Doctor of Philosophy (PhD), Wright State University, 2023, Engineering PhD

The design of Hypersonic Vehicles (HVs) requires meeting multiple unconventional and often conflicting design requirements in a hostile, high-energy environment. The most fundamental difference between ordinary aerospace design and hypersonic flight is that the extreme conditions of hypersonic flight require parts to perform multiple functions and be tightly integrated, resulting in significant coupled effects. Critical couplings among the disciplines of aerodynamics, structures, propulsion, and thermodynamics must be investigated in the early stages of design exploration to reduce the risk of requiring major design changes and cost overruns later. In addition, due to a lack of validated test data within the coupled high-dimensional design domains, concept design exploration of HVs poses unprecedented challenges, especially in terms of computational costs and decision-making under uncertainty. A common design exploration technique is to sample the expensive physics-based models in a design of experiments and then use the sample data to train an inexpensive metamodel. Conventional metamodels include Polynomial Chaos Expansion, kriging, and neural networks. However, many simulation evaluations are needed for the design of experiments because of the large number of independent parameters for each design and the complex responses resulting from interactions across multiple disciplines. Because each simulation is expensive, the total costs are often computationally intractable. Computational cost reduction is often achieved using Multi-Fidelity (MF) modeling and Active Learning (AL). MF models supplement High-Fidelity (HF) simulations with less accurate but inexpensive Low-Fidelity (LF) simulations. AL generates training data in an iterative process: rebuilding the metamodel after each HF sample is added, and then using the metamodel to select the next HF sample. Location-specific uncertainty information is critical for making this determination. To address t (open full item for complete abstract)

Committee: Harok Bae Ph.D. (Advisor); Edwin Forster Ph.D. (Committee Member); Ramana Grandhi Ph.D. (Committee Member); Mitch Wolff Ph.D. (Committee Member); Nathan Klingbeil Ph.D. (Committee Member); Jose Camberos Ph.D. (Committee Member) Subjects: Aerospace Engineering; Artificial Intelligence; Engineering; Mechanical Engineering; Statistics

Doctor of Philosophy, The Ohio State University, 2023, Educational Studies

COVID-19 pushed online learning to a greater level of prominence in 2020 faster than anyone could have ever predicted. This sudden push highlighted the level of unpreparedness of many educators to effectively teach in online learning environments possibly due to their teacher education training. Demonstrating empirical best practices for effective online education for teachers is rarely done but necessary as we emphasize that the possibilities of online learning are infinite including creating opportunities not possible, practical, or requiring different approaches than in physical classrooms. In this dissertation, I investigated if commonly recommended active learning and design best practices combined with supports to enhance Vygotsky's popular Zone of Proximal Development (ZPD) theory can create an effective online learning environment for teacher candidates (TCs) and are ZPD supports an additive theory in the design of online learning environments. Seventy-one TCs from a large mid-western university in the United States completed online learning modules designed using a leading online course design platform (Articulate 360), with learner-centered active learning activities, authentic assessments, and supports including tutorials, animations, and case studies. TCs completed a pre-post multiple-choice quiz and evaluated the online learning experience, using two validated surveys (course design and active learning) and a ZPD survey created for this study. Findings indicate that upon completing and evaluating the modules TCs perceived the active learning components (feedback, interest, interactive engagement, and problem-solving) and the online learning environment principles implemented in the intervention as present, and they had a positive perception of the design of online learning modules. Feedback and interactive engagement positively predicted problem-solving and were significant predictors however, feedback was only significant when used as a sole predict (open full item for complete abstract)

Committee: Michael Glassman (Advisor); Tracey Stuckey (Committee Member); Ana-Paula Correia (Committee Member) Subjects: Educational Psychology; Educational Technology

Specialist in Education (Ed.S.), University of Dayton, 2023, School Psychology

Physical classroom space can influence a student's sense of interconnectivity and can support learning. Social effects of the physical space have been infrequently researched regarding the role it has on student collaboration and therefore is not well understood by school personnel. This thesis shares results of a mixed method content analysis of data collected across three new law school buildings in the United States of America. Students from each law school completed a survey to determine the effects the new law school building had on their perceptions of the space, their ability to collaborate with peers and faculty, and the overall difference between their experience in the new building compared to the old law school building.

Committee: Sawyer Hunley Ph.D. (Committee Chair); Molly Schaller Ph.D. (Committee Member); Susan Davies Ed.D. (Committee Member) Subjects: Communication; Education; Higher Education; Pedagogy; Psychology; Quantitative Psychology

Master of Arts, Wittenberg University, 2022, Education

The present study investigated the effects of a project based learning (PBL) unit on the academic achievement of sixth grade math students. A group of 61 students participated in the study during which they were asked to design a garden that could help supplement the local food pantry. All students were assessed on their ratio and proportion skills before the unit began and then again after the intervention, at the conclusion of the unit. The results showed that students' academic achievement was positively affected by the intervention. This study provides valuable information to the field of PBL as it pertains to the mathematics classroom. More research is still needed on PBL and its impact on federal accountability measures to increase the use of PBL as a teaching pedagogy.

