17179 research outputs found
Sort by
Oral history interview: Steven Garza
No full textEdited and unedited transcript files (.pdf) and edited and unedited video files available with closed captioning.Oral history interview with Brian Mcleary about his brother-in-arms, Steven Garza
Toward Privacy-Preserving Eye Tracking with Applications in Cross-Platform and Extended Reality Environments
No full textAlthough its primary use has traditionally been in academic research, eye tracking has recently emerged in the consumer market as a novel interaction modality that supports several human-computer interaction applications, including healthcare, social communication, and accessibility. However, the widespread adoption of eye tracking raises significant privacy concerns, as gaze data also conveys sensitive information that can be exploited to infer personal characteristics or to uniquely identify individual users. While prior work has explored some of these risks in isolated research-grade datasets, these privacy implications have yet to be fully addressed in the context of eye tracking's growing ubiquity in the wild, particularly across platforms and in multi-sensor environments such as extended reality. To address this gap, this dissertation establishes a framework for understanding and mitigating the privacy risks emerging from widespread eye tracking, introducing novel attack models and validating practical defenses suitable for integration into consumer platforms. A comprehensive literature review identifies several sensitive attributes that may be inferable from eye tracking data, with a distinction drawn between data collected on research-grade and consumer-grade platforms. Building on this foundation, two novel privacy attacks are designed and demonstrated, targeting user identification across eye tracking platforms and in multimodal extended reality environments respectively. To mitigate these risks, a novel autoencoder-based privacy-enhancing mechanism is proposed to obfuscate personally identifying information while preserving data utility for benign, subject-agnostic applications. Beyond these foundational investigations of eye tracking privacy in real-world settings, this work also contributes several publicly available eye movement datasets that are necessary for further research in this direction.Computer Scienc
Investigation into the Processing and Centrifugation of Lunar and Martian Regolith Simulant Geopolymer
No full textAs human settlement on the Moon and Mars becomes increasingly viable, developing insitu construction methods using locally available materials like regolith is essential. This thesis investigates the feasibility of producing geopolymer concrete from lunar and martian regolith simulants, focusing on the effects of processing parameters and centrifugation on material strength. Three simulants—LHS-1, LSP-2, and MGS-1C—were tested across varying liquid-to-solid ratios and silica moduli. A novel centrifugal mixer prototype was developed to simulate the effects of artificial gravity on curing geopolymer, with the aim of reducing porosity and improving compressive strength. Results showed that centrifugation can enhance material properties, and that mix design must be carefully optimized for each simulant. This study contributes foundational data and a proof of concept for manufacturing geopolymer tiles in low-gravity environments, supporting the development of durable infrastructure for future extraterrestrial habitats.Engineerin
Relations Between Social Anxiety and Mind-Reading Motivation in Autistic and Non-Autistic Adults
No full textExtensive research has shown that autistic adults are more likely to have social anxiety than non-autistic adults, which can negatively impact well-being. There is debate, however, over the extent to which autism is associated with decreased social motivation, defined as the motivation to approach and interact with others. Understanding how social anxiety and social motivation are impacted in autism, and how social anxiety and social motivation relate to each other, has important impacts on autistic well-being. Some recent studies with non-autistic adults have found that anxiety was related to increased social motivation, but there has been limited exploration of this question in autistic populations. Particularly underexplored is the facet of social motivation that involves actively engaging with the mental states of social partners (i.e., mind-reading motivation). Using an online survey of autistic and non-autistic adults, we measured social anxiety using the Social Interaction Anxiety Scale (Mattick & Clarke, 1998) and measured social motivation using the Mind-Reading Motivation Scale (Carpenter et al., 2016). Autistic results reported significantly higher social anxiety than non-autistic adults but there were no group differences in social motivation. The relation between social anxiety and mind-reading motivation was negative but non-significant in both groups. These results have implications for understanding how these different cognitive processes are involved in autism. Planned future data collection will involve large samples and will consider links between social anxiety, mind-reading motivation, and overall well-being, including resilience, social support, and loneliness.Psycholog
Domain Adaptive Object Detection and Video Understanding from High Variance Remote Sensing Data
No full textState-of-the-art (SOTA) object detection methods, when applied to satellite and drone imagery, often struggle to identify small and densely packed objects accurately. Similar limitations are observed in gray-scale intensity imagery for pavement distress detection, where SOTA models face challenges stemming from data scarcity and class imbalance, particularly in recognizing rare distress types. Furthermore, variations in altitude, geographical conditions, and weather across datasets further degrade the performance of SOTA deep neural network (DNN) object detectors. While unsupervised and semi-supervised domain adaptation (DA) techniques have shown promise in mitigating distributional discrepancies between datasets, current pseudo-labeling strategies remain vulnerable to background noise, impeding optimal target dataset performance. Additionally, existing contrastive DA methods neglect the bias introduced by false negative (FN) target samples, which can mislead the learning process.
