71446 research outputs found
Sort by
Essays in Financial Economics
No full textThis dissertation consists of three essays in financial economics with a common focus on the consequences of firm financing frictions for labor markets, entrepreneurship, and innovation. Chapter 1 provides novel evidence on the career returns to startup employment using quasi-exogenous variation in venture capital (VC) investment. The chapter shows how VC market conditions impact the labor market for startup talent, job outcomes, and worker career progression. It then documents how these effects vary with the technology-skill specificity of workers' human capital investments. Using novel administrative data on healthcare providers, Chapter 2 demonstrates the negative externalities of financial distress for employees and consumers: filing for corporate bankruptcy increases voluntary worker turnover, leading to worsened quality of care and patient health. By collecting a novel dataset on public-sector entrepreneurial finance programs, Chapter 3 provides a new look at government efforts to finance early-stage ventures, shedding light on their global scope, scale, and consequences.Economic
A Comparison of Natural Language Models to Subtype Ischemic Stroke from Electronic Health Records
No full textIschemic stroke is a leading cause of disability and death worldwide. While the prevalence of ischemic stroke varies across race and ethnicity, it is particularly pronounced in low-resource medical institutions, where patients may experience higher rates of post-stroke complications and mortality. Despite concerted attempts to implement established interventions and explore novel treatment modalities, the coarse classification of ischemic stroke obscures the underlying heterogeneity in both pathophysiological mechanisms and clinical manifestations, rendering these efforts insufficient. As such, there is huge value in investigating the underlying etiology of ischemic stroke in large, diverse cohorts of patients that could power refined subtype discovery. While electronic health record (EHR) data represent a valuable resource for stroke subtyping, there are few studies that have utilized EHR data for building these cohorts due to the significant human effort required to label features and cases.
To this goal, the rapid progress of Natural Language Processing methods has opened the door to fully automated stroke subtyping. This study investigates the performance of two Natural Language Processing approaches, namely Logistic Regression, a statistical model, and Clinical Longformer, a pre-trained transformer model, in subtyping ischemic stroke directly from the EHR. The models were trained and tested on EHR data of about 3000 stroke patients adjudicated by board-certified neurologists. Both models achieve commendable performance, with the transformer model slightly outperforming the statistical model with recall and precision of 0.83 and 0.74 respectively. This finding highlights that Natural Language Processing offers a more consistent and scalable approach to subtyping ischemic stroke from EHR, which could significantly enhance the statistical power and facilitate large-scale stroke research. In particular, this outcome has enabled us to infer toast subtypes across a sizable cohort of 30,000 coded strokes at Massachusetts General Hospital, thereby opening up novel avenues for investigating stroke risk prediction and genetic underpinnings.Computer Scienc
Reading the Yellow Death: Pandemic, Climate, and Memory in Early Medieval Britain and Ireland
No full textThe First Plague Pandemic ravaged Europe and the Mediterranean from its appearance in 541 through the middle of the eighth century. Until recently, little evidence existed to concretely link plague to an enigmatic disease event in Britain and Ireland called the mortalitas magna in Latin and ‘the Yellow Death’ in the vernacular. Motivated by the recent emergence of molecular evidence proving plague’s presence in Britain and Ireland, this dissertation investigates the textual tradition of the Yellow Death through the framework of current advances in plague studies. The first half of this dissertation collates and assesses all attested references to mass mortality within the Insular textual record of the sixth and seventh
centuries. These events are analysed for a potential identity of plague and contextualised within the wider history of the First Plague Pandemic in Europe and the Mediterranean. Significantly, this study identifies four plague amplification events (545–554, 576, 664–670, and 680–684) and proposes models for the transmission of plague into the Insular world. The second half of this dissertation assesses the textual history of plague, examining how memory of the pandemic shaped the development of medical vocabulary and hagiographic tropes. By tracing the references to the ‘Yellow Death’ from the first Insular annals entries of the sixth century through to the writings of Gerald of Wales in the twelfth, this dissertation demonstrates that despite patchy multi-linguistic textual record, there was a continuous shared textual memory of plague across all three major Insular literary traditions (Irish, Welsh/Breton, and English) from the end of the pandemic until the beginning of the Second
Plague Pandemic in the fourteenth century.Histor
Mouse Connectome: Enhancing the Pipeline for Building a Complete Brain Circuit Map
No full textThe fundamental mechanisms underlying the operation of the brain remain one of the biggest mysteries in science. The realization of a complete wiring diagram of the mouse hippocampal formation will revolutionize research by enabling detailed investigations into cognitive functions, memory, learning, and neurological disorders.
