49755 research outputs found
Sort by
Non-Euclidean Representation Learning with Applications to Metagenomics
No full textMachine learning relies on high-quality data representations to enable accurate modeling and prediction. However, biological data, particularly in fields like metagenomics, presents unique challenges for machine learning due to its high dimensionality, noisiness, compositionality, and complex latent interactions and hierarchies. We explore the adaptation of non-Euclidean representation learning techniques to metagenomics.
First, we present two works on modeling metagenomic samples’ latent structure and dynamics: MiSDEED, a tool for generating realistic synthetic data based on generalized Lotka-Volterra dynamics, and a method for inferring microbial growth rates from 16S amplicon data by extending peak-to-trough ratio analysis.
Next, we develop novel machine learning models that can operate on non-Euclidean data representations: first, we introduce decision tree and random forest algorithms for hyperbolic spaces, then generalize these to mixed-curvature product spaces. We further introduce two major engineering efforts to improve the efficiency and accessibility of non-Euclidean machine learning: a method for speeding up non-Euclidean decision trees by three or more orders of magnitude, and a comprehensive Python library supporting end-to-end machine learning on product manifolds.
Finally, we present preliminary explorations into the role that non-Euclidean representations can play in metagenomic settings. We propose using weighted centroids to aggregate feature (species)-level embeddings into sample embeddings, aiming to learn biologically meaningful representations that can improve downstream tasks such as sample classification.
We also explore the empirical distribution of curvature measurements via Monte Carlo sampling, offering a much-needed calibration for curvature estimation in the presence of confounding variables. By adding new methods for probing and evaluating non-Euclidean embeddings, we strive to unlock the potential of representation learning to advance microbiome research and its applications in human health and beyond
Development of versatile nanosystems for protein delivery and scavenging of pro-inflammatory molecules
No full textComplex diseases often arise from intricate interactions between genetic predispositions and environmental factors, leading to cellular dysfunctions that manifest across a range of pathological conditions. Addressing these multifactorial challenges necessitates innovative therapeutic strategies capable of simultaneously targeting multiple disease pathways. This thesis explores the development and optimization of versatile nanosystems designed for protein delivery and scavenging of pro-inflammatory factors, aimed at treating diverse and complex conditions.
The research introduces three distinct approaches: (1) the development of a polymeric nanosystem for oral delivery of the pegfilgrastim protein as a radiation countermeasure for hematopoietic acute radiation syndrome (H-ARS); (2) the adaptation of this polymeric nanosystem for glucose-responsive, antibacterial, and antioxidant treatment of diabetic wounds; and (3) the creation of a 2D nanosheet for scavenging small extracellular vesicles (sEVs), aimed at mitigating metastasis in triple-negative breast cancer post-radiotherapy.
These studies utilized a combination of material synthesis, advanced characterization techniques, and delivery strategies to enhance the therapeutic functionality of the delivered cargos. For H-ARS, the polymeric nanosystem was engineered to improve bioavailability and enable controlled release, ensuring effective oral delivery of the PF protein. The glucose-responsive polymeric nanosystem, encapsulated within a P(NIPAm-co-AAc) hydrogel, facilitated responsive release under high glucose conditions, scavenging cell-free DNA and reactive oxygen species, while inhibiting bacterial growth to accelerate wound healing in mouse models. Additionally, the 2D cationic nanosheet was designed to capture and neutralize tumor-derived small extracellular vesicles (sEVs), reducing their role in metastasis and enhancing therapeutic outcomes following radiotherapy in breast cancer models.
The findings from these investigations demonstrate the significant therapeutic potential of the engineered nanosystems across three distinct disease models. The rational design of these nanocarriers addresses several key challenges in contemporary medicine, including the highly challenging oral delivery of proteins, the intricate process of diabetic wound healing, and the suppression of radiotherapy-induced metastasis. This dissertation not only bridges a critical knowledge gap regarding the strategic deployment of nanomaterials to overcome these complex pathological conditions but also contributes to the development of advanced tools for the next generation of precision medicine. By elucidating how nanosystems can be systematically optimized for specific therapeutic applications, this work provides a foundation for innovative approaches that could enhance treatment modalities across a spectrum of diseases
Posting Politics: Essays on the Supply Side of Social Media
No full textThis dissertation explores how social media engagement can shape political content on social media by shaping incentives, stimulating content creation, and steering algorithmic curation. Across three papers, I reconcile interdisciplinary findings with the unique attributes of political content.
My first paper presents an analytical framework for understanding how consumer interests and algorithmic sorting influence the types of content produced on social media platforms. Building off of a Downsian framework, I model two producers who adjust the content they create in order to maximize their reach, given the production point of their competitor. Unlike typical Downsian models, social media engagement can come both from preferences being very close to content, or very far, what I term concordant and discordant engagement, respectively. I show that polarization of content production can occur with a sufficient prevalence of discordant engagement, even without polarization in the population or producer preferences. I support this finding through interviews with content creators, including media staffers for Members of Congress.
