http://www.rssboard.org/rss-specification 720 https://escholarship.org/uc/ucsd_etd/rss Recent eScholarship items from UC San Diego Electronic Theses and Dissertations Fri, 15 May 2026 08:01:30 +0000 https://escholarship.org/uc/item/4d92w7v3 This thesis examines the spatial organization and discard practices in lithic production within the Tiwanaku colonial settlement of Río Muerto M70C archaeological site in the Moquegua Valley, Peru, to determine the function of the site. Legacy surface collection data from the 1998-1999 Moquegua Archaeological Survey were digitized and analyzed using lithic typology, raw material identification, and Geographic Information Systems (GIS) spatial analysis. The analysis examined artifact distributions, raw material patterns, and the relationship between lithic and ceramic concentrations to assess the organization and scale of production. Results show that lithic concentrations are dominated by locally sourced chert and chalcedony, with no exotic materials, and include a few complete tools. High-density lithic areas are offset from ceramic clusters, a pattern suggesting a structured discard of sharp lithic waste away from domestic areas. Ethnoarchaeological comparisons support interpreting... https://escholarship.org/uc/item/4d92w7v3 Wed, 6 May 2026 00:00:00 +0000 Gaspar, Karla P https://escholarship.org/uc/item/3189t63g Aqueous Zinc-ion batteries (AZIBs) are a promising battery chemistry for research due to their low cost and non-flammable electrolyte. One of the main problems AZIBs face is uneven Zn deposition on the electrodes, which leads to capacity loss. A conductive carbon 3D scaffold was proposed to address this issue, and four carbon scaffolds were tested: Acetylene Black (AB), Super P, Graphene, and Ketjen Black (KB). AZIB cells with each scaffold type were manufactured, cycled, and then imaged in a scanning electron microscope (SEM). The KB scaffold exhibited a “bottom-up” Zn deposition, while the AB, Super P, and Graphene scaffolds had Zn depositing only on the surface. The scaffold materials also underwent overpotential, BET porosity, electrochemical impedance spectroscopy (EIS), and cyclic voltammetry (CV) tests to better understand this deposition behavior. The overpotential and CV tests were used to calculate the carbons’ active surface areas, the EIS data was used for estimating... https://escholarship.org/uc/item/3189t63g Wed, 6 May 2026 00:00:00 +0000 Brookshire, Faith Ann https://escholarship.org/uc/item/97n035sz Intertidal tidepools experience rapid environmental fluctuations that strongly influence microbial metabolism and nutrient cycling over short timescales. However, the extent to which benthic composition structures microbial community function during low-tide isolation remains poorly resolved. Here, we examined three tidepools at Scripps Coastal Reserve that differed in benthic functional composition and heterotrophic dominance to assess how benthic structure shapes microbial diversity, biogeochemistry, and metabolic pathways during isolation periods.Benthic composition differed markedly among pools, with the mid-heterotrophic pool exhibiting the highest macro- and microbial diversity. Dissolved nutrient dynamics reflected heterotrophic dominance, with greater accumulation of ammonium in pools with higher heterotrophic biomass. Microbial community composition varied significantly by pools’ benthic composition and time of day, with time of day explaining the largest proportion of... https://escholarship.org/uc/item/97n035sz Mon, 27 Apr 2026 00:00:00 +0000 Renssen, Bibi Odilia https://escholarship.org/uc/item/7vv2v56q Background: Prenatal alcohol exposure (PAE) is a major risk factor for adverse infant outcomes, yet the role of selected maternal micronutrients (SMMNs), including choline, folate, and iron in modulating these effects remains unclear. These nutrients may influence fetal growth and early neurodevelopment, potentially mitigating or exacerbating alcohol-related deficits. Methods: Using longitudinal data from the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD) cohort in Western Ukraine, this dissertation includes three linked studies Study 1 (N=368; third-trimester subset N=111) examined associations between PAE and SMMN levels across gestation using generalized additive model. Study 2 (N=368; third-trimester subset N=111) assessed whether SMMNs mediated associations between PAE and infant growth (birth weight, length, head circumference). Study 3 evaluated infant psychomotor and cognitive outcomes (Bayley Scales PDI and MDI ≤12 months) in 290 mother–infant dyads,... https://escholarship.org/uc/item/7vv2v56q Mon, 27 Apr 2026 00:00:00 +0000 OLOLADE, STEPHEN OMOTAYO https://escholarship.org/uc/item/1694j5f0 LoRa is a new wireless technology that enables applications requiring long range, low power, low data-rate connectivity. Although LoRa has gained widespread academic interest, LoRa struggles to gain widespread, societal attention and adoption a decade after its inception. Arguably, the problem with the wireless technology’s adoption is that it is not clearly differentiated from existing wireless technologies. That is, alternative wireless technologies today (e.g. Wi-Fi, cellular, Bluetooth Low Energy) already—or will—achieve what LoRa currently has to offer. This dissertation claims that LoRa requires effective infrastructure deployment to realize the potential of LoRa technology and to deliver new applications. Specifically, deploying LoRa transceivers on non-terrestrial infrastructure (e.g. airplanes, unmanned balloons) can enable a new realm of wide-area, low power applications that can span entire metropolitan areas at fractions of the cost of traditional infrastructure.... https://escholarship.org/uc/item/1694j5f0 Mon, 27 Apr 2026 00:00:00 +0000 Yen, Alex https://escholarship.org/uc/item/8mx260r2 Over the course of my time at UC San Diego I’ve gotten the gift and grace to tackle a plethora of roles that forced me to expand and contract. I recall in my first meeting with faculty communicating my hunger to be pushed. To reach the outward borders of my craft, my voice, my body and my creativity. Something that I failed to mention and give weight to at that time was my authenticity. I got to play in the speech system and worn body of a callused hearted sailor who found companionship and love and just as quickly loses it, I got the chance to become an academic so tragically in infatuated with his fixed ideas of love and status that it blinded him to all else, I got to play a young scared boy searching for his best friend in a world of mystical odyssey and nightmare, I got to play a loyal but cowardly rogue whose words spoke a story into physical existence. I played a wrathful vulnerable father, a faux civil rights activist, a district attorney and a king on the verge of a losing... https://escholarship.org/uc/item/8mx260r2 Thu, 16 Apr 2026 00:00:00 +0000 Lebron, Germainne Joel https://escholarship.org/uc/item/3t72k6qw Solidarity manifests itself in and outside of la línea to those who cross the border. This thesis seeks to explore the sociability and solidarity molded by and for la línea by looking at two groups which have formed, developed, and grown in continued relation to this space. The WhatsApp group called The Border Fellowship, was formed by a group of four friends who shared their frustration with the long wait times at la línea. The group has since flourished into a semi-exclusive, invite only group with active members who share information, advice, favors, and, on occasion, socialize outside the confines of the mobile group. El club de la línea stands out as a small group of middle-aged women whose shared interest is crossing faster for work or getting their children to school on time. The group consisted of no more than 12 members at its peak, and extended its function to social events on weekends. Both of these groups forge deeper relationships through daily exchanges of care at... https://escholarship.org/uc/item/3t72k6qw Thu, 16 Apr 2026 00:00:00 +0000 Vargas Rangel, Yael Manuel https://escholarship.org/uc/item/9m3149dd Actor training is a crucible in which technique is refined and artistic identity is forged. I entered graduate study at UC San Diego after years of working professionally as both an actor and educator, confident that rigor and disciplined practice were sufficient for growth. What I encountered instead unsettled that belief. Technique alone cannot sustain development when the conditions of training also regulate the performer. When institutions demand vulnerability in the work while policing presence in the room, they produce a split in the actor’s process. Truth becomes something cultivated in performance but penalized in presence. Here, the process of becoming is not safe. The pursuit of deeper truth can expose the actor to misreading, restriction, and quiet forms of erasure that shape both craft and self.My early development as an actor was rooted in external transformation. I entered my undergraduate training at Howard University with little experience and a deep insecurity... https://escholarship.org/uc/item/9m3149dd Wed, 15 Apr 2026 00:00:00 +0000 Ray, Kenneth J https://escholarship.org/uc/item/7zd255gz All my life my body has been my best friend. My confidant. My lover. My storyteller. My escape. Decades of constant work for it to fail me when I need it the most. Here we go again. I thought to myself as November 2023 approached and I was in rehearsals for ‘The Promise.’ At least I am in California, I thought, even though I don’t know anyone too well I will still be okay. Well that was one of too many ‘I thought it was gonna be okay’ and it actually was not okay even in Sunny San Diego. Experiencing multiple surgeries changed how I move and inhabit my body. It royally (mime F***) messed me up. The Axons that connect everything in your brain; yeah that was obliterated. Though all hope felt lost; I still had community to support me during those times of crisis with dulces, amor, and most importantly SpongeBob. Ruby Purple, Pinkity Dinkinty, Melody red, Vivacious Green, Boop-Di-Boop, and Velocity Viscosity Black. These words, though they make no sense to the outsider, have encompassed... https://escholarship.org/uc/item/7zd255gz Wed, 15 Apr 2026 00:00:00 +0000 Lozano Jr., Ricardo https://escholarship.org/uc/item/7rw9t976 Teleoperation systems are essential for collecting diverse and high-quality robot demonstration data, especially for complex, contact-rich tasks. However, current teleoperation platforms typically lack integrated force feedback, cross-embodiment generalization, and portable, user-friendly designs, limiting their practical deployment. To address these limitations, we introduce ACE-F, a cross embodiment foldable teleoperation system with integrated force feedback. Our approach leverages inverse kinematics (IK) combined with a carefully designed human-robot interface, enabling users to capture precise demonstrations. We further propose a generalized soft-controller pipeline integrating PD control and inverse dynamics to ensure robot safety and precise motion control across diverse robotic embodiments. Critically, to achieve cross-embodiment generalization of force feedback without additional sensors, we interpret end-effector positional deviations as virtual force signals to enhance... https://escholarship.org/uc/item/7rw9t976 Wed, 15 Apr 2026 00:00:00 +0000 Paulsen, Lars Erik https://escholarship.org/uc/item/7dd778jb This dissertation details investigations into modifying the structure of aminocarbenes to achieve new functions. Chapter 1 introduces the carbene functional group, the structure and reactivity of which is highly sensitive to the nature of the substituents. In Chapter 2, the influence of a stereogenic α-carbon substituent on the oxidative addition chemistry of stable cyclic(alkyl)(amino)carbenes (CAACs) is explored. Synthetic experiments demonstrate that a relatively rigid chiral carbene, PhMeCAAC, can insert into a broad range of E-H bonds with excellent diastereoselectivity. Computational modeling provides insight into the kinetic and thermodynamic origins of the observed stereochemical outcomes. In Chapter 3, PhMeCAAC is used to construct new redox-responsive chiroptical switches with three isolable states. Aminocarbenes have long been used for designing redox-active molecules but only recently applied to structurally dynamic chiral molecules. This work merges these applications... https://escholarship.org/uc/item/7dd778jb Wed, 15 Apr 2026 00:00:00 +0000 Yorkgitis, Patrick https://escholarship.org/uc/item/61n1r3tt Higher-order ambisonics (HOA) offers a powerful set of methods for immersively recording and reproducing the three-dimensional soundfields of ecological environments, yet HOA content is significantly underrepresented in public soundscape archives. Existing ambisonic collections primarily feature first-order ambisonics or urban acoustic scenes, while the practitioners whose work most compellingly demonstrates the value of immersive ecological recording have largely not made their multichannel recordings accessible. This dissertation documents Ambisonic Earth, an ongoing project to record, process, and publicly archive the soundscapes of biodiverse ecosystems in higher-order ambisonics, while serving as a practical guide for others undertaking this work. Over 500 hours of third-order ambisonic soundscape recordings have been captured from 2022 through 2025 across 35 RESOLVE ecoregions and nine terrestrial biomes. Recording sites span 27 U.S. National Wildlife Refuges, 19 Key Biodiversity... https://escholarship.org/uc/item/61n1r3tt Wed, 15 Apr 2026 00:00:00 +0000 Deluga, Charles Augustus https://escholarship.org/uc/item/3vq5d2c9 As software systems become increasingly complex and pervasive, ensuring