Dissertations

Permanent URI for this collectionhttps://hdl.handle.net/10161/4

For access to dissertations created prior to 2006 and those not submitted electronically, please see: https://library.duke.edu/find/theses-dissertations.

Policies & procedures governing electronic theses & dissertations can be found on the graduate school website.

Browse

Now showing 1 - 20 of 6952

Access status: Embargo ,
Solid State Heat Throttle for Pyroelectrics
(2026) Keys, Christopher
The work presented here demonstrates a solid-state solution for cooling pyroelectric sensors allowing for pyroelectric miniaturization, array implementation, and increased ﬁeldability by eliminating the need for an optical chopper. Leveraging the electrocaloric eﬀect in lithium niobate, a solid-state heat throttle was used in place of an optical chopper for signal modulation to prevent saturation of a pyroeolectic sensor. The sensor itself is also novel consisting of a translucent electrode design which shows 2.5 times amplitude improvement in signal capture over more traditional coating deigns in the range between 750-1400 nm. The sensor combined with the heat throttle produced superior signal capture with no optical optical chopper required.
Access status: Embargo ,
Essays on Tax Policy Optimization under Heterogeneous Treatment Effects
(2026) Antonacci Moura, Paulo Vitor
This dissertation studies tax compliance and enforcement in developing-country settings, combining causal inference methodology with field evidence from Indonesia. It consists of three essays that address, respectively, a large-scale door-to-door enforcement campaign, a randomized messaging experiment, and the econometric methodology that underpins the first two empirical analyses.The first essay, Whom Should the Taxman Visit?, examines the effects and optimal design of a citywide door-to-door property tax enforcement campaign in Kota Gorontalo, Indonesia, in which field officers visited more than 30,000 registered properties between 2020 and 2021. To estimate average and heterogeneous treatment effects, the essay proposes a pseudo-treatment causal forest estimator that extends the R-learner framework to multi-period settings with time-invariant unobserved confounding. Verified visits raised on-time payment rates by 7.8 percentage points in the first post-campaign billing cycle, an effect that persists for at least four subsequent years and is especially pronounced among lower-value properties and taxpayers near the compliance margin. Individualized treatment-effect estimates are then used to evaluate three targeting strategies derived from explicit public-finance objectives: a participation-maximizing rule, a revenue-maximizing rule, and a conventional risk-based rule. The participation-maximizing strategy delivers 17.3 percentage-point gains at the top decile above random assignment; the revenue-maximizing strategy generates substantial additional fiscal returns; and the risk-based rule---the standard operational heuristic in tax administration guidance---performs near zero on both margins. Concentration-curve analysis shows that these rules differ not only in efficiency but in the distributional composition of whom they reach, making the equity--efficiency trade-off in algorithmic enforcement explicit and measurable.The second essay, Does Message Framing Matter for Tax Compliance?, reports the results of a randomized controlled trial conducted with the municipal government of Gorontalo City. Delinquent property owners were randomly assigned to receive a soft-tone WhatsApp message emphasizing civic duty and public benefits, a hard-tone message highlighting penalties and consequences, or no message. The soft-tone nudge raised current-year compliance by 9 to 11 percentage points from two weeks through the payment deadline, and the effect remained 9.9 percentage points higher at six months relative to a 38 percent control mean. Effects were entirely concentrated among historically low-compliance taxpayers, for whom the intervention effectively closed the compliance gap with higher-compliance peers. Framing mattered at longer horizons: the soft tone significantly outperformed the hard tone, whose effects dissipated over time in a pattern consistent with intertemporal substitution. The intervention cost less than two cents per message, making it highly scalable for resource-constrained local governments.The third essay, Pseudo-Treatment Debiasing with Orthogonal Scores in Multi-Period Panels, develops the econometric framework underlying the causal analyses in the first essay and contributes it as a general methodology for panel data settings with persistent unobserved confounding. The central insight is that pre-treatment placebo gaps---periods in which treatment has not yet occurred---provide repeated measurements of the latent selection bias in treated--control outcome comparisons. Under a bias-stability restriction equivalent to conditional parallel trends, the bias surface learned from pre-treatment periods can be subtracted from post-treatment gaps to identify the conditional average treatment effect on the treated. The essay implements this through a two-layer procedure combining within-period orthogonal score estimation with across-period pseudo-treatment debiasing, establishes large-sample inference results, introduces placebo diagnostics that directly test the identifying restriction, and proposes a sensitivity-analysis framework that quantifies how conclusions change under controlled violations of bias stability.Together, these essays advance the empirical and methodological foundations for designing, targeting, and evaluating compliance interventions in settings where administrative capacity is limited and causal identification is difficult, with direct implications for tax policy in developing countries.
