150813 research outputs found
Sort by
Shining a Light on the Nucleus: Photonuclear Measurements from Correlations to Charmonium
No full textThe atomic nucleus is comprised of a collection of nucleons (protons and neutrons), which are bound together by the nucleon-nucleon (NN) interaction that originates from Quantum Chromodynamics (QCD). While most nucleons experience the force from the rest of the nucleus as a single net “mean-field” interaction that binds them relatively weakly, a small but impactful fraction are in configurations called “Short-Range Correlations” (SRCs), in which they pair with another nucleon at very short distance to experience strong interactions, significant binding, and high momentum. Hard, high-energy scattering reactions in which an SRC pair is broken apart, knocking both nucleons out of the nucleus, provide the ability to probe the details of these SRC configurations in the nucleus. Previous measurements have had limited statistics and kinematic reach, and the theoretical tools available were insufficient to draw quantitative conclusions regarding the ground-state properties of SRCs. The studies described in this thesis represent the first global analysis of SRC breakup measurements in order to present a unified picture of SRCs within light- to medium-size nuclei. This includes the use of a novel theoretical framework, the Generalized Contact Formalism, which connects scattering cross-section measurements and the ground-state properties of the SRC pair, to quantitatively interpret a variety of electron-scattering measurements. This is brought to culmination by a report on the first measurement of SRC pairs via the use of hard meson photoproduction reactions, which, despite differing significantly from the mechanics of electron-scattering, is well-described under a common framework, pointing to a consistent and universal picture of SRCs across reaction channels. I also report on the first measurement of J/ψ photoproduction in the near- and below-threshold kinematic region, giving the first insights to the gluonic structure of bound nucleons in the large-x “valence” region and providing constraints on a gluonic “EMC effect”. In addition to these studies, I provide details on the search for Primakoff production of axion-like particles using the photoproduction data taken for this experiment, and I conclude by describing studies of nucleon spin structure measurements that will be performed at the forthcoming U.S. Electron-Ion Collider.Ph.D
Evolving Properties of Biological Materials Captured via Needle-Based Cavity Expansion Method
No full textBackground The mechanical properties of biological tissues change over time and with disease progression. Quantifying these mechanical properties can thus be instrumental for medical diagnosis and for evaluation of tissue viability for transplant. However, soft and biological materials are exceptionally challenging to mechanically characterize using conventional testing methods, which are hindered by limitations of sample size, fixturing capabilities, and sample preparation. Objective We hypothesize that Volume Controlled Cavity Expansion (VCCE) is well-suited to capture subtle mechanical differences in biological tissue. The objective of this work is therefore twofold: first, we seek to quantify how stiffness of liver and gelatin evolve with age. In achieving this understanding, we aim to demonstrate the precision of VCCE in measuring subtle changes in the mechanical properties of biological tissues. Methods Performing VCCE tests over 15 days in samples of gelatin and liver (porcine and bovine), we track the evolving pressure-volume response and deformation limits of the materials. Results In both materials, we observed time-dependent variation of the stiffness and fracture thresholds. In gelatin VCCE repeatably captured stiffening over time, which was correlated with a higher fracture stress. This was in contrast to observations in bovine liver, where stiffening corresponded to a lower fracture stress. Porcine liver initially stiffened, then reversed this trend and relaxed. Conclusion Through this work we show that liver and gelatin stiffen with age, and that this trend is measurable via VCCE. These results highlight the utility of VCCE and call attention to the need for a new class of mechanism based constitutive models that are capable of capturing variations in material over time with a minimal number of parameters
Polymer Deconstructability and Recyclability via Introduction of Cleavable Si−O Bonds
No full textThe synthesis of a new polysilylether via entropy-driven ring-opening metathesis polymerization (ED-ROMP) of cyclic bifunctional silyl ether-based monomers is reported. High molecular weight polymers (up to 100 k) with narrow dispersities were achieved at modest temperature. These polymers display excellent thermal stability and ultra-low T_g (–88 ºC). The polymers are both rapidly deconstructable via the cleavage of the labile silicon-oxygen linkages with either acid or fluoride triggers and partially depolymerizable by the addition of exogenous metathesis catalyst. Analysis of the deconstructed polymer products provided insight into the polymer microstructure, showing that the ED-ROMP process was regiorandom. Altogether, this work offers a new class of deconstructable polymers with a range of potential applications. Incorporation of these bifunctional silyl ether-based monomers into copolymers could aid in the triggered deconstruction of otherwise nondegradable hydrocarbon backbones.S.M
Instrumental uncertainties in radiative corrections for the MUSE experiment
