Research and Scholarship Repository
The Research and Scholarship Institutional Repository collects, preserves, and showcases the scholarly achievements of Texas State University's academic community. It provides open access to the diverse array of research and scholarship materials created at Texas State including articles, presentations, posters, electronic theses and dissertations, capstones, multimedia presentations, and more.
More information: https://guides.library.txstate.edu/institutional-repository
Communities in DSpace
Select a community to browse its collections.
- Research centers and institutes supporting scholarly and creative activities across the university.
- Research, creative, and scholarly works created by the university community organized by college.
- Texas State University divisions, offices, and other interdisciplinary teams across campuses.
- Academic publications and conference proceedings from members of the university community.
- Electronic theses and dissertations, and graduate and undergraduate Capstones and Directed Research.
Recent Submissions
Stories of Stonewall Warehouse: Queer Memories of Space and Place and its Potential for Queer Worldmaking
(2024-05) Manriquez, Hector; Eger, Elizabeth K.; Austin, Jasmine T.; Mandziuk, Roseann M.
This thesis examines queer public memory and how queer people communicate the significance of queer space, place, and worldmaking through their memories. Through a qualitative study of Stonewall Warehouse, the only gay bar in San Marcos, Texas that closed in 2023, I utilized autoethnography and open-ended questionnaires to investigate how former patrons, performers, and staff recall Stonewall as a queer place and space and what Stonewall has meant to them after its closure. My analysis revealed how interactive and representative artifacts depicted Stonewall as a queer place within a southern college town. For creating queer space, participants described general queer space, safe spaces, and queer friend spaces. Through interactions, patrons, performers, and staff co-created queer worldmaking at Stonewall Warehouse as a comforting, safe environment and community for queer people. Through an analysis of public memory, responses revealed how participants memorialize Stonewall by seeking out new queer places, associating themselves with local queer organizations, and through the preservations of memories and stories. By connecting my autoethnographic vignettes to participants’ responses, I combine my experiences at Stonewall Warehouse as a Mexican American, gay man from a traditional Catholic household to my participants to show how our identities shape our queer experiences and memories. Overall, this research showcases the impact of the preservation of queer public memory and calls for the creation of more local queer archives where queer people can share their experiences of queer space and place. In addition, it theorizes queer worldmaking within queer places by illustrating their interconnections and invites future researchers to continue to add to the field of Queer Communication Studies.
The Effect of Multiple Thymol Doses on Intake, Digestion, Rumen, Fermentation and Rumen Microbial Populations in Beef Castle
(2024-05) Fukuda, Emma; Drewery, Merritt; Jessup, Russell; Wickersham, Tryon; Smith, Stephen; Omana Sudhakaran, Pratheesh
Essential oils (EO) have been extensively researched for their ability to modulate rumen microbial populations and, subsequently, rumen fermentation and enteric methane production. However, biological changes from dietary EO are inconsistent due to the varying presence and concentrations of terpenes, the main bioactive component of EO. Further, mechanisms by which terpenes exert their actions are not well understood due to structural variety and complexity. Therefore, terpenes should be researched individually and at known dosages. The objective of this study was to determine the effects of thymol on forage utilization and rumen microbial populations in beef steers. Thymol is a terpene from thyme and oregano EO which is thought to reduce methanogenesis in ruminants. As thymol has potent antimicrobial activity, we aimed to identify a dose that favorably modulates microbial populations without interfering with forage utilization. Accordingly, we utilized four beef steers in a 4×4 Latin square design experiment with four 28-d periods. Four doses of thymol (0 [CON], 120, 240, and 480 mg thymol/kg intake) were administered on alfalfa cubes to a basal hay diet which was provided ad libitum. Intake and digestion were determined on d 9-12 and rumen contents were collected then separated into liquid and solid fractions on d 14 of each period. DNA was extracted from the liquid and solid rumen contents and analyzed with whole genome shotgun sequencing. Forage organic matter intake, total digestible organic matter intake, and organic matter digestibility were not affected by treatment (P ≥ 0.66). Ruminal ammonia, total volatile fatty acids, and rumen pH were also not affected by treatment (P ≥ 0.30). However, for 240 mg thymol/kg intake, molar proportions of acetate were significantly increased compared to 120 mg thymol/kg intake (P ≤ 0.02) and tended to be higher than 480 mg thymol/kg intake (P = 0.08). Propionate was significantly decreased (P ≤ 0.02) for 240 mg thymol/kg intake compared to other treatments. Numerous lactic acid bacteria (LAB) including multiple Lactobacillus, Enterococcus, and Pediococcus species increased for thymol compared to CON. Further, the ammonia producing bacteria Prevotella bryantii was greater at 240 and 480 mg thymol/kg intake compared to 120 mg thymol/kg intake (P ≤ 0.05). Methanogenic microbial species, uncultured Methanobrevibacter sp., and total archaeal abundances increased in the solid microenvironment versus CON for 240 mg thymol/kg intake (P ≤ 0.04). Ultimately, our data indicate that thymol did not negatively affect forage utilization or fermentation products, although microbial populations were affected. While methanogenic archaea were not reduced with thymol supplementation, the increases in LAB may indicate an alternative mechanism for thymol to modulate microbial populations and methanogenesis. These data provide insight into the impact of specific doses of thymol on cattle and provide a foundation for future research that specifically studies the mechanisms by which terpenes and EO exert biological actions.