To build a full wiring diagram of the mouse hippocampal formation, 12,000 semithin sections of brain tissue must be cut and collected. The integrity of each section is crucial as the loss of a single section significantly compromises the ability to trace neuronal processes from one section to another. The MagC system, a novel section collection device, offers a promising solution for cutting and collecting these 12,000 sections. However, no established workflow currently exists for utilizing MagC in long-term, large-scale cutting experiments. Specifically, there is no established method to 1) track the number of sections that have been cut, 2) systematically target the sections for detailed imaging, and 3) regularly monitor the sharpness of the knife to determine when a replacement is necessary.
This senior capstone project presents an integrated hardware–software system that enhances the workflow of MagC at three critical stages: section counting, section targeting, and knife sharpness monitoring. To this end, a dual-sensor piezoelectric force measurement system was implemented to analyze cutting dynamics, and a machine vision–based software interface was developed for identifying section locations and measuring section compression. The force system enables automatic detection of cutting cycles and provides quantitative metrics—including average force, chatter index, and onset slope—to assess knife condition preemptively. The section targeting interface leverages the Segment Anything Model (SAM) for automated section segmentation and includes manual tools for correction and coordinate export, ensuring all sections can be reliably targeted using electron microscopy.
Together, these tools form a robust, scalable MagC workflow for long-term cutting experiments, directly supporting the effort to generate the most comprehensive mouse brain connectome to date.Engineering Sciences S
Designing Networks for Educational Transformation: The Art & Science of Possibility
No full textThe American high school system is facing a crisis of disengagement, with students expressing disinterest in an outdated model that prioritizes efficiency over deep learning and personal growth. This capstone examines how learning networks can be a catalyst for large-scale educational transformation, offering a future-oriented approach to high school redesign. Anchored in the research by Tony Bryk, Paul LeMahieu, Etienne Wenger, Marshall Ganz, and Ronald Heifetz, this work explores how networks can accelerate learning among systems, build collective capacity, and inspire scalable change.
As a Strategist for the Carnegie Foundation for the Advancement of Teaching, I supported the conceptualization and launch of the Future of High School (FHS) Network, a national network of high school systems moving beyond the limitations of the Carnegie Unit to advance a new Education Architecture consisting of a broader set of goals, rigorous and engaging learning experiences, and meaningful and actionable signaling systems. This capstone documents the design and execution of the FHS Network through three phases: entering a liminal state of organizational transition, conceptualizing a network that fosters deep and sustained learning, and launching a network to influence national change. By carefully recruiting and supporting twenty-five innovative school systems, the FHS Network seeks to understand the catalyzing forces needed to root, scale, and sustain a new Education Architecture.
This work argues that designing networks for transformation requires both science and art. While both are grounded in research, the science of network design involves strategic recruitment, structured knowledge-sharing, and goal-setting processes. The art, however, lies in navigating human dynamics, fostering trust, and ensuring the network remains responsive to emerging insights. Through this dual approach, the FHS Network aims to generate knowledge that not only supports its members but also informs the broader field of education.
Ultimately, this capstone contributes to the field of educational leadership by providing insights into how networks can drive large-scale change, sustain momentum, and reimagine high school as a place where all students can thrive. While it offers a roadmap for designing networks of learning, it also serves as a compass for navigating the uncertain yet hopeful terrain of national transformation.Educatio
A Quantum Memory Network Based on Diamond Nanophotonics
No full textQuantum networks hold the potential to enable quantum-secured communication, distributed quantum computing, and non-local quantum sensing. To realize these applications, a versatile quantum network infrastructure must support entanglement between network nodes capable of storing, processing, and distributing quantum information, alongside high-fidelity photonic qubits that efficiently interface with these nodes. The silicon-vacancy (SiV) center in diamond, coupled to nanophotonic cavities, has recently emerged as a promising platform to meet these requirements. This system provides access to an electron spin as an optically active communication qubit and a 29Si nuclear spin as a memory qubit, capable of storing quantum information for extended periods.
