49082 research outputs found
Sort by
Haciendo Patria: the transatlantic construction of the Official Artist. Colombia, Ecuador, and Venezuela (1860-1890)
No full text2023This dissertation addresses the first generation of academically-trained artists from Colombia, Ecuador, and Venezuela who studied in Europe through government scholarships as a way to understand how regional fine arts institutions emerged through transatlantic dialogues with French and Italian artistic institutions. I follow the travel accounts of three artists from the Northern Andes – Epifanio Garay, Arturo Michelena, and Luis Cadena, who were the first beneficiaries of government scholarships and who played decisive roles in the foundation of art academies in their native countries. Each chapter compares the initial motivations of each government to sponsor these artists’ travels, the official participation of these governments in international events, the emerging efforts to build local artistic institutions, the individual experience of these artists in either Paris or Rome, and how the paintings produced in Europe were received in their home countries. In doing so, this study reveals a polyphony of strategies, motivations, and actors interested in importing European culture into the region of the former Nueva Granada
Modeling the health and equity impacts of climate action and air pollution control strategies at local, regional, and national scales
No full text2025Ambient air pollution poses significant health risks, with extensive research linking pollutants like PM2.5, NO2, and O3 to increased mortality and morbidity. The complex interplay between these pollutants, their sources, and atmospheric dynamics creates challenges for effective air quality management. Moreover, sociodemographic inequities in exposure to air pollution persist across multiple geographic scales, with marginalized communities facing disproportionate burdens due to historical and present-day inequities. Recent technological advancements in remote sensing, chemical transport modeling, and data integration have dramatically improved our ability to characterize air pollution exposure at fine spatial scales, even in areas lacking traditional monitoring networks. This enhanced understanding is crucial as the world grapples with climate change, presenting a unique opportunity to build solutions that simultaneously improve air quality, reduce existing inequities, and mitigate the worst impacts of our shifting climate.This dissertation explores the complex interplay between air pollution, climate change mitigation strategies, and the magnitude and distribution of health equity outcomes through three interconnected studies, each addressing fundamental aspects of air pollution exposure and health risk modeling at different geographic scales. The research examines the health benefits and equity implications of transportation emissions reduction scenarios and vehicle electrification strategies in the United States, while also providing insight regarding the health impacts of NO2 exposure in Mexico. This work collectively provides insight on alternative approaches for air pollution exposure modeling and for characterization of equity, helping to illuminate pathways for designing more impactful, equitable, and health-enhancing policies.
In Chapter Two, we explore the equity implications of various transportation emissions reduction scenarios in the northeastern United States, focusing on four distinct equity constructs: racial/ethnic exposure inequities, benefits to environmental justice communities, distribution of benefits among participating states, and rural-urban share of benefits. Using advanced chemical transport modeling, we analyze scenarios for reducing directly emitted fine particulate matter across 12 Northeast states and the District of Columbia, revealing tradeoffs among different equity constructs. Our findings highlight that scenarios resulting in greater reductions in population-weighted primary PM2.5 exposure were generally those centered in states with large urban areas, leading to greater reductions in racial/ethnic exposure inequities but higher between state or rural/urban inequality. Conversely, scenarios targeting uniform percentage emission reductions from trucks better address rural/urban inequalities but lead to smaller reductions in racial/ethnic inequity.
In Chapter Three, we evaluate the impacts of vehicle electrification strategies in the Boston metropolitan area of the Northeast United States, focusing on their potential to reduce emissions, improve health outcomes, and address existing exposure and health inequities among racial and ethnic groups. Using high-resolution chemical transport modeling, we examine a set of scenarios targeting different vehicle types within unique regions of the metropolitan area. Our findings highlight that while targeting larger vehicle fleets in suburban areas yielded greater overall health improvements, concentrating efforts on heavy-duty trucks and high-emitting vehicles in urban core areas proved most effective in reducing inequities on a per-vehicle basis. Our findings underline the importance of considering multiple pollutants and utilizing detailed health data in policy decision-making.
