attitudes and the instrumental use of religion in politics. Through a combination of historical analysis, case studies, and theoretical... https://escholarship.org/uc/item/2pb1606h Fri, 1 May 2026 00:00:00 +0000 Nadeem, Faryal https://escholarship.org/uc/item/2hp5b8sx Intonation accuracy and stability is commonly regarded as one of the most fundamentally challenging aspects of playing a bowed string instrument. While many string players are able to develop some sensitivity to intonation and pitch relations through intuition and general experience alone, their intuition is not always reliably supported by an understanding of why certain intervals sound the way they do, or how specific and informed adjustments may lead to more consistent intonation overall. These concepts can often be explained by the effect of naturally occurring musical acoustics on the perception of consonance and dissonance, as well as the implication of tuning and temperament systems on some recurring intonation issues throughout a variety of string repertoire. Even for advanced players, a lack of comprehensive understanding can create habitual inconsistencies and ineffective methods for self-evaluation while practicing that do not result in improved intonation during performances.Existing... https://escholarship.org/uc/item/2hp5b8sx Fri, 1 May 2026 00:00:00 +0000 Shang, Shirley https://escholarship.org/uc/item/24c7d242 Through the lens of approximation theory, the central problem of deep learning is to find a function that can predict the output of a target mapping given an input, with robust generalization. In this dissertation, by analyzing these individual aspects, namely i) architecture (the representation of the function), ii) data (the inputs and outputs of the function) and iii) optimization (the algorithm to find the function), we uncover a unified theme governing their interplay: structure.In the first part of this thesis, we show how the implicit structure of the architectures and the data they model can be leveraged to circumvent the Curse of Dimensionality. Applying this to the problem of image classification, we show how Convolutional Neural Networks (CNNs) achieve superior generalization compared to simple feedforward networks on high-dimensional tasks, even when trained on datasets orders of magnitude smaller than what classical lower bounds dictate. We derive an a priori generalization... https://escholarship.org/uc/item/24c7d242 Fri, 1 May 2026 00:00:00 +0000 Madala, Vamshi Chowdary https://escholarship.org/uc/item/1sz304jf Demand for high bandwidth interconnects for data centers, high performance computers, and other applications including sensing, 5G/6G mobile networks, and IoT, has driven the development of high-power on-chip light sources for silicon photonics (SiPh). Typically, laser sources are coupled externally, co-packaged, or integrated through flip chip or chiplets bonding, however, it is becoming increasingly desirable to integrate lasers and high-density gain with more scalable approaches. Micro-transfer print (MTP) integration involves the transfer of thin film coupons from native III-V substrates to silicon (Si). MTP provides a path to high throughput, low cost, high integration density and process maturity of III-V materials realized on native substrates. To date, demonstrations of MTP for SiPh have been limited to only a few mW of coupled power, therefore, improvements are required for higher coupled power and high efficiency. This dissertation focuses on the development of an MTP... https://escholarship.org/uc/item/1sz304jf Fri, 1 May 2026 00:00:00 +0000 Hu, Diya https://escholarship.org/uc/item/1pq9b32r The Wnt signaling pathway is a critical regulator of cell fate and tissue homeostasis, with dysregulation driving diseases such as colorectal cancer (CRC). Central to pathway control is the β-catenin destruction complex (DC), yet how its organization and inhibition are regulated remains unclear. Historically viewed as a static protein assembly, the DC has recently been shown to form a dynamic, liquid-like biomolecular condensate, reframing our understanding of Wnt pathway regulation. In this dissertation, I systematically mapped the DC interactome through proximity labeling of its scaffolds, APC and Axin, integrated these data with a functional CRISPRi screen for Wnt regulators and characterized the effects of CRC-associated APC mutations on DC structure.I began by characterizing the local proteome of the DC using BioID, a biotin proximity labeling technique. This approach identified 230 enriched interactors, including numerous proteins not previously associated with Wnt signaling.... https://escholarship.org/uc/item/1pq9b32r Fri, 1 May 2026 00:00:00 +0000 Pecchia, Surenna Pearl https://escholarship.org/uc/item/1hc28560 American options are ubiquitous in financial markets, and many numerical methods price them by approximating their continuous-time exercise feature with a discrete sequence of stopping opportunities. Under classical dynamic programming, a coarse grid with only a few stopping opportunities can materially undervalue the option, whereas on a very fine grid, approximation errors accumulate through the backward recursion and may ultimately break down the computation. We develop a least squares Monte Carlo (LSMC)-inspired algorithm that prices high-dimensional American options in continuous time, in settings where alternative approaches are impractical. Building on the LSMC framework, we initialize an aggregate deep neural network from regression-based timing-value data and then improve it through reinforcement-learning iterations. Starting from a coarse grid, we progressively increase the density of stopping opportunities to limit concept drift while updating the timing-value... https://escholarship.org/uc/item/1hc28560 Fri, 1 May 2026 00:00:00 +0000 Borsa, Cosmin https://escholarship.org/uc/item/1fz7r7rh This qualitative case study explores how STEM preservice teachers developed their identities as teachers of diverse students and how such identities intersected with their engagement in asset-based pedagogies (ABP) within a teacher education program. While teacher education programs aim to prepare candidates to teach in culturally responsive and equity-oriented ways, less is known about how STEM candidates build professional identities that center students' cultural, linguistic, and experiential assets while learning to apply ABP in real classrooms. This research endeavor addresses this gap by drawing on the identity framework by Gee and dimensions of asset-based pedagogy to examine how identity development and ABP emerge, evolve, and reinforce one another during teacher preparation.This dissertation was guided by two research questions:1. What structures and mechanisms were available for STEM teacher candidates to prepare to teach diverse student populations? In particular, what... https://escholarship.org/uc/item/1fz7r7rh Fri, 1 May 2026 00:00:00 +0000 Garcia, Liliana https://escholarship.org/uc/item/1d9528sn This document presents an in-depth examination of the art of pedaling in piano performance, focusing on its historical development, technical execution, interpretive function, and pedagogical application across diverse musical genres. Beginning with an introduction to the origins and evolution of piano pedals, it explores the mechanical and expressive roles of the damper, una corda, and sostenuto pedals, while highlighting the significance of foot technique and pedal notation in achieving stylistic authenticity and clarity.A core part of the study analyzes the nuanced usage of the damper pedal, addressing techniques such as connecting chords, arpeggios, and harmonies, syncopated pedaling, and the resonance of unconnected notes. Special focus is given to pedaling strategies for moderately fast melodies, dissonant passing notes, and chromatic phrases, as well as the use of the pedal to enhance color, support middle voices, and establish appropriate proportional control across textures.... https://escholarship.org/uc/item/1d9528sn Fri, 1 May 2026 00:00:00 +0000 Xu, Shiqi https://escholarship.org/uc/item/1556143k The time in which an event occurs can carry meaningful information when consideringits relation to other events. Leveraging this information for direct computation provides an alternative to the current digital paradigm, allowing for a simpler interaction with the physical world. Relationships between the temporal response created by any physical process can be directly computed upon using the same hardware substrate that supports digital logic without expensive conversions to binary representations. This offers the potential for energy efficient computation, but requires appropriate applications and careful hardware organization. This dissertation presents techniques that allow data to remain in the time domain while performing neural network operations. A general framework of temporal arithmetic is enabled through a negative log