22689 research outputs found
Sort by
THE IMPACT OF ARTIFICIAL INTELLIGENCE ON THE EFFICIENCY AND EFFECTIVENESS OF IMAGING SERVICE
No full textProblem Statement: The integration of Artificial Intelligence (AI) into radiology is significantly transforming the detection and diagnosis of lung nodules. This dissertation seeks to evaluate AI's impact on radiologists' diagnostic efficiency, the changes in their diagnostic performance, and their perceptions and attitudes toward AI-assisted lung nodule diagnosis.
Methods: Our study included a scoping review and three empirical studies at two tertiary hospitals in Beijing, China. The first study, a scoping review, evaluated AI's efficacy and acceptance in clinical radiology using the JBI guidelines for studies from March 2016 to November 2022. The second study employed a non-equivalent comparison group pretest-posttest design to evaluate changes in radiologists' reporting times for lung CT scans before and after the implementation of the AI system. The third study retrospectively examined lung nodule detection rates and measurement adjustments, comparing findings to gold standard diagnoses by senior radiologists pre- and post-AI adoption. The fourth study involved a comprehensive survey to assess radiologists' perceptions and attitudes toward AI, its usability, and its impact on their diagnostic practices and professional outlook.
Results: The deployment of the AI system initially resulted in a temporary increase in the reporting time of radiologists, indicating an adaptation period. However, this trend subsequently reversed, demonstrating improved diagnostic efficiency. Concurrently, the integration of AI-enhanced lung nodule detection reduced the incidence of missed nodules and achieved higher sensitivity but lower specificity. Radiologists generally reported positive perceptions regarding the usability of AI and its contribution to work efficiency and diagnostic accuracy. However, they expressed mixed views about the long-term impact of AI on their profession.
Conclusion: The findings of this research indicate that despite the initial challenges in implementing AI systems, the long-term integration of AI into radiology can significantly enhance diagnostic efficiency and accuracy. Radiologists' assessments of AI systems are predominantly positive, viewing AI-aided lung nodule diagnostic systems as user-friendly and appropriate for routine use. Nonetheless, these AI systems need further improvement in their learnability. While AI is widely regarded as beneficial for enhancing work efficiency and diagnostic accuracy, opinions continue to vary regarding its broader impact on the radiology profession
NARRATING BELONGING: SOMALI ADULTS’ EXPERIENCES OF ACCULTURATION AND IDENTITY
Get PDFAcculturative stress is a significant challenge for immigrants, affecting their mental well-being and integration into a new culture. This research delves into the coping strategies and cultural identity maintenance of Somali adults in Minnesota, using narrative inquiry principles. By conducting in-depth semi-structured interviews with 10 participants, this qualitative dissertation captures detailed narratives that illustrate the complexities of acculturation. The participants, varying in age, gender, and length of U.S. residence, shared their personal experiences, providing insights into how cultural, behavioral, and environmental factors influence their adaptation. Thematic analysis was employed to identify recurring themes and establish a strong theoretical framework based on the participants' experiences. Key findings highlight the significance of community support, religious practices, and personal resilience in helping Somali adults manage acculturative stress and maintain their cultural identity. These elements not only help to reduce the effects of acculturative stress but also strengthen the participants' connection to their heritage and promote community unity. This research addresses notable gaps in the existing literature by highlighting positive coping strategies often overlooked in studies that mainly focus on the adverse outcomes of acculturation, such as mental health issues and social isolation. The insights gained not only have the potential to improve academic understanding of immigrant acculturation but can also provide valuable information for future research and community initiatives aimed at facilitating smoother cultural integration
Methods and applications for large-scale pangenomic analysis