Rediscovering Three Texas Women Composers and their Music: Zulema Garcia (1870-1907), Stella Kaiser (1873-1953), and Julia D. Owen (1868-1964)
(2024-05) Douglas, Garrett; Schüler, Nico; Mooney, Kevin; Ninov, Dimitar
No abstract prepared.
Pecos River Style Atlatls: An Iconographical Analysis
(2024-05) Williams, Kenneth G.; Boyd, Carolyn Elizabeth; Reilly, Frank Kent, III; Kilby, James David
This thesis is an iconographic analysis of atlatls in Pecos River style rock art from Cedar Springs (41VV0696), Panther Cave (41VV0083) and Rattlesnake Canyon (41VV0180) and in the Mixtec Codex Nuttall. Quantifiable data regarding the context of atlatls in the artwork is presented and discussed with the aim of understanding the symbolic meaning(s) of atlatls in Pecos River style artwork. Relevant archaeology, ethnography, and iconography are presented with the intent of demonstrating that atlatls in Pecos River style rock art are associated with notions of celestial power and the cyclical process(es) of destruction and creation that creates, sustains, and maintains the existence of the universe/cosmos/reality.
Accelerating and Improving Deep Learning Inference on Embedded Platforms via Tensorrt
(2024-05) Zhou, Yuxiao; Yang, Kecheng; Qasem, Apan; Lee, Chul-Ho; Dong, Zheng
Deep learning (DL) has experienced remarkable advancement, emerging as one of the most successful machine learning techniques. Adopting DL models on a variety of computing platforms, including embedded systems, edge devices, and system-on-chips (SoCs), has led to a wide range of DL-enabled applications. Despite achieving impressive accuracy, the large size of deep neural networks may require significant execution time and computing resource consumption not only for training but also for inference, where the latter could be more time-sensitive and need to run on embedded and edge platforms with limited computing resources. To address these challenges, NVIDIA TensorRT has been developed and can be seamlessly integrated with popular DL frameworks like PyTorch and TensorFlow. On the hardware side, modern SoCs, such as NVIDIA Jetson devices, are often equipped with a diverse range of accelerators, each of which is characterized by distinct power and performance features.
This dissertation focuses on improving the execution time and power efficiency of model inference using TensorRT on embedded hardware platforms. The specific contributions are threefold.
First, we explore and evaluate several alternative workflows for the deployment of TensorRT in model inference. Each of such workflows involves steps to convert a given PyTorch or TensorFlow model to a TensorRT counterpart. We implement each of these workflows and evaluate them in several crucial aspects including model accuracy, inference time, inference throughput, and GPU utilization. Our experiment results demonstrate that TensorRT significantly enhances inference efficiency metrics without compromising accuracy. Furthermore, each of the alternative workflows for incorporating TensorRT has its pros and cons and a discussion and comparison is provided for selecting them in different circumstances.
Second, we aim to further improve inference efficiency by incorporating model quantization with TensorRT. We deploy model quantization methods in the TensorRT framework and assess their effectiveness through experiments on the NVIDIA Jetson Orin SoC. There are also different workflows for implementing quantization with TensorRT. We investigate and profile them with respect to various DL models and batch sizes. Our experiments demonstrate that employing quantization within TensorRT significantly enhances the efficiency of inference metrics while maintaining a high level of inference accuracy. Additionally, we explore these workflows for implementing quantization using TensorRT and discuss their advantages and disadvantages. Based on our analysis of these workflows, we provide recommendations for selecting an appropriate workflow for different application scenarios.
Third, we investigate the approaches to efficiently leverage multiple accelerators on an SoC to execute model inference using NVIDIA TensorRT. We profile the execution time and energy characteristics for neural network layers executing on various accelerators. We examine various factors influencing layer execution. We propose two algorithms to assign individual layers of a single model to execute on multiple accelerators, aiming to minimize energy consumption while adhering to a predetermined target NN inference execution time. We implement the proposed approaches using the ResNet50 model on the NVIDIA Jetson Orin platform. Our experiments demonstrate that adopting a coarse-grained layer grouping strategy and properly assigning layer groups to different accelerators can yield greater benefits in terms of energy consumption while preserving desired end-to-end inference time.