transformation with delay-based approximations. Fully temporal large scale, programmable architectures can leverage this framework through the use of... https://escholarship.org/uc/item/1556143k Fri, 1 May 2026 00:00:00 +0000 Gretsch, Rhys https://escholarship.org/uc/item/0wp3z703 Law enforcement agencies face a recruitment and retention crisis with significant implications for public safety and organizational effectiveness. Traditional police culture has contributed to strained police-community relations, negative public perception, and poor officer retention. In response, agencies are attempting cultural transformation through reformed socialization practices aimed at improving public interactions, officer well-being, and retention outcomes. This crisis has coincided with broader cultural transformation within agencies, yet little research examines how Generation Z officers (Born 1996-2012), the newest cohort entering law enforcement, experience organizational socialization amid these shifts.This study examines this cultural evolution through a generational framework, investigating the Generation Z officers and their organizational socialization (OS). While High-Reliability Organizations like policing traditionally require newcomers to shed previous identities... https://escholarship.org/uc/item/0wp3z703 Fri, 1 May 2026 00:00:00 +0000 Cavaness, Sofia Allegra https://escholarship.org/uc/item/0dd8m7g8 Granular materials are collections of discrete, macroscopic solid particles that interact primarily through dissipative contact forces. In industrial processing—ranging from pharmaceutical wet granulation to the handling of construction materials and food powders—these systems are rarely dry. Instead, they are frequently mixed with liquid binders to control dust, induce agglomeration, or modify flowability. These interstitial binders are often rheologically complex fluids containing dissolved polymers, surfactants, or suspended fines, rather than simple Newtonian liquids. Similarly, in soils, complex materials such as extracellular polymeric substances (EPS) or mucilage secreted by bacteria and plants help stabilize soil. The presence of such viscoelastic additives fundamentally alters the cohesive forces between grains, thereby modifying bulk transport behavior. At the particle scale, these complex fluids bind grains together through the formation of capillary bridges.... https://escholarship.org/uc/item/0dd8m7g8 Fri, 1 May 2026 00:00:00 +0000 Rajesh, Sreeram https://escholarship.org/uc/item/092706b9 The diffusion of ideas, innovations and behaviors is a complex sociological phenomena for which various computational models have been proposed by economists, algorithmists and ML researchers. However, it remains insufficiently understood due to lack of propagation data on networks and the complex interplay between dynamic network structure and node-level features. Moreover, real-world networks exhibit skewed degree distributions and propagation size distributions, thereby necessitating learning models that account for imbalanced distributions and data scarcity. The goal of this thesis is two-pronged: (1) How do coupled representation and classifier learning models cope with imbalanced data distributions, particularly with minority classes, and how can their weaknesses in limited data settings be mitigated?, (2) How can dynamics of propagation on real-world networks, themselves admitting a natural power law degree distribution, be understood using the machinery of gradient-based... https://escholarship.org/uc/item/092706b9 Fri, 1 May 2026 00:00:00 +0000 Sharma, Saurabh https://escholarship.org/uc/item/4jf447jd <p>This dissertation introduces a set of microfluidic systems for on-demand detection and quantification of drugs of abuse. We demonstrate sensitive detection of methamphetamine in saliva and detection of fentanyl in the presence of heroin, and establish an analysis strategy for quantification of drugs of abuse using surface-enhanced Raman spectroscopy (SERS). </p><p>Drug abuse results in thousands of overdose deaths and billions of dollars in direct and indirect annual economic cost. Rapid, reliable, and miniaturized chemical detection systems for detection and quantitation of such drugs, especially in biological fluids, are urgently needed for forensics, healthcare, and law enforcement applications. </p><p>SERS is a vibrational spectroscopy technique capable of detecting chemicals at very low concentrations. The phenomenal sensitivity and label-free detection capability make SERS an attractive chemical detection modality for miniaturized drug sensing systems. However, detection... https://escholarship.org/uc/item/4jf447jd Tue, 21 Apr 2026 00:00:00 +0000 Salemmilani, Reza https://escholarship.org/uc/item/6nn947w5 Recent success in scaling large language models (LLMs) has achieved remarkable capabilities in a wide range of real-world applications. Despite these achievements, LLMs face two central challenges: trustworthiness and inference efficiency. Given their impressive performance, it is crucial to ensure that LLMs are reliable and aligned with human intent. At the same time, deploying such large-scale models introduces significant challenges in achieving computational efficiency. This dissertation presents my study of trustworthy and efficient LLMs, advancing both reliability and inference efficiency as for the next generation of foundation models.The first part of this dissertation focuses on trustworthiness, encompassing adversarial robustness and hallucination. I propose TEXTGRAD, a gradient-driven framework for evaluating model robustness, followed by Self-Denoised Smoothing and Semantic Smoothing, two methods that enhance robustness against adversarial and jailbreak attacks by... https://escholarship.org/uc/item/6nn947w5 Wed, 11 Feb 2026 00:00:00 +0000 Hou, Bairu https://escholarship.org/uc/item/0mh086tc The detrimental impacts of invasive plant species on ecosystems globally is a well- studied area within ecology. Groundwater-dependent ecosystems (GDEs) are one of these systems, but they specifically remain understudied. These GDEs are important resources to both humans due to their water use, and wildlife due to their habitat use, especially in areas surrounded by anthropogenic structures. The purpose of this study was to assess how mammals use the habitat within a GDE and how the invasive plant, Arundo donax, impacts this mammalian habitat use. The impacts of habitat structure and seasonal conditions were also explored. Trail cameras, invertebrate sampling, vegetation sampling, microclimate sampling, and remote sensing tools were all used to gather information on the Santa Clara River Watershed. These data were incorporated into generalized linear mixed models (GLMMs) to model mammal occupancy with the various ecosystem covariates.Numerous cases of significant plant-animal... https://escholarship.org/uc/item/0mh086tc Wed, 11 Feb 2026 00:00:00 +0000 Fairbanks, Eric https://escholarship.org/uc/item/3q27h65z We applied mathematical methods from control theory to inform the optimal design of Deep Brain Stimulation (DBS) protocols and to model physiology time series dynamics. As DBS emerges as an effective treatment for Parkinson’s disease and other neurological disorders, neuro-engineering perspectives are growing to inform the optimal design of stimulation protocols. First, we derive analytical approximations for optimal stimuli with a control objective of desynchronizing neural oscillators. Based on phase response curve modeling of neuron spiking dynamics, we see that the approximations achieve nearly optimal chaotic desynchronization. Next, we calculate calculate magnitude-constrained optimal stimuli for DBS in cases where there is a biological or electronic constraint on the amount of electrical current that can safely be applied over time. We specify the control objective with a Hamiltonian formulation for desynchronizing a population of neurons by maximizing the Lyapunov exponent... https://escholarship.org/uc/item/3q27h65z Tue, 10 Feb 2026 00:00:00 +0000 Zimet, Michael https://escholarship.org/uc/item/2d626420 Mitochondrial dysfunction and metabolic derangement have emerged as pathological features across diverse kidney diseases. The kidney is one of the most metabolically active organs in the body, requiring continuous ATP generation to sustain regulated reabsorption and secretion across the renal tubular epithelium. Beyond their canonical role in energy production, mitochondria are critical regulators of apoptosis, redox balance, and responses to cellular stress. Mitochondrial injury therefore has profound consequences for renal physiology and is increasingly recognized as a driver of both acute kidney injury (AKI) and chronic kidney disease (CKD). This dissertation investigates how mitochondrial dysfunction contributes to renal pathology in two mechanistically distinct yet conceptually related settings: (I) oxidative stress–regulated cleavage