137091 research outputs found
Sort by
Towards Increased SSD Bandwidth with GPU-Initiated SSD Access with SSD I/O Scheduler
No full textThis paper presents a hybrid SSD-GPU architecture to address escalating demands for faster data processing.
This reduces data transfer overhead and enhances storage bandwidth. Our research includes the implementation
of a modified SBIOS algorithm that exploits internal parallelism and reduces cross-block latency incurred
from write requests.UndergraduateComputer Scienc
Adapting to change: the impact of environmental and labor market shocks on small businesses in the United States.
No full textThis dissertation examines the economic fragility of small businesses in the face of localized environmental and labor market shocks. Small businesses are a vital component of the U.S. economy, accounting for over 90% of all establishments and contributing nearly half of non-farm GDP. Despite their macroeconomic importance, small businesses operate with thinner profit margins, limited access to credit, and reduced pricing power. These characteristics make them more vulnerable to external disruptions compared to larger firms. Yet, research on small business resilience has been constrained by data limitations and has often focused on large, publicly listed firms.
Motivated by these gaps, this dissertation leverages a proprietary dataset from a major U.S. payment processor that captures real-time transaction activity for over 15 million establishments. With high-frequency, granular data across multiple years and geographies, the dissertation sheds light on how small businesses respond to transient but economically meaningful shocks, specifically extreme temperature events and localized minimum wage increases. By focusing on both the direct and spillover effects of these shocks, the essays provide novel evidence on how temporary events can generate persistent inancial consequences for small firms, with broader implications for income, employment, and regional economic activity.
The first essay analyzes the economic impact of temperature shocks on 15 million small businesses. For unusual heat (cold) events, based on local historical temperatures, I find a decrease in sales by 7.2% (11.6%). Young, small, and discretionary merchants and those with higher sales volatility experience a stronger negative impact. Negative effects are amplified during weekends and prolonged cold spells. A one standard deviation increase in hot days (8.4 days) in a year raises the exit rate by 4.7%. Lost sales are not recovered on normal days, indicating a permanent loss rather than a delayed recovery. This decline further contributes to macroeconomic reductions in total income and employment levels. In contrast to the previous literature, my results show that heat and cold shocks can adversely impact small businesses. My results highlight the economic fragility of small businesses in the longer-term due to temporary, short-lived temperature shocks.
The second essay studies the effects of city- and county-level minimum wage increases on small business performance, focusing on both direct impacts and regional spillovers. Using a novel dataset covering 15 million small businesses from a major U.S. payment processor and a stacked event study design, I find that local minimum wage hikes lead to sharp and persistent declines in sales and customer traffic, with effects ranging from 5% to 10% for dining restaurants in directly impacted ZIP codes. These revenue losses attenuate with distance and extend up to 50 miles beyond policy boundaries, consistent with regional wage spillovers. Leveraging establishment-level variation, I find that newer, smaller, and more volatile businesses experience significantly larger revenue declines, as do those in highly competitive markets with limited pricing power. County-level wage data further confirm that dining-sector wages increase not only in policy-implementing counties but also in adjacent ones, particularly those with lower income, tighter labor markets, and lower marginal commuting costs. These results highlight the broader economic reach of local labor policies and the trade-offs between wage regulation and small business sustainability.Ph.D.Business Administratio
Antibody-omics for biomarker discovery in vaccinology and infectious diseases
No full textAntibodies play a critical role in the immune response by neutralizing toxins, thereby modulating vaccine response and disease outcome. Their protective effector functions are primarily regulated by the antigen-binding domain (Fab) and the crystallizable domain (Fc). While traditional methods of antibody profiling have successfully applied top-down approaches for mainly Fab analysis indicative of overall isotype titer, the Fc interactions between adaptive and innate immunity remain overlooked. As a result, a poorly defined antibody profile of heterogenous individuals can lead to incomplete conclusions regarding disease state or vaccine efficacy. Thus, the long-term objective of this research was to fully characterize the antibody profile using both top-down and bottom-up approaches in a rapid, cost-effective, and efficient manner for translational applications in biomarker discovery, point-of-care diagnostics, vaccine development, and therapeutics. The first specific aim focused on measuring Fc N-linked glycosylation, which is a common post-translational modification well known for its influence on pro- and anti-inflammatory antibody functionality. Currently, gold-standard methods of glycosylation analysis, such as mass spectrometry, are often sample and time-intensive, which can present a critical bottleneck in biomarker discovery. Here, I systematically validated the use of cost-effective sugar-binding lectins with glycoengineered antibodies for rapid (~2.5hr) high-throughput screening of antigen-specific antibody glycosylation. Ultimately, our established multiplexed antibody-omics platform was expanded for deep biophysical characterization on a broad set of antigen-specific antibodies including isotype, subclass, FcR-binding, and glycosylation analysis. The second specific aim was to apply our robust antibody-omics platform to characterize vaccine and adjuvant efficacy. While the role of adjuvants in improving the immunogenicity of vaccines has long been recognized and developed, maintaining and assessing vaccine efficacy over extended periods continues to be a critical challenge. Here, I evaluated the immunogenicity of two adjuvanted vaccine cohorts: HIV-1 and Onchocerciasis. Univariate statistical analysis indicated an IgG1 and IgG4 polarization in both vaccine studies across adjuvant groups, amongst other unique vaccine/adjuvant antibody correlates of protection. A similar experimental approach, with the addition of complement analysis, was performed in the third specific aim to assess and predict the risk of antibody-mediated rejection (ABMR) from sera of kidney transplant patients. While the chance of experiencing rejection is increased by the presence of donor-specific antibodies (DSAs) against human leukocyte antigens (HLAs) on the graft, not all transplant recipients experience rejection. Thus, multivariate LASSO-SVM analysis was performed between ABMR+DSA+ and ABMR-DSA+ patients to elucidate the antibody signatures that led to rejection versus non-rejection. Ultimately, this work provides a foundation for rapid, multiplexed humoral profiling with the potential to screen antibody biomarkers across a range of disease and vaccine efficacy studies.M.S.Bioengineerin