Access status: Embargo ,
Seeing the King of the Judaeans: Reading the Gospel of John in the Shadow of the Flavian Roman Imperial Dynasty
(2026) Dunstan, Tyler Allen
Why does the Judaean ethnonym (Ἰουδαῖος, ioudaios, Jew/Judaean) appear so frequently within the Gospel of John? More specifically, how does the Judaean ethnonym function within the Gospel’s stated aim to persuade readers, or deepen the commitment of those already persuaded, that Jesus of Nazareth is the Messiah and Son of God? This dissertation proposes a visual intervention into these well-known and consequential questions. The visual approach is twofold. Interpretation of the Gospel of John’s use of the Judaean ethnonym is enriched and clarified by a largely untapped reservoir of meaning in the Flavian Roman imperial dynasty’s visual rhetoric about the Judaean ethnos ca. 70–100 CE. Vespasian and his son Titus, both of whom held military commands in Judaea during the revolt of 66–70 CE, later leveraged their “foreign” victories to consolidate imperial legitimacy in the aftermath of the Roman civil wars of 68–69 CE, the year of four emperors. This imperial “brand campaign” included Judaean-associated events and monuments such as the double triumphal procession over Judaea by the father and son Flavian duo, the construction of the Flavian Amphitheater (Colosseum), triumphal victory arches, a new Temple of Peace that also housed the Jerusalem temple sacred objects/furnishings, along with Judaea Capta and numerous other coin series that spread these visuals far and wide across the well-traveled Roman roads and protected seas. Flavian and Josephan studies have produced several invaluable monographs and edited volumes bringing the Flavian era and the writings of Josephus to life in new ways and drawing fresh awareness to the centrality of the Judaeans within the Flavian dynasty’s imperial identity construction. Ancient art historians have also creatively mapped ancient rhetoric-related discussions about ancient sight onto some of the visual material still available to us from Graeco-Roman antiquity. Not only do these works provide fresh resources for what to look at, but also new, rhetorically rooted, analytical frameworks for trying to “see” as the ancients saw.Second, this study reads the Gospel of John itself as a visually oriented ancient Greek narrative with persuasive rhetorical intent. Given the close relationship between visuality and rhetoric, reading may be understood as a form of seeing, and seeing as a form of reading. This project argues that the Gospel of John is a carefully arranged sequence of images guiding the reader’s eye to a concluding theophanic diptych of the crucified and risen Jesus. It is here that we find the Judaean ethnonym in one of the Gospel’s climactic images. Jesus is crucified as “Jesus of Nazareth, the king of the Judaeans” (Ἰησοῦς ὁ Ναζωραῖος ὁ βασιλεὺς τῶν Ἰουδαίων, John 19:19). Prevailing readings of John’s Ἰουδαῖος language prior to this moment in the Gospel, however, do not provide compelling readings of this leitwort when it appears in one of the Gospel’s defining images. This dissertation argues that the Gospel of John’s Judaean rhetoric, or the way that the Gospel utilizes the Judaean ethnonym, is best understood as a dialectic of accusation and apology, ultimately functioning as a rhetoric of Judaean solidarity.
Access status: Open Access ,
Essays on Firm Behavior and Market Dynamics under Supply and Technology Shocks
(2026) Bai, Boyang
This dissertation consists of three essays in industrial organization that examine how firms and markets respond to major economic shocks and how these responses shape market outcomes. Specifically, it studies firm and market adjustments under supply, credit, and technology shocks, with a focus on pricing, production, financing, and labor-market outcomes. Across chapters, I combine newly constructed datasets with reduced-form and structural empirical approaches to quantify firm responses and their implications for market structure and welfare.Chapter 2 studies the welfare effects of supply disruptions in the context of the 2020--2023 automobile semiconductor shortage. Using detailed automobile industry data, I document that automakers reallocated production toward higher-priced models as supply constraints tightened, while prices rose across all segments, with disproportionately larger increases for lower-priced vehicles. To rationalize these patterns, I develop a stylized framework incorporating supply constraints and heterogeneous demand elasticities and estimate a structural model to quantify welfare effects. The results show that firms partially offset production losses through product-mix reallocation and higher prices, but consumers faced higher prices and reduced access to affordable vehicles, leading to aggregate welfare losses.Chapter 3 (co-written with Deheng Xu and Yiyuan Zhang) examines how bank consolidation affects industry concentration outside the financial sector, focusing on the U.S. homebuilding industry. We construct a novel dataset tracking financing, construction, and sales activities of homebuilders and exploit mergers among large banks as an exogenous source of credit supply shocks. Using within-firm variation and staggered difference-in-differences designs, we show that bank consolidation disrupts lending relationships, reduces borrowing and construction activity among affected firms, increases exit rates, and ultimately contributes to higher industry concentration in regions more exposed to consolidation.Chapter 4 (co-written with Zhangchi Ma) investigates the labor-market effects of generative AI on high-skilled workers, leveraging the H-1B visa program as a natural laboratory. H-1B occupations are heavily concentrated in fields most exposed to AI, and the program's employer-sponsorship structure creates mobility frictions that make these workers particularly sensitive to demand-side shocks. Combining a novel, usage-based measure of AI exposure from real human--AI conversation data with administrative H-1B wage records, we document that AI usage is extremely concentrated across occupations and distinguish between automation and augmentation channels of human--AI interaction. Employing a difference-in-differences design around the release of ChatGPT, the chapter aims to estimate how AI exposure affects wages, the wage premium over prevailing wages, and firms' demand for high-skilled labor.Collectively, these essays highlight the central role of economic shocks in shaping firm decisions, industry dynamics, and consumer welfare. By integrating rich data with structural and reduced-form approaches, this dissertation provides new evidence on the mechanisms through which shocks affect market outcomes and offers insights relevant for policy discussions on resilience, competition, and technological change.