To address these challenges, we propose a \textbf{support-guided debiased contrastive learning} framework that enhances the labeling of unlabeled target datasets and mitigates bias in target detection. Our main contributions are as follows: (i) a support-set curated approach for generating high-quality pseudo-labels from target dataset proposals; (ii) a novel domain alignment strategy that reduces distribution gaps across datasets by aligning local, global, and instance-aware features; and (iii) a debiased contrastive loss function that improves model robustness to intra-class variability across images and domains. Our experiments span several remote sensing datasets. We further extend our investigation to CCTV and drone video datasets for video understanding and anomaly detection, with a focus on the transition from static images to spatiotemporal video data. Our findings indicate that video anomaly detection is particularly challenging due to difficulties in preserving long-range temporal information and feature misalignment during fusion. The dissertation demonstrates that multimodal learning, leveraging textual data and transformer-based fusion, can substantially enhance object detection and recognition, and anomaly detection performance. Experimental validation is conducted on two CCTV benchmarks and two drone datasets. Overall, this dissertation advances the field of object detection and video understanding in remote sensing and surveillance by addressing key challenges in domain adaptation, bias mitigation, and multimodal learning.Computer Scienc
Teachers' Technology Use in the World Geography Classroom
No full textThis directed research project sought to document practices of World Geography (WG) teachers in Texas and explore their attitudes toward educational technology use across varied school settings, particularly in the Dallas-Fort Worth (DFW) metroplex. The study aimed to address gaps in understanding how digital integration unfolds in WG classrooms where curricular guidance, professional development, and school-level expectations vary widely. While prior research identifies barriers to technology uptake in geography instruction, fewer studies have investigated how WG teachers apply educational and geospatial technologies within the constraints of their daily instructional context in Texas. Using an interpretive qualitative case study approach, data were collected through individual interviews with five public high school WG teachers representing urban, suburban, and rural districts within DFW. Interview transcripts were thematically analyzed to identify patterns in technology use, barriers to implementation, and teacher perceptions. Findings show that teachers relied on educational technologies for structuring lessons and delivering content, while geospatial tools were used more selectively and mainly in physical geography units. While all participants held positive views of technology’s instructional value, their implementation was shaped by contextual factors, such as time, training, tool complexity, and classroom management demands. This study concludes that meaningful technology integration in WG classrooms does not depend on access alone. Rather, implementation reflects the intersection of contextual constraints, individual confidence, and pedagogical intention. While individual interviews offered valuable insights into these dynamics, I recommend expanding the sample and incorporating focus groups to capture a broader range of instructional experiences across school contexts.Geography and Environmental Studie
Risk Factors Associated with Beach Drowning and Tourist Death Susceptibility in Bangladesh
No full textSeaside beaches are popular tourist destinations, offering unique recreation opportunities and ecologically diverse environments. Ensuring the safety and security of visitors is vital for developing beach tourism. Yet these physically dynamic, expansive landscapes pose unique hazards to tourists. Bangladesh, with approximately 710 kilometers of coastline along the Bay of Bengal, holds significant potential for beach tourism. However, beach drowning incidents, particularly in popular destinations like Cox’s Bazar and Kuakata, frequently make media headlines. This poster describes a study that aims to explore the risk factors associated with beach drownings in Bangladesh. Additionally, it analyzes the geographic locations of tourists’ vulnerabilities based on reported drowning fatalities. Consistent data on beach drowning deaths and injuries in Bangladesh are not available through government or non-government databases. Therefore, daily news reports on beach drownings published in 11 mass media outlets have been analyzed for the years 2021 to 2024. Quantitative data on drowning events were extracted from the media reports. The study found 49 beach drowning deaths, with 82 percent occurring from June to November, and men accounting for 92 percent of the fatalities. A content analysis of the news articles highlighted three primary contextual influences on reported beach drownings: (a) coastal environment, (b) beach safety management, and (c) public awareness and behavior. This study can contribute to policy formulation and implementing measures to minimize the risk of beach drownings. Moreover, it may help promote sustainable beach tourism, especially in developing countries like Bangladesh.Sociolog
Factors of Individuals that Sustained Weight Loss after Discontinuing a GLP-1 RA [poster]
No full textGlucagon-like peptide-1 receptor agonists (GLP-1) have been shown to be effective for weight loss; however, weight regain after discontinuation is common. This systematic review aimed to review characteristics that lead to sustained weight loss after discontinuing a GLP-1. A systematic search was conducted in Medline, CINHL, and Web of Science. Data extraction was performed using a PRISMA flow diagram. Five studies were included in the final analysis. Risks of bias and validity were assessed using a rapid critical appraisal tool. Three themes were found among the five articles. Themes included including exercise during GLP-1 use and continue post treatment to sustain weight loss, diet after GLP-1 use to sustain weight loss, and being in a group with social support can yield to better adherence. The systematic review demonstrates that weight regain after GLP-1 discontinuation is common but there are factors that have been found to prevent it. Diet, exercise, and social support can help maintain weight loss after stopping GLP-1. The findings emphasize the need for implementing a long-term treatment strategy to maintain weight loss that includes diet and exercise. More research on the discontinuation process is needed.Nursin
Oral history interview: Luis Grajeda
No full textEdited and unedited transcript files (.pdf) and edited and unedited video files available with closed captioning.Oral history interview with the Grajeda family about their family member, Luis Grajeda
Ozone for Industrial Wastewater Treatment: Recent Advances and Sector Applications
No full textOzonation and ozone-based advanced oxidation processes, including ozone/hydrogen peroxide and ozone/ultraviolet irradiation, have been extensively studied for their efficacy in treating wastewater across various industries. While sectors such as pulp and paper, textile, food and beverage, microelectronics, and municipal wastewater have successfully implemented ozone at full scale, others have yet to fully embrace these technologies’ effectiveness. This review article examines recent publications from the past two decades, exploring novel applications of ozone-based technologies in treating wastewater from diverse sectors, including food and beverage, agriculture, aquaculture, textile, pulp and paper, oil and gas, medical and pharmaceutical manufacturing, pesticides, cosmetics, cigarettes, latex, cork manufacturing, semiconductors, and electroplating industries. The review underscores ozone’s broad applicability in degrading recalcitrant synthetic and natural organics, thereby reducing toxicity and enhancing biodegradability in industrial effluents. Additionally, ozone-based treatments prove highly effective in disinfecting pathogenic microorganisms present in these effluents. Continued research and application of these ozonation and ozone-based advanced oxidation processes hold promise for addressing environmental challenges and advancing sustainable wastewater management practices globally.Engineerin