fragments of polycystin-1 (PC1) in autosomal dominant polycystic kidney disease (ADPKD), and (II) cGAS-STING signaling in calcium oxalate (CaOx)-crystal... https://escholarship.org/uc/item/2d626420 Tue, 10 Feb 2026 00:00:00 +0000 Sharpe, Elizabeth Hikmet https://escholarship.org/uc/item/96r9t4pq Despite structural integration in higher education, students often fail to form cross-social class relationships. This dissertation applies a sociocultural lens to investigate how social class similarity shapes interpersonal connection through two complementary studies. Study 1, a field study of randomly assigned college roommates, shows that income and first-generation status mismatch predict reduced perceived financial and interpersonal similarity, which in turn undermine relationship quality. Study 2, a laboratory experiment examining initial interactions, reveals a counterintuitive pattern: first-generation status mismatch sometimes leads to higher perceptions of common ground, emotional similarity, empathic accuracy, and interest in the partner, perhaps due to impression-management when class differences are made salient in a laboratory setting. Together, these findings suggest that social class similarity fosters sustained relationship quality over time, while class dissimilarity... https://escholarship.org/uc/item/96r9t4pq Mon, 9 Feb 2026 00:00:00 +0000 Papadakis, Viki https://escholarship.org/uc/item/88n4174t This dissertation investigates three questions about consent: what moral difference consent makes, whether consent is necessary for sex to be morally benign, and which kinds of actions and arrangements fall within the domain of consent at all. I defend an authority-based view on which consent is a distinctive normative power by which individuals can make certain actions compatible with their authority. On this picture, consent is an exercise of “allowing authority” that can coexist with residual duties and other moral constraints. I argue that, properly understood, consent is morally necessary for benign sexual activity, and that recent attempts to sideline consent in sexual ethics rest on mistaken assumptions about what consent presupposes and how it operates in ideal cases. Finally, I address a puzzle about consent’s domain, focusing on requested actions and on the use of another’s labor or agency: I argue that it can be meaningful and important to speak of a person’s consenting... https://escholarship.org/uc/item/88n4174t Mon, 9 Feb 2026 00:00:00 +0000 Davies, Morgan Edward https://escholarship.org/uc/item/846270v0 The Ethereum blockchain and its decentralized finance (DeFi) ecosystem have fundamentally transformed financial infrastructures, with over $100 billion in total value locked across thousands of interconnected contracts. With the growth of DeFi, the interactions between smart contracts have become increasingly complex, enabling advanced financial protocols and applications that offer unprecedented functionality -- from automated market makers to lending platforms and synthetic assets. Nonetheless, bugs in smart contract interactions are a common cause of critical vulnerabilities: many services interact with contracts that must be trusted to manage digital assets, creating a web of dependencies where a single vulnerability can cascade across multiple protocols. As a result, hundreds of millions of dollars are stolen every year through exploits that target the subtle semantics of inter-contract communication.The core security challenge lies not in simple coding errors, but in these... https://escholarship.org/uc/item/846270v0 Mon, 9 Feb 2026 00:00:00 +0000 Ruaro, Nicola https://escholarship.org/uc/item/7rw0n0hz Physical activity is well-established as a promoter of individual health and well-being, yet its role in fostering positive social relationships remains underexplored. Building on theory from exercise science and close relationships research, this dissertation tests a developing theoretical model proposing that physical activity restores emotional, cognitive, and physiological resources, which in turn enable individuals to engage in pro-relationship behaviors that enhance relationship and social outcomes. Across three studies, I examined both intrapersonal and interpersonal outcomes of physical activity using complementary field and laboratory designs. Studies 1a and 1b were field studies that examined changes in psychological well-being and pro-relationship cognition and behavior from before to after participation in leisure activity classes. In Study 1a, participants were recruited from both exercise and art classes offered through the UCSB Department of Recreation. In... https://escholarship.org/uc/item/7rw0n0hz Mon, 9 Feb 2026 00:00:00 +0000 Harris, Paige https://escholarship.org/uc/item/7nj1g1v4 Quilombolas in the Iguape Basin and Valley, Bahia, Brazil struggle to counter land and environmental problems affecting their means of subsistence. Through sixteen semi-structured interviews with adult current and former members of the Quilombola Council of the Iguape Basin and Valley (founded in 2005) this study investigated threats posed by land dispossession and environmental rifts to quilombolas’ social reproduction. Quilombolas were contacted through snowball sampling and shared memories regarding their relationship with the land- and waterscapes of the Basin associated with labor practices. In the memories, they foregrounded spatialized daily acts of survival (McKittrick 2013), such as the cultivation of plots and fishing activities. While the land question harks back to the late 19th century, the environmental question has been more noticeable since the early 2000s. The interaction between land encroachment (Bowen 2021; Cruz 2014, 2022) and environmental and corporeal... https://escholarship.org/uc/item/7nj1g1v4 Mon, 9 Feb 2026 00:00:00 +0000 Goncalves dos Santos, Anderson Henrique https://escholarship.org/uc/item/4v4150kb Heterogeneous catalysts are central to countless aspects of modern life, from providing transportation fuels, pharmaceuticals, and food products to mitigating harmful environmental emissions. In particular, zeolites are a class of aluminosilicates that are commonly used in heterogeneous catalytic applications due to their ion-exchange capacity, well-defined microporous crystalline frameworks, and high surface areas. Dispersed metal species are often distributed throughout a zeolite support to form bifunctional (or dual-functional) catalysts which contain both metal and acid sites, enabling them to catalyze coupled reactions, such as hydrogenation/dehydrogenation and hydroisomerization/cracking reactions. The reactive properties of metal-containing zeolite catalysts depend strongly on their atomic-scale compositions and structures. For example, ultra-stable Y-zeolites (USY) containing dispersed platinum (Pt) are widely used industrially for hydrocracking and hydroisomerization... https://escholarship.org/uc/item/4v4150kb Mon, 9 Feb 2026 00:00:00 +0000 Pischer, Anna Lane https://escholarship.org/uc/item/3sx7f1m3 Johannes Brahms’s first Violin Sonata in G Major, Op. 78, is deeply influenced by two of his earlier compositions: the lieder “Regenlied” and “Nachklang” (Nos. 3 and 4 from 8 Lieder und Gesänge, Op. 59). These lieder supply the melodic and motivic foundation from which the entire three-movement violin sonata is derived. Brahms’s compositional process goes far beyond mere quotation. He masterfully expands and evolves the lieder’s motivic material into an entirely new work.Regenlied’s principal theme introduces three key motives that unify all three movements of the sonata: 1. The Rain Drop Motive (RDM): The most notable motive, it is a dotted quarter-eighth note rhythm found in the anacrusis of the melody. 2. Motive Z (z): An appoggiatura found on the text “Re -gen”. 