Get PDFRecent technological and algorithmic advances have propelled genome sequencing from a multi-billion dollar, decades long endeavor to common and available research practice. Thanks in large part to long-read, ultra-long-read, and high fidelity sequencing, assembling complex genomic regions from non-model organisms is now possible. While telomere-to-telomere assemblies are not yet common-place and perfect genome assemblies are not yet possible, there have been great strides on both fronts. However, it is increasingly apparent that in order to fully appreciate the rich genomic diversity of a population, and thus accurately map phenotypes to genotypes, a single reference genome assembly is insufficient. Too much genetic variation is lost with a single reference, and thus pangenomes, defined as the entire set of genetic information within a clade, are necessary to resolve complex genotype-to-phenotype relationships. This thesis presents several pieces of work related to the pangenome problem. First, we review and present a tutorial for k-mer based applications in genomics, specifically for efficiently modeling genomes from non-model species, a key first step in genome and pangenome assembly of these species. Next, we detail the construction and analysis of a genus-wide pangenome of Solanum (nightshades), with a focus on utilizing this pangenome for biological insights. We then detail a novel, alignment-free method for efficiently analyzing and visualizing large pangenomes. Finally, we discuss several applications of pangenomes, with a focus on the plant kingdom. Taken together, these chapters underscore the necessity of pangenomes to capture the full spectrum of genetic diversity and provide innovative methods and applications for their assembly and analysis, particularly within the plant kingdom
Minds Amid the Cosmos: The Ethics of Psychiatric Treatment During Space Travel
Get PDFWith the rise of commercial space programs, intentions for future planetary exploration and hopes of deep space travel in the far future, there is a growing need to examine current bioethics systems and frameworks and consider whether they are sufficient for future extreme situations. This is particularly relevant with regards to the bioethics surrounding psychiatric care during space missions. It is argued here that current guidelines are insufficient, and there is a need both for further examination and for the development of an effective bioethics regarding diagnosis and treatment (voluntary or otherwise) of emergency and chronic psychiatric conditions in the extreme environments beyond Earth’s orbit. Novel mental health concerns are addressed, as well as the need to develop a useful formulary for psychotropic medications, and the potential directions for effective future research. Additionally, this paper discusses the ethics and pertinent considerations of various mission scenarios and types, and considers questions of unwanted psychiatric treatment in space. Mission scenarios are broken down by both distance-based categories and purpose-based categories, with specific focus given to military missions, commercial missions, and missions required for humanity’s survival. Considering the risks inherent in space travel, it is likely that complex scenarios with significant bioethical tensions will be encountered by astronauts. Understanding the extent to which such tensions will involve truly novel ethical issues is important. These considerations are discussed, and an argument is offered that such scenarios may indeed involve true bioethical novelty when compared to similar scenarios on Earth
SPSA-augmented Hamiltonian Monte Carlo
Get PDFHamiltonian Monte Carlo (HMC) is an efficient Markov Chain Monte Carlo (MCMC) algorithm with fast convergence and scales well in the dimension of the target density. Nevertheless, its gradient requirement challenges implementation in large-scale problems. To compute a gradient for one leapfrog integration update, 2d measurements of the density are required. This study introduces HMC where gradients are obtained by means of the Simultaneous Perturbation Stochastic Approximation algorithm (SPSA). This approximation scheme requires only 2 density measurements per gradient evaluation and thus can facilitate simulations in high-dimensional settings. We prove convergence of the SPSA-HMC algorithm by extending the general framework in Zou and Gu [1] for unbiased gradient estimates. Furthermore, we analyze how two variance reduction methods further improve computational efficiency of the SPSA-HMC algorithm
Transition prediction in high-speed boundary layers using Bayesian deep operator networks
Get PDFTransition in high-speed boundary layers is sensitive to uncertainty in the oncoming disturbance waves. Therefore, a transition model that predicts both transition and its distribution is desirable. Such a model can be learned from direct numerical simulation data. One approach is to train an ensemble of deep operator networks (DeepONets), and to make an ensemble of predictions for each condition of interest. This strategy provides a measure of epistemic uncertainty of the network model. Alternatively, a Bayesian approach is introduced, where a single Bayesian DeepONet can quantify the uncertainty of predictions. The loss function in this case is modified to account for the aleatoric uncertainty of transition. Our results are demonstrated for a flat plate boundary layer at Mach 4.5, which is forced by a primary planar instability wave that undergoes subharmonic secondary instability and breakdown to turbulence
Rational Design of Next-Generation Organic Semiconducting Materials via Electronic Structure Methods and Machine Learning