The final study in Chapter Four assesses the public health burden of NO2 exposure in Mexico, highlighting uncertainties in health impact assessment modeling. This work utilizes two globally modeled ground-level NO2 datasets alongside TROPOMI satellite-derived tropospheric NO2 data to analyze spatial patterns in the pollutant across Mexico and their effects on population exposure estimates and health impact calculations, with different concentration-response functions also evaluated. The analysis reveals tens of thousands of premature deaths annually attributable to ambient NO2 exposure across Mexico annually. The study finds that health estimates vary more with the choice of concentration-response function at the national scale than the exposure dataset, though it is important to note only two exposure datasets were compared. Notable differences emerge between these exposure datasets, however, at the state level, particularly near Mexico City. While demographic patterns are consistent, differences are observed for smaller subpopulations like Indigenous language speakers. This work describes the notable health impacts of NO2 across Mexico, which were previously challenging to define due to limited air monitoring networks. It also highlights the complexities involved in selecting the most appropriate inputs for air pollution health impact assessments at different geographic scales.
In conclusion, this dissertation underscores the importance of applying air pollution modeling techniques that fit both the pollutants of interest and the policy context, with heightened importance when considering local or regional contexts. Although the three chapters encompass diverse geographic scales and methodological frameworks, a recurring theme relates to the potential tradeoffs between overall public health improvements and targeted equity gains. Analyses that elucidate these tradeoffs and describe the attributes of policies that perform best across multiple endpoints will be maximally informative. These insights lay a groundwork for future research and policy development that simultaneously address air quality, climate change, and health equity
艾迪綏 Mary Ann Aldersey – Missionary in China (1839-1860)
No full textHistorians have been revising the history of Christian missionaries in China over the last few decades. In particular, they have been taking a closer look at the role and activities of women missionaries, as well as the missiology of the leading mission agencies. Mary Aldersey deserves more attention, not only because she was a more effective missionary than we knew but also because she was part of a major struggle in the mission agencies over the role and fitness of women for this work. One of the most effective mission stations was in Ningbo (formerly Ningpo) and it appears that this was due in no small part to her work. Several missionaries testified to how they were influenced by her, not the least of which was Hudson Taylor, the founder of the China Inland Mission (CIM).Boston University - China Historical Christian Database (CHCD) Project: https://www.bu.edu/history/research/china-historical-christian-database
Neural machine translation for low-resource conditions
No full text2023Neural Machine Translation (NMT) has seen significant advances in recent years and many efforts have succeeded in creating efficient and trustworthy NMT models which perform remarkably well. Yet, many issues such as lack of monolingual or parallel data for certain Languages and Language Pairs, and constraints in compute resources call for further analysis of the NMT pipeline, to understand model behavior and how different methods affect NMT results; and for a focus in the development of bilingual and multilingual models and data augmentation techniques. Our research aims to enhance the performance of NMT models in low-resource conditions by unifying multiple strategies to address these challenges comprehensively.To this end, we present a series of approaches, which attempt to improve NMT results in a wide range of Low-resource scenarios: 1. Firstly, we develop a low-resource NMT pipeline that leverages code-switching and comparable data extraction. Utilizing unsupervised, semi-supervised, and supervised training methods, we substantially improve translations for under-represented languages like Gujarati, Somali, and Kazakh when paired with English. 2. Building on these technical advances, we then conduct an empirical analysis focused on French and Gujarati translations to and from English. This investigation not only benchmarks the performance of unsupervised and supervised NMT models but also delves into model behavior, output quality, and robustness. 3. The insights gained previously inform our third approach, where we introduce an explainability-based method specifically tailored for low-resource NMT settings. 4. We extend the low-resource paradigm from the bilingual to a multilingual setup, using a Transformer-based multilingual and conditional computation-inspired model, namely a Task-level Mixture of Experts model, to boost results in Direct (non-English) NMT of a large number of Language Pairs. Our work provides a valuable understanding of NMT and lays the ground for further expansion of proposed methods to other languages and low-resource conditions