their correctness becomes even more critical since a single bug can ripple into catastrophic consequences. This growing demand for reliability, together with significant advances in automated reasoning, has driven renewed interest in formal verification. Auto-active verifiers are now used in industry to verify components, such as cryptographic code, authorization libraries, and operating systems. Despite these successes, however, formal verification remains a niche due to its substantial cost in time and required expertise. In this dissertation, I focus on those experts that verify code using auto-active verifiers to better understand their workflows, needs, and pain points, and to develop techniques that lower the barrier to adopting formal verification. To this end, I first interviewed practitioners who have verified real-world systems to characterize their expectations, practices and perceived opportunities... https://escholarship.org/uc/item/3vq5d2c9 Wed, 15 Apr 2026 00:00:00 +0000 Mugnier, Eric https://escholarship.org/uc/item/0zt5w9zn My graduate training at the University of California, San Diego, has been a crucible for my development as an artist. It was a journey on two parallel tracks. One focused on mastering classical acting techniques. The other focused on building a career. The synthesis of these two paths, craft and hustle, has forged in me not just a skilled actor, but a resilient and intentional artist.The core of my time at UCSD was rooted in the essential components of classical acting. I focused heavily on speech, voice, and movement. These techniques are designed to help you project to the back of a theater while maintaining an embodied, truthful presence. This training strengthened the cornerstone of my craft: emotional authenticity, embodiment, and transformation. It helped me tackle the challenges of sharing stories with larger audiences. Learning to fill a large performance space without losing intimacy or emotional truth has been central to my development as an actor. This foundation was... https://escholarship.org/uc/item/0zt5w9zn Wed, 15 Apr 2026 00:00:00 +0000 Tolson, Kennedy Makayla https://escholarship.org/uc/item/5bw4w35w Internal waves play a fundamental role in ocean mixing and the global overturning circulation, yet their energy pathways remain poorly quantified in energetic high-latitude regions. This dissertation investigates internal wave energy transport in the Southern Ocean and coherent internal tide detection in the Southern Ocean and the high-latitude North Atlantic, combining high-resolution ocean modeling and satellite altimetry.The first part quantifies the meridional redistribution of internal wave energy in the Southern Ocean using the 1/48° global MITgcm LLC4320 simulation. By decomposing fluxes into tidal, near-inertial, and continuum bands, I demonstrate that the semidiurnal tide is the dominant component of meridional energy transport, driving a net poleward flux of O(10) GW that accounts for over 80% of the total. The strong flux convergence near Antarctica implies diapycnal diffusivities consistent with observed background interior mixing rates in the Southern Ocean, suggesting... https://escholarship.org/uc/item/5bw4w35w Sat, 11 Apr 2026 00:00:00 +0000 Li, Youran https://escholarship.org/uc/item/0vk8m45c Intermittent computing systems harvest limited energy from the environment to power their sensing and computing work. Due to low and irregular power harvesting, these systems contend with unpredictable interruptions. Research has focused on ways to cope with the intermittent energy, leveraging soft and hard intermittency, and driving advances in intermittent computing hardware architecture, time-keeping, and software runtimes. However, intermittent computing systems are still difficult to design, program, and deploy. This work explores ways to simplify intermittent computing state management by exploring low power energy sources and analyzing boot and sleep performance. I present a module for real wall-clock timekeeping, leveraging extremely low power sources to maintain time indefinitely. I then analyze the true costs of hard intermittency for intermittent computing hardware platforms, and develop a model and analysis demonstrating that in most cases, especially where computing... https://escholarship.org/uc/item/0vk8m45c Sat, 11 Apr 2026 00:00:00 +0000 Marcano, Gabriel Eugenio https://escholarship.org/uc/item/7k91z6b1 Artificial intelligence (AI) systems are increasingly embedded in educational and cognitive support technologies, with growing ambitions to augment human intelligence through adaptive feedback. Advances in deep learning and large language models have enabled systems capable of generating context-aware responses and interacting dynamically with users. However, despite rapid algorithmic progress, the development of scalable and principled Human–AI integration systems has lagged behind expectations. A primary obstacle is the absence of systematic methodologies for modeling cognition under intervention and for training adaptive feedback policies without extensive human-in-the-loop data collection. Three fundamental challenges hinder scalable cognitive augmentation. First, collecting sufficient Human-AI interaction data to train such models directly from real users is costly and difficult to scale. Adaptive Human–AI systems thus require computational models that simulate how... https://escholarship.org/uc/item/7k91z6b1 Fri, 10 Apr 2026 00:00:00 +0000 Xu, Songlin https://escholarship.org/uc/item/5617r14f Understanding how neural populations give rise to behavior requires analytical frameworks that can disentangle high-dimensional, distributed, and context-dependent neural dynamics. Modern recording technologies now enable measurements across thousands of neurons, brain regions, and experimental conditions, but extracting interpretable structure from such data remains a fundamental challenge. In particular, neural computation is rarely localized to a single region or static across conditions; instead, it emerges from distributed interactions, selective engagement of neural subpopulations, and reorganization of population structure across behavioral contexts and task demands. This dissertation develops a set of interpretable machine-learning frameworks to study neural population dynamics from three complementary perspectives: reorganization across conditions, context-dependent feature relevance, and distributed computation across brain regions. First, we introduce Differential... https://escholarship.org/uc/item/5617r14f Fri, 10 Apr 2026 00:00:00 +0000 Sristi, Ram Dyuthi https://escholarship.org/uc/item/4ms0d1fb Large Language Models (LLMs) have demonstrated remarkable few-shot generalization, yet their capacity for autonomous self-improvement remains a critical frontier. This work investigates meta-learning paradigms across three fundamental dimensions: algorithmic emulation, continuous representation, and agentic self-refinement. First, we introduce MetaTree, a transformer trained via meta-learning to generate interpretable decision tree models rather than direct predictions. By learning from both greedy and globally optimized trees, MetaTree adaptively selects its splitting strategy based on data context. This approach consistently outperforms classical decision tree algorithms on held-out datasets and successfully generalizes to deeper trees than those seen during training. Second, we propose Vector-ICL, which extends the in-context learning capabilities of LLMs from discrete tokens to continuous vector spaces across diverse domains. Through lightweight embedding projectors... https://escholarship.org/uc/item/4ms0d1fb Fri, 10 Apr 2026 00:00:00 +0000 Zhuang, Yufan https://escholarship.org/uc/item/44x2t5kw While view synthesis has achieved photorealism at high framerates, existing methods like 3D Gaussian Splatting are primarily optimized for visual appearance. This makes them limited to observation of static scenes: they excel at VR walkthroughs but struggle in interactive domains like robotics and simulation, where scenes must react to physical changes. In parallel, the rise of video generation model challenges 3D view synthesis to find a niche, with it’s power to produce highly dynamic images, where the entire scene shifts moment to moment. For 3D view synthesis to fill a niche in robotics, it must become a bedrock of consistency. In this dissertation, we identity three underlying attributes that should remain valid even as the scene is manipulated, namely: material, density, and structure, and address each of them.First, we address Material Consistency. Traditionally, view synthesis has prioritized appearance over geometric accuracy, as retrieving high-fidelity geometry is an... https://escholarship.org/uc/item/44x2t5kw Fri, 10 Apr 2026 00:00:00 +0000 Mai, Alexander Tuan-Kiet https://escholarship.org/uc/item/1sv2w2fc As machine learning models increasingly influence high-stakes decisions in healthcare, finance, and criminal justice, the demand for interpretable decision logic has grown. This dissertation develops Boolean logic methods for learning and explaining rule-based classifiers on tabular data, addressing three core challenges: the discrete nature of logical structure, sensitivity to label noise, and the combinatorial complexity of explanation spaces. The first contribution introduces differentiable neural architectures that learn global decision rule sets through gradient-based optimization. By formulating Boolean operators as neurons with continuous weights and dynamic biases, the approach enables end-to-end training of disjunctive normal form classifiers while regularizing for sparsity. The resulting models achieve accuracy competitive with black-box methods while producing compact, human-auditable rule sets. The second contribution revisits two-level logic minimization as a machine... https://escholarship.org/uc/item/1sv2w2fc Fri, 10 Apr 2026 00:00:00 +0000 Qiao, Litao https://escholarship.org/uc/item/06k3t67h Adapting our thoughts and behaviors in response to changing environments or internal states (cognitive flexibility) is critical for achieving our goals. Prior work links beta oscillations (12-30 Hz) to motor inhibition during action-stopping and suggests analogous roles for beta in cognitive inhibition—such as removing irrelevant content from mind—to support switching. This dissertation examines the role of cortical beta in inhibiting competing motor tendencies and task rules across a variety of switching demands. Chapter 1 analyzes magnetoencephalography recorded during response conflict and implicates a signature of successful action-stopping—right-inferior frontal gyrus beta power—in potentially inhibiting incorrect actions during response-switching. Chapter 2 uses scalp electroencephalography (EEG) recorded during a rule-switching experiment to test whether after using a task rule, beta activity clears that rule from working memory to facilitate switching. Unique features... https://escholarship.org/uc/item/06k3t67h Fri, 10 Apr 2026 00:00:00 +0000 Daniel, Pria Lauren https://escholarship.org/uc/item/56b1q12w Aging is accompanied by widespread molecular changes, yet the relationships among these changes – and their ultimate causes – remain incompletely understood. This dissertation ix investigates the epigenetic and genomic alterations that accumulate during mammalian aging, their mechanistic origins, and potential strategies for intervention.First, I characterize the landscape of epigenetic aging across multiple layers of the epigenome. Through comprehensive analysis of six histone modifications and DNA methylation in >1,000 humans and mice, I demonstrate that age-related epigenetic changes are not isolated phenomena but rather reflect a coordinated remodeling process. Epigenetic changes across all layers converge upon a common set of genes, enabling the construction of a "pan-epigenetic" clock capable of predicting age from any epigenetic layer across species.Second, I establish that somatic mutations and epigenetic change are intimately coupled: mutations coincide with extensive... https://escholarship.org/uc/item/56b1q12w Wed, 8 Apr 2026 00:00:00 +0000 Koch, Zane https://escholarship.org/uc/item/4sp1z000 Modeling—the construction of selective, purpose-relative surrogates—connects scientific inquiry, engineering construction, and musical practice. This dissertation asks how models condition what becomes musically possible: what can be composed, played, built, and experienced. Drawing on the author’s work in real-time physical modeling sound synthesis, modular system design, sound installation, and performance, it follows technical practice into the conceptual territory it opens, developing a vocabulary that moves between formal description and musical experience without reducing either to the other.Time operates throughout as a premise rather than a topic. Temporal articulation is never simply given; it becomes accessible through concrete mediations—instruments, interfaces, notations, synthesis methods, and feedback loops. Audibility is constituted through time–frequency entanglement; space and material structure become available as temporally organized response. Musical coherence... https://escholarship.org/uc/item/4sp1z000 Wed, 8 Apr 2026 00:00:00 +0000 Wang, Zehao https://escholarship.org/uc/item/4485s854 To transition from highly controlled industrial settings to scalable deployment in the real world, robots must be capable of safe interaction in uncertain, unstructured and even unknown environments. Classical robotics has excelled in structured settings, where environments are well modeled. Yet, the most impactful applications for modern robotics, such as the cramped deformable environments in surgery to the unpredictable layout of the