Get PDFSignificant and persistent advances in computing power in the 21st century have enabled high-fidelity physics-based simulations like never before. Rigorous ab initio methods like density functional theory (DFT) provide valuable insight into a material’s electronic structure. As the artificial intelligence (AI) and machine learning (ML) fields develop, the ways in which we can use data from these resource-intensive calculations multiply. This thesis first demonstrates the fidelity of first principles calculations in characterizing the thermodynamic behavior of novel n- and p-type semiconducting polymer systems, some of which exhibited record-breaking performance as thermoelectric materials. It establishes, for the first time, computational evidence supporting doping via preformed Lewis acid-base complexes, and analyzes the associated frontier molecular orbitals. The focal point of this thesis is a ``building-block'' approach to semiconducting polymer design, which samples a large (10^4) combinatorial space spanned by various functional groups and solvents, in an effort to find a relationship between what can be observed at the electronic scale and thin-film semiconducting polymer device performance. Using electronic structure data to establish a physics-informed Gaussian process surrogate model, it provides insight into the complex solution chemistry that directly impacts semiconducting polymer performance. It thus demonstrates a data-driven approach to rational materials design of novel next generation materials
Essays on Covid and Capital
Get PDFCovid-19 was an unprecedented shock to the global economy, and the historic levels of uncertainty during the pandemic caused massive disruptions in financial markets around the world. The essays in this thesis are offered to provide insight into international finance and behavior in international capital markets during the Covid-19 pandemic. The fiscal policy response to Covid-19 included income support for households, and the regulatory response included forbearance on asset quality rules in the banking sector, and in Chapter 1 of this thesis, a counterfactual analysis within an expected loan loss framework was used to estimate the impact of these policies. I find that Covid-era regulatory forbearance and income support introduced a structural break in the relationship between loan performance and macroeconomic conditions, and I estimate that these policies have mitigated at least between 380 billion in expected loan losses throughout the global banking system during the pandemic. Furthermore, regulatory forbearance and income support policies in other countries have mitigated between 15 billion in expected loan losses in the cross-border loans portfolios of U.S. banks.
International capital flows are influenced by factors at their country of origin as well as their destination country, and Covid-19 was a shock to both. In Chapter 2 of this thesis I use a Gravity model of trade in financial assets to estimate Covid-19’s “push” and “pull” on cross-border portfolio holdings, and I find that Covid acted as a negative pull factor on the long-term debt of countries in Asia and the short-term debt of countries in the MENA region. I also find that Covid acted as a negative push factor on the short-term debt from Advanced Economies, but also as a positive push factor on the long-term debt of Advanced Economies, suggesting possible substitution effects among investors. These findings have policy implications, as investment dominated by pull factors can be influenced by domestic policy, whereas investment dominated by push factors cannot.
I continue my analysis of international capital markets in Chapter 3 of this thesis. Theoretically, countries with greater international capital mobility should be more likely to experience a capital shock, especially following a major disruption such as Covid-19, and in this essay I empirically investigate the relationship between international capital mobility and sudden stops and surges in portfolio flows. In general, I find that the primary effect of capital mobility reduces the probability of a stop in portfolio debt flows and a surge in equities. However, this effect is offset by a country’s risk premium, as proxied by the difference between a country’s central bank policy interest rate and the U.S. federal funds rate. These empirical results imply that a capital management policy designed to minimize the probability of a sudden stop or surge in portfolio flows while maintaining access to international markets should focus on reducing the country’s risk premium rather than imposing additional capital controls
FORMATIVE RESEARCH: UNDERSTANDING HOW TO REDUCE BREASTFEEDING INEQUITIES AMONG LATINE FAMILIES LIVING IN THE UNITED STATES-MEXICO BORDER REGION IN ARIZONA
No full textBackground
The benefits of exclusive breastfeeding are well-documented, but current practices do not meet national public health goals, and disparities in optimal infant feeding practices exist.
