Alliance One Tobacco (Malawi)
Academic Research Repository at the Institute of Developing EconomiesNot a member yet
52305 research outputs found
Sort by
Housing Price Prediction with Computer Vision and Image Features
No full textArtificial Intelligence (AI) is transforming nearly every facet of human life, and the housing market is no exception. This thesis explores AI to enhance housing price predictions, developing computer vision and machine learning techniques to achieve greater accuracy. Through a series of experiments, our research demonstrates that integrating Multiple Listing Services (MLS) data with property image features and image-based scores improves the performance of housing price prediction models.Extension Studie
From Silos to Systems: A Qualitative Analysis of Braiding Funds for HIV and Syndemic Service Delivery in the United States
No full textThis doctoral project explores the practice of braiding funds as a strategic approach to improving the coordination and integration of public health funding streams—specifically within the U.S. Centers for Disease Control and Prevention’s National Center for HIV, Viral Hepatitis, STD, and Tuberculosis Prevention (NCHHSTP). With growing attention to syndemic conditions that intersect with HIV, there is an urgent need to align categorical funding in ways that enhance service delivery, reduce fragmentation, and improve outcomes for disproportionately affected communities. Braiding funds offers a promising solution, allowing public health agencies to coordinate multiple funding sources while maintaining accountability to each.
Using a qualitative design informed by implementation science, this study employed 19 key informant interviews with CDC officials and leaders from state and local health departments across the United States. A stratified purposive sampling approach was used to ensure geographic, epidemiological, and jurisdictional diversity. Data were analyzed using a hybrid coding framework, themes were identified to highlight facilitators, barriers, and opportunities in implementing braided funding. The project was guided by three conceptual frameworks: the EPIS (Exploration, Preparation, Implementation, Sustainment) framework, the Positive Deviance model, and Kotter’s Eight Steps for Leading Change.
Findings underscore the administrative complexity of managing multiple funding streams and the tension between flexibility and compliance. Participants identified both structural and cultural challenges, as well as innovative strategies used by jurisdictions to align resources. The study concludes with a set of targeted recommendations to strengthen braided funding efforts within CDC’s NCHHSTP and lays the groundwork for broader adoption across federal public health initiatives.Public Healt
Revolution of the Heart: The Alternative New Woman in Early Soviet Media
No full textThe early Soviet government of the 1910s and 1920s was one of the first countries in the world to grant women equal rights with men. Among the Bolshevik Party members was a belief that so-called “new people” would come into being through the material and economic changes that the Revolution accomplished. For women and their representation in fictional narrative arts, this change in political and economic systems provided them with new opportunities for self-determination. State propaganda campaigns sought to bolster and harness women’s empowerment for political goals. Against this, a variety of authors of the 1920s instead explored a more individual-focused application of new opportunities for women. The authors analyzed in this project are Aleksandra Kollontai, the noted Bolshevik advocate for women’s rights, author Evgeni Zamyatin, avant-garde playwright Sergei Tret’yakov and Abram Room, filmmaker. I identify a fictional construct in works by these four artists which I call the “Alternative New Woman” through a feminist-informed close reading of key fictional works by these authors. Vasilisa Malygina by Kollontai, We by Zamyatin, I Want a Baby! by Tret’yakov, and Bed and Sofa by Room all feature a female protagonist who pursues her own desires (rather than conforming to societal expectations) throughout her narrative arc, which forms the basis for my definition of the Alternative Woman. The Alternative New Woman challenges both traditional and Bolshevik understandings of gender, femininity, and individuality. The Alternative New Woman of the 1920s also explores the changing social and material conditions of everyday life during the New Economic Policy (NEP), especially on topics that are traditionally understood as part of the feminine domain: family, marriage, sexuality, and reproduction.Slavic Languages and Literature
House Rules: Urgency, Gridlock and Institutional Culture in North Shore Massachusetts
No full textCape Ann located north of Boston on the Atlantic Coast is widely recognized as a coherent place and cultural destination. Yet, it lacks a co-located administrative boundary within the structures of United States federalism and home rule in Massachusetts. Unlike many other states, Massachusetts does not maintain intermediary governance such as counties. As a result, Cape Ann faces a fragmented landscape for addressing the intensifying pressures of climate uncertainty and aging infrastructure. The absence of the ‘middle scale’ provides a critical lens to understand how territorial identity operates independently of formal jurisdictions. Through an investigation of regionalism and the institutional patchwork in Cape Ann, an organizational logic is developed to assess how these dynamics shape the conversation around practical challenges. This thesis not only describes Cape Ann’s institutional culture but also proposes novel perspectives on regional capacity to navigate gridlock in the face of shared risk and accelerating change.Department of Urban Planning and Desig
The View from Down Here: The Airfield Beyond Empire
No full textNorth Field, a World War II-era airfield on the island of Tinian, embodies the enduring impact of military infrastructure on land, people, and memory. Once the busiest airfield in the world, its relentless flatness and placelessness reflect a larger system of logistical landscape left frozen in time.