3. Motive T (t): The ascending interval of a third found on the text “wal”. Consistent with the theme of rain (“regen”) and tears “Nachklang”), Brahms uses these three motives in the minor mode throughout the lieder... https://escholarship.org/uc/item/3sx7f1m3 Mon, 9 Feb 2026 00:00:00 +0000 Eben, Jared https://escholarship.org/uc/item/2zc5z7q5 The West African Monsoon (WAM) is a significant climatic feature for many nations because it influences water resources, agriculture, public health, and hydropower generations across many African nations. Understanding the WAM variability from intraseasonal to interannual time-scales is therefore of major socioeconomic importance. This study contrasts two metrics for characterizing the WAM large-scale features (onset, demise and duration) and spectral variability using two gridded daily precipitation datasets and 850 hPa daily data from the European Centre for Medium-Range Weather Forecasts reanalysis version 5 (ERA5) at 0.25° resolution (1979–2024). The first index applies Combined Empirical Orthogonal Functions (CEOF) to ERA5 precipitation, 850hPa circulation, temperature, and specific humidity. The annual northward migration of precipitation bands is explained by the first and second CEOFs with distinct spatial patterns. CEOF-1 correlates with unimodal rainfall in the Sahel,... https://escholarship.org/uc/item/2zc5z7q5 Mon, 9 Feb 2026 00:00:00 +0000 Olatunde, Ayodeji Oluwabusayo https://escholarship.org/uc/item/2x09v3vk This dissertation embarks on an in-depth examination of the kafala temporary labor migration system and the variation in the dyadic relationships between sending and destination countries. Temporary migration is a growing subfield within political science, migration, and regional studies; as such, the first two chapters seeks to contribute theoretical frameworks to explain why sending states choose to intervene in policy decisions regarding their expatriate populations abroad, and how migrant workers’ experiences under kafala can shape domestic political attitudes, respectively. In contrast, the third chapter shifts to a more traditional, deductive approach to analyze the financial effects of migration policy restrictions while building upon the theoretical frameworks introduced earlier. To holistically capture the macro-, meso-, and micro-level dynamics that exist within the kafala system, I employ a mixed-methods research design. Through a combination of process-tracing,... https://escholarship.org/uc/item/2x09v3vk Mon, 9 Feb 2026 00:00:00 +0000 Uddin, Noosha https://escholarship.org/uc/item/2ns319x8 Failures of sustained attention are pervasive in everyday life. In the lab, these states of attention can be tracked unobtrusively using response times (RTs) in monotonous continuous performance tasks requiring frequent responses. Here, we used a RT-based real-time triggering procedure to investigate the interplay between fluctuations of attentional state and another attentional phenomenon: experience-dependent capture. We tested two accounts. One proposes attentional lapses could reduce one’s ability to suppress capture by the distractor, leading to poorer performance (capture suppression account). The other proposes capture “snaps” attention back into focus and reduces commission errors (perceptual recoupling account). Participants first completed a training task that differentiated colored targets by reward and then performed a sustained attention task that required they execute visual search on each trial. On rare trials, the colors relevant during training served as distractors... https://escholarship.org/uc/item/2ns319x8 Mon, 9 Feb 2026 00:00:00 +0000 Shelat, Shivang https://escholarship.org/uc/item/251980jr Gender norms reflect a uniquely cultural expression of what evolutionary anthropologists refer to as sexual conflict, wherein the interests of either sex diverge and each develops strategies in effort to gain the upper hand. With men seemingly winning the conflict across societies, this thesis aims to better understand the forces that uphold and challenge patriarchal norms, focusing specifically on the role of men. Drawing on interviews, focus groups, experimental vignettes, and survey data in a rural but fast urbanizing Tanzanian community, I investigate (i) local understandings of men’s support and opposition to gender equality, (ii) the consequences men face when challenging patriarchal norms, (iii) who enforces these consequences, and (iv) how socioecological change may incentivize support for women’s empowerment. Consistent with sexual conflict theory, men and elders held limited notions of empowerment that largely preserved men’s authority, and were more likely than women... https://escholarship.org/uc/item/251980jr Mon, 9 Feb 2026 00:00:00 +0000 Kilgallen, Joseph https://escholarship.org/uc/item/239193t6 Neurodegenerative diseases are often linked to proteinaceous species that are found in misfolded and insoluble form. An example is amyloid plaques, which are a hallmark of Alzheimer’s disease. These insoluble clumps of protein are not only found in AD diseased brains, but in 30 to 50% of patients with HIV/AIDS as well. This disruption of amyloid β metabolism has been linked to viral proteins such as neurotoxic HIV1-Tat. This viral component is understood to have a direct and indirect effect on amyloid β aggregation, making it difficult to pinpoint a molecular mechanism and develop a treatment.&nbsp;Ion mobility mass spectrometry in concert with AFM and molecular dynamics is used to characterize the oligomerization and coaggregation of Tat and amyloid β peptides. The work then switches gears to the ALS related protein, SOD1, and characterizes the effect of SOD1 point mutations on its aggregation. Looking through the lens of ion mobility, we can gain insight into the structure of... https://escholarship.org/uc/item/239193t6 Mon, 9 Feb 2026 00:00:00 +0000 Prokopovich, Dmitriy https://escholarship.org/uc/item/1pf5m65x Dynamic Nuclear Polarization (DNP) is a powerful technique for enhancing the intrinsically poor sensitivity of solid-state Nuclear Magnetic Resonance (ssNMR) by up to 660x, allowing experiments that would ordinarily take days to be run in mere minutes, thus enabling NMR characterization on multiple classes of materials from in-situ biological samples to solar cells. The success behind DNP lies in the transfer of high spin polarization of electron spins to the otherwise poorly polarized nuclear spins, making use of microwaves at or near the electron spin resonance (ESR) frequency to drive the transfer.However, DNP’s full potential has yet to be realized, as there are many systems, from live cells at in vivo temperature to molecular orientations in semiconductor thin films, where DNP is currently far from its theoretical maximum. Current commercially available DNP radicals require high microwave powers, sample spinning, and ultralow temperatures to come close to this maximum, translating... https://escholarship.org/uc/item/1pf5m65x Mon, 9 Feb 2026 00:00:00 +0000 Chaklashiya, Raj Kumar https://escholarship.org/uc/item/16f8k3bn The nervous systems of animals perform computations that support adaptive behavior. In the case of insect navigation, a region of the insect brain, called the central complex, performs vector computations that facilitate the representation of space in a world-centered (allocentric) reference frame. While there is an in depth understanding of the anatomy of the central complex in the fruit fly, Drosophila melanogaster, the function of most of its cell types remains a mystery. I leveraged the theoretical framework of vector computation in combination with electron microscopy anatomy data to hypothesize a function for a previously uncharacterized cell type called PFL1. Specifically, I propose that it transforms a vector stored in the fly’s memory from an allocentric reference frame to an egocentric (fly-centered) reference frame. Through 2-photon functional imaging experiments, I show that PFL1’s input output relationship is consistent with this hypothesis. This finding leads me... https://escholarship.org/uc/item/16f8k3bn Mon, 9 Feb 2026 00:00:00 +0000 Gorko, Benjamin https://escholarship.org/uc/item/0mq7j3m7 One of the core debates in the philosophy of time concerns the debate between A-theorists and B-theorists. A-theorists maintain that there are absolute, objective tensed features of events and that time robustly passes in the sense that events undergo changes in these tensed features over time. In contrast, B-theorists deny that tensed features, such as pastness, presentness, or futurity are absolute, objective features of events and deny that time passes in a robust sense. The A-theory is often charged with problems of logical coherence or consistency with the Special Theory of Relativity. However, one place where the A-theory has been thought to fare better than the B-theory concerns experience. The idea is that it seems as though we experience events as present and that it appears to us in experience as though time robustly passes. From this claim about experience, it has been suggested that the best explanation for these aspects of experience is that the A-theory is true.... https://escholarship.org/uc/item/0mq7j3m7 Mon, 9 Feb 2026 00:00:00 +0000 Neidlinger, Jordan https://escholarship.org/uc/item/0kh0s8br Computer models and simulations are widely used to understand scientific and social phenomena. It is common to have high-dimensional outputs from computer models on spatial or spatial-temporal grids for each input parameter set or initial conditions. To reduce the computational cost of expensive simulations, statistical surrogate models have been developed for predicting computer model outputs based on a small number of simulation runs. Among these surrogate models, Parallel Partial Gaussian Process (PPGP) is an efficient framework for predicting high-dimensional outputs from computer simulations with quantified uncertainty. However, the computational costs of PPGP increase cubically with respect to the number of training simulation runs due to computing the inversion and logarithm of the