Knowing oneself in action: an account of self-knowledge of beliefs and commitments
No full text2024How do we know our beliefs and commitments which have moral significance and shape our character as advised by the Oracle of Delphi “Know Thyself”? We ordinarily both take ourselves to have, and aspire to have, certain beliefs and commitments. It is also very important to us to that we get these facts right about ourselves. Knowing which beliefs and commitments we hold, which we aspire to hold, and whether there is an ontological gap between them, is an important component of personal integrity and wellbeing. I identify two central challenges in achieving Delphic Self-Knowledge. One challenge is the problem of indifference, which arises when agents know their dispositional beliefs and actions but do not care about them. I argue that empiricist views, according to which we know our beliefs by observing how we reason, act, and react, faces the problem of indifference. The second challenge is the problem of epistemic irresponsibility, which arises when agents believe that they have a belief or a commitment on the basis of insufficient evidence from their conscious judgements and decisions. I discuss that the transparency theorists face the problem of epistemic irresponsibility because they claim that forming a judgement about what is true, or a decision about what to do, are sufficient for self-knowledge of having a belief and a commitment. I argue that these are often not sufficient. I argue that the challenges they encounter prevent the current theories on self-knowledge from providing a satisfactory account of a phenomena I call “epistemic aspiration,” which arises when moral agents aspire to have certain beliefs. I offer a self-knowledge account that explains both the significance of caring about one’s beliefs and actions, but also the significance of taking epistemic responsibility for knowing one’s beliefs and actions. To achieve this, I argue that commitments are expressive of our value-driven self. I go on to argue that knowledge of commitment requires external evidence and we need knowledge of fit between what we take ourselves to be committed to and our actions to know whether we in fact act in the way we are committed to.2027-02-11T00:00:00
Causal investigations of rhythmic electrophysiological mechanisms underlying healthy cognition and disease using transcranial alternating current stimulation
No full text2024Learning from favorable feedback is fundamental for adaptive behavior. This learning is hypothesized to be facilitated by high beta-low gamma frequency (20-35 Hz) rhythmic activity, potentially originating from the orbitofrontal cortex (OFC), but no causal evidence currently exists. In Study 1, I tested this hypothesis using electroencephalography (EEG)-guided high-definition transcranial alternating current stimulation (HD-tACS) of OFC beta-gamma rhythms. In a randomized, double-blind, sham-controlled, between-subjects experiment with 60 healthy young adults (mean age 25.8, standard deviation [SD] 5.8 years), I showed that modulation of OFC beta-gamma rhythms selectively modulates reward-guided behavior without affecting punishment-guided behavior, supporting the hypothesis. Obsessive-compulsive (OC) behaviors involve abnormalities in reward processing and OFC activity. If OFC beta-gamma rhythms facilitate reward processing, then their modulation may be a strategy for improving OC symptoms. In Study 2, I investigated this hypothesis in 64 young adults (mean age 23.9, SD 3.8 years) using a randomized, double-blind, sham-controlled experiment. These participants did not have any neuropsychiatric diagnoses but exhibited a wide range of subclinical OC tendencies, as measured using the Obsessive-Compulsive Inventory – Revised (OCI-R; baseline scores: mean 20, SD 10.3; ≥16 indicates moderate OC symptoms). I found that repetitive entrainment of OFC beta-gamma rhythms in 30-minute sessions over five consecutive days rapidly reduced OCI-R scores. Improvements sustained for three months and were stronger for individuals with more severe symptoms at baseline. These findings set the foundation for novel rhythmic neurophysiological theories and therapeutics for OC behaviors. As tACS is an emerging technology, its overall efficacy remains a matter of debate. In Study 3, I examined whether tACS reliably modulates cognitive function by performing a statistical meta-analysis of 102 peer-reviewed studies. I found evidence for improvements in several cognitive domains (such as attention, working memory and long-term memory), with improvements also evident in subgroups of older adults (age > 60 years) and clinical populations. Using meta-regression analyses, I showed the importance of using current flow models and parameters such as modulation intensity and the timing of assessment of cognitive function. These findings suggest the promise of this tool for both causal investigational and translational purposes, and identify avenues for future improvement
Statistical methods for evaluating treatment effect in the presence of multiple time-to-event outcomes
