33760 research outputs found
Sort by
Decision-Making Aid for Human-Machine Teaming in Multiplayer Pursuit-Evasion Games
No full textEffective decision-making in adversarial multiplayer games is complicated by the large number of degrees-of-freedom in state-action pairs by the players. Difficulty is exacerbated by complex objectives that often develop in strategic deployment of tactics that coordinate multiple objectives. In this paper, we consider the decision-making problem for human-machine teaming to facilitate strategic command and control. We formulate decision-making problems for multiplayer pursuit-evasion games with multiple strategic objectives of varying complexity, such as capturing an evader, or steering it away or towards parts of the domain. We test the complexity of decision-making tactics through human-subject experiments. Operators are able to command a team of pursuers by selecting formations and motion primitives from a library of tactics to influence an evader's behavior. The designed interface allows the operator to perform practice trials in simulation that enable testing various strategies and examine time-risk before deployment on hardware platforms. Through surveys and metrics, we show that such an interface is found to be useful for decision-making by users. Performance was not impacted in any significant manner by the number of practice trials. This result suggests the impact a decision-making aid could have on improving performance in command and control of human-machine teams.
Previous chapterhttps://doi.org/10.2514/6.2025-227
Essays on Content Management on Digital Platforms: Design and Provision Implications of Online Reviews, Multimedia, and AI-Generated Content
No full textDigital platforms host vast amounts of content in different formats, such as text, image, and video, generated from diverse sources including human and artificial intelligence (AI). The sheer volume and heterogeneity of content in formats and sources introduce opportunities, as well as complexities and challenges for platforms to manage the experience of users, who may act as content consumers or content creators. This dissertation consists of three essays that examine current practices and strategies for content design and provision, and their implications on user behavior across multiple empirical settings. Essay 1 focuses on a common type of content: online reviews. I investigate the impact of a unique review display strategy on the decision-making of potential consumers in a consumer review platform. The findings show that simply adjusting the review display structure to highlight a reviewer’s consumption history helps identify repeat buyers, which serves as a signal of quality and boosts business performance. Essay 2 studies knowledge communication through two distinct communication media, specifically text and video, and compares their impact on the psychological processes underlying knowledge transfer in a social Q&A platform. I find that videos, compared to text, stimulate affective, social, and perceptual processes to a greater extent, while both communication media are comparably effective in triggering cognitive processes. Interpretable video analytics further provides granular insights into the content design of knowledge-intensive videos. Essay 3 focuses on an online mental health forum that connects support seekers with counselors and explores how human counselors respond to the integration of a chatbot that delivers AI-generated support. The findings indicate a noticeable decline in the baseline level of services provided by human counselors, marked by their reduced engagement and less frequent use of psychotherapeutic strategies. The three essays collectively provide theoretical insights and practical guidance for managing complex and heterogeneous content on digital platforms
The Body Yet Becomes the Light
No full textTHE BODY YET BECOMES THE LIGHT examines the mind’s fluidity, how it fills in gaps in family history, encounters racial alterity, and what a writer brings to the poetic line even if they do not intend to. These poems investigate the self as a site of interruption, how a bird’s eye can reflect unspoken violence—how a candle can capture the dream burning in it. The collection explores how we constitute ourselves through our surroundings, whether that be through the child working at the restaurant or the writer at the desk. These poems scrutinize the edges of dream, work, and nature through a situation of personal diaspora; how does labor affect a child, a family? What does it mean to want to become a myth, or animal-like, or to write at all? The collection contends with escape—exploring how language collides with a fraught world and in turn reinvigorates, relinquishes, and reveals parts of us that lay unsprung
Essays on Race in Political and Public Institutions
No full textThis dissertation investigates the following features of liberal-democratic society in political and public institutions through the lens of race: the freedoms of speech and expression, the right to vote, and humane engagement with law enforcement.
The second chapter, "The Impact of Conservative Media on Southern Partisan Realignment", examines the most listened to pro-segregation radio broadcaster of the 1960s, Carl McIntire, to gauge conservative media's impact on Southern partisan realignment. I exploit irregular topographical contours as a source of exogenous variation in exposure to McIntire's political program. I incorporate these signal strength measures in a triple difference-in-differences design for estimation. A one standard deviation increase in predicted signal strength is associated with a 2.31 p.p. decline in Democratic vote share in the South. This effect corresponds to an 18% increase in overall realignment. His impact on realignment is driven by local ideological competition on preserving segregation.