Access status: Embargo ,
Independent Trivariate Bicycle Codes and Hyperbolic Codes
(2026) Galimova, Aygul Azatovna
Quantum error correction is necessary for fault-tolerant quantum computation. The surface code achieves a threshold of approximately $1\%$ under circuit-level depolarizing noise and planar connectivity, but it encodes only one logical qubit per patch, with an encoding rate $k/n$ that approaches zero as code size increases. Finding codes with better encoding rates, higher distances, and practical thresholds under realistic noise models is a central problem in quantum error correction.This dissertation constructs and evaluates quantum error correcting codes using algebraic, topological, and dynamic approaches.Bivariate bicycle codes, introduced by Bravyi et al., use pairs of polynomials over two cyclic shift matrices to build quantum LDPC codes. In Chapter~\ref{ch:trivariate}, we extend this construction to three independent cyclic shift matrices, introducing independent trivariate bicycle codes. Using asymmetric polynomial pairs on three-dimensional tori, we construct four codes including a $[[140,6,14]]$ code with $kd^2/n = 8.40$. In the code-capacity setting, the $[[140,6,14]]$ code achieves a pseudothreshold of $8.0\%$, exceeding the best multivariate bicycle code of Voss et al.\ ($7.9\%$). With circuit-level depolarizing noise, pseudothresholds reach $0.59\%$ for $[[140,6,14]]$ and $0.53\%$ for $[[84,6,10]]$. On the SI1000 superconducting noise model, the $[[140,6,14]]$ code achieves a per-round per-observable rate of $5.6 \times 10^{-5}$ at $p = 0.20\%$. We additionally present self-dual codes with $kd^2/n$ up to $10.0$.Hyperbolic surface codes achieve constant encoding rate by embedding qubits on negatively curved surfaces. In Chapter~\ref{ch:css_erasure}, we construct hyperbolic CSS surface codes from $\{8,3\}$, $\{10,3\}$, and $\{12,3\}$ tessellations via the Wythoff kaleidoscopic construction with the Low-Index Normal Subgroups algorithm. We extend the circuit-level erasure noise model and Wang et al.\ quadratic expansion fitting of Chang et al.\ from planar surface codes to these hyperbolic codes. Applied to the Bolza $k = 4$ semi-hyperbolic family, this pipeline yields erasure thresholds of $3.6\%$ (general model) and $4.7\%$ (tailored model) at erasure fraction $R_e = 1.0$, with erasure-to-Pauli ratios of $5.0\times$ and $6.5\times$ matching the surface code values to within $5\%$. Per-observable crossing-point analysis at $R_e = 1$ (Models 1--3) independently yields a ratio of $5.3\times$, consistent with the surface code's $5.2\times$.Floquet codes replace stabilizer measurements with periodic sequences of weight-2 measurements that dynamically generate the code space. In Chapter~\ref{ch:monolithic_floquet}, we build hyperbolic Floquet codes from the same lattices and evaluate them under correlated EM3, SDEM3, and erasure noise models. Under SDEM3 depolarizing noise motivated by Majorana tetron architectures, all three tessellation families achieve per-observable thresholds of ${\sim}1.0$--$1.2\%$. Under erasure noise motivated by spin-optical architectures, fine-grained $\{8,3\}$ codes achieve thresholds of ${\sim}8.5$--$9\%$, exceeding the ${\sim}6.3\%$ planar honeycomb threshold while encoding $k = 4$ logical qubits.Scaling quantum computers beyond a single chip will likely require distributing qubits across networked quantum processing units. In Chapter~\ref{ch:distributed}, we distribute the Floquet codes across processing units via spectral bisection and evaluate them under depolarizing and erasure noise with asymmetric local and non-local error rates. With depolarizing noise at local fidelity $99.97\%$, fine-grained codes achieve non-local pseudothresholds up to $3.0\%$. Under distributed erasure noise at $1\%$ local loss, thresholds reach $35$--$40\%$ for $\{8,3\}$.Overall, this dissertation explores how algebraic, topological, and dynamic constructions can produce practical quantum error correcting codes with improved encoding rates, distances, and thresholds under hardware-motivated noise models.
Access status: Embargo ,
Tuning the Sound Field of an Electrodynamic Loudspeaker Using a Horn-like Acoustic Metamaterial Lens
(2026) Hernandez, Gregory Mark Angel
Electrodynamic transducers, which are commonly called loudspeakers, are electromechanical devices that convert an electrical current into an acoustic wave. These technologies, for example, are found in headphones, cellphones, and televisions---essentially any public address system or airborne sound communication instrument. Some disadvantages of a singular transducer are the inability to control the radiated field generated by the loudspeaker, and the directional nature of the source as the emitted wavelength becomes comparable to the size of the oscillating surface. Sometimes several transducers are implemented in an array to overcome these limitations, but this inherently increases the geometric size of the radiating array, consumes more electrical energy, and requires complex signal processing to control the sound waves, while still not addressing the directional nature of the sources. Alternatively, acoustic lenses have been introduced to control the sound field of a loudspeaker, but their complex, narrowband, and bulky designs have not found an impactful application since their inception. However, with the increasing efficiency of 3D-printing and the conceptualization of acoustic metamaterials, compact, broadband, cost effective, and complex acoustic lenses can be engineered. This thesis derives and details the creation of such an acoustic metamaterial lens that is constructed of thousands of tiny acoustic horns to control the radiated sound field in various ways from one geometrical design for a singular electrodynamic transducer.
Access status: Embargo ,
Getting Out: Mixing Outdoor Recreation, Place-Making, and War in Iraqi Kurdistan
(2026) Steenberg, Summer
This dissertation explores outdoor recreation as a means of making political claims, restructuring memory, and reimagining the Iraqi Kurdish Region (Başûr), also colloquially known as Kurdistan. Existing scholarship has explored Kurdish sovereignty, trail-building in the Middle East, and war-curated landscapes across the region. Building on this work, my research addresses twin questions: What does getting outside enable in Kurdistan and how do ideas of care and exercise as leisure create relations across varied Kurdish contexts? Outdoor recreation in Kurdistan expanded with the formalization of the Kurdish Regional Government (KRG) in 2005. Reading through two years of fieldwork, I contend that getting outside (be it in mountaineering, paragliding, or hiking), and the economies these activities create, permit plasticity and reinvention amid the tensions of family, ongoing violence, and regional geopolitics. This rise in Kurdish outdoor recreation involves expansive claims; the KRG claims to govern all of Kurdistan. The Kurdish Mountaineering and Climbing Federation claims to represent all Kurdish mountaineers. Participants claim the “right way” to get outside. Trails link nomadism and war to claim histories of belonging. Together, these claims construct a historical entitlement to place as well as new horizons for connections. The dissertation approaches getting outside as a creative process. Hiking makes land Kurdish through traversals that create embodied histories. Building Kurdish trails makes places Kurdish. Envisioning getting outside as a return manifests a history in which that past existed. Far from peripheral to matters of state-making and belonging, outdoor recreation materializes them in its performance. At the same time, combining recreation, environmentalism, and discipline creates hierarchies of difference between the semi-autonomous state and federal Iraq, between hikers and picnickers, and between Kurdish and Arab participants. In exploring all of this, I suggest that outdoor recreation opens new ways to discuss Kurdish sovereignty and temporalities of war. With these proposals, this study contributes to ethnographies of Kurdistan and mountaineering, as well as to theories of ecologies of war, outdoor sport, and nationalism.