urban pantry, are all defined by inherent uncertainty. One of the core challenges in autonomy is enabling robots to reason about these uncertainties, both in their own dynamics, and also in the external world, to ensure safety without sacrificing performance during complex tasks.This dissertation builds towards safe intelligent autonomy in unstructured and unknown settings by focusing on three motivating domains: Surgical Automation, Autonomous Navigation and Physical Manipulation. The core contributions of this work is a set of frameworks that... https://escholarship.org/uc/item/4485s854 Wed, 8 Apr 2026 00:00:00 +0000 Shinde, Nikhil Uday https://escholarship.org/uc/item/2tn6h91f This dissertation charts a trajectory in which advances in artificial intelligence (AI) play a central role in pushing the high-energy physics frontier, complementing progress driven by higher collision energies and larger colliders. The discovery potential of the LHC and future colliders relies on accurate reconstruction of increasingly complex particle collision events. In the CMS experiment, this task is performed by the particle-flow (PF) algorithm. This dissertation presents the first implementation of a machine-learning-based particle-flow (MLPF) reconstruction in the CMS detector based on transformer architectures. In simulated top quark–antiquark pair (ttbar) events under LHC Run 3 (2023–2024) conditions, MLPF improves jet energy resolution by 10–20% compared to standard PF for jets with transverse momentum between 30–100 GeV. Runtime performance is evaluated using simulated multijet events, with a median inference time of 20 ms per event on an NVIDIA L4 GPU, compared... https://escholarship.org/uc/item/2tn6h91f Wed, 8 Apr 2026 00:00:00 +0000 Mokhtar, Farouk https://escholarship.org/uc/item/2793v1v1 Bacterial transcriptional regulatory networks (TRNs) coordinate global gene expression to enable adaptation across diverse environmental and physiological contexts. The rapid expansion of publicly available RNA-sequencing data driven by decreases in sequencing cost provide unprecedented opportunities to study TRNs at scale. Leveraging these large transcriptomic collections, new analytical approaches are required to derive meaningful biological insight from high-dimensional data. In this thesis, we use iModulon analysis, a data-driven framework based on independent component analysis, to characterize TRNs and complex regulatory behaviors across multiple bacterial systems. First, we reconstruct the global TRN of Synechococcus elongatus PCC 7942, revealing regulatory organization underlying photosynthesis, circadian processes, and light sensing. We then investigate early antibiotic responses in Escherichia coli MG1655, identifying distinct transcriptional phases and highlighting... https://escholarship.org/uc/item/2793v1v1 Wed, 8 Apr 2026 00:00:00 +0000 Yuan, Yuan https://escholarship.org/uc/item/5bk8c1fn Li-ion batteries are a predominant power source for modern life, renowned for powering everything from electric vehicles to consumer electronics. However, these materials are vulnerable to lifetime capacity fade, fatigue, and degradation by electrode particle cracking with regular usage. Energy storage systems, particularly Li-ion batteries, withstand tightly coupled electro-chemo-mechanical processes that induce strain localization, drive microstructural cracking, and lead to eventual capacity fade and mechanical failure. This dissertation presents a modular computational framework that integrates human-constructed and machine-learned solution enrichment for predictive multiphysics modeling of heterogeneous microstructures. To embed microstructure-informed weak and strong discontinuities and facilitate direct image-based model construction, an interface-modified reproducing kernel particle method (IM-RKPM) is proposed. Accuracy and convergence properties of IM-RKPM for modeling... https://escholarship.org/uc/item/5bk8c1fn Tue, 7 Apr 2026 00:00:00 +0000 Susuki, Kristen Kimi https://escholarship.org/uc/item/7ff0m2tf Rocking structural systems incorporating unbonded post-tensioning (PT) tendons and supplemental energy dissipation (ED) devices represent a promising low-damage alternative to conventional seismic design strategies. While these systems exhibit favorable self-centering and damage-control characteristics, their nonlinear response, particularly after interface decompression and gap opening, poses challenges for simplified analytical modeling. The Monolithic Beam Analogy (MBA) provides a member-level compatibility relationship that, when combined with sectional equilibrium and material constitutive models, enables prediction of the monotonic moment–rotation response of jointed concrete members. However, its accuracy across varying geometries, axial load levels, and energy-dissipation mechanisms has not been rigorously established at ultimate limit state (ULS) conditions, particularly for the diverse connection details developed over the last decade.This thesis presents a comprehensive... https://escholarship.org/uc/item/7ff0m2tf Mon, 6 Apr 2026 00:00:00 +0000 Dadashi Jordehi, Shayan https://escholarship.org/uc/item/1d32k7w1 The rapid electrification of aviation and the rise of urban‑air‑mobility have moved lithium‑ion battery packs to the heart of next‑generation electric vertical take‑off and landing (eVTOL) aircraft development. These packs must simultaneously deliver high power, achieve high energy density, maintain thermal uniformity under aggressive discharge profiles, dissipate transient heat safely, and withstand mission‑dependent mechanical loads, all within stringent mass budgets. Conventional design strategies based on parametric sweeps or design‑of‑experiments explore only narrow regions of an immense design space and usually impose fixed, oversimplified heat‑source assumptions. Couplings among electrochemical, thermal, fluid‑dynamic, and structural behaviors are also typically neglected and treated in isolation by separate teams, leaving major gains in safety, lifetime, and mass efficiency untapped. This dissertation develops multiphysics level‑set topology optimization methods... https://escholarship.org/uc/item/1d32k7w1 Mon, 6 Apr 2026 00:00:00 +0000 Guibert, Alexandre Thomas Romain https://escholarship.org/uc/item/8vg8360k The XENONnT experiment, located at the INFN Laboratori Nazionali del Gran Sasso (LNGS), is primarily designed to search for rare events such as dark matter (DM) interactions in a liquid xenon (LXe) time projection chamber (TPC), with a particular focus on weakly interacting massive particles (WIMPs). However, its large target mass of Xe-134 and Xe-136 also enables sensitive searches for neutrinoless double-beta decay (0νββ), a process whose observation would conclusively demonstrate the Majorana nature of neutrinos and provide direct insight into the absolute neutrino mass scale. Building on techniques developed for direct DM searches, XENONnT can leverage its large target mass, excellent energy resolution, low background level, and robust calibration program to perform dedicated 0νββ analyses. In this thesis, we present a comprehensive sensitivity projection framework for 0νββ searches in XENONnT, together with a machine-learning-enhanced analysis strategy designed to improve... https://escholarship.org/uc/item/8vg8360k Wed, 1 Apr 2026 00:00:00 +0000 Zhong, Min https://escholarship.org/uc/item/74t8f2vx This dissertation details the intentions, logistical details, and reflections surrounding the events of MitchFest 2025. Specifically, this document is concerned with the way in which the sound installation; Shifting Textures; came into existence, the interactions between the installation and participants, and future usage of the ideas explored during the event.The opening chapter details general information concerning the festival, and how difficulties realizing the planned elements of the event were solved. This is followed by a detailed description of all the technical components that were used to construct the sound installation. To complete the documentation of MitchFest some time is taken to reflect on the artistic intentions and results that occurred during and after the festival.The final section briefly expounds a few potential projects that could be continuations, adaptations, or reiterations of the ideas presented and tested in MitchFest 2025. And to complete the document... https://escholarship.org/uc/item/74t8f2vx Wed, 1 Apr 2026 00:00:00 +0000 Carlstrom, Mitchell Kraig https://escholarship.org/uc/item/6cv8v30n Physically based rendering has achieved broad adoption and shows even greater potential across industries, powering film production, gaming, and high-fidelity synthetic data generation for training physical AI models. The film industry began transitioning to path tracing around 2015, followed by early adoption in game engines and interactive systems. At the same time, emerging world models and robotics systems increasingly face limitations from scarce or constrained real-world data, highlighting the need for scalable, photorealistic simulation. Rendering now sits at the intersection of entertainment, computer graphics, and machine learning. Growing pressure on frame budgets in games and visual effects, alongside rising demand for large-scale synthetic datasets, presents both challenges and opportunities for advancing rendering research. This dissertation responds to these challenges through complementary Monte Carlo and neural approaches spanning multiple scales of light... https://escholarship.org/uc/item/6cv8v30n Wed, 1 Apr 2026 00:00:00 +0000 Xu, Bing https://escholarship.org/uc/item/6b15s73c What if composing were a way of mapping out your life? This dissertation explores music composition as a form of sonic cartography—a journaling and place-making practice shaped by movement, listening, and encounters with place. Through field recordings, chamber works, electronic media, and multimodal projects, it traces the development of a compositional “way” across childhood influences, international study, and site-based creative work in Australia, the United States, and Thailand. Rather than advancing a single linear argument, the dissertation follows how ideas emerge through walking, recording, and responding to inhabited environments. Across these chapters, microphones, DAW sessions, and notation function as journals that capture experience and transform it into musical form. The project positions composition as a practice-informed mode of inquiry—where method grows through making, reflection, and return—and offers sonic cartography as a framework for understanding how sound... https://escholarship.org/uc/item/6b15s73c Wed, 1 Apr 2026 00:00:00 +0000 de Filippo, Stephen https://escholarship.org/uc/item/49h8h587 As the total number of refugees and asylees rises on a global scale, higher education is becoming more important to the resettled refugee and asylee population in the United States and other host countries (UNHCR, 2023a). As people from refugee backgrounds do not make the choice to resettle, have a choice of their new homes, and oftentimes do not have the same rights and opportunities as citizens, the ability to access higher education can be particularly impactful for this vulnerable population (Ramsay & Baker, 2019). Building on previous research, and utilizing Berry’s Model of Acculturation, this study will focus on the experience of students from refugee backgrounds in accessing higher education, while including the perceived barriers and unique support systems, as to better inform future policy and practice (Berry, 2005). https://escholarship.org/uc/item/49h8h587 Wed, 1 Apr 2026 00:00:00 +0000 Fujikawa, Corey Wylde https://escholarship.org/uc/item/3zf6g15s RNA interference (RNAi) is a common inhibitory regulation mechanism in mammalian cells where RNA molecules inhibit gene expression by mRNA degradation or translation inhibition. While many fungi naturally possess this machinery, the model yeast Saccharomyces cerevisiae (budding yeast) famously lost it over evolutionary time. By introducing DCR1 and AGO1, RNAi system was reconstituted in BY4741, a commonly used strain in S. cerevisiae yeast research. Prior to this, a mini-library of yeast shuttle plasmids were constructed by combining sfGFP fluorescence in addition to antibiotic resistance to indicate potentially positive plasmid construction after E. coli transformation. The “double-screening” strategy highly improves the efficiency of plasmid construction.Upon reconstructing the RNAi machinery in S. cerevisiae (BY4741), we observed intriguing single-cell bistability within the population, despite the cells being genetically identical and cultured under the same condition. While... https://escholarship.org/uc/item/3zf6g15s Wed, 1 Apr 2026 00:00:00 +0000 Shao, Zihang https://escholarship.org/uc/item/3vv634qf The PnC is a brainstem hub that receives cortical input and influences medullary respiratory centers. Because of its impact on respiratory modulation, the PnC is a region of interest when studying the effects of fentanyl. Opioid-induced respiratory depression (OIRD) is the primary cause of overdose mortality, yet the upstream neural mechanisms remain poorly understood. Could the PnC be implicated in the onset of OIRD after fentanyl treatment? To explore this, transcriptomic changes in the PnC will be analyzed from various experimental groups, which should provide some clues as to how the PnC is impacted by fentanyl exposure. This project aims to answer the following question: how does fentanyl exposure alter the transcriptomic profile of neurons in the pontine caudal nucleus (PnC)? Spatial single-cell transcriptomics data was used to compare gene expression in control mice and mice exposed to fentanyl, and specifically examined changes in VGAT