No full text2024Contemporary randomized trials frequently assess treatment effects across multiple time-to-event outcomes. In scenarios involving competing risks, prioritized outcomes, or informative censoring, alternatives to conventional methods to estimate and test for treatment effects are needed. For competing risks data, we proposed a doubly robust estimator for the difference in the restricted mean times lost to a specific cause. The estimator relies on non-parametric pseudo-observations of the cumulative incidence function, and therefore does not rely on the proportional hazard assumption. We evaluated the performance of the estimator in different scenarios of model misspecification. We applied the estimator to compare the event-free time lost to disease progression in the POPLAR and OAK studies for non-small-cell lung cancer. For prioritized time-to-event outcomes, we compared the performance of novel tests that prioritize events with higher clinical importance to traditional tests that do not. None of the tests was uniformly best when component-wise treatment effects varied. As these tests differ in how they characterize the treatment effect over the entire disease course, we proposed a generalizable framework to quantify the information used and ignored by each test. Under the Gumbel survival copula model, we also derived analytically the true value of the treatment effect corresponding to each test. We illustrated these methods using a five-component prioritized outcome in the SPRINT randomized trial. For informative censoring, we considered the issue of differential censoring between randomization groups in oncology trials. We assessed the impact of informative censoring on the treatment effect estimation, as well as on the performance of generalized log-rank tests under a delayed effect setting. We showed how to generate informative censoring data from survival copulas with piece-wise exponential marginals. We also derived the relationship between the copula rank correlation and the probability of informative censoring. We showed how to use this relationship to guide the choice of an adequate copula model to analyze informative censoring data.2027-02-12T00:00:00
Exploring free energy landscapes in complex biomolecular systems with advanced computer simulations and neural networks
No full text2023Recent advances in computer simulation and experimental techniques have motivated computational chemists and biophysicists to better understand the function of complex biomolecular systems by exploring the underlying free energy landscapes with extensive sampling and/or accurate potential functions. With creative application of existing techniques and continuing development of new methodologies, increasingly complex mechanistic problems can now be solved with computational techniques. In this dissertation, we take advantage of state-of-the-art molecular dynamics simulations and free energy approaches, as well as modern neural networks to tackle two major problems in the area of computational biophysics. The first topic was inspired by recent deep mutational scanning experiments on a transcription factor, the Tetracycline repressor (TetR), which revealed an unexpected distribution of allostery hotspots that cannot be explained by existing models. Accordingly, we have developed a new computational framework to understand the molecular basis of allostery and the broad distribution of hotspot residues in TetR. The key was to integrate long timescale molecular dynamics simulations, free energy computations and analyses of the structural and dynamical properties of TetR at both local and global scales. The mechanistic framework and multifaceted analysis strategy is expected to be applicable to many allostery systems. In the second part, we aim at improving the computational efficiency and accuracy of multi-level free energy simulations so that accurate quantum mechanical potential functions can be applied to complex biomolecular systems at the cost of an inexpensive method, such as a semi-empirical quantum mechanical approach. The solution we propose is an innovative combination of modern neural networks and enhanced sampling simulations, resulting in a computational framework that greatly improves the convergence and accuracy of multi-level free energy calculations for condensed phase systems.2025-08-05T00:00:00
Adapting to a warming climate: electricity demand, air conditioning, and the health impacts of extreme heat
No full text2023The increasing incidence and intensity of days and spells of extreme heat is expected to continue with climate change, with interconnected and cascading consequences across multiple scales and sectors. In particular, high temperature exposures directly affect population health (e.g., increased risk of hospitalization and death) and cooling energy demand (i.e., the use of residential air conditioning (AC) as adaptation). Heat extremes are often amplified in urban areas due to the thermodynamic properties of the built environment. While we have a strong understanding of the relationship between heat and energy demand, energy and AC, and the impacts of heat on morbidity and mortality, there remain notable knowledge gaps in the dynamics that underpin these relationships, and only a handful of studies are able to explore their linkages together, especially at fine spatial scales. In this dissertation, I combine econometric and