Access status: Embargo ,
Designing for Degradation: Strategies for Resorption in 3D Printed Medical Devices
(2026) Augustine, Emily
Rib fractures and critical-sized segmental bone defects represent significant clinical challenges in orthopedic trauma due to high morbidity, limited regenerative capacity, and shortcomings in current treatment options. The standard treatment of most rib fracture cases is limited to pain control and respiratory support, with the surgical stabilization of rib fractures (SSRF) using titanium plates reserved for severely injured patients. Although SSRF has been shown to improve long-term patient outcomes, its expanded use has been limited by the invasiveness of the procedure and a lack of safe and effective resorbable fixation materials. Similarly, existing clinical standards for the treatment of segmental bone defects, such as autografts and allografts, are limited by tissue availability, donor site morbidity, and risk of infection, presenting a substantial unmet clinical need for bone graft alternatives. Additive manufacturing of resorbable polymeric resins represents a promising approach to address unmet needs in medical device fabrication as a route to access customizable, complex geometries. Herein, we first present a materials strategy based on fumarate-based copolyester–hydroxyapatite (HAp) composite thiol-ene networks enabled by continuous liquid interface production (CLIP) for the fabrication of implantable orthopedic devices and demonstrate preclinical efficacy in vivo. 3D printed fumarate-based copolyester-composite osteosynthetic plates were assessed for the surgical stabilization of rib fractures (SSRF) in a rabbit rib fracture model. Compared to rigid titanium fixation plates, ribs fixed with 3D printed composite plates elicit fracture calluses with decreased inflammatory response, enhanced osseointegration, and bone morphometry at 2- and 4-weeks post-fracture comparable to clinically used titanium plates. Additionally, fumarate-based copolyester-composites were fabricated into fully resorbable triply periodic gyroid scaffolds, and the mechanics under compressive loading were tuned by varying gyroid porosity and design. Bone ingrowth and remodeling was assessed in a load sharing rabbit radial segmental defect model at 4 and 12 weeks. Microcomputed tomography and histological evaluation show copolyester-composite scaffolds lead to significantly improved bone deposition and endochondral ossification across the defect sites compared to untreated controls, unfilled fumarate-based copolyester and 3D printed titanium scaffolds. Taken together, this work represents a marked improvement over commercially available orthopedic implants with respect to fixation devices and synthetic bone graft alternatives for the restoration of native bone volume in segmental defects. To address limitations in achievable degradation rates of 3D printed devices to date, silyl ether-based multifunctional thiol crosslinkers were developed as generalizable approach to accelerating degradation in 3D printed thiol-ene copolyester networks. Silyl ether-crosslinked networks demonstrate tunable degradation profiles dependent on both crosslinking chemistry and copolyester backbone composition with modular control over mechanical properties. Additionally, assessments of network degradation by mechanical property loss over time are shown to offer a framework for predictable degradation in 3D printed copolyester networks.
Access status: Open Access ,
Effects of environmental variation and stand development on water use and production of five structurally-diverse Pinus species
(2026) Boroski, Chainey Alexandria
For five economically and ecologically important southern U.S. Pinus species, I explore: (1) the impact that canopy architecture has on needle-scale and plant-scale hydraulic traits; (2) how these differences may manifest in patterns of water use at the daily scale in response to environmental variation; and (3) how canopy and stand characteristics, as well as stand development, control leaf-scale and stand-scale transpiration on interannual timescales. For two common-garden, multi-year sap-flux monitoring studies, I quantified sensitivity to vapor pressure deficit and soil moisture among species or genotypes within the same environment, compared their sap-flux density and leaf transpiration under standardized reference conditions, and explained long-term variation in water use as a result of increasing sapwood area, leaf area, and tree height. Daily-scale sap-flux density and leaf-level transpiration were broadly conservative within a study for a given planting density, primarily reflecting local environmental conditions even in the presence of established hydraulic trait differences. Stand-scale variation in sap-flux and transpiration at longer timescales broadly reflected the impacts of stand development and intraspecific competition, causing water use per unit sapwood or leaf area to decline over time. This led to a decoupling of stand transpiration from sapwood area per unit of ground area, occurring within the first 8-10 years of rapid stand development for some species. The interacting, often convoluted effects of within-canopy shading and increasing pathlength resistance were each explicitly considered as drivers of long-term change in transpiration. Tree height (pathlength resistance) explained interannual decline in sap-flux and leaf-scale transpiration as well as or better than variation in leaf area index. These findings suggest that Pinus transpiration may be able to be modelled assuming a single physiology, accounting for environmental variability and stage of development. Further, the addition of hydraulic architecture – especially height – to transpiration models could serve to support more accurate estimates of water use in developing stands than those that consider leaf area in isolation. Finally, due to the effects of both competition and development on leaf-scale and stand-transpiration, lower-density plantings may be associated with higher water yield to surrounding ecosystems, as well as broader-crown genotypes.