“The View from Down Here” critiques the U.S. Department of Defense as the world’s most prolific landscape architect, shaping global terrains through both infrastructural transformation and its representation. Using archival and field research, counter mapping, and speculative design, this project unsettles the airfield’s rigid flatness. Taking on North Field’s four runways, site-specific interventions expose buried histories still held in the landscape, re-presenting the infrastructure of empire at a multitude of scales and temporalities, where the present intertwines with the past and future. This project re-imagines technical landscapes as sites of remembrance, resistance, and life; asking the question―what happens after empire and wasteland?Department of Landscape Architectur
Analyizing Parcel Level Prioritization and Cost Benefit Analysis as Private Land Conservation Tools
No full textPrivate conservation land trusts play a critical role in the attainment of important global climate and biodiversity-related goals, so it is important they make well-informed, strategic decisions about which lands to protect, considering a variety of factors including the ecological, social, and economic. This thesis research evaluated the robustness of a land acquisition prioritization tool designed for Kentucky Natural Lands Trust (KNLT) as an academic exercise by students in a practicum for a Harvard Extension School Land Conservation course. The prioritization tool aimed to guide KNLT’s land acquisition strategy by ranking 23 potential acquisitions (parcels) in the Cumberland Mountain region of Kentucky, USA. The study first conducted a sensitivity analysis of the original prioritization tool by testing the impact of changing importance scores of selection criteria on the parcel rankings. Then parcel level cost analysis was completed using The Nature Conservancy’s Stewardship Calculator to obtain the parcel level cost of acquisition and long-term stewardship. Finally, statistical correlations between parcel rank and cost were examined. Results revealed a robust parcel prioritization tool that experienced minimal changes in rankings despite altering the importance weight of selection criteria. Results also demonstrated a weak, negative correlation between parcel rank and cost, challenging the assumption that the higher-priority parcels would be more costly to protect. These findings underscore the importance of incorporating ecological and economic considerations into conservation planning and provide KNLT with insights to inform actions that can help further protect Cumberland Mountain.Extension Studie
Computationally Speaking: The Mathematical Foundation of Large Language Models and An Exploration Into How They Tell Stories
No full textOver the past year, large language models such as ChatGPT have gained immense popularity, with hundreds of millions of active users. The adoption of these models in everyday tasks marks a significant shift in how we perceive and interact with tech- nology, making it all the more crucial to understand how these new tools work. This thesis aims to elucidate the inner workings of large language models, starting from first principles. We begin with an introduction to foundational machine learning concepts. Next we analyze the underlying architecture of neural networks, focusing on the evo- lution from basic feed-forward networks to Recurrent Feed-Forward networks, Long Short-Term Memory networks, and most importantly Transformer networks. In this analysis, we highlight key components such as the residual stream vector space and at- tention block. We then explore the optimization algorithms used to train autoregres- sive Transformer networks, including deterministic gradient descent, stochastic gradi- ent descent, and Adam, with an emphasis on their convergence properties. Finally, we present current research on Transformer network interpretability, including an ongoing research project about differentiating storytelling modes in the popular large language model Llama2. This thesis underscores that the first step to using machine learning responsibly is to understand it mathematically.Computer Scienc
My Brother’s Teacher: Strengthening Morehouse’s Alumni Educator Pipeline Amid Educational Violence Through Mentoring, Capacity Building, and Morehouse’s Network of Black Male Educators
No full textAfrican American communities must take proactive measures to support the development of Black educators through strategic professionalization and capacity-building to achieve their educational goals for Black students. Addressing this development requires an institution-building approach, an intentional process of strengthening the institution, and a commitment to fostering growth. It represents active resistance and deliberate healing from the endemic educational and social violence inflicted upon Black educational communities. This capstone examines Morehouse Center for Excellence in Education’s efforts to establish institutional support for students and alumni returning to classrooms where they experienced educational violence and taking on the role of educators. It offers potential next steps for the Morehouse Center for Excellence in Education to support alumni within their educational communities despite the violence that disproportionately affects Black educators, particularly Black males. MCEE’s TIME and My Brother’s Teacher initiatives tackle both issues while addressing the professional isolation of Morehouse alumni educators. Networking among Morehouse alumni educators and unpacking their professional experiences reveals the profound impact Black male educators from Morehouse College continue to have in the field. Mitigating the educational violence that Morehouse’s alumni educators face through a mentorship approach while emphasizing their invaluable educational contributions fosters a more substantial and sustainable pipeline of Morehouse alumni educators.