Committee: Amy McGuffey (Advisor); Marlo Schipfer (Committee Member); Hillary Libnoch (Committee Member) Subjects: Education; Education Philosophy; Educational Psychology; Educational Theory; Middle School Education

Doctor of Philosophy (PhD), Ohio University, 2021, Instructional Technology (Education)

The current context of learning spaces is guided by several factors but most importantly, from student opinions and perspectives. This study explores and addresses constructivist pedagogical practices and students learning experiences in higher education. Precisely, the dissertation attempts to find out and to understand the importance and the roles of collaborative learning spaces and the challenges students face in those learning spaces. First, the study uses relevant theoretical frameworks of constructivism (constructivist theory of learning) and distributed cognition (collective intelligence) to highlight the importance of understanding the big picture of focusing on students developing their own knowledge and understanding of the world in an environment that encourages engagement and collaborative learning. To further understand the concept of these constructivist practices, this study directed its investigation to the analysis and the interpretation of documentation that relates directly to the study both physical and online. The documentation inquiry includes the mission statements of the space, flyers, posters, training materials and official websites affiliated to the collaborative learning space. The next inquiry was through an online open-ended questionnaire that was directed to learners in the collaborative learning space. Here, the study was interested in understanding learners' experiences in the space from their point-of-view. This was done by first, probing on the physical structure and the design of the learning space, safety, comfort and impact the space has to their learning and engagement in search for knowledge and skills. Secondly, the examination was on various ways in which students learn, interact, and actively collaborate and the challenges involved in the space. Seven themes emerged after an inductive data analysis from the documentation and online qualitative survey: Collaborative learning, flexible learning environ (open full item for complete abstract)

Committee: Greg Kessler (Advisor); Jesse Strycker (Committee Member) Subjects: Adult Education; Education; Educational Technology; Instructional Design

Doctor of Philosophy, The Ohio State University, 2021, Computer Science and Engineering

Ranking is a fundamental problem that has been extensively studied in the machine learning community due to its broad applicability to different application areas such as recommendation, web searching, social choices, and crowd sourcing. The problems of ranking focus on finding the full ranking or partial rankings (e.g., top-$k$ selection) of a set of items from noisy observations. The items may refer to choices, movies, pages, and advertisements and the observations may refer to the queries about the users' preferences on these items. In many systems, the observations can be done in an active (or adaptive) manner, i.e., for each observation, the learner can adaptively choose items to observe according to past observations. An interesting problem is to understand the number of observations needed (aka sample complexity) for finding the ranking under the active setting. In this dissertation, we present new algorithms and derive new sample complexity upper and lower bounds that improve the state-of-the-art for three significant and fundamental problems in this topic. This dissertation explores the following problems: I. Exploring top-fraction arms in stochastic bandits. The stochastic multi-armed bandit (MAB) model is a classical online learning model for studying sequential decision making under uncertainty. In an MAB model, there are multiple arms (each refers to an action or an item), and each sample (aka pull) of an arm returns an independent reward according to this arm's unknown latent distribution. The first problem is to find $k$ distinct arms that are approximately within the top-$\rho$ fraction with respect to the mean rewards of the arms with error tolerance $\epsilon$ and confidence $1-\delta$. We develop algorithms and analyze the sample complexity lower and upper bounds for four variations (the number of arms is infinite or finite; the threshold of the top-$\rho$ fraction is known or unknown). The upper bounds and lower bounds of all four variations (open full item for complete abstract)

Committee: Ness Shroff (Advisor); Jia Liu (Advisor); Raef Bassily (Committee Member); Shaileshh Bojja Venkatakrishnan (Committee Member) Subjects: Artificial Intelligence; Computer Science

Master of Science (M.S.), Xavier University, 2020, Psychology

This study examined the impact of ethics training engagement (i.e., active learning vs. passive learning) on unethical decision-making in the workplace. Participants were randomly assigned to one of the two conditions. Next, a baseline measurement of ethical ideology was collected using the Ethics Position Questionnaire (EPQ) and participants then engaged in ethics training based on the condition to which they were randomly assigned. They then had the option to read along or listen to a hypothetical scenario about an employee faced with the opportunity to make an unethical decision, and completed the Unethical Decision Questionnaire (UDQ). Results showed that participants in the passive learning condition were significantly more likely to perceive an unethical situation as ethical compared to participants in the active learning ethics training condition. Additionally, participants in the passive learning condition were significantly more likely to engage in unethical decision-making than participants in the active learning condition. The current findings contribute to the existing literature by providing evidence that active learning in ethics training programs could reduce unethical decision-making within the workplace.