and VGLUT (markers of inhibitory... https://escholarship.org/uc/item/3vv634qf Wed, 1 Apr 2026 00:00:00 +0000 Preska, Ryan https://escholarship.org/uc/item/3tm922h7 Background: Despite persistent tobacco control efforts, the prevalence of smoking among low-income populations served by Federally Qualified Health Centers (FQHCs) remains high. While evidence-based interventions for tobacco cessation exist, they are often implemented inconsistently across priority populations, highlighting the need to explore multilevel factors that influence their adoption. Methods: This dissertation utilized a mixed and multi-method approach across three studies guided by the Practical, Robust Implementation Sustainability Model. Study 1 was a cross-sectional analysis from 13,462 tobacco using patients which examined the relationship between patient, provider, and clinic-level characteristics and the receipt of tobacco cessation counseling. Study 2 employed a descriptive qualitative design to contextualize quantitative findings and explore barriers and facilitators to implementing tobacco cessation efforts through semi-structured interviews with 16 diverse... https://escholarship.org/uc/item/3tm922h7 Wed, 1 Apr 2026 00:00:00 +0000 Salgin, Linda https://escholarship.org/uc/item/1h44768t The timers in an operating system underlie everything including scheduler interrupts, radio timeouts, and user alarms. Despite their critical nature, timers are notoriously difficult to test for correctness. This thesis formally verifies the correctness of a timer virtualization layer in an embedded operating system. We prove that all timers are scheduled correctly, and every callback fires according to the OS specification. We introduce a framework to model timers and present an experience report on the state of retrofitted formal verification. https://escholarship.org/uc/item/1h44768t Wed, 1 Apr 2026 00:00:00 +0000 Rashid, Samir https://escholarship.org/uc/item/149695cq Most genetic studies have historically centered on single-nucleotide polymorphisms (SNPs) in individuals of European ancestry, overlooking both complex genomic variants and global diversity. This thesis focuses on improving our understanding and analysis of tandem repeats (TRs)—a major, yet underexplored, source of genomic variation—through a multi-ancestry lens. First, we developed EnsembleTR, a genotyping framework that harmonizes TR calls across methods, and used it to construct a reference panel of 1.7 million TRs from 3,550 individuals in the 1000 Genomes Project and H3Africa cohorts. This panel enabled various analyses, such as identifying ancestry-specific repeat expansions, expression quantitative trait loci (eQTL) mapping, and investigating sequence determinants of TR stability. To further improve our understanding of TRs, we built LongTR, a new tool for accurate TR quantification from both PacBio and Oxford Nanopore data, demonstrating superior accuracy and scalability... https://escholarship.org/uc/item/149695cq Wed, 1 Apr 2026 00:00:00 +0000 Ziaei Jam, Helyaneh https://escholarship.org/uc/item/0vq8c7jd The end-to-end (E2E) autonomous driving paradigm has achieved near-human performance on open-loop benchmarks by leveraging large-scale imitation learning. However, the evaluation landscape remains highly fragmented across disparate simulation engines. Open-loop evaluation protocols correlate weakly with closed-loop reliability, failing to account for the compounding kinematic errors that occur when policies are deployed in reactive environments. Furthermore, evaluating these models across different simulators introduces confounding variables related to visual rendering and vehicle dynamics, making it difficult to isolate genuine planning failures from simulator-specific overfitting.To resolve these structural deficiencies, this thesis introduces a unified, cross-simulator evaluation architecture built upon the MetaDrive physics engine. The framework enforces strict structural invariance by explicitly decoupling vehicle dynamics from the rendering pipeline. This is achieved by... https://escholarship.org/uc/item/0vq8c7jd Wed, 1 Apr 2026 00:00:00 +0000 Ravichandran, Abhijit https://escholarship.org/uc/item/0dx2q5vf Microplastics are increasingly recognized as behavioral stressors capable of changing predator– prey interactions in aquatic ecosystems, yet how different plastic types influence foraging performance remains poorly understood. This study tested the short-term effects of two microplastic types—conventional ELA plastic and biodegradable TPU plastic—on the feeding behavior of the mosquito fish Gambusia affinis. Fish were exposed to three concentrations of each plastic type and tested under two prey densities (5 and 10 mosquito larvae) following a three-day exposure period. Feeding performance was quantified from standardized five-minute video recordings of eating.Under low prey density (5 prey), fish exposed to ELA consistently showed reduced prey-capture rates across concentrations, whereas TPU treatments produced greater behavioral variability and concentration-dependent changes in feeding performance. A two-way ANOVA revealed a significant main effect of plastic type under the... https://escholarship.org/uc/item/0dx2q5vf Wed, 1 Apr 2026 00:00:00 +0000 Wang, Xinyu https://escholarship.org/uc/item/4z87j9vf Recent advances in machine learning have produced highly capable models; however, their ability to generalize beyond training data remains poorly understood. In practice, models are evaluated on predefined datasets but deployed in open-ended real-world environments, where out-of-distribution generalization—the ability to make reasonable predictions on previously unseen inputs—is critical. Understanding whether and how a model can generalize requires insight into its underlying decision logic. For example, when distinguishing handwritten digits such as 0 and 1, it is unclear whether a model relies on meaningful shape features or on spurious correlations that may not extend beyond the training distribution. Identifying such mechanisms is essential for assessing the reliability of model generalization to new tasks and domains. To address this challenge, we introduce OmniInput, a framework for evaluating model behavior in an input space that is far larger than the scope of traditional... https://escholarship.org/uc/item/4z87j9vf Mon, 30 Mar 2026 00:00:00 +0000 Liu, Weitang https://escholarship.org/uc/item/2qv9h26p Vision–language model (VLM)–based GUI agents have shown promise for automating computer tasks from natural-language instructions, yet they still struggle with long-horizon planning under partial observability, interface diversity, and compounding errors. We tackle this problem with an efficient, memory-augmented multi-agent framework that decomposes GUI automation into a high-capacity Planner and lightweight auxiliary modules—an Actor for execution and a Memory Manager for retrieval—so that compute and learning are concentrated where they matter most: high-level decision making.Viewing the system through a compute-allocation lens, we find that planner capacity is the dominant factor for long-horizon success, while execution and retrieval can be handled by substantially smaller models with minimal performance loss. Building on this insight, we introduce a planner-centric reinforcement learning method that optimizes only the Planner, using trajectory-level rewards provided by a... https://escholarship.org/uc/item/2qv9h26p Mon, 30 Mar 2026 00:00:00 +0000 WU, WENYI https://escholarship.org/uc/item/8475v2hd As strategic competition between the United States and China persists, Washington has repeatedly called on “like-minded” governments to coordinate their foreign policies with its own. What explains variation in the extent to which democratic states align their China policies with those of the U.S.? Traditional international relations scholarship treats states as unitary actors whose foreign policies reflect objective threat perceptions. I argue that this approach underestimates the political foundations of alignment decisions. Instead, I theorize that how a democracy navigates great power competition depends on two dimensions of its security relationship with the United States: its reliance on U.S. security support (dependence) and its confidence in U.S. commitment (precarity). Through a most-similar comparison of Australia and New Zealand—two antipodean democracies that diverge markedly in their alignment with Washington—I show that convergence with a confrontational U.S. stance... https://escholarship.org/uc/item/8475v2hd Sat, 28 Mar 2026 00:00:00 +0000 Liu, Yanchuan https://escholarship.org/uc/item/27h309nh Hypertension remains under-diagnosed and under-managed in precisely the communities least served by traditional devices. This dissertation advances a physics-guided, on-device approach to cardiovascular monitoring that prioritizes access, signal fidelity, and interpretability over black-box convenience. I contribute three systems that pair measurement physics with machine learning and validate them across devices and settings. First, I introduce BPClip, an ultra-low-cost mechanical attachment that enables oscillometric blood pressure measurement using commodity smartphone camera/flash hardware. A spring-loaded pinhole encodes applied finger pressure while the brightness fluctuation encodes pulsatile blood volume, allowing calibration-free oscillometry on-device. In a feasibility study (N=29, with 24 included in accuracy analysis), the system achieved mean absolute errors of 8.72/5.49 mmHg (SBP/DBP) and a small-batch material cost of about \$0.80 per unit, indicating feasibility... https://escholarship.org/uc/item/27h309nh Sat, 28 Mar 2026 00:00:00 +0000 Xuan, Yinan https://escholarship.org/uc/item/25p408ct Non-invasive imaging-based assessment of blood flow plays a critical role in evaluating heart function and structure. Computed Tomography (CT) is a widely used imaging modality that can robustly evaluate cardiovascular anatomy and function, but direct methods to estimate blood flow velocity from movies of contrast evolution have not been developed. This study evaluates the impact of CT imaging on Physics-Informed Neural Networks (PINN)-based flow estimation and proposes an improved framework, SinoFlow, which uses sinogram data directly to estimate blood flow. We generated pulsatile flow fields in an idealized 2D vessel bifurcation using computational fluid dynamics and simulated CT scans with varying gantry rotation speeds, tube currents, and pulse mode imaging settings. We compared the performance of PINN-based flow estimation using reconstructed images (ImageFlow) to SinoFlow. SinoFlow significantly improved flow estimation performance by avoiding propagating errors introduced... https://escholarship.org/uc/item/25p408ct Sat, 28 Mar 2026 00:00:00 +0000 Guo, Jinyuxuan https://escholarship.org/uc/item/5f43x9fz Typically, occurrences of precipitation are judged to be "unusual'' or "extreme'' based on an accumulation over a fixed time interval. Since long-term historical precipitation observations have been collected once per day, the most basic analysis assigns a probability to individual daily accumulations. In engineering applications, sums over fixed numbers of days (often, 2 through 30) are considered to be the basic unit of analysis. This work presents an alternative which may be considered in parallel to its predecessors.Precipitation events themselves may be defined as having random durations and the start and end times of the event are defined by stopping rule. This is motivated by the meteorology and history of California, which together inform us that it is oftentimes the persistence and apparent relentless of precipitation events which bring about disaster. Rather than iterate on existing methods, we may consider their shortcomings, and address the core assumptions which brought... https://escholarship.org/uc/item/5f43x9fz Thu, 5 Mar 2026 00:00:00 +0000 Weyant, Alexander James https://escholarship.org/uc/item/56x7r8qc Navigation through the space that surrounds us is essential to survival. Movement within an environment can serve a multitude of goals for mammals: migration, foraging, chasing prey, evading predators, and so on. Success in these tasks often involves complex interactions with one’s surroundings that must be computed and translated into goal-directed actions. Neural circuits supporting these interactions must dynamically integrate external information with existing spatial representations and self-movement processes. They must also be able to adapt their functionality to suit changing task demands, which may sometimes require rapid reference frame shifts: for example, from single-minded, high-speed pursuit to reorientation and finding one’s way back to the pack or den. To better understand the mammalian brain’s representation of goal-directed spatial cognition, we focused on key structures in the spatial system: the hippocampal formation (areas CA1, CA3) and the subiculum as wells... https://escholarship.org/uc/item/56x7r8qc Wed, 4 Mar 2026 00:00:00 +0000 Tung, Janet https://escholarship.org/uc/item/63p0p9b4 This dissertation is a comprehensive, multi-fidelity study of the aerospike nozzle flow field. An in-house method of characteristics (MoC) solver models the linearized, inviscid field. Euler and Reynolds-averaged Navier–Stokes (RANS) solutions are obtained with the finite-volume solver ANSYS Fluent. Time-resolved simulations of the turbulent flow field at supersonic and hypersonic conditions are conducted using a novel, high-fidelity Navier-Stokes solver. Viscous effects on a truncated aerospike nozzle flow are systematically investigated by comparing a linearized, inviscid flow model (MoC), an inviscid Euler flow model, and a viscous RANS model with