The third chapter, "Voter Mobilization or Electoral Oversight?: Exploring Determinants of Black Re-Enfranchisement", analyzes Black political participation at the height of the Civil Rights Movement. While the Voting Rights Act of 1965 (VRA) is credited with the large-scale "re-enfranchisement" of Blacks following some 75 years of restrictive test devices, state-level registration rates call into question if the VRA alone is responsible for reigniting Black political participation. The political economy literature has conjectured that pre-VRA gains in black registration could be attributed to mobilization campaigns funded by the Voter Education Project (VEP). This chapter is the first to put this theory to the test. I estimate a triple difference-in-differences design that permits a statistical horse race between VEP, VRA, and VEP-by-VRA counties. Results indicate that the VEP was equally effective in bolstering Black turnout relative to the VRA from 1964 to 1980. Further suggestive evidence links pre-VRA gains to the VEP. White resistance to Black registration is predictive of long-run gains, but organization type is most associated with pre-VRA gains.
The fourth chapter, "Understanding Policing in the Aftermath of Gun Violence: Examining Investigatory Stops and Crime in Chicago", examines investigatory stops conducted by Chicago police officers following recent gun violence. I use a staggered difference-in-differences design to dynamically compare stops between neighboring police beats. In the immediate aftermath of gun violence, I find a rise in total stops that corresponds with a decline in index crime incidents and results in more seized weapons and/or contraband. Over the next month, officers continue to conduct additional stops, but these stops prove unproductive in yielding illicit goods, and seemingly have no impact on crime. Stops resulting in enforcement action persist three months after gun violence. Non-White stops in regions with considerable gang concentration account for much of the baseline rise in total stops.
The concluding chapter, "Insights into American Black Journalism: From the Freedom's Journal to the 21st Century", examines trends in Black print media from 1820 to 1997. Black media access accelerated in industrial Northern centers around the timing of the Great Migrations. Regional path dependency was deterministic of the rate of print publications over time: differences between smoothed South and Non-South rates never converged over this roughly 180 year window. Evidence suggests that Black authors wrote on a diverse set of subject matter. Finally, burgeoning literary rates in Black communities are surprisingly not correlated with the founding of new Black newspapers and periodicals
Uneven Access: Exploring Computer Science Course Availability Across U.S. States
No full textThe goal of this research is to explore the relationship between the number of middle schools offering computer science and state-level factors like funding, urbanization, and government policies. We’re using Everett Rogers’ Diffusion of Innovation theory, which explains how new ideas and innovations spread between populations. Key concepts like complexity and time may help explain differences between states. Our research questions explores: How does the number of middle schools offering computer science vary across different states in the United States, and what factors may contribute to this variation? For qualitative data, we’re reviewing individual state government policies around CS education. Our quantitative data includes the number of middle schools offering CS within each state and the total number of middle schools in each state. We’ll compare these to identify patterns, disparities, and possible causes. Based on policy data, common barriers include a lack of resources, certified teachers, enrollment, and time. Complexity impacts adoption. And since computing is still relatively new, this may explain why offerings vary. States like Alabama, Arkansas, and Maryland are above the 75%, while Wisconsin, Montana, South Dakota, DC, Oklahoma, and New York are below the 25%. Factors may include: policy, funding, teacher supply, school district size, location, access to tech, and even CS job availability after college. Interestingly, states with higher total enrollment, like California and Texas, don’t always have the highest CS offering rates. These findings highlight the need to expand CS access nationwide, which could help shape future policy or research
SPATIAL ANALYSIS OF SPOT (LEIOSTOMUS XANTHURUS) POPULATION DYNAMICS