determinant of the covariance matrix, which prohibits using a large number of simulations to train our model. To address this computational challenge of the PPGP surrogate model, we employ... https://escholarship.org/uc/item/0kh0s8br Mon, 9 Feb 2026 00:00:00 +0000 Liu, Tianhong https://escholarship.org/uc/item/9x65h82s This thesis investigates the asymptotic behavior and bias properties of spike eigenvector estimators in high-dimensional statistical models. Motivated by challenges in estimating latent structures from noisy, high-dimensional data, it focuses on how shrink-age and model specification shape the geometry and accuracy of estimated eigenvectors across different asymptotic regimes.Chapter 1 provides the theoretical background and mathematical tools, introducing connections between random matrix theory, spectral geometry, and shrinkage estimation. Chapter 2 develops and analyses a James-Stein type shrinkage framework for eigenvector estimation under rank misspecification in the factor model Y = BX + ε. In high-dimensional settings, empirical eigenvectors are inherently inconsistent even under correct model specification. The analysis characterizes the additional optimization bias that arises when the estimated rank K differs from true latent dimension q. Theoretical results show that... https://escholarship.org/uc/item/9x65h82s Sat, 7 Feb 2026 00:00:00 +0000 Jin, Zhuoli https://escholarship.org/uc/item/9vt6825t This dissertation investigates three of Claude Debussy’s Vasnier Songs, which were unknown to the musical world until they were rediscovered twelve years ago, offering fresh insight into a chapter of his development that had been forgotten. The Three Songs for Madame Vasnier (2013)—Caprice, Rondel Chinois, and La Fille aux Cheveux de Lin–composed between 1880 and 1881 for soprano Marie-Blanche Vasnier, reveal the origins of the stylistic and harmonic language that would later define his mature mélodie. This study is written primarily for classical singers and voice students seeking a bridge between analysis and lived experience. It proposes an interpretive framework grounded in the acting techniques of Uta Hagen—specifically her Six Steps: identity, circumstance, relationship, objective, obstacle, and tactic. While traditional art song pedagogy often privileges intellectual or historical interpretation, Hagen’s method invites performers to root those insights in physical truth... https://escholarship.org/uc/item/9vt6825t Sat, 7 Feb 2026 00:00:00 +0000 Barker, Olivia Lyon https://escholarship.org/uc/item/9v07q8p4 An estimated minimum of 40% of the global population lives within 50 km of the coasts and are at the forefront of climatic and land use impacts to our coasts, such as sea level rise and developmental pressure. This has led to increased scientific and economic interest in boosting our coastal resilience by supporting natural ecosystem processes that mitigate the impacts of climate change to the coasts. Coastal wetlands can serve as nature-based solutions to climate change and provide climate resiliency through their many ecosystem services, such as carbon sequestration and soil accretion. However, these ecosystems are highly spatially reduced or disturbed across the globe, and what remains of them are vulnerable to further disturbance, such as salinization. Coastal salinization patterns are changing due to environmental impacts of climate change and human development on sea level, freshwater inflow, precipitation patterns, and sediment supply. These changes are expected to have... https://escholarship.org/uc/item/9v07q8p4 Sat, 7 Feb 2026 00:00:00 +0000 Silva, Germán David https://escholarship.org/uc/item/9fm1p1ch I. Palladium-catalyzed aminations can be efficiently run, under plug flow conditions on water, with low levels of a recyclable palladium pre-catalyst. General protocols to couple both aryl and aliphatic amines with aryl/heteroaryl bromides have been developed. Further highlights include “on water” conditions, short reaction times, recycling of the reaction medium, low levels of residual Pd in the isolated products, and low E-Factors as a green metric, all these features showcasing this alternative to amination reactions typically run in batch mode.II. New plug flow opportunities for carrying out SNAr reactions are disclosed using water mixed with the green and separable/recyclable co-solvent 2-MeTHF. This approach enables C-O, C-N, and C-S bond formations. Several features characteristic of this sustainable technology are highlighted, such as recycling of the aqueous reaction medium, applications to several intermediates en route to numerous APIs, and a new “flow-to-batch-to-flow”... https://escholarship.org/uc/item/9fm1p1ch Sat, 7 Feb 2026 00:00:00 +0000 Wong, Madison https://escholarship.org/uc/item/8hk1h4m1 This dissertation examines the critical role of geosocial determinants of health in shaping HIV risk and resiliency in the United States. A geographic perspective is key for understanding how structural and social factors interact with individual characteristics to produce distinct patterns of risk and resiliency across space and populations. Despite advances in biomedical treatment and prevention for HIV, persistent disparities remain, particularly for racial and ethnic minority groups. Each chapter in this dissertation examines how place-based contexts, ranging from demographic composition to neighborhood characteristics, produce distinct spatial patterning for HIV risk and resilience and underscores the importance of a spatial framework to end the HIV epidemic in the United States. The first study develops and extends a county-level imputation methodology to estimate racially stratified population densities for sexual minority men (SMM). Using these estimates, the analysis... https://escholarship.org/uc/item/8hk1h4m1 Sat, 7 Feb 2026 00:00:00 +0000 Reid, Sean Casey https://escholarship.org/uc/item/8fj8v5qv The emergence of a superconducting phase within the charge density wave (CDW) order in the CsV3Sb5 kagome superconductors has spurred investigations to elucidate the nature of their interactions. Moreover, the ambiguous origin of this unique CDW order emphasizes the novel complexity of this phase. Doping and pressure-dependent experiments reveal that these orders compete for the low-energy states near the Fermi level (EF ). However, hole-induced changes to the microscopic environment of the CDW order remain unexplored.In this work, the evolution of the CDW order in CsV3Sb5−xSnx across 0 ≤ x ≤ 0.68 is tracked using coherent phonon spectroscopy (CPS)—an ultrafast pump-probe technique. The sensitivity of this probe to minute reflectivity changes enables the detection of a CDW-induced single-particle gap opening at EF . This electronic renormalization causes a sign flip in the reflectivity response, extending the CDW transition from x ≈ 0.06 to x ≈ 0.15, with TCDW ≈ 61 K at x = 0.15.Decomposing... https://escholarship.org/uc/item/8fj8v5qv Sat, 7 Feb 2026 00:00:00 +0000 Kongruengkit, Terawit https://escholarship.org/uc/item/8249472c With the growing prevalence of neural networks, especially in safety critical domains such as medical applications, evaluating their correctness has become a crucial problem. While accuracy on a previously unseen dataset provides one metric for correctness, neural networks have been shown to have numerous and often unexpected faults. These faults include lack of robustness against small changes in input, learned biased behavior based on existing biases in the training dataset, and generation of unsafe outputs for a given application context. Evaluation of neural networks thus should not just encompass an accuracy measure, it must also take into consideration other negative outcomes of the usage of neural networks in practice when used as part of a software system.In this dissertation, I will describe my work on two key types of neural network evaluation: robustness and safety. For both of these criteria, there is a common thread of quantitative analysis. It is not enough to know... https://escholarship.org/uc/item/8249472c Sat, 7 Feb 2026 00:00:00 +0000 Downing, Mara https://escholarship.org/uc/item/7zb601mb Scientists devote years to the model development cycle, which is the process of finding a model that describes a process, using data to fit parameters to the model, analyzing uncertainties in the fitted parameters, and performing additional experiments to refine and validate the model. This dissertation addresses these challenges by introducing machine learning–based tools that accelerate and automate key steps in the model development cycle, enabling the experimentalist to focus on what they do best: scientific discovery. First, we show how Neural Differential Equations can be used for data-driven modeling of time-series data and dynamical systems found in science and engineering. Building on this foundation, we introduce our state-of-the-art symbolic neural ordinary differential equation (neural