slip and no-slip walls.A comparison of MoC and Euler solutions reveals that expansion waves originating from the thruster-module throat alter the jet expansion over the aerospike, potentially leading to flow separation. Viscous model simulations show that the development of a shear layer and boundary layer delays wake closure,... https://escholarship.org/uc/item/63p0p9b4 Wed, 4 Feb 2026 00:00:00 +0000 Pyle, Zachary Quinten https://escholarship.org/uc/item/5d71m4s3 Salsa choke echoed on Colombia’s streets during the 2021National Strike, uniting marginalized communities through rhythm, movement, and resistance. In Cali—dubbed the Salsa World Capital—a powerful social movement emerged from the historically neglected Aguablanca District: La Primera Línea, a grassroots collective that stirred a new generation to confront systemic inequality and violence. This essay explores salsa choke as both a cultural expression and a political tool, situating it within the historical continuum of son and salsa as mediums of Afro-diasporic resistance. By analyzing how La Primera Línea and protesters nationwide used music to reclaim public space and affirm cultural identity, this study highlights the enduring role of Afro-Colombian sonic traditions in the struggle for justice and inclusion. https://escholarship.org/uc/item/5d71m4s3 Wed, 4 Feb 2026 00:00:00 +0000 Zamudio Romero, Hernán Camilo https://escholarship.org/uc/item/5md676qt This dissertation looks at how the sonic identity of the oboe has been shaped and constrained by intertwined histories of pedagogy, aesthetics, and geopolitics, and how these legacies can be reimagined through listening, embodiment, and creative practice. I suggest that the Enlightenment ideal of the beautiful or singing tone, formulated alongside the category of timbre, still shapes expectations for how the oboe should sound and which bodies are allowed to sound it. In response, I turn to the idea of voice/timbre, treating oboe sound as material, embodied, and affective rather than as a fixed aesthetic goal. I follow the oboe from the French court of Louis XIV to present day discussions of timbre, close listening, and my own autobiographical reflections on Latin American canciones and the voice of Mexican singer Chavela Vargas (1919-2012). At the heart of the project is an ongoing collaboration with Mexican singer Mariana Flores Bucio, in which reimagined canciones and improvisations... https://escholarship.org/uc/item/5md676qt Wed, 21 Jan 2026 00:00:00 +0000 Gaona Villamizar, Juliana https://escholarship.org/uc/item/612831r6 Cellular signaling enables organisms to respond to environmental cues through intricate biochemical networks that regulate cell growth, differentiation, and survival. Among these, the MAPK pathway is a key regulator of cellular homeostasis, and its dysregulation is implicated in numerous cancers. The first part of this dissertation focuses on elucidating the molecular mechanism by which CRAF kinases interact with membranes for their activation. Using quantitative fluorescence microscopy on RAS-functionalized supported lipid bilayers, we reveal that the RAS-binding (RBD) and cysteine-rich (CRD) domains of CRAF exhibit positive cooperativity, in which RBD–RAS binding facilitates CRD–lipid engagement, and vice versa. This cooperative interaction establishes a two-step membrane-binding mechanism that strictly depends on RAS activation. Moreover, we discovered a lateral rebinding mechanism, wherein RAF extends its membrane residence time by rebinding to nearby RAS molecules. A weak... https://escholarship.org/uc/item/612831r6 Mon, 19 Jan 2026 00:00:00 +0000 Jimenez Salinas, Andres https://escholarship.org/uc/item/0dh3v9np Up to 80% of cancer patients suffer from a wasting syndrome, also known as cachexia, leading to ~30% of the cancer-related deaths. Currently, no pharmacological therapy has been approved in the United States for the treatment of cancer-induced cachexia. Cancer is frequently associated with chronic inflammation that leads to the disruption of normal metabolic homeostasis. Immunotherapies provide opportunity to modulate immunity and inflammation. In particular, virus-like nanoparticles (VLPs) have been developed as a promising vaccine platform and can be engineered with various compounds for targeted drug delivery against specific diseases, such as cancers, infectious diseases, and inflammatory diseases. VLPs are nanomaterials isolated from plant, mammalian viruses or bacteriophages are assembled without their infectious nucleic acid. Unlike traditional vaccines that often require aluminum-based compounds as adjuvants, VLPs can stimulate the immune system and act as a built-in adjuvant.... https://escholarship.org/uc/item/0dh3v9np Mon, 19 Jan 2026 00:00:00 +0000 Song, Ning-Yu https://escholarship.org/uc/item/8xn0698k RNA structural biology is a field dedicated to studying highly flexible macromolecules that can be thousands of nucleotides long. The best method to address these dynamic molecules is to use cryo-EM. Described here is an initial story investigating the flexibility of a group II intron RNP, to better understand the mechanism behind lariat formation in the first step of RNA splicing catalysis. The branch-site adenosine was found to interact in a base triple with another base pair in the branch-site helix, which positioned the 2’OH of the adenosine for splicing. This newfound understanding of RNA splicing was then compared to the spliceosome, the macromolecular machine that processes all pre-mRNA splicing within eukaryotes, and we explained how some severe human diseases are caused by disruption of this base triple. The second story is an investigation of riboswitch mechanics, using a newly developed group II RNA scaffold to facilitate this. The driving force behind this is that... https://escholarship.org/uc/item/8xn0698k Wed, 14 Jan 2026 00:00:00 +0000 Rudolfs, Boris https://escholarship.org/uc/item/75z1501z Over the past decade, deep learning models have grown by more than five orders of magnitude, pushing distributed training systems to their limits and driving major advances in system design. However, despite significant progress, a fundamental issue rooted in the nature of the input data, Input Irregularity, remains largely overlooked. Input irregularity refers to non-uniform, skewed, or highly variable patterns in the training data and workload, which can lead to severe system inefficiencies.This dissertation argues that addressing input irregularity is critical to the next generation of scalable deep learning systems and develops a set of irregularity-aware algorithmic and system-level optimizations to eliminate inefficiencies across the entire training stack. First, we tackle input-level workload variability in large-scale LLM training. We introduce WLB-LLM, a workload-balanced 4D parallelism framework that mitigates GPU imbalance caused by variable-length inputs through irregularity-aware... https://escholarship.org/uc/item/75z1501z Wed, 14 Jan 2026 00:00:00 +0000 Wang, Zheng https://escholarship.org/uc/item/6g54p6pk Sustainable computing efforts have traditionally focused on runtime efficiency. However, a significant fraction of the carbon emissions associated with computing systems are incurred not during use, but rather manufacture. These embodied emissions are responsible for 40% of the lifetime carbon footprint for server-class hardware, and as much as 80% for consumer electronics like smartphones. This work presents a new way of thinking about sustainable computing, in terms of both operational and embodied emissions. It explores one consequence of this thinking—that reducing the demand for newly manufactured hardware is an important strategy for emissions reduction. To this end, I propose the repurposing of consumer-class hardware as general purpose computing or sensing hardware. This is explored through several real-world deployments of repurposed smartphones, the results of which show that repurposed devices can provide a computing platform that is several times more carbon-efficient... https://escholarship.org/uc/item/6g54p6pk Wed, 14 Jan 2026 00:00:00 +0000 Switzer, Jennifer https://escholarship.org/uc/item/6bs9k46j In this dissertation, I investigate the rise and global diffusion of digital authoritarian stability, a durable institutional equilibrium in which authoritarian regimes leverage internet technologies to achieve domestic peace and entrench their rule. Building on the failed hypothesis that the internet would universally foster liberalization, this study examines how China pioneered an illiberal model of “internet sovereignty” that successfully decouples economic prosperity from digital freedom. I advance a multi-level theory explaining this phenomenon. First, I formalize a rational-choice model demonstrating that regimes rationally invest in internet censorship because the benefits of reducing opposition coordination and enhancing coercive capacity outweigh the costs of increased information asymmetry. Second, I propose an ideal point theory of internet censorship, arguing that the capability gaps between countries’ ideal and actual levels of internet censorship create a global... https://escholarship.org/uc/item/6bs9k46j Wed, 14 Jan 2026 00:00:00 +0000 Hoffman, Geoffrey https://escholarship.org/uc/item/65q560mt Ribosomes are large ribonucleoprotein complexes responsible for the process of translation, wherein the genetic information stored within an mRNA sequence is converted from the chemical language of nucleic acids into that of amino acids, resulting in a protein. During translation, amino acids are joined together in the peptidyl transferase center of the ribosome. This active site is formed exclusively by ribosomal RNA and the P-site and A-site tRNA substrates that hold the growing peptide chain and the incoming amino acid, respectively. The peptidyl transfer reaction forms a new peptide bond between the last amino acid added to the nascent chain and the amino acid held by the A-site tRNA: the alpha-amine of the incoming amino acid performs a nucleophilic attack on the carbonyl of the ester that links the nascent chain to the 3′ end of the P-site tRNA. The P-site tRNA thus acts as a leaving group while the nascent chain is handed off to the A-site tRNA via the newly formed peptide... https://escholarship.org/uc/item/65q560mt Wed, 14 Jan 2026 00:00:00 +0000 Hingey, Jason M https://escholarship.org/uc/item/6051k6ft The study of open quantum systems represents a frontier in modern physics, crucial for understanding diverse domains such as quantum optics, molecular physics, and quantum thermodynamics. These systems exhibit rich dynamical processes such as non-equilibrium driving, coherent dynamics, and dissipative environmental couplings. However, this complexity poses profound theoretical challenges. This thesis aims to establish the non-equilibrium Green’s function (NEGF) formalism and its many-body flavor, pseudo-particle (PP-) NEGF formalism as rigorous theoretical frameworks and demonstrate their capacity to provide a more accurate, and consistent understanding of open quantum systems by applying these methods to the field of optoelectronics, molecular physics and quantum thermodynamics, and critically evaluating the limitations of the competing traditional theoretical approaches previously used to study these fields.First, the NEGF technique is employed to evaluate approximations that... https://escholarship.org/uc/item/6051k6ft Wed, 14 Jan 2026 00:00:00 +0000 Li, Anqi https://escholarship.org/uc/item/4zp0q4pb Aerosols are small particles in the atmosphere made up of solid and/or liquid matter that come from many sources, natural and anthropogenic alike. They are a major cause of concern due to both their influence in environmental impacts, such as climate change through cloud seeding, as well as in human health through both pollution and airborne illnesses. The first section of this work focuses on understanding the chemical complexity of Sea Spray Aerosols, specifically, 1) the influence of air/water interfaces on medium chain fatty acid pKas and how that impacts their surface activity; 2) the role of surfactant type and organization on Burkholderia cepacia lipase’s behavior in aerosols with particular regards to water evaporation; 3) the impacts of chemical complexity in model SSA particles in terms of morphology and molecular organization. Building up towards human health related studies, the next section of this work investigates membrane protein systems including 1) mechanistic... https://escholarship.org/uc/item/4zp0q4pb Wed, 14 Jan 2026 00:00:00 +0000 Wauer, Nicholas https://escholarship.org/uc/item/3g35x415 Hardware has become an increasingly indispensable backbone of modern society, underpinning every layer of technological infrastructure from consumer electronics to large-scale industrial automation. The economic model behind hardware design and manufacturing is driven by the pursuit of productivity, cost efficiency, rapid development cycles, and uncompromising quality assurance. To achieve these goals, several industry trends have emerged, most notably the widespread integration of third-party intellectual property (IP) cores and the outsourcing of design and fabrication processes across a globalized supply chain. While these practices accelerate innovation and reduce production cost, they also introduce new vulnerabilities and trust concerns, as the origin, authenticity, and integrity of embedded hardware components become more difficult to verify and uphold. These concerns are amplified by a parallel shift in the deployment landscape, where custom hardware components that were... https://escholarship.org/uc/item/3g35x415 Wed, 14 Jan 2026 00:00:00 +0000 Li, Leon https://escholarship.org/uc/item/3fs0d4z5 Modern estimation theory for