No full textThe MUSE experiment at the Paul Scherrer Institute is measuring elastic lepton-proton scattering cross sections in a four-momentum transfer range from Q2 of approximately 0.002–0.08 GeV2 using positively and negatively charged electrons and muons. The extraction of the Born cross sections from the experimental data requires radiative corrections. Estimates of the instrumental uncertainties in those corrections have been made using the ESEPP event generator. The results depend in particular on the minimum lepton momentum that contributes to the experimental cross section and the fraction of events with hard initial-state radiation that is detected in the MUSE calorimeter and is excluded from the data. These results show that the angular-dependent instrumental uncertainties in radiative corrections to the electron cross section are less than 0.4% and are negligible for the muon cross section
Design and Manufacture of a Modular Continuous Unit Dose Pharmaceutical Lyophilizer
No full textPharmaceutical lyophilization (freeze-drying) enables long term storage and simplified transportation for aqueous vaccines and protein formulations. Modern industrial pharmaceutical freeze-driers rely on large batch and open loop formulation processing, limiting supply chains and resulting in variable quality products. This work describes the design and manufacture of a modular continuous lyophilization machine for pharmaceutical production. Additionally, the scaling and design methodology outlined in this work enables the development of both smaller systems for laboratory testing and larger machines to fit the needs and requirements of individual facilities. This machine introduces three new technologies to the pharmaceutical freeze-drying process. The first innovation is a continuous flow lyophilization topology which separates the lyophilization steps spatially rather than temporally. This layout allows product to travel through the system in smaller batches for increased product uniformity and quality control. The second innovation is a weight-based sensor for monitoring residual water content. This sensor enables in-situ monitoring of product during sublimation, and it resolves mass measurements as small as 5mg. The third innovation is the implementation of a thermal shock method of inducing controlled nucleation. The convective cooling and spatial non-uniformity within the machine allow vials to experience a 40°C temperature drop in less than 30 seconds. This nucleation front starts on the vial walls, rather than at the top surface of the solution in the vial, potentially increasing the water sublimation rate during drying compared to current nucleation methods. The machine designed and built for this work integrates into modern factory processes and can be scaled from the lab bench to a production line. The manufactured prototype demonstrates improvements on the production rate, flexibility, and quality of existing machines.Ph.D
Evaluating the reliability of a microperimetry-based method for assessing visual function in the junctional zone of geographic atrophy lesions
No full textPurpose To assess the repeatability of a microperimetry methodology for quantifying visual function changes in the junctional zone of eyes with geographic atrophy (GA) in the clinical trial context. Methods A post hoc analysis of the OAKS phase III trial was conducted, which enrolled patients with GA secondary to age-related macular degeneration. Microperimetry using a standard 10 − 2 fovea centered grid was performed at baseline and follow-up visits. GA regions were traced on fundus autofluorescence (FAF) images. Two graders independently registered baseline microperimetry images with baseline FAF images in a sampling of 30 eyes from the OAKS study. Agreement between the two graders’ assessments of mean sensitivity and the number of scotomatous points within a ± 250 m GA junctional zone was assessed. Results The intraclass correlation (ICC) and coefficient of repeatability (CoR) for the mean junctional zone sensitivity were 0.987 and 0.214 dB, respectively. The ICC and CoR for the total number of scotomatous points within the junctional zone were 0.991 and 1.42, respectively. Conclusions The repeatability of the methodology and its compatibility with standard MP acquisitions appear to make it well-suited for identifying and analyzing retinal sensitivity within high-risk areas of the retina. Summary We present a microperimetry-based methodology for assessing visual function changes in the junctional zone of geographic atrophy lesions using a standard 10 − 2 fovea centered grid in a clinical trial context. The approach’s repeatability and compatibility with standard microperimetry grids may make it useful for assessing the effects of GA therapeutics
Uniacute Spherical Codes
No full textA spherical L-code, where L ⊆ [−1,∞), consists of unit vectors in Rd whose pairwise inner products are contained in L. Determining the maximum cardinality NL (d)
of an L-code in Rd is a fundamental question in discrete geometry and has been
extensively investigated for various choices of L. Our understanding in high dimensions is generally quite poor. Equiangular lines, corresponding to L = {−α, α}, is
a rare and notable solved case. Bukh studied an extension of equiangular lines and
showed that NL (d) = OL (d) for L = [−1, −β]∪{α} with α, β > 0 (we call such
L-codes “uniacute”), leaving open the question of determining the leading constant
factor. Balla, Dräxler, Keevash, and Sudakov proved a “uniform bound” showing
lim supd→∞ NL (d)/d ≤ 2p for L = [−1, −β]∪{α} and p = α/β + 1. For which
(α, β) is this uniform bound tight? We completely answer this question. We develop a