We first demonstrate that the SiV center's electron spin can efficiently generate single photons with complex spatiotemporal waveforms, facilitated by a novel asymmetric nanophotonic cavity design. We then establish a two-node quantum network between two SiV centers housed in separate laboratories and connected via optical fiber. To do so, we use photonic qubits to mediate entanglement between two spatially separated electron spins, as well as between two spatially separated nuclear spins. Finally, using bidirectional quantum frequency conversion, we convert the photonic qubits to telecommunication frequencies, enabling entanglement generation between the two nuclear spins via a 35 km fiber deployed in the Boston metropolitan area. These advancements mark progress toward large-scale, deployable quantum networks using SiV centers coupled to nanophotonic cavities.Physic
Architected Liquid Crystal Elastomers with Spatially Programmed Alignment, Shape Morphing, and Mechanics
No full textLiquid crystal elastomers (LCEs) are responsive soft materials that undergo reversible shape morphing when cycled above their nematic-to-isotropic transition temperature. This property, coupled with their programmable alignment and mechanics, makes LCEs ideal for advanced applications in adaptive structures, energy absorption, and artificial muscles. However, fabricating monolithic LCEs with spatially varying director alignment in arbitrary architected structures remains a significant challenge. To address this, my Ph.D. thesis focuses on developing a universal framework to correlate printing conditions with director alignment, laying the groundwork for fabricating architected LCE lattices with spatially programmed alignment, shape morphing, and mechanics.
During extrusion-based 3D printing, LCE inks experience coupled shear and extensional flows, that enable spatial control of nematic director alignment along prescribed print paths. Combining experiments and computational modeling, we investigated the effects of ink composition, nozzle geometry, and printing parameters on flow-induced alignment. Rheological measurements revealed that the Weissenberg number (Wi) strongly predicts alignment, with uniform alignment achieved at Wi >> 1. COMSOL simulations and in-operando X-ray measurements confirm that hyperbolic nozzles produced printed LCE architectures with improved alignment compared to tapered nozzles, resulting in enhanced stiffness and actuation strain. By varying Wi during printing, LCE architectures with uniform composition yet locally encoded degree of alignment, and hence shape-morphing transitions were realized.
Next, we fabricated architected LCE lattices with a high degree of flow-induced alignment via direct ink writing and systematically characterized their shape morphing, stiffness, and energy absorption across strain rates spanning six orders of magnitude. Compared to non-mesogenic elastomeric (silicone) counterparts, LCE lattices exhibit superior energy absorption, with energy absorption ratios up to 18-fold higher at the highest strain rates. A finite element model capturing their shape-morphing response shows excellent agreement with experimental data. In summary, this work demonstrates the potential of architected LCEs as programmable soft materials for myriad applications that require stimuli-responsive, tunable properties.Engineering and Applied Sciences - Engineering Science
Design of custom CRISPR-Cas9 PAM variant enzymes via scalable engineering and machine learning
No full textCRISPR-Cas nucleases have facilitated the widespread adoption of precise genome editing in the laboratory. However, CRISPR-Cas mediated genome editing is constrained by the strict requirement for a protospacer adjacent motif (PAM) flanking the genomic target site. This limits the use of SpCas9 to genomic positions which are flanked by an NGG motif. A suite of SpCas9 variants each targeting a distinct PAM would expand the range of accessible genomic sequences while maintaining the specificity and allele-discrimination properties imparted by a PAM requirement. This dissertation describes the development of a range of SpCas9 variants with altered PAM specificities. PAM-altered SpCas9 variants are identified by structure/function-informed saturation mutagenesis followed by bacterial selections. Next, full PAM requirements are characterized for a set of ~1000 enzyme variants and used to train a machine learning model to relate PAM specificity to amino acid sequence. This PAM ML algorithm (PAMmla) is used to predict the PAM requirements for 64 million enzyme variants, leading to novel SpCas9 variants with unique nucleotide preferences at the third and fourth position of the PAM. PAMmla-predicted enzymes outperform evolution-based enzymes and highly optimized SpCas9 variants (e.g. SpG and SpRY) as nucleases and base editors across various sites in human cells and show consistently fewer genome-wide off targets. A second rational engineering strategy, termed “SpRYbridization”, is used to further engineer PAMmla-derived enzymes to relax nucleotide preference at the second position of the PAM, expanding targeting range beyond the NG PAM space while maintaining more specific PAM preferences than SpRY. Together, this work establishes the feasibility of integrating ML with protein engineering to derive a catalog of bespoke SpCas9-based enzymes, achieving a greater plasticity of the PAM interacting domain than previously explored. This framework for quickly identifying safe and effective SpCas9 variant enzymes motivates a shift away from generalist genome editing technologies towards custom editors for a wide range of genome editing applications.Biological and Biomedical Science