In my second paper, I investigate how engagement signals affect the production of comments and original posts, in both political and non-political subreddits. I conduct a series of field experiments on Reddit, contrasting both commenting and posting behaviors. I find that Reddit Awards seem to incentivize increased comments for new users, but do little to move veteran Redditors. I find weak evidence for a relationship in the opposite direction for individuals who post, rather than comment. These results suggest that engagement can affect certain types of original content production, including political content. However, posting and commenting are different behaviors that appear to have distinct relationships with engagement and user tenure.
My third paper presents two TikTok experiments designed to highlight how algorithms respond to engagement signals. Again, my aim is to highlight how political content is treated. I conduct an algorithmic audit to show how engagement signals can alter initial recommendations. I find that effect sizes are conditional on the topic of interest, noting that engagement with political content appears to trigger relatively high rate of related recommendations. I support these audit results with a lab experiment, examining how initial engagement signals persist over time. I observe algorithmic behavior over 40 minutes of browsing by treatment-blind users. I find that political recommendations persist for treated accounts, even after significant browsing time. I also present preliminary results for algorithmic effects on user attitudes and experiences
Three Essays on Food Insecurity in the United States
No full textThis dissertation includes three papers that examine the programs, policies, and contexts that explicitly or implicitly relate to food insecurity in the United States.
Paper I estimates participation in social welfare programs across disability status and examines how access to participation differs by level of unmet food need in the household. Paper II estimates the causal effects of state minimum wage increases on household food security status across varying levels of wage increases and household characteristics. Paper III explores the relationship between food security status and metro-area classification, investigating how distributional differences in individual, household, and state-level characteristics across geographic location influence this relationship
The Impact of Concept Based Inquiry on Clinical Reasoning
Get PDFClinical reasoning (CR) is an essential skill for nursing and is necessary for safe and effective nursing practice. This skill is critical in how well nurses make clinical decisions and take action to assist their patients. However, this skill is significantly lacking in new nurses, and an academic-practice gap is thought to be the reason. This study investigated the impact of a new concept-based inquiry teaching approach on nursing students’ CR and its relationship with demographics, retention, critical thinking (CT), and metacognition. A quasi-experimental pretest and posttest study design using a non-equivalent comparison group was conducted with undergraduate students enrolled in a baccalaureate pre-licensure program using a concept-based curriculum.
The concept-based inquiry (CBI) approach was introduced to students in the intervention group, while the control group participated in a flipped classroom teaching approach. Both groups engaged in content related to the concept of gas exchange. Clinical reasoning was measured using an NCLEX Next Generation (NGN) style exam created by the author, the clinical reasoning exam (CRE). Test analysis on the exam revealed challenges in writing and using these types of exams to assess CR and were discussed. The relationships of CBI to demographics, critical thinking, and metacognition were also evaluated. A demographic survey assessed age, gender, race/ethnicity, learning disabilities, work experience, military experience, and education level. The Health Sciences Reasoning Test assessed CT, and the Metacognitive Awareness Inventory assessed metacognition.
The results of the study revealed no significant difference in CR between the CBI intervention group and the Flipped Classroom comparison group after the gas exchange class. No significant relationships to demographics or knowledge retention were noted. However, the study revealed that the CBI significantly affected CT and metacognition, critical elements of the CR process. This study demonstrates that the CBI approach may be a promising new teaching method that should be investigated further to determine if using it for a more extended time period would have a greater impact on CR
Navigating the Relational Work and Emotional Labor of Relationship Building in Teaching: Forming Teacher-Student Relationships in the Context of Perceived Challenging Student Behaviors
Get PDFTeachers engage in relational and emotional work as they develop relationships with their students, forming teacher-student dyads that impact the lives and experiences of both the teacher and student. Within their schooling contexts, some teachers become known for their ability to build teacher-student relationships, particularly with students who demonstrate behavior that may be challenging or deviate from the explicit and/or implicit socially accepted norms of formal schooling – students who are formally or informally labeled as demonstrating misbehavior. The purpose of this research was to explore the professional experiences, knowledge, and skills of these teachers – who consistently build relationships with students whose behaviors may be described as challenging by others – with particular emphasis on the approaches the teachers enacted in building relationships with students, what experiences informed their approaches, and how these teachers navigated the emotional labor of teaching and being in relationship with their students. Utilizing a qualitative practitioner inquiry-based research approach, participants were 13 inservice elementary educators teaching in the United States, whose classrooms ranged from first grade through fourth grade (Cochran-Smith & Lytle, 2009).
Extant research within the field of student misbehavior is difficult to locate, frequently lacks the perspective of the student or the teacher, and adopts a behavioral stance with the goal of attributing the source of the misbehavior. In addition, while research studies illuminate the ways relational approaches are beneficial, this relational approach has not yet been applied to the area of student misbehavior or to identify specific professional skills, knowledge, and experiences that inform how educators who consistently build relationships with these students approach their work, how they learned these skills, and how they think about and navigate the emotionality embedded in their work.