engineering applications is increasingly characterized by multimodal, high-dimensional state distributions and nonlinear, high-accuracy measurements with complex data associations. In this dissertation, we address the challenges of sampling-based multi-object tracking (MOT) and localization in complex 3D environments. I first focus on enhancing sampling efficiency in passive MOT scenarios. To this end, I incorporate particle flow (PFl) techniques into a belief propagation (BP) framework for sequential estimation of an unknown number of states in the presence of measurement-origin uncertainty. In particular, I develop a "flow-induced'' proposal PDF for importance sampling that consists of a weighted mixture of parallel flows performed for multiple candidate measurements. Particle flows are either modeled using deterministic or stochastic spatio-temporal processes. The resulting method for MOT is evaluated with synthetic data and in a challenging passive... https://escholarship.org/uc/item/3fs0d4z5 Wed, 14 Jan 2026 00:00:00 +0000 Zhang, Wenyu https://escholarship.org/uc/item/2zm2c1kt This dissertation develops a unified theoretical framework for understanding how collective light–matter interactions shape optical and chemical phenomena in molecular systems. It establishes how classical linear optics emerges from a fully quantum electrodynamical (QED) description, reconciling microscopic quantum models with the experimental success of classical optical methods in modeling polaritonic spectra. By employing large-N expansions and field-theoretic response theory, the work identifies clear boundaries between classical and genuinely quantum regimes of cavity electrodynamics. In high-Q photonic architectures, it predicts that vacuum fluctuations can imprint nonlinear spectroscopic signatures onto linear spectra, revealing a new class of QED effects accessible through linear optics. Extending beyond collective coupling, the dissertation introduces a cavity-enhanced Raman platform for achieving single-molecule vibrational strong coupling and demonstrates how few-molecule... https://escholarship.org/uc/item/2zm2c1kt Tue, 13 Jan 2026 00:00:00 +0000 Koner, Arghadip https://escholarship.org/uc/item/7mb6h9wh Our subjective impression is of a richly detailed visual world, yet empirical results reveal that observers can perceive and remember only limited information at any given moment. One resolution to this paradox is that people integrate observations across time and space to form ensemble representations – summary statistics that provide missing or noisy object information and afford observers a broad perceptual experience. This dissertation explores how people integrate information over space and time to circumvent resource limitations, what constraints govern these processes and what paradoxes emerge when ensemble perception mechanisms outperform explicit statistical displays. Chapter 1 investigates constraints on temporal integration by examining how people use kinematic properties to track multiple objects. When position information is ambiguous, observers can use velocity direction to maintain correspondence across time, but cannot use acceleration, revealing a hierarchy in... https://escholarship.org/uc/item/7mb6h9wh Mon, 12 Jan 2026 00:00:00 +0000 Wang, Yang https://escholarship.org/uc/item/6816w0d8 We prove that the boundary Dehn twist on K3#K3 \ B 4 is nontrivial in the smooth mapping class group, which provides another example of an exotic diffeomorphism on a simply connected spin four-manifold. We do so by finding an algebraic criterion which must be satisfied if the two maps are smoothly isotopic. The main tools involved are the Pin(2)-equivariant families Bauer-Furuta invariant, equivariant topological K-theory, and the Atiyah-Hirzebruch spectral sequence to show this algebraic criterion cannot be satisfied, and this establishes the result. As a corollary, we find any smooth bundle K3#K3 ,→ E ↓ S 2 has w2 (T vE) = 0, so E is spin. https://escholarship.org/uc/item/6816w0d8 Mon, 12 Jan 2026 00:00:00 +0000 Tilton, Scott Commodore https://escholarship.org/uc/item/5vq4d81t Cerebrospinal fluid (CSF)—the clear fluid surrounding the brain and spinal cord—plays a key role in maintaining the health of the central nervous system, yet many aspects of its motion remain poorly understood under both physiological and pathological conditions. This dissertation, organized in three parts, addresses open questions through a combination of theoretical modeling, magnetic resonance imaging (MRI), and computational fluid dynamics (CFD).Part I investigates the modeling and hydrodynamic effects of arachnoid trabeculae, the microanatomical structures populating the CSF spaces. Because modeling these fine structures explicitly is computationally prohibitive, homogenized descriptions are commonly used to represent their influence on flow and transport. However, most of these models are applied outside their strict range of validity, underscoring the need to assess their accuracy. We address this by solving canonical problems and comparing model predictions against CFD... https://escholarship.org/uc/item/5vq4d81t Mon, 12 Jan 2026 00:00:00 +0000 Lopez Nozaleda, Guillermo https://escholarship.org/uc/item/4hb1c6j6 Along the California coast, eelgrasses (Zostera spp.) are marine foundational species that support fisheries, stabilize coastal sediment and sequester carbon. Eelgrass beds are conserved and continually replanted to restore ecosystem function following disturbance from human activities such as dredging and pollution. This dissertation investigates a Zostera hybrid for suitability in restoration, testing the hypothesis that the hybrid population is resilient to low light stress. We take a genomic approach, assembling de novo genomes of California eelgrass species Zostera marina, Zostera pacifica and hybrid Z. marina x Z. pacifica, and use these resources to conduct a growth experiment to test the hybrid’s low light tolerance.The marine biome presents the next frontier for the genomic era. As long read sequencing technologies improve in accessibility and quality, we may envision a genomic catalogue of diverse marine life to support these kinds of conservation projects, along with... https://escholarship.org/uc/item/4hb1c6j6 Mon, 12 Jan 2026 00:00:00 +0000 Moore, Malia https://escholarship.org/uc/item/2ws3x4hx Large Language Models (LLMs) have become central to modern artificial intelligence, excelling at natural language understanding, multi-step reasoning, and open-ended dialogue. Yet despite their impressive capabilities, today’s LLMs remain fundamentally constrained by the static nature of their parameters. Once training concludes, these models cannot continually incorporate new information, refine skills, or adapt their internal representations over long time horizons. This limitation prevents current systems from functioning as persistent, autonomous agents—roles that demand lifelong learning, selective forgetting, long-term retention, and contextually grounded decision-making.This dissertation proposes the concept of LifeSpan Cognitive Systems (LSCS): agents that inhabit complex environments, accumulate multimodal experiences, and make decisions grounded in both recent and distant history. To unify the diverse approaches to long-term memory in intelligent systems, we propose... https://escholarship.org/uc/item/2ws3x4hx Mon, 12 Jan 2026 00:00:00 +0000 Wang, Yu https://escholarship.org/uc/item/0nm9w89b This project examines how identity stigmatized minorities make sense of state legibility projects in the context of welfare promises to legibility, and the extent to which such individuals engage with state categories in terms delineated by the state versus interpreting the categories through the lens of their own experiences. To do so, I draw on a first-of-its-kind, in-person survey of 530 working-class hijra/transgender women across three North Indian cities, various survey and vignette-based experiments, and ethnographic field insights. The first chapter asks why stigmatized minorities may seek to be legible to the state, an entity that historically oppressed them. Conventional wisdom suggests that marginalized groups, with strained relationships with the state, avoid being “seen” by it. Yet, legibility to the state also entails recognition, and recognition may lead minorities to access state welfare. The chapter finds that while fears of legibility persist, albeit in... https://escholarship.org/uc/item/0nm9w89b Mon, 12 Jan 2026 00:00:00 +0000 Baral, Siddhartha https://escholarship.org/uc/item/64j5c63k Octocorals are metazoans that prolifically produce terpenoid natural products, rivaling the chemical diversity of plants and microbes. However, until recently the genomic basis of this chemistry was unresolved. With the discovery of terpene cyclase (TC) genes encoded in octocoral genomes, a framework has emerged to trace the evolution of specialized metabolism in these animals. This thesis follows the genetic twists and turns that enable octocorals to generate novel chemistry and shows how functional characterization both validates genomic trends and provides mechanistic insight into enzyme evolution. Chapter 2 explores deep-sea octocorals, sequencing and profiling specimens of both major taxonomic orders collected at depths of ~1000 m. Chemical analyses revealed broad sesquiterpene diversity, while diterpenes were largely confined to five families (XBECK). Phylogenetic and functional analyses of deep-sea coral TCs, alongside mined public data sets, revealed rich sesquiterpene... https://escholarship.org/uc/item/64j5c63k Sat, 10 Jan 2026 00:00:00 +0000 Grayson, Natalie Eve https://escholarship.org/uc/item/5kf0n3jf Biological systems can be represented as networks of interacting molecules, cells, and tissues, where nodes denote components and edges their functional, physical, or regulatory relationships. Within these networks, communities are groups of nodes that interact more strongly with one another than with the rest of the network, providing a natural way to decompose biological organization across multiple scales. Systems biology seeks to explain how phenotypes emerge from this network architecture and its community structure. I organize my work around three questions: (Q1) Given a set of genes, what is their shared function? (Q2) How do different contexts reshape the organization of network communities? and (Q3) What do these principles reveal in a real disease, focusing on the hierarchical community structure of autism spectrum disorder (ASD)? In Aim 1, I evaluate large language models (LLMs), showing that GPT-4 can act as an omics assistant for inferring gene-set function and complementing... https://escholarship.org/uc/item/5kf0n3jf Sat, 10 Jan 2026 00:00:00 +0000 Alkhairy, Sahar https://escholarship.org/uc/item/0pt6j61x Deploying machine learning in high-stakes applications requires principled uncertainty quantification to ensure reliability when incorrect predictions carry undesired consequences. This dissertation develops novel methods for uncertainty quantification in deep learning and demonstrates their critical role in enabling safe decision-making across diverse domains.The thesis makes contributions in two parts. Part I develops novel uncertainty quantification (UQ) methods for time series data. We introduce two methods that leverage structure in probabilistic modeling: PECCO, a probabilistic equivariant neural architecture that exploits symmetries in multi-agent trajectory forecasting to improve calibration and sample efficiency; and Domain Adversarial Neural Process (DANP), a neural process model that can learn across different domains. Then, we introduce two works that leverage structure for post-hoc calibration in the framework of Conformal Prediction. CopulaCPTS, which uses copulas... https://escholarship.org/uc/item/0pt6j61x Sat, 10 Jan 2026 00:00:00 +0000 Sun, Sophia Huiwen https://escholarship.org/uc/item/9n32t20z Advances in wearable sensing technologies continue to transform the assessment of human movement, muscle engagement, and physiological function across sports, clinical, and military settings. Traditional systems such as optical motion capture (mocap) and surface electromyography (sEMG) provide valuable biomechanical and neuromuscular information, yet they remain constrained by their laboratory requirements, susceptibility to motion artifacts, and limited suitability for everyday or field applications. To address these limitations, this work introduces a comprehensive framework for flexible, self-adhesive, textile-based skin-strain sensors created by integrating thin, piezoresistive graphene nanosheet (GNS) films with unidirectionally stretchable kinesiology tape (K-Tape). These “Motion Tape” sensors conform to the skin, offer high sensitivity and linearity, and leverage the inherent compliance of K-Tape to provide stable, repeatable, and movement-artifact-resistant strain measurements... https://escholarship.org/uc/item/9n32t20z Wed, 7 Jan 2026 00:00:00 +0000 Lin, Yun-An https://escholarship.org/uc/item/8100s31v The increasing intensity and frequency of ecological disturbances are disrupting cycles of recovery on global coral reefs. A lack of broad-scale studies with parametrized wave impact data for tsunamis and cyclones has led to a lack of understanding of broad-scale hydrodynamic disturbance on reefs relative to our understanding of localized impacts. This study uses quantified metrics of wave strength to explore the broad-scale community effects of hydrodynamic disturbance events on the coral reefs of Samoa and Tonga after the 2018 Cyclone Gita and the 2022 Hunga Tonga-Hunga Ha'apai eruption and tsunami. We found that reefs were affected heterogeneously by these disturbance events, and by comparing the trends in change among