The Role of Community Gardens on Environmental and Human Health in Urban Settings in the United States
No full textTo feed a growing population, the global food system relies on High Yield Variety crops (HYVs), fertilizers, pesticides, controlled irrigation, heavy machinery, and numerous other cultivation practices that have been shown to harm the health of the environment and humans alike. The United States is the world’s largest food exporter totaling over $174 billion in 2023. This production of agricultural products, however, is focused largely on commodity crops while 60% of the fruits and 35% of the vegetables consumed in the United States are imported. The long distances that U.S. food travels have significant environmental impacts and also make the U.S. food system highly vulnerable to external shocks. This vulnerability was demonstrated during the COVID-19 pandemic when supply chains collapsed because migrant farm workers could not reach their destinations due to travel restrictions; and large-scale, highly centralized food processing facilities had to shut down due to high COVID-19 infection rates.
My research explored one aspect of local food systems that can guard against supply chain shocks and stabilize the availability of fresh, unprocessed fruits and vegetables in U.S. cities where 80% of U.S. consumers live, namely community gardens. Research questions included: (1) how do the individual, social/community, and environmental benefits of community gardens, identified in the peer-reviewed literature published between 2000-2024, motivate gardeners; and (2) did Washington, D.C. residents’ attitude toward gardening and urban food systems change during and since the end of the COVID-19 pandemic? I examined this second research question through survey of gardeners which asked about their attitude toward gardening and possible changes in attitude during and since the end of the COVID-19 pandemic. Broadly, I hypothesized that community gardens play in improving individual, community, and environmental health; and that Washington, D.C. gardeners sought the benefits of gardening during the pandemic, shifting to finding alternative local food solutions.
A review of the literature between 2000-2024, which was conducted in 5-year increments, revealed a growing interest in community gardens and a broadening scope of research to determine the full extent of benefits that community gardens can have on their communities. One of the core themes that emerged was that community gardens can positively impact human, environmental, and community health when the role of the garden meets the needs of the community members. The review indicated that the motivating factors for gardeners evolved as urban gardeners responded to societal trends. In the early 2000s, the primary research focus was on community building and the physical health benefits of gardening. This shifted after 2010, as researchers and gardeners alike were interested in the environmental and public health benefits of gardening. After the COVID-19 pandemic, researchers focused more on the mental health and wellbeing and community resilience benefits of urban gardening.
To understand the role of community gardens in Washington, D.C., and the factors that motivated gardeners, both pre-and post-pandemic, I surveyed local gardeners. Initial results indicated that participants primarily garden to feel closer to nature and increase their access to healthy food. The pandemic highlighted their desire to connect more closely with where their food was coming from and generally increased their interest in gardening.Extension Studie
Learning to Fragment Molecular Graphs for Mass Spectrometry Prediction
No full textMass spectrometry is a common technique used to identify metabolites, small molecules that are yielding important new insights into a number of important biological and physiological processes, including organ function, nutrient sensing, and gut physiology. However, identifying these metabolites is a difficult task. Most methods for metabolite identification compare the spectrum of an unknown compound against a database containing referential spectra of known compounds. However, this approach fails if the true compound is not in the reference database, motivating approaches to generate a larger standard library by directly learning to simulate the forward fragmentation process. In this thesis, I propose a machine learning framework that learns the fragmentation pathways for each molecule to predict its mass spectrum, drawing upon both advances in machine learning and domain insights from analytical chemistry.Computer Scienc