Access status: Open Access ,
Beam Spin Asymmetry in Dihadron Final States with Kaons in Semi-inclusive Deep Inelastic Scattering
(2026) Pecar, Connor
The substructure of nucleons and how internal strong force dynamics generate their properties, such as mass and spin, is a topic of major experimental and theoretical investigation. One probe of nucleon substructure is dihadron semi-inclusive deep inelastic scattering (SIDIS), where a lepton exchanges a virtual photon with a quark in the nucleon and the scattered electron is measured along with two hadrons in the final state. The SIDIS cross section is sensitive to three-dimensional nucleon momentum structure encoded in Transverse Momentum Dependent Parton Distribution Functions (TMD-PDFs). It is also sensitive to the hadronization process, where a struck quark forms color-neutral hadrons, via Dihadron Fragmentation Functions (DiFFs). DiFFs provide unique sensitivity to the polarization of the struck quark and its impact on the final distribution of hadrons. In this dissertation, the beam spin asymmetry in dihadron SIDIS with charged kaon final states is measured for the first time with data collected using CLAS12. Comparison of this measurement to the same asymmetry for pion final states provides novel information on flavor dependence in hadronization. This measurement is also sensitive to twist-3 PDFs and DiFFs, which encode quark-gluon correlations. These results therefore provide further insight into the dynamic nature of nucleon substructure. This dissertation will also describe the development of an alignment method for the CLAS12 Ring Imaging Cherenkov detector using Bayesian optimization, extending the usable acceptance of the detector significantly for future particle ID in physics analyses with CLAS12.
Access status: Embargo ,
Tailoring Quantum Error Correcting Protocols to Biased Noise Models
(2026) Campos, Julie Abigail
Quantum systems are inherently vulnerable to asymmetric noise processes which manifest as biased noise channels in quantum computing systems. To mitigate the effects of these noise mechanisms efficiently, quantum error correction (QEC) protocols can be tailored to prioritize the suppression of these dominant errors. In this dissertation, we characterize the benefits of noise-tailoring methods by evaluating their qualitative effects on the structure of QEC codes and performing a comparative analysis of the thresholds and logical error rates between tailored and standard codes. In particular, the utility of Clifford deformations are explored in both the circuit-based and measurement-based quantum computing frameworks. The primary strength of this technique lies in its ability to introduce additional symmetries to the stabilizers of a QEC protocol. These symmetries make the code more perceptive to dominant errors while providing the decoder with information on the underlying structures of syndrome configurations, thereby enabling more accurate decoding. Despite the trade-offs that often come with these modifications, our results demonstrate gains in error tolerance thresholds and more efficient suppression of logical error rates. The methodologies and perspectives presented in this work can provide guidance for designing and optimizing QEC protocols tailored to biased noise models.
Access status: Embargo ,
Deep Learning Applications in Light-based Ultrasound Imaging: Photoacoustic Microscopy, Photoacoustic Tomography, and Passive Cavitation Mapping
(2026) DiSpirito, Anthony
Medical imaging systems inevitably encounter fundamental physical limitations that have traditionally been addressed either through costly, time-intensive hardware innovation or by accepting tradeoffs in resolution, image quality, imaging speed, or overall system usability and commercialization potential. The emergence of advanced computational methods, particularly deep learning, offers a fundamentally different paradigm: rather than working around hardware constraints, these limitations can be mitigated through software. This shifts the primary challenge in imaging development from physical tradeoffs to problems of dataset generation, model design, optimization, and inference robustness. In this thesis, we present deep learning frameworks that address long-standing physical constraints in light-based ultrasound imaging, specifically photoacoustic microscopy (PAM), photoacoustic computed tomography (PACT), and passive cavitation mapping (PCM). In PAM, we investigate the limitation imposed by laser repetition rates and the resulting tradeoff between spatial sampling density and imaging speed for a fixed field of view. We develop a deep learning-based upsampling approach that restores high-quality images from sparsely sampled data without sacrificing acquisition speed. In PACT, we demonstrate that conventional resolution limits can be surpassed by coupling a digital model of a 512-element ring-array transducer with supervised deep learning deconvolution. In PCM, we use deep learning to characterize the complex interactions of cavitation collapses during kidney stone lithotripsy and generate predictive visualizations of stone material loss, enabling improved treatment monitoring and effectiveness.
Access status: Embargo ,
Poly(Allyl Glycidyl Ether Succinate) (PAGES) for Solvent Free Vat Photopolymerization 3D Printing of Degradable Medical Devices
(2026) Segal, Maddison
Additive manufacturing has enormous potential to change the medical industry with its inherent customization and ability to produce complex geometries. Vat photopolymerization (VPP), a type of additive manufacturing that cures liquid resins into solid 3D printed parts using light, is especially renowned for its high resolution and efficiency. While there are many commercially available resins for VPP printing, nearly all of them produce non-degradable 3D printed parts. Current efforts towards degradable resins are limited to functionalizing degradable polymers with photopolymerizable end groups, which results in a network that is only partially degradable. Other efforts utilizing thiol-ene reactions can produce fully degradable networks, but struggle with slow degradation rates and significant amounts of part shrinkage from the use of solvents in the resin to achieve a printable viscosity. The development of new, degradable resins for 3D printing could allow VPP to be used to manufacture temporary medical implants that could be degraded and resorbed into the body. Herein, poly(allyl glycidyl ether succinate) (PAGES) polyester was synthesized, characterized, and used to formulate solvent free 3D printable resins for both digital light processing (DLP) and continuous liquid interface production (CLIP) 3D printing on the kg scale. PAGES resins are unique because they are inherently low viscosity due to the comb-like polymer architecture, allowing them to be printed completely solvent free at room temperature, unlike any other thiol-ene resins currently reported. Solvent free printing not only improves part tolerancing by eliminating the part shrinkage issue that plagues other resins but also improves the mechanical integrity of the 3D printed parts by promoting the formation of a more ideal network and avoiding cracking, warping, and residual stress formation associated with solvent evaporation. 3D printed PAGES networks displayed tunable mechanical properties similar to that of soft tissue and adjustable degradation rates ranging from 1 week to over 1 year. Large animal preclinical studies in a sheep vascular model indicated that PAGES 3D printed networks were both degradable and biocompatible in a living system, showing complete resorption of a 3D printed sheath device after four months in vivo with similar levels of inflammation as occurred in control procedures. Overall, this work highlights the development of a library of solvent free PAGES resins for VPP 3D printing that are high resolution, scalable, tunable, degradable, biocompatible, and resorbable, positioning them as an ideal candidate for manufacturing temporary medical implants.