While situated within Morehouse and aligned with the college's mission and values, MCEE’s focus on developing Black male educators necessitates creating something new. The Morehouse Center for Excellence in Education's institution-building efforts aim to capture the breakthroughs made by Morehouse alumni educators through a qualitative narrative approach. These narratives are the foundation for developing strategies to advance the center’s priorities. The negative interactions and experiences faced by Morehouse alumni educators reflect the challenges many Black male educators encounter across the sector. The Impact of a Morehouse Educator initiative provides narratives to support the targeted mentoring initiative, My Brother's Teacher. Redefining Black male educators from victims of educational violence to contributors to academic success fosters both resistance and resilience while directly confronting the pervasive violence directed towards them and the marginalized students they serve.Educatio
An mHealth application framework to achieve personalized health integrating n-of-1 clinical trial with digital twins
No full textThe past few decades have marked a technological upheaval of the entire healthcare industry. The mandate to transition from paper-based medical records to electronic-based records has digitized patients’ entire medical histories, acting as a pivotal catalyst for ushering healthcare into the digital era. The explosive growth in digital health continues to be fueled by technology-driven solutions including mHealth, telemedicine, electronic health records (EHRs), AI-powered diagnostics, and health data analytics. Digital health includes any digital tool or platform used to enhance healthcare delivery and management. Armed with an arsenal of copious data, the transition from population-based health practices to personalized health is starting to take hold. Traditional evidence-based medicine and its key concepts of randomization and generalizability are the backbone of advancing medical treatments. However, the very pillars of randomized clinical trials that support population health contradict the aim of personalized medicine. Traditional randomized clinical trials are not viable approaches for personalized medicine, especially given that clinical trials often fail to meet enrollment criteria and face a dearth of participant diversity. Recent emerging trends and technologies have evolved to support and allow a more personalized approach to medicine. Patients with chronic conditions may benefit the most from or experience an improvement in the understanding of their condition by participating in a different type of evidenced-based practice - specialized n-of-1 clinical trials.
This paper aims to examine digital health trends that have shaped the landscape to view n-of-1 clinical trials as an attractive alternative. The rise in the use of artificial intelligence, machine learning, and digital twins has allowed for a more granular look at an individual patient. This paper explores and provides a comprehensive framework for building a digital health application designed to support n-of-1 clinical trials integrated with digital twins technology and incorporating AI/ML components as a mechanism to achieve personalized medicine
Advancing Deep Learning for Multiagent AI: Mechanisms, Organizations, and Dynamics
No full textEffective and safe multi-agent AI holds the promise to solve important problems across domains such as digital markets, multi-robot coordination, human-AI alignment, and modular methods for monolithic learning systems.
This thesis aims to advance deep learning methods so that they scale to and provide certifiable alignment guarantees in real-world complexity. Technically, it approaches this goal by encoding structure into deep learning that mediates multi-agent dynamics and enables rigorous analysis. This agenda explores systems of independent agents as well as embodied systems, modeled compositionally as multiple agents. This approach has produced a series of deep learning methods that are the first of their kind, achieving provably aligned and effective multi-agent AI that synergizes insights from computational economics, game theory, reinforcement learning, and robot learning.
Specifically, this agenda operates at three macro-to-micro levels of system abstraction, focusing on incentive structures that shape learning outcomes, interaction structures that simplify dynamics, and statistical structures that guide agent learning, respectively. Key contributions are as follows:
Incentive structure: At this level, I design principled incentives that make a target behavior an equilibrium to which rational agents converge, avoiding the need to analyze dynamics in detail. To achieve this, I encode a menu-based architecture and function discontinuity into deep learning to address landmark problems in incentive design, including auction and contract theory. Applied to computational economics, this method yields the first provably aligned (truthful reporting) and effective (revenue optimizing) neural mechanism for multi-bidder, multi-item auctions and single-bidder combinatorial auctions.
Interaction structure: At this level, I accelerate learning by encoding structures that mediate agent interaction, such as groupings, symmetry, and near decomposability into deep learning. These structures simplify multi-agent dynamics and boost learning efficiency in the control and design of a single embodied agent, and also promote multi-agent cooperation.
Statistical structure: At this level, I identify statistics that characterize the dynamics of a system. The problem of agent learning is formulated as modeling and thereby steering these dynamics. I propose a unified approach for dynamics modeling using a generative neural ordinary differential equation that encodes the dependency structure among the statistics. I also study agent modeling using large language models, with the aim of optimizing decision-making policies based on these models. These studies are supported by principled analyses that rigorously assesses the effectiveness of dynamics modeling and policy learning.Engineering and Applied Sciences - Computer Scienc