ODE) tools for the symbolic regression of dynamical systems. To account for noise and uncertainty, we further develop Bayesian inference techniques for symbolic neural ODEs, enabling... https://escholarship.org/uc/item/7zb601mb Sat, 7 Feb 2026 00:00:00 +0000 Fronk, Colby https://escholarship.org/uc/item/7fm10592 Oil-in-water emulsions allow for the transport and deposition of oil droplets in a more convenient and less viscous aqueous phase. For example, in agriculture, horticultural oils that help with the control of pests and fungi can be emulsified in an aqueous phase and sprayed on crops. Similarly, in cosmetic products, oils that help with skin hydration can be emulsified in an aqueous phase that has a more pleasing, lightweight feel. Emulsions are often sprayed on surfaces. The atomization or destabilization of the volume of fluid into drops plays a critical role in the examples above, as the size of the drops affects deposition efficiency. Smaller drops increase surface coverage, while larger drops reduce off-target drift. An understanding of the fragmentation, which until now has been primarily based on a trial-and-error approach, is beneficial as emulsion composition can vary greatly depending on the application. We use a classical model experiment where a drop of emulsion... https://escholarship.org/uc/item/7fm10592 Sat, 7 Feb 2026 00:00:00 +0000 Gonzalez, Sara https://escholarship.org/uc/item/7fj0v96n Biodiversity is declining at an alarming pace, driven by the effects of habitat loss and fragmentation, climate-change, invasive species, and emerging diseases1. Among emerging diseases, fungal pathogens are particularly concerning: they remain poorly understood, possess long-lived saprobic stages, evolve rapidly, and harbor cryptic genomic elements that facilitate adaptation2. One emergent fungal disease, chytridiomycosis, caused by Batrachochytrium dendrobatidis (Bd), has been implicated in the decline or extinction of over 500 amphibian species globally.In my dissertation, I examine factors that drive the differences in population responses to Bd. Using populations across California, I examined predictors of infection dynamics, the role of adaptive immune markers, and Bd lineage diversity. I found that local, environmental context is the strongest driver of infection prevalence and load, while certain immune markers are linked to past epizootics, suggesting an adaptive response.... https://escholarship.org/uc/item/7fj0v96n Sat, 7 Feb 2026 00:00:00 +0000 Russell, Imani https://escholarship.org/uc/item/7db8w3vs Early wireless technology began from the first discovery of radio waves by HeinrichHertz and utilization in the first wireless communication system by Guglielmo Marconi evolving to the modern day millimeter-wave phased arrays for 5G/6G applications and sensing. As these modern systems move to higher frequencies to accommodate high bandwidth applications, efficiency and power are limited and thermal issues arise due to smaller array element spacing and lower efficiency. Much research has been done to improve the efficiency for these millimeter-wave systems. In particular, the power amplifier consumes the most power and in effect, much research is focused on maximizing power amplifier efficiency. To address this problem, this thesis explores methodologies and design of high-efficiency power amplifiers above 100 GHz utilizing III-V based semiconductor device technologies including GaN HEMTs and InP HBTs. The structure of this thesis begins with theoretical limitations in power... https://escholarship.org/uc/item/7db8w3vs Sat, 7 Feb 2026 00:00:00 +0000 Lam, Eythan https://escholarship.org/uc/item/6nk4t4h1 This research investigates how party identification, race, and immigration generation, individually and in interaction, structure Americans’ willingness to escalate U.S. involvement in response to Russia’s invasion of Ukraine. Using the 2022 Cooperative Election Study (CES), I operationalize escalation in two ways: (1) an interpretable max-escalation ladder (0-4) that maps “check-all-that-apply” responses to ordered policy rungs, and (2) as latent escalation from a 2-parameter explanatory item-response theory (2PL X-IRT) model that weights items by difficulty and discrimination. This intersectional, measurement-aware approach clarifies when, and for whom, foreign policy cues stick, and where lived histories and diasporic ties shift the calculus. Across models, I find that partisanship matters modestly. Democrats are generally more supportive of moving up the escalation ladder than Independents or Republicans, yet the gap can shrink at the most kinetic step. Furthermore, immigration... https://escholarship.org/uc/item/6nk4t4h1 Sat, 7 Feb 2026 00:00:00 +0000 Esteban, Rhoanne https://escholarship.org/uc/item/6fm617h4 The Andrews–Curtis conjecture, proposed in 1965, asks whether any balanced presentation of the trivial group can be transformed to the trivial presentation using a specific set of moves. This thesis investigates both the standard conjecture and its stable variant from algorithmic and topological perspectives.We develop a state-of-the-art algorithm for trivializing group presentations and apply it to resolve several open problems. We prove that all but two presentations in the Akbulut–Kirby series can be reduced in length, and we resolve various potential counterexamples in the Miller–Schupp series, including three infinite subfamilies.The stable Andrews–Curtis conjecture allows additional stabilization moves and is equivalent to a topological statement: any contractible 2-dimensional polyhedron can be 3-deformed to a point. We investigate this formulation through the systematic study of fake surfaces, which are generic 2-polyhedra. We present a complete classification of acyclic... https://escholarship.org/uc/item/6fm617h4 Sat, 7 Feb 2026 00:00:00 +0000 Fagan, Lucas https://escholarship.org/uc/item/5wt4f2hw Insects have been framed as the future of food. This dissertation examines the emergence of the U.S. edible insect industry and its diverse community of social actors including educators, entrepreneurs, farmers, researchers, and advocates who seek to normalize insect consumption through cultural change. Drawing on over six years of multi-sited ethnographic research, I trace the social contours of this advocacy community as it envisions, navigates, and attempts to enact food systems transformation through processes of normalization. While advocates frame insect agriculture as essential to achieving more sustainable, just, and resilient food systems, these imagined food futures hinge on the community’s ability to transform American cultural aversions and reconfigure moral boundaries for what is considered “good to eat.” I approach the U.S. edible insect advocacy community’s work as a future-oriented cultural project in which food futures become sites of negotiation over sustainability,... https://escholarship.org/uc/item/5wt4f2hw Sat, 7 Feb 2026 00:00:00 +0000 Wade Osleger, MacKenzie https://escholarship.org/uc/item/5cw477nw Marine heatwaves (MHWs) are persisting and increasing in frequency and duration, and they are occurring with greater severity than in past decades. MHWs are a widespread occurrence, and events can span thousands of miles. As anthropogenic climate change worsens, we will be experiencing more intense marine heatwave events, so investigation into their biological impacts is necessary. My research focus was centered on the impacts of marine heatwaves, on the purple sea urchin, Strongylocentrotus purpuratus. S. purpuratus is an important sedentary herbivore in Santa Barbara Coastal kelp forests. Previous research identifies transgenerational plasticity in urchins, where the environment of the adults impacted the phenotype of offspring. Specifically, maternal and paternal provisioning in eggs and offspring assist with successful fertilization and offspring physiology. In my dissertation, I had three aims: 1) investigate how exposure to fluctuating MHWs during gametogenesis in adults... https://escholarship.org/uc/item/5cw477nw Sat, 7 Feb 2026 00:00:00 +0000 McDonald, Adriane Michelle https://escholarship.org/uc/item/5bp5h09n Between 2011 and 2013, a series of hunger strikes originated in Pelican Bay State Prison in Northern California and spread throughout the California Department of Corrections and Rehabilitation (CDCR) in protest of long-term solitary confinement for “validated” (presumably verified) male gang members. The main strike from July 8 to September 5, 2013, ultimately involved more than thirty-thousand male prisoners in nearly every state prison and county jail in California. It is not uncommon for prisoners to protest prison abuses, but they are often violent forms of protest that involve assaulting prison staff or destroying state property. For example, prisoners will organize mass “cell extractions” in which they cover their windows and board their cell doors to lure in correctional officers to assault