Access status: Open Access ,
Essays on Empirical Asset Pricing
(2026) Wang, Rong
This dissertation studies the economic interpretation and empirical implications of the zero-beta rate in asset pricing models. In linear factor models, the zero-beta rate is defined as the expected return on a portfolio with zero exposure to all priced systematic risk factors. Although closely related to the risk-free rate in theory, empirical research typically proxies the zero-beta rate using Treasury bill yields without directly estimating it from asset pricing restrictions. This dissertation asks three main questions: What does the data imply about the zero-beta rate across a wide range of factor models? How should the high empirical estimates of the zero-beta rate be interpreted? And what are the implications of zero-beta portfolios for measuring the risk-free rate, testing and comparing asset pricing models, and constructing investment strategies?To answer these questions, the dissertation proceeds in four chapters.Chapter 2 documents and characterizes the zero-beta rate across a broad class of asset pricing models. Using portfolio-based and cross-sectional regression approaches, the chapter estimates the zero-beta rate and evaluates its stability in- and out-of-sample. The key finding is that the estimated zero-beta rate is persistently high—typically around 10–12 percent per year—and remarkably robust across model specifications. This empirical regularity constitutes the zero-beta rate puzzle.Chapter 3 provides a theoretical and quantitative resolution of this puzzle by linking the zero-beta rate to factor model misspecification. Using mean-variance geometry, analytical derivations, and simulation analysis, the chapter shows that when a factor model is strongly misspecified, pricing errors mechanically distort the estimated zero-beta rate. In such settings, the zero-beta rates converge to the mean return of the global minimum-variance portfolio. The central conclusion is that the high zero-beta rate reflects model incompleteness rather than the true intertemporal risk-free rate.Chapter 4 turns to bounding the frictionless risk-free rate directly. Combining no-arbitrage restrictions with forward-looking measures of expected market returns and lower bounds on the market risk premium, the chapter derives economically meaningful upper bounds on the unobserved risk-free rate. The results indicate that the frictionless risk-free rate is substantially lower than the factor model-implied zero-beta rate. This evidence further reinforces that the zero-beta rate should not be interpreted as the true risk-free rate.Chapter 5 develops practical applications of zero-beta portfolios for model testing, comparison, and investment. The chapter constructs zero-investment zero-beta portfolios to design diagnostic tests of model validity and formal procedures for comparing competing factor models. It also evaluates factor-neutral investment strategies under realistic transaction costs. The findings show that zero-beta portfolio properties provide informative measures of model performance and that factor-neutral strategies are economically meaningful because they are more resilient to trading costs than conventional factor-investing approaches. Overall, the dissertation clarifies the economic meaning of the zero-beta rate, resolves the zero-beta rate puzzle through the lens of factor model misspecification, and demonstrates how zero-beta portfolios can be used to evaluate asset pricing models and design implementable investment strategies.
Access status: Embargo ,
Investigating whether epistasis constrained Threonine Deaminase 2’s novel function evolution in Solanaceae, and how selfing affects transposable element dynamics in Ipomoea
(2026) Cao, Gongyuan
Gene duplication is one of the primary sources of new genes in evolution, and, on rare occasions, novel functions can arise from newly duplicated genes. Much attention has been paid to the evolutionary mechanisms of novelty through gene duplication, which could take from dozens of years to dozens of millions of years. However, it remains understudied what factors could prolong this process. One possible explanation for a prolonged process generating novel functions in a copy is that the novel function needs many additive substitutions to occur. Alternatively, the number of function-switching substitutions could be few, but their effects are epistatic and require other neutral permissive substitutions to be fixed by drift first, which could take a long time. A third possibility is that selection pressure on the novel function temporarily fluctuates, and the novel function could be lost through maladaptive substitution accumulation when selection is relaxed, requiring more substitutions to leverage the effects of these maladaptive substitutions when selection resumes. In the first part of my dissertation, I examine the relative contributions of these constraining factors, especially epistasis, to the prolonged adaptive evolution of Threonine Deaminase copy 2 (TD2) in the family Solanaceae. TD2 in tomato (TD2S) is reported to have evolved a novel function over 30-50 million years of adaptive evolution following duplication. The novel function constitutes two aspects that are considered adaptive: 1) having a higher catalytic efficiency, and 2) being resistant to protease. To test whether epistasis prolongs TD2S’s evolution of catalytic efficiency, I identify three key substitutions that could significantly enhance its catalytic efficiency and examine whether their effects depend on the presence of other substitutions. To test whether epistasis prolongs TD2S’s evolution of protease resistance, I divide its sequence into three segments and design a dozen chimeric constructs by fusing one or two segments of TD2 with two or one segments of TD1, which is non-resistant. I quantitatively examine their protease resistance and ask whether any segment of TD2 could confer resistance regardless of the presence of other segments in the chimeric constructs. I find that epistasis, especially high-order epistasis, is involved in TD2S’s evolution in both catalytic efficiency and protease resistance. I then compare the temporal patterns of when key substitutions and potential permissive and epistatic resistance-enhancing substitutions occur in the phylogeny. I find that this pattern matches the expectation of adaptive evolution constrained by epistasis, as earlier data suggests. I also find that fluctuating selection may have constrained TD2S’s resistance evolution, but of less importance than epistasis. Together, these results suggest that epistasis and potential fluctuating selection prolonged TD2S’s adaptive evolution.The transition from outcrossing to selfing is one of the most important evolutionary transitions in plants, which is well studied to be accompanied by changes in various phenotypic traits and genomic features. Many theoretical studies have predicted how these traits and features would change after the transition, and many empirical studies have verified most of these predictions. Among genomic features, one that is less studied and is expected to behave differently under different models when the transition occurs is transposable elements (TEs). The two major models on how a transition from outcrossing to selfing affects TEs’ dynamics are 1) the ectopic recombination model and 2) the deleterious insertion model. The first model predicts that selfers will have more TE in their genomes than their closely related outcrossers, because increased homozygosity in selfers could reduce TE-induced ectopic recombination. The second model predicts that selfers will have less TE than outcrossers because selfers have smaller genomes and stronger background selection, which will help purge deleterious TEs more quickly. In the second part of my dissertation, I examine which model better describes the TE differences in the predominantly selfing species Ipomoea lacunosa (LAC) and its closely related predominantly outcrossing species I. trifida (TRIF). To do so, I annotated the TE content, identified all LTR-RT elements, and calculated the average TE-gene distance and insertion age of all intact LTR-RTs in the two genomes. I find that LAC has higher TE content, fewer solo-LTRs, and shorter average distance to genes than TRIF. These results suggest that the ectopic recombination model, not the deleterious insertion model, better describes changes in TE activity in LAC than TRIF.