epidemiological methods to provide further insights into several dimensions of the intersection of heat, electricity, AC, and health in urban populations, and holistically assess these linked relationships together. In my first chapter, I characterize the response of urban electricity demand to temperature at fine temporal resolution across a subset of world cities, and quantify the impacts of future heat adaptation on net and peak energy demand under mid-century warming. Temperature-demand response functions and future demand impacts are heterogeneous across temperate and tropical cities, highlighting the important role that the structure of electricity demand plays alongside distributional temperature shifts in evaluating the impacts of climate change on future energy demand. In my second chapter, I construct fine spatial resolution estimates of any residential AC across a large set of US metropolitan areas. Inter-urban availability of AC exhibits a strong latitudinal gradient, while intra-urban AC is systematically unequally distributed within cities. This inequality is also negatively correlated with social vulnerability (SVI) and surface urban heat island intensity (SUHI), suggesting that differential AC compounds existing heat health disparities. In my third chapter, I additionally compute individual and ZCTA-level estimates of AC use on extreme heat days alongside individual probability of AC in California cities, and evaluate the differences in the moderating effects of these related attributes of heat vulnerability on heat-related hospital admissions. AC prevalence and AC use are correlated, but both measures of adaptation are only weakly correlated with social vulnerability within cities. The spatial distribution of health risks from extreme heat echoes spatial patterns of increasing social vulnerability, and both AC prevalence and use significantly modify the association between extreme heat and a number of health outcomes. However, effect estimates differ between AC prevalence and AC use, suggesting that AC ownership does not necessarily reflect AC usage, and, crucially, that there remain additional unobserved dynamics driving the heat-adaptation-health relationship. Identifying the underlying factors and determinants of population heat health vulnerability at the local scales in which impacts and adaptation decisions take place is necessary as cities and municipalities develop and refine heat resilience policies and climate adaptation strategies aimed at reducing heat health inequities and improving community well-being
Essays in industrial organization and political economy
No full text2023My dissertation studies two topics in industrial organization and political economy. The first two chapters investigate the difference between joint and solo bidders in their bidding strategies and winner’s curse levels in common value auctions. The third chapter examines how the political strategies of the ruling party change in response to the weakening of its control over localities. Chapter 1 performs a reduced-form analysis of joint bidding and the winner’s curse in first-price common value auctions, using data for the Outer Continental Shelf oil and gas leases auctions. My coauthor, Kippeum Lee, and I introduce a novel instrument utilizing a new dataset of firms' office addresses recorded in lease contract agreements. We then examine the impact of joint bidding on bid levels, revealing that joint bidders submit approximately 75% higher bids than solo bidders on average. We further study the effect of bidder structure and competition on average bids and show that one more joint bidder (and hence one less solo bidder) leads to an around 30% increase in the average bid for an auction, controlling for the overall competition level. Chapter 2 constructs a model of common value auctions with bidder asymmetry arising from joint bidding and measures the winner’s curse. My coauthor, Kippeum Lee, and I build a model of asymmetric common value auctions by dividing bidders into joint and solo types. We then consider a myopic-bidder model where bidders do not account for the winner’s curse. To quantify the winner's curse for both joint and solo bidders, we compare the expected common value of myopic bidders to that of rational bidders who internalize the “bad news” associated with winning for each bidder type. Based on the estimation of the winner’s curse, we find that solo bidders experience a more substantial winner’s curse relative to joint bidders. Chapter 3 studies the political manipulation of central-to-local transfers in the context of Japan. I build a theoretical model of budget allocation with imperfect local monitoring, predicting changes in allocation strategies in response to municipal consolidation. The empirical analysis finds that the long-ruling Liberal Democratic Party adjusted its distributional politics in response to the weakening of its control over localities, as observed during massive municipal mergers. The party favored locally aligned villages and competitive constituencies before the merger period. However, the party shifted its allocation strategy by providing more funds to highly competitive constituencies during the merger period without differentiating between villages based on local alignment