framework for studying uniacute codes, including a global structure theorem showing
that the Gram matrix has an approximate p-block structure. We also formulate a notion
of “modular codes,” which we conjecture to be optimal in high dimensions
Parsimonious Principles of Deep Neural Networks
No full textAt the core of human intelligence lies an insatiable drive to uncover the simple underlying principles that govern the world’s complexities. This quest for parsimony is not unique to biological cognition but also seems to be a fundamental characteristic of artificial intelligence systems. In this thesis, we explore the intrinsic simplicity bias exhibited by deep neural networks — the powerhouse of modern AI. By analyzing the effective rank of the learned representation kernels, we unveil the observation that these models have an inherent preference for learning parsimonious relationships in the data. We provide further experimental results to support the hypothesis that simplicity bias is a good inductive bias for finding generalizing solutions. Building upon this finding, we present the Platonic Representation Hypothesis — the idea that as AI systems continue to grow in capability, they will converge toward not only simple representational kernels but also a common one. This phenomenon is evidenced by the increasing similarity of models across domains, suggesting the existence of a Platonic “ideal” way to represent the world. However, this path to the Platonic representation necessitates scaling up AI models, which poses significant challenges regarding computational demand. To address this obstacle, we conclude the thesis by proposing a framework for training a model with parallel low-rank updates to effectively reach this convergent endpoint.Ph.D
Impact of Environmental Regulation on Data Center Valuation
No full textArtificial intelligence has become one of the defining trends of modern society, with applications spanning virtually every industry. This societal shift has also influenced the real estate landscape. While data centers have existed for decades, it is only in recent years that they have garnered significant attention, demonstrated by their strong rent growth and compressed cap rates.1 Along with the attention over data centers, there also has been extensive research on how data centers impact the environment, such as "Quantifying the Sustainability Impact of Data Center Availability" by Manish Marwah et al. which present how data center power architecture may impact the environment and "The Environmental Footprint of Data Centers in the United States" by Md Abu Bakar Siddik, Arman Shehabi, and Landon Marsto. This research delves into quantifying the environmental impacts of data centers, specifically focusing on carbon and water footprints. However, what remains unexplored is how environmental regulations influence the valuation of data centers as a distinct real estate property type. This thesis examines how data center valuations could be impacted if existing environmental regulations were applied to regions where data centers are concentrated. The findings reveal a complex dynamic: while penalties under these regulations would reduce net operating income (NOI), potentially devaluing these assets, the same regulations would discourage new development, exacerbate the already constrained supply, and ultimately drive-up market rents for these properties. As a result, these opposing forces create ambiguity regarding the net impact of such regulations on data center valuations, with the outcome depending on which force prevails. What is clear, however, is that tenants would bear the brunt of these regulations, as landlords are likely to pass on increased costs through higher rents. On the other hand, while the environmental impacts of data centers and AI applications is critical to achieving sustainability goals, the societal benefits of AI solutions—ranging from advancements in healthcare to increased operational efficiencies—must also be considered. Balancing these competing priorities presents a unique challenge for policymakers and investors, with significant implications for the future of real estate and the digital economy.S.M
Data-Driven General Purpose Foundation Models for Computational Pathology
No full textThe field of computational pathology has undergone a remarkable transformation in recent years. Researchers have leveraged supervised and weakly-supervised deep learning with varying degrees of success to address a wide range of tasks, including cancer subtyping and grading, metastasis detection, survival and treatment response prediction, tumor site-of-origin identification, mutation prediction, biomarker screening, and more. However, traditional task-specific models often require extensive labeled data and struggle to generalize across diverse pathology tasks. This limitation motivates the exploration of foundation models, which promise a more scalable, versatile solution by learning broad representations that can be adapted to various downstream applications. In this thesis, I will investigate the capabilities and limitations of data-driven foundation models in computational pathology. Specifically, I will explore two frameworks for developing general-purpose encoder models for pathology images: one using paired image-text data, and another leveraging self-supervised learning on large-scale unlabeled images. Additionally, I will examine downstream applications of these foundation models, including zero-shot transfer to gigapixel whole slide images and the development of an interactive multimodal AI assistant for pathologists.Ph.D