Access status: Embargo ,
Beyond Intended Mechanisms: Immune Interactions of Trop2-Directed Antibody-Drug Conjugates and a HER2-Directed Self-Replicating RNA Vaccine in Breast Cancer
(2026) Ma, Xingru
Breast cancer immunotherapeutics are generally developed and evaluated according to their intended mechanisms of action, yet their interactions with the tumor-immune microenvironment are often broader and less well understood. This dissertation addresses the problem that unintended or overlooked immune interactions may substantially influence therapeutic efficacy, but remain insufficiently characterized for both antibody-drug conjugates (ADCs) and self-replicating RNA cancer vaccines. To investigate this issue, I studied two breast cancer immunotherapy platforms: Trop2-directed ADCs and a Venezuelan equine encephalitis virus (VEEV)-based viral replicon particle self-replicating RNA (VRP-srRNA) vaccine targeting HER2.A combination of in vitro, in vivo, and translational approaches was used. For studies of Trop2-directed ADCs, human Trop2 knock-in and knockout breast tumor cell lines and mouse tumor models were used to evaluate antigen-dependent and antigen-independent antitumor effects of Sacituzumab Govitecan (SG) and Datopotamab Deruxtecan (Dato-DXd). Additional functional assays were performed to assess Fcγ receptor activation, macrophage-mediated antibody-dependent cellular phagocytosis, and treatment-induced adaptive antitumor immunity. For studies of the VRP-srRNA HER2 vaccine, patient samples from a phase I clinical trial and complementary mouse vaccination models were analyzed to characterize vector-directed humoral and cellular immune responses and to determine how repeated homologous vaccination influenced subsequent boosting. These studies demonstrate that the activity of breast cancer immunotherapeutics is shaped not only by their intended targets, but also by unintended interactions with antigen-negative tumor cells, innate immune cells, and vaccine vector components. Trop2-directed ADCs exhibited context-dependent antigen-independent effects while retaining Fc-mediated immune function, and VRP-srRNA vaccination induced anti-vector immunity that limited the benefit of repeated homologous boosting on a short dosing schedule. Together, these findings show that bidirectional interactions between immunotherapeutic agents and the immune system are important determinants of therapeutic outcome and should be more deliberately considered in the design and optimization of breast cancer therapies.
Access status: Open Access ,
Modeling, Analysis, and Interpretation of High-Throughput Genomic Perturbation Assays
(2026) Doty, Richard
High-throughput genomic perturbation assays provide large-scale measurements of molecular response to targeted genetic interventions, but the resulting data are typically sparse, noisy, and characterized by rare or weak signals. In this regime, biological ambition outpaces naïve statistical practice: detecting meaningful regulatory effects requires procedures that remain calibrated under dilution, heterogeneity, and limited signal strength. This dissertation develops statistical foundations for calibrated inference in such perturbation systems. We first introduce a generative framework for gRNA-level count data from single-cell CRISPR assays that represents observed counts as the superposition of signal and nuisance components, each governed by an occurrence mechanism and a conditional magnitude distribution. This concurrent-process formulation reproduces key aggregate features of empirical UMI count distributions and provides a structured statistical representation of sparsity and noise in the data-generating process. Building on this foundation, we develop FRACTEL, a bounded-rank aggregation procedure for combining multiple gRNA-level tests targeting a shared regulatory element. Rather than diffusing evidence across all guides or relying solely on the single most extreme result, FRACTEL evaluates the least likely among a bounded set of leading order statistics, concentrating evidentiary weight where signal is most plausibly localized while controlling the depth of aggregation. Empirical analyses characterize how power depends on signal architecture, guide multiplicity, and baseline expression, illustrating the impact of dilution and heterogeneous activity on detection.To enable principled calibration beyond empirical resampling, we derive exact and asymptotic distributional results for finite-rank order-statistic aggregation statistics. We obtain explicit null distributions and critical values in finite samples, establish fixed-rank asymptotics when the aggregation depth remains small relative to the number of tests, and show how evidentiary weight is distributed non-uniformly across ordered p-values as the rank bound varies. These results provide calibrated inference in small-sample and truncated regimes and clarify the operating characteristics of bounded-rank testing procedures. Finally, we apply rank-based and allele-controlled statistical analyses to in silico saturated mutagenesis experiments using a predictive regulatory model. By constructing local perturbation landscapes in constitutively closed genomic regions and in candidate cis-regulatory elements, we identify context-dependent directional asymmetries consistent with regulatory constraint: variants observed in the human population are relatively depleted among mutations predicted to increase activity in closed regions, whereas in active regulatory elements they are relatively depleted among mutations predicted to decrease activity. Together, these findings link calibrated statistical methodology to biologically interpretable patterns of regulatory constraint in high-throughput perturbation experiments.