Master of Science in Biomedical Engineering
No full textRotator cuff tear (RCT) is one of the most common causes of shoulder pain and disability, and surgical repairs of RCT have a high risk of re-tear, and they are unable to address the disrupted supraspinatus muscle microenvironment, which then leads to muscle degeneration phenotypes such as muscle atrophy, fatty infiltration, fibrosis and inflammation. Our team hypothesized that delivering anti-inflammatory drugs in a controlled manner into the injured supraspinatus muscle would mitigate muscle inflammatory response, shift cellular phenotypes toward a pro-regeneration state, and promote muscle regeneration. In this study, we aimed to evaluate two anti-inflammatory drugs that have the potential to shift the spectrum of macrophage polarization from inflammatory M1-like macrophages to anti-inflammatory M2-macrophages. We used a heparin-containing poly(ethylene glycol) hydrogel system for the in vivo controlled delivery of those therapeutic drugs. In this study, muscle cross- sectional area, embryonic myosin heavy chains and centrally located nuclei were evaluated as markers of muscle regeneration. We found the sustained release of anti-inflammatory drug TSG-6 from fully sulfated heparin hydrogel fragments significantly promoted muscle regeneration compared to other groups. We also identified spatiotemporal identities of muscle regeneration that muscle regeneration was significantly more pronounced at muscle regions that were closer to the tendon, and the regeneration markers peaked at their unique timepoints. Additionally, a proof-of-concept study was conducted for the drug IL-10, and the effect of this drug on muscle regeneration will be evaluated in the future.M.S.Biomedical Engineerin
Narrative Pathways: A Player-Driven Approach to Interactive Fiction Text Game Generation
No full textWe detail the development of Story2Game, an approach to dynamic text game generation through Large Language Models within interactive fiction games. The approach involves a multi-stage process: firstly by writing a story, populating the elements of a representative game world, and code-building actions that emulate the events in the story. Outside of a short and rudimentary list of default actions, there are no constraints on valid actions, removing a major restriction of the diversity of story generation and translation into a game-based setting. Action generation lies at the key of this process, and is built upon identifying key preconditions and effects to connect items and characters through causal effect reasoning, which can then be interpreted into game state changes whenever a user acts. Additionally, we introduce a technique to allow for dynamic action generation upon a player’s desire to perform an action not recognized by the game state. This dynamic action generation process may require novel adjustments to the representation of certain items in the game state and also checks for logical consistency within other relevant actions and objects. Our evaluation hinges on the measurement of accurate and successful generations of user playthroughs of an entire story.UndergraduateComputer Scienc
The Scholars Event Theater
No full textInterview portion of Lost in the Stacks, episode 654. In this 9th edition of the Georgia Tech Library Guidebook series, we speak to Catherine Manci, Public Programming Librarian, and Felicia Kornegay, Events Manager, about the GT Library's Scholars Event Theater, including what happens in the theater, and why it's a valuable campus space.Interview portion of Lost in the Stacks, episode 654. In this 9th edition of the Georgia Tech Library Guidebook series, we speak to Catherine Manci, Public Programming Librarian, and Felicia Kornegay, Events Manager, about the GT Library's Scholars Event Theater, including what happens in the theater, and why it's a valuable campus space
Characterization of Thermal Conductance Between a Ball and Parallel Plates for Spacecraft Bearing Applications
No full textBearings are essential to rotational spacecraft components, so a proper understanding of bearing behavior is necessary for reliable spacecraft operation. Characterization of bearing thermal conductance provides accurate temperature predictions, which are important to a successful design so that critical mechanical fit is maintained and lubricant degradation and migration is avoided. This thesis describes a static “Ball-on-Flat” (BoF) test platform that has been developed to simplify bearing conductance testing, specifically for low-speed bearings where dynamic effects are not considered. The test consists of a single, lubricated ball between two plates that use the similar material properties as traditional bearing raceways. The goal of this thesis is to use thermal conductance data from the simplified bearing geometry to validate analytical conductance models for a complete bearing, including Aerospace’s BBTherm tool. The motivation for adding data-driven truth to the BBTherm model with test data is to ensure proper modeling of the effect that the lubricant meniscus has on the bearing conductance. The BoF test data has shown good agreement between theoretical models indicating that the combination of the BoF test and BBTherm tool will allow for ease of testing and quicker thermal characterization of complete bearings for various ball and lubricant combinations.M.S.Mechanical Engineerin
Developing Shelf Stable Formulations to Improve the Industrial Utilization of Lignin in Automotive Tires
No full textThe disposal of used automotive tires presents a significant environmental challenge, compounded by the tire industry's reliance on non-renewable petrochemical materials such as carbon black. This research explores the development of shelf-stable natural rubber-lignin formulations as sustainable alternatives for tire manufacturing. Specifically, the study investigates how varying pH and ionic conditions affect the stability of natural rubber (NR) latex mixed with plant-derived lignin. Dynamic Light Scattering (DLS) analysis revealed that mildly acidic conditions improved the colloidal stability of the rubber-lignin matrix, slightly reducing particle size distribution and polydispersity, whereas basic and ionic conditions generally destabilized the formulation. These findings contribute toward the industrial viability of lignin-based, environmentally friendly tire materials and inform future work focused on additive optimization and 3D printing compatibility of uncured rubber systems.UndergraduateMaterials Science and Engineerin