them; another form of protest is to clog the sinks and toilets to flood the entire prison pod which on average causes thousands of dollars of property damage. These types of incidents... https://escholarship.org/uc/item/5bp5h09n Sat, 7 Feb 2026 00:00:00 +0000 Terrell, Clint J https://escholarship.org/uc/item/57j833hj Coastal ecosystems worldwide face intense development pressure. Consequently, undisturbed coastal regions are scarce and continuously diminishing. The loss of intact coastal habitat may negatively affect terrestrial coastal wildlife, many of which use marine areas and resources. There could be important cascading effects on the ecosystems these wildlife species inhabit. Yet, we have surprisingly limited data on the frequency and nature of marine resource use by terrestrial animals, as well as on the importance of wild coastlines in facilitating this behavior. Here, we aim to fill this gap by studying the diets of three species of large terrestrial mammals (coyote, bobcat, wild boar) using scat DNA metabarcoding at two coastal reserves – one remote, and one suburban – along the central California coastline. We investigate (1) how the diets of some of the region’s most common wild large mammals living near the coast compare to one another; (2) to what extent each of these mammals... https://escholarship.org/uc/item/57j833hj Sat, 7 Feb 2026 00:00:00 +0000 Lewin, Grace Beatrice https://escholarship.org/uc/item/5799c428 The rapid scaling of large language models (LLMs) has led to unprecedented advances in natural language and multimodal understanding, but at immense cost in data, compute, and energy. Recent frontier LLMs are generally trained on tens of trillions of tokens with tens of thousands of GPUs, raising critical concerns about sustainability and efficiency. This dissertation addresses these challenges by advancing the data- and training-efficiency of LLMs through novel data curation techniques and efficient architectural designs. On the data efficiency side, we aim to enhance pre-training data efficiency by developing robust pipelines that curate high-quality pre-training data. Classifier-based data curation approaches such as RedPajama and DCLM have established strong efficacy for text-only data filtering but remain limited in multimodal contexts. In particular, existing methods like CLIPScore primarily address image-text caption data curation and cannot tackle the more complex... https://escholarship.org/uc/item/5799c428 Sat, 7 Feb 2026 00:00:00 +0000 Wang, Weizhi https://escholarship.org/uc/item/4s77s1cg The 2023-2024 academic year saw a surge in student activist organizing across the U.S., with over 130 encampments established at university campuses to protest the war in Gaza (Ulfelder, 2024). Activists, attempting to create political change through social actions, face a unique set of organizational communication challenges, bringing together unpaid members under a shared decision-making structure during times of sociopolitical upheaval. While the communication between activists constitutes these organizations, their formation and function remain understudied, motivating this research. This case study examines the emergence, structure, and dissolution of two such activist organizations at one university campus. Through interviews with activists and administrators, a detailed timeline is constructed and analyzed through the framework of Communicative Constitution of Organizations (CCO) to understand the communicative flows that comprised this organizing (McPhee, Myers &amp; Iverson,... https://escholarship.org/uc/item/4s77s1cg Sat, 7 Feb 2026 00:00:00 +0000 Oak, Bedlam https://escholarship.org/uc/item/4kg2n3cb Software bugs continue to pose significant challenges to modern society, causing considerable economic impact, slow-down of fundamental research, and, in the worst case, catastrophic physical consequences. As bugs evolve into security vulnerabilities, the risk of intentional exploitation carried by malicious actors escalates, potentially leading to dire consequences for human rights and national security. Thus, identifying and addressing the root causes of software vulnerabilities (at scale) became crucial. However, automated vulnerability identification is an inherently complex task. First, the diversity and complexity of modern software systems require an understanding of many domain-specific details, making it impossible to create a one-size-fits-all solution. Secondly, automated security analyses need to strike an optimal balance between precision and efficiency: catching as many instances of a class of vulnerability as possible, while reducing false positives. This dissertation... https://escholarship.org/uc/item/4kg2n3cb Sat, 7 Feb 2026 00:00:00 +0000 Gritti, Fabio https://escholarship.org/uc/item/4j72s68g Multicellular systems are a specific class of biological organization where cells no longer act independently but as a community. These systems strongly benefit from cell specialization, allowing for growth, structural complexity, and adaptation to changing environmental conditions. This thesis presents work on how structure and function are linked in two examples of multicellular systems: microbial granule composites and glioblastoma tumors. First, we examine the relation between structure and function in "pink berry'' granular biofilms with multiscale structural and mechanical characterization, to determine the extent to which microscale heterogeneity influences macroscale material properties. Our methods reveal that by computing the structural composition of cell microcolonies, we can relate individual microscale mechanical contributions from cell microcolonies and the extracellular matrix to the macroscopic properties crucial for pink berry biofilm function. This multiscale... https://escholarship.org/uc/item/4j72s68g Sat, 7 Feb 2026 00:00:00 +0000 Maansson, Lisa https://escholarship.org/uc/item/4f074505 Layered sodium transition-metal oxides are promising cathodes for sodium-ion batteries. One concern with sodium layered oxide materials is the existence of multiple crystalline phases, as controlled by sodium content. 6-coordinate sodium can occupy either octahedral or prismatic sites between the metal oxide layers, which affects anion packing and symmetry. As such, the electrochemical behavior of the material is fundamentally altered. Careful control of elemental composition and corresponding structure is therefore key to unlocking advanced electrodes for sodium-ion batteries. This work investigates whether rapid microwave-assisted synthesis can reliably produce phase-pure layered NaxMn1/2Fe1/2O2 and compares furnace- and microwave-prepared P2 analogs. Microwave synthesis yielded phase-pure P2-Na2/3Mn1/2Fe1/2O2 and uniquely enabled the direct synthesis of a metastable, phase-pure P3-NaMn1/2Fe1/2O2, which is not accessible by conventional high-temperature routes. Products were... https://escholarship.org/uc/item/4f074505 Sat, 7 Feb 2026 00:00:00 +0000 Moya, Enrique Canales https://escholarship.org/uc/item/49p6q47z This dissertation traces the creation of the National Society of the Colonial Dames of America (NSCDA) from its founding in 1891 ending in the 1960s with the California State Society of the NSCDA’s work furnishing La Casa de Estudillo in San Diego. The NSCDA is a patriotic hereditary society open to women who can trace their lineage to a colonial ancestor who resided in one of the Thirteen Colonies prior to 1750. Unlike other patriotic hereditary societies, the NSCDA identified itself primarily as a historic preservation organization and operates a large number of historic house museums across the United States. I argue that the NSCDA should be central characters in historians’ emerging understandings of nineteenth- and twentieth-century historic preservation and development of the field of public history. The NSCDA’s trajectory into historic preservation disrupts current conceptions of the field “professionalizing” through the exclusion of amateur women’s societies at the turn... https://escholarship.org/uc/item/49p6q47z Sat, 7 Feb 2026 00:00:00 +0000 John, Emma Lou https://escholarship.org/uc/item/3vk63799 Dynamic nuclear polarization (DNP) is a powerful technique designed to overcome the inherently low signal sensitivity of solid-sate nuclear magnetic resonance (SSNMR). DNP enhances signal by transferring polarization from highly polarized unpaired electron spins to nearby nuclei via microwave irradiation near the electron spin resonance (ESR) line, leading to enhancements of up to several orders of magnitude. However, widespread adoption of DNP remains limited by instrumentation accessibility, commercially available polarizing agents, challenges with sample preparation, and an incomplete understanding of underlying spin dynamics governing polarization transfer.This dissertation advances the understanding of the electron spin dynamics underlying efficient, low-power DNP. We investigate two promising polarizing