No full textSpot (Leiostomus xanthurus) is a forage fish that supports important commercial and recreational fisheries along the U.S. Atlantic coast. Assessment approaches to estimate abundance and mortality of spot for management on the U.S. Atlantic coast have previously not passed peer review. Our objectives were to (1) estimate spatial differences in biomass and mortality rates over time using fisheries-dependent indices and total mortality rates, and 2) estimate Spot abundance and fishing mortality rates in the Chesapeake Bay and the Atlantic coast using spatial population models. We calculated indices of biomass using commercial trip level data from gillnet fisheries and estimated total mortality rates using the Chapman-Robson approach with ageing data from fishery independent and dependent monitoring programs from Maryland to South Carolina. Dynamic factor analysis (DFA) was used to test for spatial differences in trends in biomass and mortality. We developed a two-region statistical catch-at-age model using the Template Model Builder package in R to distinguish the portion of the population along the Atlantic coast (New Jersey to Florida) from the portion in the Chesapeake Bay. We modeled ages 0-3+ for 2002-2019 with age-specific occupancy probabilities for each region. The data inputs for our model included multiple survey indices, commercial and recreational catch, shrimp trawl bycatch, and age composition of each data source. We implemented two models: one where fishing mortality varied between regions (SVM) and another where occupancy probabilities varied over time (TVO). Indices of biomass varied widely across space and time. Spot appeared to have high total annual mortality rates (>80% per year) that have increased in recent years. DFA indicated that stock biomass and mortality rates appeared to follow similar trends across regions. The TVO model was a better fit to the data and produced more biologically plausible estimates than the SVM model. In the TVO model, spot abundance was similar in both regions and has increased since 2015, which was a very poor year class. Patterns in fishing mortality differed between fleets with the commercial fishery declining substantially over time. The results from this work will aid in the development of an upcoming stock assessment for spot to inform management
STRUCTURAL AND EXCITONIC PROPERTIES OF QUANTUM DEFECT-TAILORED ULTRASHORT CARBON NANOTUBES FROM DENSITY FUNCTIONAL THEORY
No full textFluorescent ultrashort nanotubes (FUNs) have recently received renewed interest as a class of luminescent nanomaterials, owing to developments in sp^3 quantum defect chemistry in single-wall carbon nanotubes (SWCNTs). By tailoring ultrashort SWCNTs with quantum defects, their typically quenched photoluminescence is re-activated and brightened. Since the potential uses of FUNs for applied and fundamental science have only begun to be explored, there exists a need for a theoretical understanding of their basic optical and electronic properties. To that end, we have performed quantum chemical calculations within a density functional theory framework to explore these properties. By performing ground and excited electronic state calculations on molecular models of ultrashort SWCNTs, including those with sp^3 quantum defects, we predict the following properties: 1) there exist natural molecular models of ultrashort SWCNTs, in the sense that they have the lowest ground state energy amongst their isomers, and exhibit the correct absorption spectra, 2) the lowest bandgap excitonic absorption in SWCNTs has a length-dependent energy shift ∆E that follows a scaling law ∆E ∼ L^{−1/2}, for a SWCNT model of length L, departing from a standard particle-in-a-box view of quantum confinement in SWCNTs, and 3) ultrashort metallic SWCNTs can host organic color centers made by sp^3 defects due to the quantum confinement effect, contrasting with the usual view of metals as fluorescent quenchers. These predictions have implications for the edge-structure of SWCNTs and the length-dependence of their optical properties, providing a theoretical picture of FUNs that points to a gap in understanding of these materials and suggests a need for greater experimental focus on the properties of ultrashort nanotubes
Willingness to Pay for Express Lanes, Heterogeneity Effects, Reliability Measures, and Optimal Control Strategies from Passively Collected Data
No full textIn March 2020, governments worldwide implemented various social restriction measures, including lockdowns, to mitigate the spread of the COVID-19 virus. This led to a profound shift in travel behavior, accompanied by a notable reduction in congestion. This dissertation investigates travel demand on Express Lanes (ELs) during two time periods: before the pandemic (from January 2020 to mid-March 2020) and during the pandemic (from mid-March 2020 to the end of May 2020). A large data set of trips observed on Express Lanes derived from transponders is combined with vehicle probe measures of travel times and used to estimate the willingness to pay (WTP) for travel time savings and reliability. This dissertation also addresses system-wide performance optimization by integrating multiple control strategies.