Access status: Open Access ,
Competition Between Transcription Factors and Pol-δ During Okazaki Fragment Maturation Drives Site-Specific Mutagenesis
(2026) Boivin, Olivier
Mutagenesis is a fundamental and complex biological process. Despite the remarkable fidelity of DNA replication, diverse mutational processes succeed at introducing new genomic variation, thereby providing a substrate for natural selection and genetic disorders. A thorough understanding of the nature of mutations, as well as their origins, is thus of high relevance to evolutionary biology, as well as mechanistic studies of human disease. During DNA replication, the lagging strand is synthesized as a set of discontinuous Okazaki fragments (OFs) initiated by the error-prone polymerase-α (Pol-α). During OF maturation, the higher-fidelity polymerase-δ (Pol-δ) displaces and re-synthesizes the Pol-α-created portions of OFs, removing mis-incorporated bases. However, transcription factors (TFs) that rapidly re-associate with DNA after replication fork passage can impede this displacement, potentially leading to the retention of error-prone Pol-α-synthesized DNA in the mature genome. Prior work established that three general regulatory factors (GRFs), Abf1, Reb1, and Rap1, strongly bias Okazaki fragment junction (OJ) positions, but whether this mutational mechanism extends broadly across TFs, and how specific TF-polymerase collisions influence mutagenesis at the nucleotide level, remains poorly understood. This study combines computational genomics and in vitro biochemistry in Saccharomyces cerevisiae to comprehensively characterize the relationship between TF binding and replication-associated mutagenesis. In the computational analysis, OJ frequencies and interspecies nucleotide substitution rates were mapped at the binding sites of 35 sequence-specific TFs. This genome-wide survey identified four GRFs whose binding sites exhibit a significant correlation between OJ accumulation and elevated substitution rates, extending prior observations to a broader set of factors and establishing the specificity of this phenomenon. To directly test the mechanistic basis of TF-mediated replication impediment, an in vitro strand displacement assay was developed using reconstituted yeast replication components and the TF Cbf1 as a model DNA-bound barrier. This assay was adapted for high-throughput measurement of strand displacement efficiency in the presence of Cbf1, enabling systematic evaluation of how specific nucleotide mismatches within the Cbf1 binding motif modulate the interplay between TF binding and Pol-δ processivity. The results demonstrate that Cbf1 binding impedes Pol-δ strand displacement synthesis, and that certain mismatches within the Cbf1 recognition sequence further reduce strand displacement efficiency in the presence of bound Cbf1. Importantly, these mismatches correspond to the nucleotide substitutions enriched at Cbf1 binding sites in the computational genomic analysis and are consistent with patterns reported in the literature. The convergence of the in vitro biochemical data with the genome-wide substitution patterns provides direct evidence that TF-mediated impediment of OF maturation contributes to site-specific mutagenesis at regulatory elements. These findings support a model in which the mutational cost of TF binding shapes the evolution of regulatory DNA, and they offer a mechanistic explanation for the enrichment of mutations at functionally important genomic sites.
Access status: Embargo ,
The Faith of Abraham and the Patient Perseverance of the Saints in Hebrews
(2026) Leyva, Julie Michelle
While many studies on faith in the New Testament focus on the character of Abraham within Paul’s letters, these studies have not turned their attention to Abraham’s faith in the book of Hebrews. This dissertation will accomplish two goals: first, it will provide a needed treatment of Abraham in the book of Hebrews; second, it will contribute to the ongoing scholarly discussion on what faith is in Hebrews. Methodologically, this dissertation will focus on the literary and theological rendering of Abraham in the book of Hebrews. Aligned with this task, I will examine the intertextuality of Hebrews and its interpretation of LXX texts. Hebrews employs a distinct method of interpreting Old Testament texts and analyzing this use of the OT in order to better understand the letter’s argument will be an important part of the dissertation. By providing a close reading of Abraham’s faith in Hebrews, the unique understanding of faith that Hebrews offers will be illuminated. Faith in Hebrews will be described as the conviction that strengthens those patiently awaiting God’s promises, even as they endure trial and uncertainty. Rather than an understanding of faith as merely cognitive belief, faith will be seen as a way of life. Abraham’s long journey of discipleship represents faith as the conviction by which Abraham acted throughout his life, not just at a particular moment. We will see that Abraham’s faith is the most important example of faith in Hebrews, after Jesus himself. In addition to providing this needed analysis of Abraham’s role in Hebrews, this dissertation will also contribute to the scholarly understanding of the audience of the letter. Through a review of the scholarly literature and a focus on Abraham’s role in this discussion, I will conclude that the audience of Hebrews is best understood as Jewish Christians.
Access status: Embargo ,
Superconductivity in 2D Gate-Tunable Heterostructures
(2026) McCourt, Jordan Thomas
Electronic states confined to 2-dimensions (2Ds) display a rich array of phenomena notfound in their 3-dimensional counterparts. In 2D superconductors, the reduced dimen- sionality enhances the role of fluctuations, giving rise to Berezinskii–Kosterlitz–Thouless (BKT) transitions and rich vortex phenomena with a strong dependence on the su- perfluid density. Remarkably, the dilute 2D superfluid can be modified by an electric field, allowing for the study of density and disorder driven quantum phase transitions. Furthermore, this offers a promising route to field-effect superconducting electronics. In this dissertation, I will present measurements of novel devices in the 2D super- conducting state at the interface of the complex oxide KTaO3. Explicit electrostatic modification of the superconducting state is demonstrated by a total suppression of supercurrent, before revealing the interplay with a Coulomb blockade. The influence of charging energy in the Coulomb blockade regime results in an unusual periodic supercurrent and quantum dot behaviour under an applied magnetic field. The neg- ligible magnetic screening of 2D superconductivity forms a critical state of vortex matter which results in the observation of quantum tunneling of individual vortices from pinning sites in the device. Analysis of the tunneling rate suggests overdamped tunneling, as may be expected for a topological defect consisting of many electrons. These results demonstrate the rich array of quantum phases that can be realised in a single heterostructure, and furthermore inform the development of quantum devices and fundamental studies of quantum dynamics.

Browse

Recent Submissions