islands, these events demonstrated the ability to increase turf cover, decrease coral cover, decrease reef structural complexity, cause significant community compositional change, and reverse regional trends in reef recovery. Further studies into broad-scale... https://escholarship.org/uc/item/8100s31v Wed, 7 Jan 2026 00:00:00 +0000 Flattery, Emma Katherine Ellise https://escholarship.org/uc/item/74p0s1vp A common failure that occurs in polymer matrix composites (PMC) is delamination which is a separation between the plies that can be induced by various phenomena including manufacturing defects and damage incurred during service and maintenance. Resin injection is an industry-standard technique used to repair delaminated composite materials using laminating resins (i.e. Henkel EA9396 and EA956 is specified by some organizations). This repair method is lower cost, less invasive, and faster than patch repairs, making it appealing for reduction of labor and material costs. However, the efficacy of this method is undercut by the difficulty of filling the delaminations completely and into small cracks (< 1 micron) due to the relatively high laminating resin viscosity (~3500 cP at 25°C). Current injection repair processes are considered non-structural due to repetitive drilling and incomplete crack filling. An investigation is ongoing into epoxies that have significantly lower viscosity... https://escholarship.org/uc/item/74p0s1vp Wed, 7 Jan 2026 00:00:00 +0000 Perera, Hiruni Hettiarachchige https://escholarship.org/uc/item/6rc0314s This study investigates the impact of the Large Eastern Pacific Marine Heatwave (LEPMH) from 2013-2016 on larval rockfish assemblages within the Southern California Bight (SCB). Marine heatwaves (MHWs), characterized by prolonged periods of anomalously high sea surface temperatures, have increased in frequency and intensity over recent decades, with significant ecological consequences. This research utilizes data from the California Cooperative Oceanic Fisheries Investigations (CalCOFI) program, which has been collecting plankton samples, including rockfish larvae, since 1949. The study specifically examines changes in species richness, evenness, and the abundance of various rockfish species groups (northern vs. southern, deep vs. shallow, targeted vs. non-targeted) before and during the LEPMH.Contrary to initial hypotheses, results indicated a significant decrease in both species richness and evenness during the heatwave. Additionally, declines were observed in the abundance... https://escholarship.org/uc/item/6rc0314s Wed, 7 Jan 2026 00:00:00 +0000 Bulkeley, Lucille https://escholarship.org/uc/item/6k3737d2 Ocean alkalinity enhancement (OAE) through carbonate mineral dissolution has the potential to increase the oceans storage capacity for atmospheric CO2 while mitigating ocean acidification. Here, using field observations and laboratory experiments, the reactions and dynamics of calcium carbonate sediments suspended in seawater were investigated as potential OAE applications. Seawater alkalinity anomalies were characterized along a visible plume of resuspended carbonate sediments formed by the passing of a commercial cruise ship across the Bermuda carbonate platform. Alkalinity anomalies along the plume were undistinguishable from the natural background variability with the seawater remaining at equilibrium or supersaturated with respect to the bulk Mg-calcite composition of the resuspended sediments. Laboratory free-drift experiments at different CO2 levels showed that dissolution of mixed sediments collected near the plume site were initiated at a seawater aragonite saturation... https://escholarship.org/uc/item/6k3737d2 Wed, 7 Jan 2026 00:00:00 +0000 Flemig, Steven https://escholarship.org/uc/item/3j20r7nk When crabs and other crustaceans molt, they face multiple challenges resulting from the transition to a hydrostatic skeleton. In the soft, hydrostatic state crabs are more vulnerable to predators because their claw defenses become impotent, and their mobility is reduced. This state is especially challenging for crabs that live and molt on land, where biomechanical constraints are greater and increase with body mass. The speed at which animals are able to recover from this process directly impacts their chance of survival. Here we examined postmolt mobility of the Atlantic ghost crab, Ocypode quadrata, which inhabits sandy beaches and is known for their fast crawling. We hypothesized that small, juvenile crabs would harden their new exoskeleton and resume crawling at high speeds quickly following ecdysis. We measured cuticle material properties (hardness and stiffness), podomere mechanical properties (flexural stiffness, EI, and critical buckling strength, σcr), and crawling kinematics... https://escholarship.org/uc/item/3j20r7nk Wed, 7 Jan 2026 00:00:00 +0000 Champagne, Natalie Ruth https://escholarship.org/uc/item/8k63h9cx Cells continually rebalance limited transcriptional and proteomic resources to survive and grow across environments, but the analytical (omics-scale) and mechanistic (model-based) tools that explain this balance often operate in isolation. This dissertation bridges that divide by integrating knowledge-enriched multi-omics data analytics with proteome-constrained genome-scale modeling to enable interoperable discovery and prediction. First, we show that the bacterial proteome is modularized in a manner that mirrors transcriptomic modules: matched proteome and transcriptome module components share gene content and condition-dependent activities, revealing how transcriptional and post-translational regulation shape proteome composition. Crucially, these modules enable inference of absolute proteome allocation directly from transcriptomic signals, quantifying resource trade-offs without requiring proteomics in every condition. Second, we integrate experimental data with Metabolism... https://escholarship.org/uc/item/8k63h9cx Thu, 25 Dec 2025 00:00:00 +0000 Patel, Arjun https://escholarship.org/uc/item/7d42j4hk Safety is a fundamental requirement in the control of autonomous and engineered systems operating in complex real world environments. The CBF method for safe control design was found helpful in many safety critical applications such as robotics, autonomous driving, unmanned areal vehicles, power systems, and multi-agent systems. This dissertation develops new theoretical and practical advances for safety-critical control using the framework of Control Barrier Functions (CBFs). While CBFs provide a powerful methodology for certifying and enforcing safety, a number of challenges remains in using CBFs as a safety framework. One challenge is guaranteeing CBF-based controllers are regular enough (continuous/ bounded). Without continuity some safety results are not applicable, and without boundedness practical applicability of such controllers is undermined. Another challenge is handling nonsmoothness within the CBF approach, which is exceedingly important in enforcing complex... https://escholarship.org/uc/item/7d42j4hk Thu, 25 Dec 2025 00:00:00 +0000 Alyaseen, Mohammed https://escholarship.org/uc/item/6w9154pr Electroencephalography (EEG) has been a fruitful tool to investigate brain mechanisms involved in cognition and disease. EEG analysis is dominated by the event-related potential (ERP) technique and the study of rhythmic oscillations. However, novel analysis techniques made it possible to also investigate aperiodic activity, characterized by its seemingly unstructured fluctuations in time. However, in the frequency domain it manifests as a clear, inverse relationship between frequency and power, measured from its spectral decay. Physiologically, aperiodic activity is hypothesized to reflect the aggregate excitatory and inhibitory drive in local brain regions. Historically, aperiodic activity was considered noise, masking its potential as a signal of interest.The aim of this thesis is to examine the functional relevance of aperiodic activity in relation to cognition, using causal neurostimulation methods. To do this, I first investigated how seizure-based treatments for patients... https://escholarship.org/uc/item/6w9154pr Thu, 25 Dec 2025 00:00:00 +0000 van Engen, Quirine https://escholarship.org/uc/item/5z05t218 Electron spins are promising resources for quantum technologies, and molecules offer additional advantages of higher tunability and scalability. In particular, organic diradicals host electron spins that are relatively long-lived, offering more avenues for engineered interactions with their environment. In the first few chapters of this Dissertation, I will discuss our recent works on designer organic diradicals that can function as optically addressable molecular spin qubits in quantum information science. These light-initiated platforms have traditionally been hosted in metallic complexes, and our design principles for an organic analogue have been supported by ab initio calculations and experimental implementations. Subsequently, I will discuss how triplet carbenes, which also exhibit diradical-like electronic structures, can display larger spin polarisations and improved spin readout mechanisms. This will be illustrated with a few realistic carbene candidates, designed with... https://escholarship.org/uc/item/5z05t218 Thu, 25 Dec 2025 00:00:00 +0000 Poh, Yong Rui https://escholarship.org/uc/item/42573614 Numerical models used to understand subduction zones and earthquake rupture are subject to uncertainty because the Earth’s interior cannot be directly observed. Robust sensitivity analysis can quantify the influence of model parameter variability on model outcomes, but this is challenging due to the computational expense of the models combined with the high number of parameters of interest. To address this challenge, we present a framework for model order reduction in geophysical applications. We use the interpolated Proper Orthogonal Decomposition to build fully data-driven, non-intrusive reduced order models (ROMs). We build ROMs that accurately approximate output from subduction zone thermal models and earthquake dynamic rupture models and are orders of magnitude faster to evaluate. We present ROMs for 2D kinematic-dynamic models of subduction zone temperature, and use them to quantify the variability in slab interface temperatures that results from model input variability... https://escholarship.org/uc/item/42573614 Thu, 25 Dec 2025 00:00:00 +0000 Hobson, Gabrielle Miranda https://escholarship.org/uc/item/2609w4m2 We prove local well-posedness as well as singularity formation for the g-SQG patch model on the plane (so on a domain without a boundary), with α ∈ (0,1/6] and patches being allowed to touch each other. We do this by bypassing any auxiliary contour equations and tracking patch boundary curves directly instead of their parametrizations. In our results, which are sharp in terms of α, the patch boundaries have L2 curvatures and a singularity occurs when at least one of these L2-norms blows up in finite time. https://escholarship.org/uc/item/2609w4m2 Thu, 25 Dec 2025 00:00:00 +0000 Jeon, Junekey https://escholarship.org/uc/item/1hz6b714 Neural networks have been used in a wide variety of practical settings to great success. To understand their success, much research has been done to develop fundamental principles for deep learning, whether by theory or empirics. A research area of particular interest is that of interpolation, in which a model achieves 100% accuracy on the data it was trained on. Classical statistical learning theory cautions against interpolation, in particular when noisy data is used, as it can lead to overfitting. Modern theories, however, propose that overfitting may in fact be benign, meaning the model can generalize despite overfitting to noisy training data. This thesis develops a refined picture of interpolation and overfitting in two parts. The first part of this thesis studies the in-distribution and out-of-distribution generalization behaviors of modern interpolating methods. We propose a taxonomy of overfitting for neural networks and kernel methods which characterizes the overfitting... https://escholarship.org/uc/item/1hz6b714 Thu, 25 Dec 2025 00:00:00 +0000 Mallinar, Neil Rohit https://escholarship.org/uc/item/9wd4g8qq We examine the fundamental physics of SQG and SQG+ theories through the study of point vortex systems, stability analysis of Kolmogorov flows, numerical simulation of turbulent flows. For SQG+ point vortex systems, we found that most of the conserved quantities in SQG are destroyed by the introduction of ageostrophic terms, but we found that inter-vortex distances are still fully conserved in a two-vortex system. In addition, we found stable polygonal and collinear configurations for SQG and SQG+ point vortex systems. In particular, it appears that the stability condition of $N < 7$ for a $N$-sided vortex polygon from~\cite{havelock1931} still applies in SQG and SQG+ based on our numerical results. Vertical movement of passive tracers are amplified with a large Rossby number when advected by a vortex polygon, and the tracer distribution at long times becomes more spread out toward larger depths. Furthermore, the far-field behavior of the vertical excursion scales with a power... https://escholarship.org/uc/item/9wd4g8qq Wed, 24 Dec 2025 00:00:00 +0000 Lee, Mac https://escholarship.org/uc/item/9mw3z2nq Understanding the category Modk(G) of the smooth representations of a p-adic Lie group G over a field k of characteristic p is integral in developing the p-adic and mod-p Langlands programs. However, the work of Peter Schneider, Matthew Emerton and others have suggested that we might need to shift our focus to the derived category D(G) of Modk(G) in order to make further progress. Noting that D(G) is a tensor triangulated