This dissertation is structured into four main chapters. The first chapter explores the change in travel behavior on Express Lanes and assesses changes in WTP before and during the pandemic across different temporal phases. The second chapter investigates the heterogeneity of among Express Lane users’ WTP under the impact of lockdown policy. The third chapter provides a comprehensive analysis of users’ value of travel time (VOT) and reliability (VOR) on the Express Lane before and during the pandemic. The fourth chapter extends the focus to system-wide optimization, presenting an integrated traffic management system that combines real-time ramp metering with dynamic toll pricing strategies to alleviate congestion and enhance freeway performance
“NOBODY WILL SEE PAST MY BLACKNESS; THEY WILL ONLY SEE A LOUD BLACK WOMAN”: EXAMINING HOW GHANAIAN IMMIGRANT WOMEN IN THE UNITED STATES COMMUNICATE THEIR MATERNAL HEALTHCARE EXPERIENCES
No full textMaternal mortality in the United States continues to rise, with Black and immigrant women facing the highest risks due to mistreatment, structural racism, and systemic health disparities. Yet, research often overlooks the distinct experiences within these broad categories. This dissertation focuses on Ghanaian immigrant women, a growing African immigrant population in the U.S., to examine how they communicate their maternal healthcare experiences. Using the Culture-Centered Approach (CCA) and the Situational Theory of Publics (STP), the study explores how culture, structure, and agency shape communication in a healthcare system not designed for their needs.Based on 23 in-depth interviews with Ghanaian immigrant women across the United States, the study investigates five core questions: (1) what problems Ghanaian immigrant women face when receiving maternal care, (2) what constraints they encounter when attempting to address their maternal healthcare problems, (3) how they communicate their problems and the barriers in addressing these problems, (4) how Ghanaian immigrant women perceive and make sense of their maternal healthcare experiences, and (5) how they perceive their involvement in maternal healthcare and how does this perception influence their communication.
Findings show that women frequently encounter structural and interpersonal barriers, including high costs, limited provider access, long wait times, inadequate communication, and experiences of racial bias. Cultural factors, such as deference to authority, modesty norms, and community-based decision-making, also shape how women respond to these challenges. Many participants described enduring poor treatment without complaint, not because they lacked awareness but because they perceived speaking up as unsafe or futile. Others sought support through informal networks, such as family members, friends, and online groups, rather than directly engaging with healthcare providers.
The study reveals that although women feel deeply invested in their maternal care, their involvement is often expressed through quiet, cautious, and strategic forms of communication. Silence, for many, is not a sign of disengagement but a protective response rooted in cultural norms, past experiences, and fear of judgment or immigration-related consequences.
By integrating CCA and STP, the study demonstrates how culture, structure, and perceived agency interact to shape maternal healthcare communication. It challenges dominant models that equate empowerment with assertiveness and emphasizes the legitimacy of less visible forms of agency. The findings underscore the need for healthcare systems to recognize diverse communication styles, reduce structural barriers, and involve immigrant communities in the design of care. Ultimately, the study contributes to a more inclusive and context-sensitive understanding of maternal health equity
Engaging Communities in Stormwater Solutions: Lessons from Edmonston, MD
No full textFinal report for URSP 372: Diversity and the City (Spring 2025). University of Maryland, College ParkEdmonston, Maryland is a small and diverse town situated in a floodplain alongside the Anacostia River. The town faces repeated flooding at homes and streets, damaging property and urban infrastructure, with major events in 2020 and 2022. To deal with this challenge, the town is part of projects such as the Green Street Project on Decatur Street, Sustainable Maryland, the Port Towns Walkable Watershed, and the Rain Check Rebate Program.
Students in “Diversity and the City” (URSP 372), developed a community-based participatory research approach to better understand how residents experience flooding, what kinds of solutions they implemented at their homes, and what support they need to mitigate the events.
The approach was developed in five phases: 1. developing survey questions; 2. technical site visit; 3. survey application with community members; 4. analysis; 5. reporting to the stakeholders. The students engaged with 22 residents and completed a survey and visual inspection of each property. This process considered qualitative and quantitative questions and residents’ stories to understand the community needs and flooding experience.
Out of 22 residents, 10 (47.8%) classified the 2020 flood as the biggest flood event. When asked which part of the house flooded in 2020, 54.5% (12/22) had basement flooding and 40.9% of residents (9/22) said their yards flooded. Floods were caused mainly by pump failure (36.4%, 8/22) and weather-related conditions (31.8%, 7/22).
Through surveying and collecting data, it became clear that while many residents express interest in flood mitigation interventions, only 22.7% (5/22) residents participate in stormwater programs and only 54.5% (12/22) had one or more flood mitigation interventions in their homes.
Based on residents’ responses, this study recommends the town take four actions: creating a bilingual education campaign, conducting more engagement with residents exposed to flooding, using surveys and workshops to educate on flood mitigation strategies and programs, and developing a resources webpage. These recommendations can increase participation in flood mitigation efforts and promote climate justice, adaptation, and equity across the community.Prince George's County, M