Committee: Dalia Diab Ph.D. (Committee Chair); Morrie Mullins Ph.D. (Committee Member); Anna Ghee Ph.D. (Committee Member) Subjects: Ethics; Psychology

Master of Computer Science, Miami University, 2020, Computer Science and Software Engineering

While classification performance has improved with the adoption of Neural Network models, the cost of acquiring and labeling the data required to outperform other classification methods is often prohibitively high. Semi-Supervised learning attempts to incorporate unlabeled data in the learning process which can improve performance, however such methods assume preexisting, static sets of labeled and unlabeled data, which are often difficult to attain for novel problems. Active learning addresses these problems by determining which unlabeled samples will, when labeled, best improve a supervised model's performance. Existing methods to prioritize samples have primarily been considered in isolation, despite the existing Information Density framework to combine these methods together. We employ this framework to combine the current state of the art uncertainty based method with a novel similarity based method to improve performance. We also extend the framework itself by considering a dynamic combination of these two methods that shifts priority from one to the other. This iterative process of increasing the labeled set with data prioritized by our acquisition function enables the creation of powerful classification models at greatly reduced costs.

Committee: John Femiani (Advisor); Karsten Maurer (Committee Co-Chair); Daniela Inclezan (Committee Member); Zaobo He (Committee Member) Subjects: Computer Science

Doctor of Philosophy (PhD), Wright State University, 2019, Computer Science and Engineering PhD

Information Extraction (IE) techniques are developed to extract entities, relationships, and other detailed information from unstructured text. The majority of the methods in the literature focus on designing supervised machine learning techniques, which are not very practical due to the high cost of obtaining annotations and the difficulty in creating high quality (in terms of reliability and coverage) gold standard. Therefore, semi-supervised and distantly-supervised techniques are getting more traction lately to overcome some of the challenges, such as bootstrapping the learning quickly. This dissertation focuses on information extraction, and in particular entities, i.e., Named Entity Recognition (NER), from multiple domains, including social media and other grammatical texts such as news and medical documents. This work explores the ways for lowering the cost of building NER pipelines with the help of available knowledge without compromising the quality of extraction and simultaneously taking into consideration feasibility and other concerns such as user-experience. I present a type of distantly supervised (dictionary-based), supervised (with reduced cost using entity set expansion and active learning), and minimally-supervised NER approaches. In addition, I discuss the various aspects of the knowledge-enabled NER approaches and how and why they are a better fit for today's real-world NER pipelines in dealing with and partially overcoming the above-mentioned difficulties. I present two dictionary-based NER approaches. The first technique extracts location mentions from text streams, which proved very effective for stream processing with competitive performance in comparison with ten other techniques. The second is a generic NER approach that scales to multiple domains and is minimally supervised with a human-in-the-loop for online feedback. The two techniques augment and filter the dictionaries to compensate for their incompleteness (due to lexical variat (open full item for complete abstract)

Committee: Krishnaprasad Thirunarayan Ph.D. (Advisor); Keke Chen Ph.D. (Committee Member); Guozhu Dong Ph.D. (Committee Member); Steven Gustafson Ph.D. (Committee Member); Srinivasan Parthasarathy Ph.D. (Committee Member); Valerie L. Shalin Ph.D. (Committee Member) Subjects: Artificial Intelligence; Computer Science

Doctor of Philosophy, The Ohio State University, 2019, Anatomy

Many undergraduate students enroll in gross anatomy courses to support future academic success. Therefore, gross anatomy education at the undergraduate level is tasked, in part, with preparing students for subsequent graduate and professional anatomy instruction. A current trend in anatomy education at the medical professional level is a reduction of hours allotted to gross anatomy curricula. Alternative pedagogies are becoming increasingly necessary as time devoted to gross anatomy education declines and hours are disproportionately allocated across gross anatomy lecture and laboratory components. Modernizing gross anatomy instruction by adapting alternative pedagogies at the undergraduate level may function to facilitate reduced required curricular hours while maintaining effective anatomy instruction that supports future gross anatomy encounters for students at all levels of experience. The purpose of the current study was to apply the instructional design principles of a blended active learning intervention to the gross anatomy education of undergraduate students. Blended learning models are characterized by a combination of traditional in-person instruction and technology-mediated online learning. The blended learning model implemented in the current study was derived from inverted lecture delivery, flipped classroom, and flipped learning. A flipped classroom typically consists of a pre-recorded lecture delivered online as preparatory work prior to attending class. In the flipped classroom students are restricted to one form of recorded lecture selected by the instructor. This differs from the inverted lecture delivery method, which employs multiple means of lecture transmission self-selected by the student based on their perceived learning style. A multi-modal approach to gross anatomy educational research was utilized in which the ADDIE instructional model and action research stages were integrated for the current study. Action research was of interest for th (open full item for complete abstract)

Committee: Kirk McHugh (Advisor); Melissa Quinn (Committee Member); