UARK (University of Arkansas )
Not a member yet
19829 research outputs found
Sort by
How Novice Arkansas Teacher Corps Teachers in the Arkansas Delta Stabilize in their Profession: A Dissertation
Get PDFFoundational research exists on the teacher life cycle, novice teacher struggles, and the role of burnout on public sector employees, including teachers (Huberman, 1989; Veenman, 1984; Maslach & Jackson, 1981). However, less research explores teachers\u27 perceptions about what allows them to move through the profession’s life cycle and past initial struggles and avoid burnout. There is a significant gap in the literature about why some teachers persist while others leave the profession at alarming rates (U.S. Department of Education, 2007; Alliance for Excellent Education, 2005). Seven phases exist in a teacher’s life cycle: survival/discovery, stabilization, experimentation/activism, reassessment/self-doubts, serenity, conservatism, and disengagement (Huberman, 1989). This qualitative case study explores what teachers feel contributes to their professional stability and burnout in the Arkansas Delta as they attempt to move from the survival/discovery phase to the stabilization phase in a teacher’s life cycle.
The Arkansas Delta comprises ten million acres of land in Eastern Arkansas and is one of the six natural subregions of Arkansas (Gatewood, 1993). The Arkansas Delta is rural and poor, and “virtually all statistical indices relating to education have persistently pointed to the Delta’s educational deficiencies” (Gatewood, 1993, p. 13). I hope a better understanding of how novice teachers achieve stabilization while working in challenging environments can inform the teacher education process, allow more teachers to remain in the profession, and improve K-12 student achievement scores
Only You Can Prevent the Apocalypse: Donald Trump, the Environmental Movement, and Egotism at the Brink
Get PDFThis thesis explores the rhetorical style termed \u27Apocalyptic Egotism,\u27 focusing on its manifestation in American political discourse and environmental activism. By analyzing the rhetoric of Donald Trump and prominent climate advocates, including Al Gore, King Charles III, David Wallace-Wells, and Greta Thunberg, the study identifies four key stylistic elements: imminent crisis, historical insulation, epistemic certainty, and hyperbolic agency. The analysis reveals how this rhetorical style represents a distinct departure from traditional apocalypticism and perpetuates Anglo-centric power structures and marginalizes non-Western voices, with significant implications for understanding race, rhetoric, and global inequality in the context of modern crises. The thesis concludes by proposing opportunities for further research on intersectionality, marginalized perspectives, and alternative rhetorical strategies in addressing urgent global challenges
Comparison of RT-qPCR and RT-ddPCR on Assessing Model Virus in Wastewater
No full textThere is an increasing demand for quantifying viral loads in diverse wastewater systems using polymerase chain reaction (PCR). This study evaluates the performance of two commonly used workflows: reverse transcription quantitative PCR (RT-qPCR) and reverse transcription droplet digital PCR (RT-ddPCR) in wastewater. We compared the two methods by measuring the viral ribonucleic acid (RNA) of a model virus Phi6 in samples collected from various treatment stages at the Westside Wastewater Treatment Facility in Fayetteville, AR. RNA was extracted from real and synthetic wastewater samples and analyzed in parallel using both RT-qPCR and RT-ddPCR. Findings reveal that both methods demonstrated similar performance for detecting high and medium viral loads. However, RT-ddPCR showed significantly greater sensitivity for low viral loads, reliably detecting trace levels of viral particles where RT-qPCR struggled with detection. For direct viral detection without RNA extraction, RT-ddPCR\u27s performance was more impacted by water quality, whereas measurement on extracted samples demonstrated improved performance against inhibitors. Although RT-ddPCR entails higher costs and longer processing times, its superior sensitivity and resilience to sample contaminants, when used with RNA extraction, underscore its value for precise viral monitoring in wastewater applications
Contrastive Learning Techniques for Fraud Detection
Get PDFDetecting fraud in computing platforms involves identifying malicious user sessions, often using deep learning models, but several challenges hinder effective deployment. Attackers can craft diverse malicious sessions that closely resemble normal ones, complicating the learning of robust decision boundaries. While supervised contrastive learning offers a promising solution through class-specific clustering, its potential remains underexplored. Real-world datasets typically contain few labeled malicious sessions and many normal ones, creating an open-set anomaly detection challenge. Costly expert annotation further limits labeled data, especially for smaller organizations, leading to Positive Unlabeled (PU) learning and noisy label learning issues. Organizations are increasingly turning to LLMs for their adaptability, minimal retraining needs, and the capability of In-Context Learning (ICL) to adapt to evolving fraud patterns, though this potential is not yet fully realized. This dissertation addresses these challenges using supervised contrastive learning to develop practical fraud detection frameworks, with key contributions outlined below.
We present a robust supervised contrastive learning based fraud detection framework which operates in the open-set anomaly detection setting;
We present a supervised contrastive learning based fraud detection framework which operates in the PU learning setting;
We present a supervised contrastive learning based fraud detection framework which operates in the noisy label setting.
We present an ICL based framework for coded fraud text and hate speech detection which leverages supervised contrastive learning for demonstration selection
Conservation Agriculture Practice Effects on Greenhouse Gas Emissions from Fine-textured Soils in Arkansas
Get PDFBiochar application and reduced tillage (RT) are both practices within the conservation agriculture framework, but specific impacts of these practices on direct greenhouse gas (GHG) emissions still require investigation in varied agricultural systems. This research aimed to evaluate the effects of biochar source (i.e., powder- and pellet-sized) and application rate (i.e., 0, 2.5, and 5 Mg ha-1) on GHG production in simulated furrow-irrigated rice (Oryza sativa) in a greenhouse experiment, and to evaluate the effects of RT relative to conventional tillage (CT) on GHG production in soybean (Glycine max) in southeast Arkansas. Both studies evaluated carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) fluxes and season-long emissions, emissions intensity (EI), and global warming potential (GWP), as well as soil physical and chemical properties, and plant response, such as yield. Results showed an almost 10-fold reduction (P \u3c 0.05) in N2O season-long emissions from the 5 Mg biochar ha-1 application rate of pellet-sized biochar (4.9 kg N2O ha-1 season-1) compared to the control (i.e., 0 Mg ha-1; 41.7 kg N2O ha-1 season-1), without any yield penalty, indicating the ability of pellet-sized wood chip biochar to mitigate N2O production in upland rice. In contrast, season-long N2O emissions did not differ (P \u3e 0.05) among powder-sized biochar treatments. Bulk density was lower (P \u3c 0.05) in the CT (1.29 g cm-3) than in the RT (1.44 g cm-3) treatment at the beginning of the growing season, but the greater bulk density in the RT treatment did not result in decreased yield compared to CT. Season-long CO2, CH4, and N2O emissions and EI, as well as GWP, did not differ (P \u3e 0.05) between RT and CT, indicating that more time is needed after practice implementation to realize GHG mitigation benefits from RT. These results contribute to the wider knowledge of potential conservation agriculture benefits in Arkansas by quantifying direct climatic impacts of RT and biochar amendment
Evaluation of Fluridone in Rice
No full textThe increasing pressure of Palmer amaranth [Amaranthus palmeri (S.) Watson] in rice (Oryza sativa L.) fields demands new chemical control options to manage this weed. The recent labeling of fluridone on this crop from the three-leaf rice stage offers an additional residual herbicide for battling Palmer amaranth; however, research is needed to evaluate its effectiveness and rice tolerance. Field experiments were conducted to evaluate Palmer amaranth control and rice tolerance to fluridone across different locations in Arkansas in 2022, 2023, and 2024. Preemergence (PRE) applications of fluridone at 168 g ai ha⁻1; (1× label rate) and 336 g ai ha⁻1; (2× label rate) caused severe crop injury to several rice cultivars, leading to grain yield reductions of up to 49% in eight of the most commonly grown cultivars in Arkansas in a flooded rice production. The cultivar DG263L exhibited injury levels of up to 50% and 32% following PRE and three-leaf applications at the 1× rate, respectively, with yield reductions observed at both application timings, indicating low tolerance to fluridone. When comparing ten application timings from 20 days preplant to postflood applications, the treatments near planting caused the greatest injury to rice grown in a delayed-flood system and the applications at PRE and delayed-preemergence decreased rice grain yield. Therefore, fluridone should not be used at early rice stages, as indicated by the label. Additionally, fluridone applied in mixture with standard rice herbicides at the three-leaf growth stage increased injury by up to 8 percentage points compared to the standard herbicide alone. No yield or groundcover reductions were detected with the addition of fluridone, indicating this herbicide can be applied with other rice herbicides for improved weed control while posing minimal risk to the crop. In a furrow-irrigated rice system, PRE applications of fluridone at the 2× label rate caused 8% to 34% injury; however, no yield reduction was observed. In the same system, fluridone at the 1× and 2× label rates reduced Palmer amaranth density by at least 65% and 88%, respectively, four weeks after treatment compared to the nontreated control. Furthermore, the addition of florpyrauxifen-benzyl postemergence following a PRE application of fluridone at 0.5×, 1×, or 2× label rates reduced Palmer amaranth escapes and decreased seed production by at least 94% compared to fluridone alone at rice maturity. Although florpyrauxifen-benzyl resulted in lower weed densities at the end of the season, the presence of remaining weeds likely contributed to lower rice grain yield compared to treatments where no weeds were present. Thus, sequential applications of other effective herbicides are necessary for season-long Palmer amaranth management
Adversarial Machine Learning: Methods for Attacks and Defenses
Get PDFWith the rapid development of machine learning in real-world applications, enhancing security plays an important role. Adversarial machine learning focuses on understanding malicious actions from attackers and developing defensive techniques against such threats when deploying machine learning systems. An attack can occur in different scenarios, such as poisoning attacks during the training stage and evasion attacks during the testing stage. Although extensive research has explored defense strategies to deal with these harmful attacks, there is a need for further research into areas such as how to counteract malicious attacks with healthy noise or how to train an adaptive defense against a mixture of attacks. Additionally, developing novel attack methodologies is essential for uncovering underexplored vulnerabilities in the training and testing pipelines, thereby providing defenders with deeper insights into the inherent weaknesses of model architectures. Most adversarial attacks primarily aim to degrade overall classification accuracy; however, there is a notable lack of attack strategies that target models designed for fair prediction or multimodal retrieval. Furthermore, malicious users continuously devise subtle methods to disseminate harmful content on social media, necessitating the development of intelligent systems capable of detecting and mitigating such content. Vision-Language Models, which have been widely used in real-world applications, hold significant potential for fostering safer and more respectful online environments.
The goal of this dissertation is to address critical challenges in ensuring safety in machine learning models, focusing on the development of novel attack and defense methods. We begin by investigating two defenses against specific attack types: poisoning attacks and evasion attacks. Next, we examine the potential threats posed by adversaries targeting the fairness of machine learning models, introducing a novel poisoning attack on fair machine learning systems. We then analyze the vulnerabilities of multimodal pre-trained models under adversarial attacks. Finally, we explore methods for detecting and mitigating hateful content in multimodal memes utilizing Vision-Language Models. In this dissertation, we present the following frameworks and algorithms. We develop a defense against data poisoning attacks by leveraging the influence function, which helps the model reduce the harmful effect of poisoned training data; We introduce a defense against evasion attacks via adaptive training, which makes the model adaptive and robust to unseen attacks at the testing stage; We design an attack on fair machine learning models, which not only degrades model accuracy but also hinders the model\u27s fairness objective; We investigate adversarial attacks that degrade the multimodal retrieval capability of pre-trained models; We introduce a definition-guided prompting-based method for detecting hateful memes; We develop a unified framework to transform hateful memes into non-hateful versions
Flexible Microelectrode-Based Impedance Immunosensor for Rapid Detection of \u3ci\u3eE. coli\u3c/i\u3e O157:H7
No full textAbstract Food contamination poses a significant threat to public health, the economy, and human health worldwide, occurring at any stage of the food supply chain, from farm to fork. Escherichia coli 0157:H7, recognized as one of the principal causes of foodborne illness, poses a considerable risk to food safety. The primary objective of this investigation is to devise a flexible microelectrode-based immunosensor capable of swiftly identifying E. coli O157:H7 cells without the need for incubation in a pure culture environment. In the development of the biosensor, the gold electrode (composed of 50% Au) underwent an initial coating process involving 2-aminoethanethiol/cysteamine, followed by functionalization with glutaraldehyde to serve as a linker, and subsequent immobilization with anti-E. coli polyclonal antibodies (pAbs). The determination of the optimal concentration of cysteamine and glutaraldehyde for sensor development revealed an optimal concentration of 20 mg/mL. Upon interaction between anti-E. coli pAbs and E. coli O157:H7, a substantial increase in impedance amplitude from 2.7 to 6.93 kΩ was observed, highlighting the efficacy of the immunosensor in detecting the target pathogen when compared to a bare electrode. Furthermore, the developed immunosensor exhibited the capability to detect E. coli O157:H7 cells with a detection limit ranging from 101 to 103 CFU/mL without the requirement of an incubation period. Visual confirmation of successful binding of E. coli O157:H7 onto the flexible microelectrode-based immunosensor was achieved through scanning electron microscopy imaging, providing insights into the adherence of bacterial cells to the microelectrode surface. Keywords: Foodborne pathogens, flexible biosensors, electrochemical immunosensor; rapid detection; Escherichia coli O157:H
Real time Adaptive Control of a PID via Genetic Algorithm Machine Learning Systems
Get PDFThe most common control method that is utilized by all industries across the world is the proportional-integrative-derivative controller (PID) due the relatively low cost and complexity of the system. However, there are draw-backs with PIDs, it is not adaptative to a changing system, so it works on nominal systems, and it starts breaking down when a system begins to have a non-linear response. The method chosen to overcome both is the utilization of machine learning with the use of genetic algorithms.
This method allows any PID system to be capable of adapting in real-time, while not adding significant additional cost and not requiring specialized equipment. In this paper a PMSM AC motor was set-up with a simplistic calculation on settling time, % overshoot and % error programmed in Python with PyTorch. With MATLAB being utilized to plot the results and provide additional analysis. The purpose of this is not to generate a 1-1 realistic motor but to demonstrate that if a system is able to output settling time, error, and overshoot parameters the algorithm attempts to drive it down to 0 while outputting the up-to-date PID values
The American Lowrider as a Form of Art
Get PDFThis paper argues that lowriders, a style of heavily customized vehicle deeply rooted in Chicano culture, ought to be recognized as a legitimate American art form. After establishing the historical and cultural background of lowriders, this paper identifies a cluster of essential aesthetic features which define lowriders, such as lowered suspension, custom wheels, elaborate paintwork, and opulent interiors. Several paradigmatic lowriders exhibiting these features are then examined against the background of various philosophical definitions of art. The analysis demonstrates that lowriders with their elaborate body modifications and paintwork do align with the essentialist views such as Clive Bell’s formalism and R.G. Collingwood’s expressionism. Lowriders are also demonstrated to align with the various non-essentialist views such as Weitz’ “family resemblance” theory, George Dickie’s institutional theory, and Berys Gaut’s cluster theory of art. Additionally, this paper examines how lowriders may also be admitted as art under the historical-intentional definitions proposed by Jerrold Levinson and Robert Stecker, as well as Arthur Danto’s evolving views. The cumulative examination of paradigmatic lowriders against the various definitions of art supports their inclusion as a legitimate American art form