Objectives
This research aims to characterize prenatal breastfeeding intentions, describe infant feeding practices, and identify facilitators and barriers to exclusive breastfeeding in the first two months among Latine families in a rural county on the U.S.-Mexico border in Arizona.
Methods
This formative research is part of an implementation research project to co-design and implement a program to support breastfeeding among Latine families in a rural U.S.-Mexico border county. This study utilized the Social Ecological Model and the Theory of Planned Behavior. In-depth interviews were conducted with third-trimester pregnant (n=8) and ≤2 months postpartum (n=28) Latina women and healthcare (n=5) and nutrition service providers (n=7). Thematic analysis was completed with deductive and inductive coding in Dedoose. A secondary analysis of clinical pediatric data from a Federally Qualified Health Center was conducted, including a descriptive analysis and chi-squared tests to describe infant feeding practices in the first two months (n=214).
Results
In the sample of 214 infants, 20.1% were fed human milk exclusively in the first week, which decreased to 7.9% for one month and 6.1% for two months. Among the pregnant and postpartum women, their prenatal breastfeeding intentions were characterized as exclusive breastfeeding (n=4), breastfeeding (i.e., without specifying partial or exclusive) (n=22), partial breastfeeding (n=8), or formula-only (n=1). Most women initiated breastfeeding (n=22), but the majority fed their newborns formula at the hospital (n=18). Perceived insufficient milk, coupled with the hospital staff’s suggestion and provision of formula, were barriers to exclusive breastfeeding at the birthing hospitals. Family support facilitated exclusive human milk feeding in the first two months, while perceived or actual milk insufficiency and employment were barriers.
Conclusions
Future programs should collaborate with existing healthcare and nutrition services to strengthen counseling on breastfeeding intentions in the prenatal period, provide timely support while reducing formula provision at birthing hospitals, continue to address perceived milk insufficiency and encourage family support throughout the first two months. Additionally, workplace protections for continued breastfeeding should be ensured
PHOTON PROPAGATION IN INFANT NECK TISSUE: A MONTE CARLO AND RAY-TRACING SIMULATION STUDY
Get PDFNon-invasive central venous pressure (CVP) monitoring is critical for managing hemodynamic conditions in pediatric patients, particularly in infants where traditional invasive methods pose significant risks. This study investigates photon propagation through a simulated three-dimensional cross-section of infant neck tissue measuring 10mm x 10mm x 16.15mm, targeting the internal jugular vein (IJV). We employed traditional Monte Carlo and ray-tracing simulations at a wavelength of 630nm. Simulations were conducted in TracePro and with a custom Python model, utilizing a virtual CAD representation of the SFH 7072 OSRAM biomonitoring device for emitter and detector placement. This setup allowed us to assess photon interaction within a six-layer tissue medium.
We present detailed mathematical formulations underlying the Monte Carlo simulations specific to photon propagation in this anatomical region. Results indicate that photon detection rates at the IJV remain consistently low, with approximately 1.5% of emitted photons reaching the detector, regardless of variations in IJV diameter or total photon count. This detection percentage aligns with expected values, underscoring the challenge of achieving sufficient photon penetration and capture for accurate measurements. However, positional adjustments of the emitters and detectors demonstrated that optimal device orientation can moderately improve photon capture efficiency, highlighting a potential avenue for device optimization. Findings reveal both the limitations of photon penetration in achieving an optimal signal-to-noise ratio and the challenge of maximizing photon capture at the IJV. This study provides critical insights into optimizing device design for more effective, non-invasive CVP monitoring in infants, emphasizing the importance of rigorous mathematical modeling in understanding photon-tissue interactions