category, we follow a common practice in studying tensor triangulated categories by attempting to classify the localizing tensor ideals of D(G). In this thesis, we obtain such classifications when G is compact and when G is abelian with a Noetherian augmented Iwasawa algebra. https://escholarship.org/uc/item/9mw3z2nq Wed, 24 Dec 2025 00:00:00 +0000 Rankothge, Bharatha Madusanka https://escholarship.org/uc/item/897290mh Effective natural language processing systems typically require extensive human annotations, creating a major bottleneck for deploying models on new tasks. This thesis develops methods that reduce the dependence on human supervision by exploiting the inherent structure of the data, task, and language models themselves.First, we present X-Class, which performs text classification using only class names by exploiting corpus-level distributional structure. Rather than requiring labeled examples, the method learns adaptive document representations that align with the given classes through clustering, allowing the corpus itself to provide supervisory signal. Specifically, X-Class estimates class representations by incrementally adding similar words, obtains document representations via class-attention mechanisms, and trains classifiers on confident pseudo-labeled documents. Extensive experiments demonstrate that X-Class can rival and even outperform seed-driven weakly supervised... https://escholarship.org/uc/item/897290mh Wed, 24 Dec 2025 00:00:00 +0000 Wang, Zihan https://escholarship.org/uc/item/85h7489m The human microbiome is composed of a vast number of microorganisms which change over the course of our lives. Some changes in microbiome composition are expected as we age, however, deviations from healthy microbiome trajectories could be indicative of accelerated aging or disease. Due to study heterogeneity, methodological limitations, and the need for advanced computational approaches to handle multi-omics data, analyzing the microbiome’s involvement across different body sites, disease states, and host associated factors remains challenging. The following dissertation addresses these challenges by leveraging machine learning and multi-omics integration to analyze microbiome datasets. The work outlines a multi-study analysis for identifying microbial signatures associated with signs of skin aging, multi-omics analysis revealing patterns in the progression of acne, and Transformer-based Robust Principal Component Analysis (TRPCA) for chronological age prediction from microbiome... https://escholarship.org/uc/item/85h7489m Wed, 24 Dec 2025 00:00:00 +0000 Myers, Tyler https://escholarship.org/uc/item/76t7z7n9 This dissertation traces the distinct identity of grunge as a unique subculture, a signifying practice and a genre of hard rock music from its temporal and geographical locations in the Pacific Northwest. In outlining three integrands of grunge, the project also defines The Seattle Sound within the grunge genre through textual and sonic mappings of a specific cultural geography affected by the urban cartography of Seattle and the dynamic meteorology of the surrounding Puget Sound. A soundscape of Seattle grunge is defined in this project through a small interconnected community of artists in the mid 1980s through the early 1990s situating its heavy sound as uniquely connected to place and communal practice.The terrains of style, artistic praxis, and vibrational acoustics are continually examined within this work paying careful attention to economies of noise that are affectual foundations of place. Urban improvement projects, commercial aerospace industries, and marine commerce... https://escholarship.org/uc/item/76t7z7n9 Wed, 24 Dec 2025 00:00:00 +0000 Zeiner, Linnea https://escholarship.org/uc/item/56f66397 Correlated oxides are known for their rich, stoichiometry-dependent phase diagrams, particularly in the context of structure, electronic transport, and magnetism. Various external stimuli, including temperature, voltage, and magnetic field, can drive phase transitions in these systems. A notable example is resistive switching (RS): a sudden change in a material’s resistivity under the application of an electric field. In RS materials, the presence of crystallographic defects can alter electronic and magnetic properties, producing changes in physical mechanisms driving the electronic phase transition. Functional design of disorder enables novel functionalities and energy efficient devices. In this dissertation, focused ion beam (FIB) irradiation was utilized to selectively induce defects in VO2, V2O3, and La0.7Sr0.3MnO3 to examine changes in their equilibrium and nonequilibrium transport behaviors. First, selective Ga+ irradiation was utilized to localize metallic filamentary formation... https://escholarship.org/uc/item/56f66397 Wed, 24 Dec 2025 00:00:00 +0000 Ghazikhanian, Nareg https://escholarship.org/uc/item/4bk5s0b0 Noble liquid detectors have a proven track record in searches for neutrino physics and dark matter searches. These include a series of world leading upper limits on WIMP-nucleon spin-independent cross sections in the GeV range, the most stringent lower limit on the neutrinoless double beta decay half life, and a detection of the elusive Coherent Elastic Neutrino Nucleus Scattering (CEvNS). This thesis is on the research and development of low-energy sensitive noble liquid detectors for nuclear recoil based physics searches. The two primary motivations for the development of such detectors are Coherent Elastic Neutrino Nucleus Scattering (CEvNS) from reactor antineutrinos, and low-mass dark matter in the few-GeV mass range. Both CEvNS and low-mass dark matter detectors need: low energy sensitivity – preferably to single-electron (SE) ionization signals, a large amount of target nuclei, and a low background. Such requirements are satisfied by both liquid xenon (LXe) and liquid... https://escholarship.org/uc/item/4bk5s0b0 Wed, 24 Dec 2025 00:00:00 +0000 Qi, Jianyang https://escholarship.org/uc/item/3m66x2wb The rapidly increasing requirements for higher data consumption and largely available spectrum at mm-wave frequencies demands higher data-rate communication systems. Although 5G is at its early stage in the current market, advanced research in sub-THz and THz communication has enabled beyond 5G(B5G) and 6G communication over short range. However, systems at D-band and higher frequencies suffer from large space path loss and demand much higher system complexity and performance requirements which leads to higher cost and design difficulty. As a result, phased-arrays and MIMO systems are needed and have been demonstrated for D-band short-range communication. To lower the cost of phased-arrays and MIMO systems for massive production, highly integrated chips based on silicon technologies (CMOS or SiGe), low-cost PCB designs and antennas are required.This dissertation presents the D-band receiver circuits, system architectures and measurement techniques to demonstrate the feasibility... https://escholarship.org/uc/item/3m66x2wb Wed, 24 Dec 2025 00:00:00 +0000 Wang, Changtian https://escholarship.org/uc/item/27g0s74r In theory the security of Internet communication depends on a variety of cryptographic protocols including ones for symmetric and asymmetric encryption, key exchange, signing messages, and generating random numbers. The security of these cryptographic protocols is based on underlying mathematical problems that are thought to be hard. Ideally, the protocols cannot be broken without attacking the underlying hard problems. However, in practice security depends on these protocols being correctly and securely implemented to keep them from leaking valuable information to an active, or even passive, attacker. While the security of implementations can be tested by a variety of means, including using code analysis and formal verification to improve source code, we focus on how existing security on the Internet can be probed by beginning with passive collection and analysis of real network traffic. We show that by passively looking at cryptographic data from real world connections, and... https://escholarship.org/uc/item/27g0s74r Wed, 24 Dec 2025 00:00:00 +0000 Sullivan, George Arnold https://escholarship.org/uc/item/1t20z9fc Explanation can be a powerful tool for learning, but the effectiveness of explanation may vary across learning contexts depending on what is being explained and to whom. We explored these questions through seven coordinated studies, examining explanations in the classroom and laboratory with varying levels of peer presence and interaction. In Chapter 3, we manipulated peer discussions and found that discussions were associated with stronger student performance, particularly when students had to apply discussed content to new contexts. Chapter 4 focused on self-explanation and prompt type in both synchronous and asynchronous classroom and laboratory settings. Self-explanation yielded more mixed results; although we did not find a consistent advantage for a written prompt to self-explain, we did find evidence in some studies that explaining one's own thinking was linked to better later performance than explaining the correct answer. This pattern also depended in part on whether... https://escholarship.org/uc/item/1t20z9fc Wed, 24 Dec 2025 00:00:00 +0000 Ibrahim, Dania https://escholarship.org/uc/item/1pn6h3r3 Modern optical sensing increasingly relies on detector platforms that combine high sensitivity, broad spectral reach, and compatibility with large-scale CMOS fabrication. Applications in quantum information, time-resolved imaging, autonomous systems, and biomedical diagnostics require reliable sensing under extremely low photon flux as well as effective imaging in the short-wave-infrared region. This dissertation develops two separate CMOS-compatible pixel technologies that address these needs across different operating regimes: a silicon-based avalanche device with controlled self-quenching and rapid recovery for high-throughput single-photon detection, and a monolithically integrated quantum-dot photodiode with thin-film-transistor amplification for scalable short-wave-infrared imaging. A complementary circuit-modeling framework captures the essential electrical behavior of these detectors, including avalanche dynamics, junction operation, noise processes, and voltage-mode... https://escholarship.org/uc/item/1pn6h3r3 Wed, 24 Dec 2025 00:00:00 +0000 Jiang, Yunrui https://escholarship.org/uc/item/15n3k2gh Connection to nature (CTN)—the extent to which individuals perceive themselves connected to the natural world—has emerged as a key construct in environmental psychology, predicting pro-environmental behavior (PEB) and well-being. However, existing measures of CTN face two critical limitations: 1) Most are uni-dimensional despite theoretical recognition that CTN encompasses cognitive, emotional, and behavioral aspects; and 2) many inadvertently position nature as separate from humanity—perpetuating the human-nature binary. This dissertation addresses these limitations through two studies. Study 1 (Chapters 1-3) systematically reviews 26 published CTN measures, identifies gaps in measurement, and develops the DEEP CTN Scale—a multi-dimensional measure emphasizing humans-as-integrated-with-nature. Using exploratory factor analysis across three samples (N = 1,637) and confirmatory factor analysis with both a student (N = 341) and general population sample (N = 315), a four-factor... https://escholarship.org/uc/item/15n3k2gh Wed, 24 Dec 2025 00:00:00 +0000 Lindsay, Debra Jane https://escholarship.org/uc/item/0rk3h9fr Signal processing techniques for audio often operate independently on incoming signals. However, it is increasingly useful to analyze and process signals in the context of other signals in an audio mix. This practice is most commonly employed in manual or automated audio engineering, but there are other applications in which signals might be processed so that a particular auditory or musical effect is achieved upon their combination. This work presents two such applications focusing on modulation control in a mix, outlines signal processing methods for each application and evaluates them both analytically and creatively using examples of music composition and sound design.Part I explores auditory roughness as a controllable parameter in music. Roughness most often manifests in mixtures of sounds with sinusoidal partials that are close in frequency. While roughness control is often linked to musical tunings and intervals, the methods presented here control the roughness of sound... https://escholarship.org/uc/item/0rk3h9fr Wed, 24 Dec 2025 00:00:00 +0000 Hyrkas, Jeremy https://escholarship.org/uc/item/0qj2226n Since the landmark adverse childhood experiences (ACEs) study in 1998, ACEs have remained a significant public health problem in the United States. Nearly two-thirds of adults report at least one ACE, with children living in poverty and children of color at disproportionately higher risk of experiencing ACEs and subsequent poor health and social outcomes. Latino children and families face compounding adversities including housing instability, discrimination, and immigration-related stressors. While strengthening protective factors through preventive care offers promising pathways to improve child trajectories, interventions face complex barriers to implementation and sustainability, particularly in low-resource settings.This dissertation advances understanding of how to address ACEs through three studies examining: 1) whether social and environmental protective factors can moderate adversity experiences in low-income families and build child resilience, 2) how multi-sector... https://escholarship.org/uc/item/0qj2226n Wed, 24 Dec 2025 00:00:00 +0000 Viglione, Clare