agents: water-soluble BDPA (SABDPA) and P1 center diamond. SABDPA, a narrow-line radical exhibiting similar properties to the common polarizing agent BDPA, has demonstrated... https://escholarship.org/uc/item/3vk63799 Sat, 7 Feb 2026 00:00:00 +0000 Tobar Tobar, Celeste Aydee https://escholarship.org/uc/item/3jm8f2pj Southern California shrublands are highly diverse and include many endemic species found nowhere else on earth. These ecosystems are also increasingly at risk of degradation and loss due to urbanization, altered disturbance regimes, and the spread of invasive non-native annual grasses. Following disturbances like fire, non-native annual grasses can become abundant and outcompete shrub seedlings for resources leading to the decline or extirpation of the shrub community. Although the drivers of this vegetation conversion have been well studied, strategies for restoring degrading, non-native invaded shrublands, particularly on remote hillslopes, remains less understood. Additionally, vegetation conversion has been treated as an “all or none” condition with little attention paid to native species that might be able to persist within or recruit into these degrading sites. To address these knowledge gaps, this dissertation examines how shrubland communities change along a gradient of... https://escholarship.org/uc/item/3jm8f2pj Sat, 7 Feb 2026 00:00:00 +0000 Lucero, Stephanie https://escholarship.org/uc/item/3898w9fd In response to the growing need for climate change education, this study explores how the Global Competence framework, in conjunction with the SAGE framework (Scientific Content Knowledge, Action-Oriented Learning, Global-Local Connectivity, and Equitable Climate Advocacy), supports teachers in integrating climate literacy into their classrooms. Through a yearlong professional development program, 12 teachers from California coastal counties engaged in workshops, follow-up sessions, and collaborative learning experiences designed to strengthen their instructional capacity. This qualitative study draws on multiple data sources—including surveys, interviews, teacher-developed pedagogical materials, and action plan blueprints—to assess how teachers’ understanding of climate literacy evolved and whether their proposed approaches aligned with their actual classroom implementation.Findings indicate that teachers deepened their climate literacy knowledge through Scientific Content Knowledge,... https://escholarship.org/uc/item/3898w9fd Sat, 7 Feb 2026 00:00:00 +0000 McNish, Donald W https://escholarship.org/uc/item/32m9g313 This thesis examines India’s pervasive and asymmetrical state control of Hindu institutions as a unique, state-led extractive framework over its majority faith – a global anomaly in secular democracies distinct from typical patronage or separation models. It argues that this system is sustained by three interlocking incentives: ideological control by maintaining a “custodial logic,” political patronage by appointing loyalists to temple boards, and financial extraction by diverting devotee donations to state projects. I argue that this “custodial state” is a hybrid, fusing colonial bureaucratic tools with the claimed moral authority of traditional kings. This work traces the system’s genealogy from its colonial foundations to its judicial codification as “Constitutional Hinduism.” Using legislative analysis and empirical case studies from Uttarakhand and Rajasthan, this thesis substantiates this extractive reality, calling for a new research agenda based on proven, resilient models... https://escholarship.org/uc/item/32m9g313 Sat, 7 Feb 2026 00:00:00 +0000 Shah, Jignyasu https://escholarship.org/uc/item/30v4s4np This doctoral dissertation examines the effects of income redistribution on electoral participation, voter preferences, and crime reduction. Empirically, it exploits the rollout of minimum wage increments in northern Mexico to operationalize the positive variation of the income distribution. Using cross-sectional data from INEGI, this work finds that income redistribution depresses turnout by incentivizing outmigration, stimulates right-wing voting in legislative elections, and discourages municipal-level crime incidence. These results have important implications for assessing the prospects of equitable democracy, the emergence of anti-redistributive preferences, and the effectiveness of nonviolent security policies. https://escholarship.org/uc/item/30v4s4np Sat, 7 Feb 2026 00:00:00 +0000 Moreno Plascencia, Jorge Raúl https://escholarship.org/uc/item/2s08d1v7 Cells and biomaterials often rely on electrostatically driven phase separation to organize macromolecules, yet the physical principles linking polymer structure to condensation remain incomplete. This dissertation investigates how flexibility and secondary structure control complex coacervation of charged polymers through two complementary approaches: one in which flexibility is modulated by sequence-encoded secondary structures and the other in which it is constrained by externally applied force. DNA–poly-L-lysine (PLL) mixtures were used to isolate the effects of nucleic acid conformation. Comparisons of single-stranded, duplex, and hairpin constructs revealed that structural differences alter both dissolution salt concentration and dense-phase composition. Rigid duplexes packed more tightly within the dense phase, producing higher concentrations but dissolving at lower salt concentrations, whereas more flexible single strands formed less dense but more stable coacervates.... https://escholarship.org/uc/item/2s08d1v7 Sat, 7 Feb 2026 00:00:00 +0000 Nguyen, Anna https://escholarship.org/uc/item/2k87z363 California’s 2012–2015 drought imposed severe water limitations on agriculture, with uncertain long-term impacts on perennial crops. Using NASA Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) imagery, we examined crop canopy responses in southern Santa Barbara County between 2013 and 2018. AVIRIS mosaics were calibrated and classified with Random Forest to identify stable parcels of avocados, citrus, nut trees, and wine grapes. Vegetation indices (VIs) representing canopy greenness (NDVI, SAVI, EVI), water status (NDII, NDWI, MSI, GVMI), and pigment dynamics (CI-Red Edge, MCARI, PSRI) were extracted and compared across years.All crops declined in canopy condition during drought, but trajectories differed by species. Wine grapes showed the steepest and earliest declines, while avocados declined more gradually and nut trees and citrus showed intermediate responses. Greenness indices consistently tracked canopy decline, water-sensitive indices were variable and influenced... https://escholarship.org/uc/item/2k87z363 Sat, 7 Feb 2026 00:00:00 +0000 Begin, Renee https://escholarship.org/uc/item/22t9n951 This thesis was prepared in collaboration with the University of California, Santa Barbara Special Research Collections and is hugely indebted to the preservation work performed by Gitta Alpár’s family. Gitta was a Jewish Hungarian opera and operetta singer who reached the heights of her career in early 1930s Berlin just before she was forced to flee. The donation of Gitta Alpár’s papers to UCSB has allowed me to compile the most complete and comprehensive biography of this glamorous and influential figure, enriching our understanding of performing life in 1920s and 30s Europe, the transatlantic emigration of persecuted Jewish stars, the radio scene in Buenos Aires, Argentina, and the fate of emigre stars in post-WWII United States. The rich evidence supporting this project includes Gitta’s own diligently saved newspaper clippings and magazine articles from Europe, Argentina, and the United States, as well as a wealth of photos, personal correspondence, and legal documents. By... https://escholarship.org/uc/item/22t9n951 Sat, 7 Feb 2026 00:00:00 +0000 Park, Caitlyn https://escholarship.org/uc/item/1zt88480 This dissertation uses the simple, stereotyped nervous system of the tunicate Ciona robusta to investigate how cell identities are specified during embryogenesis and how ancestral brain regions evolved across chordates.The Ciona larval central nervous system (CNS) comprises 177 neurons and is the only chordate CNS with a complete connectome; its CNS derives from a 40-cell neural plate with highly reproducible lineage. First, I develop and validate a framework to assign spatial identity to single-cell RNA sequencing (scRNA-seq) clusters from the Ciona neural plate. By integrating published Ciona scRNA-seq datasets and applying differential expression analysis, I generated candidate marker sets for discrete neural plate territories based on a manually curated compendium of in situ expression patterns. Targeted in situ hybridization confirmed the predicted spatial domains for select clusters, revealing transcriptional continuity from early blastomeres to their descendants and providing... https://escholarship.org/uc/item/1zt88480 Sat, 7 Feb 2026 00:00:00 +0000 Miao, Yishen