Using data from multiple sources (semi-structured interviews, focus groups, ongoing anecdotal exchanges, fieldnotes, and reflective journals) and analyzing a collective 222 years of teaching experience, the study demonstrates how teachers who enact relational approaches when fostering teacher-student relationships adopt the stance of researchers of the lives and stories of children. In these teacher-student relational dyads, teachers described their emotional experiences and the emotional labor of the teaching profession, including the systemic challenges that impeded their relational work. Teachers in the study used strategies to navigate their emotional labor, and these strategies were often developed from their early life experiences, observing other educators, experiential learning in the field, and practicing self-reflection. In particular, teachers negotiated the profession's emotional demands through emotional and relational modeling, or what emotional labor theory poses as surface and deep acting (Hochschild, 2012). The enactment and use of relational and emotional modeling ultimately enabled teachers to be authentically engaged as researchers of the lives and stories of children, forming teacher-student dyads with students whose behaviors were perceived as challenging.
The research findings have significant implications for preservice teacher education and inservice teacher professional development, particularly in supporting teachers as they develop teacher-student relationships and navigate the emotional labor of teaching through relational and emotional modeling
Age-Related Changes in Information-Seeking Behavior about Morally Relevant Events
Get PDFWith age, people increasingly emphasize intent when judging transgressions. However,
people often lack information about intent in everyday settings; further, they may wonder about
reasons underlying pro-social acts. Three studies investigated 4-to-6-year-olds', 7-to-9-year-olds',
and adults' (data collected 2020-2022 in the northeastern United States, total n=669, ~50%
female, predominantly White) desire for information about why behaviors occurred. In Study 1,
older children and adults exhibited more curiosity about transgressions versus pro-social
behaviors (ds=.52-.63). Younger children showed weaker preferences to learn about
transgressions, versus pro-social behaviors, than did older participants (d=.12). Older children's
emphasis on intent, but not expectation violations, drove age-related differences (Studies 2-3).
Older children may target intent-related judgments specifically toward transgressions, and doing
so may underlie curiosity about wrongdoing
Search for same-charge top-quark pair production in collisions at √ = 13 TeV with the ATLAS detector
Get PDFA search for the production of top-quark pairs with the same electric charge or ̄̄) is presented. The analysis uses proton-proton collision data at √ = 13 TeV, recorded by the ATLAS detector at the Large Hadron Collider, corresponding to an integrated luminosity of 140 fb⁻¹ Events with two same-charge leptons and at least two -tagged jets are selected. Neural networks are employed to define two selections sensitive to additional couplings beyond the Standard Model that would enhance the production rate of same-sign top-quark pairs.
No significant signal is observed, leading to an upper limit on the total production cross-section of same-sign top-quark pairs of 1.6 fb at 95% confidence level. Corresponding limits on the three Wilson coefficients associated with the ⁽¹⁾_, ⁽¹⁾_, and ⁽⁸⁾_ operators in the Standard Model Effective Field Theory framework are derived
Vanishing of quadratic Love numbers of Schwarzschild black holes
Get PDFThe induced conservative tidal response of self-gravitating objects in general relativity is parametrized in terms of a set of coefficients, which are commonly referred to as Love numbers. For asymptotically-flat black holes in four spacetime dimensions, the Love numbers are famously zero in the static regime. In this work, we show that this result continues to hold upon inclusion of nonlinearities in the theory for Schwarzschild black holes. We first solve the quadratic Einstein equations in the static limit to all orders in the multipolar expansion, including both even and odd perturbations. We show that the second-order solutions take simple analytic expressions, generically expressible in the form of finite polynomials. We then define the quadratic Love numbers at the level of the point-particle effective field theory. By performing the matching with the full solution in general relativity, we show that quadratic Love number coefficients are zero to all orders in the derivative expansion, like the linear ones
The Legacy of Antiblackness: Intergenerational Narratives of Black Students in Philadelphia’s Public Schools
No full textThis dissertation centers the complicated interplay of history, racial politics, and community sensemaking to explore how antiblackness is reproduced over time and in different public school contexts within Philadelphia. By exploring the intergenerational educational experiences of Black Philadelphians, I document current and former Black students’ understandings of the different manifestations of antiblackness in educational policy and practice over several decades.
This research, therefore, brings the voices of Black current and former students into a historical analysis of the role of antiblackness in shaping schooling experiences and what it means (and meant) to be Black in different school contexts during different time periods. I examine this meaning of Blackness over time and across school buildings through interviews with current and former Black high school students, followed by more in-depth intergenerational interviews of three families to understand how memories of schooling shape intergenerational understandings of and experiences in schools. I contextualize this qualitative data through a review of published historical literature and archival research on the history of Black education in Philadelphia. Studying education through the intergenerational experiences of families is an under-utilized approach within educational research.
Yet, through this study, I demonstrate that it is an approach that reveals how various processes of social reproduction—particularly antiblackness—both persist and morph within the educational system over time as well as how family schooling memories and perspectives shape students’ educational journeys. Through the eyes of multiple generations of Black families in a northern city that was a destination for millions of Black people who moved from the South during the Great Black Migration, my dissertation contributes to the research on the enduring racial inequality in our schools from the early-20th century to today by showing the evolution of antiblackness over time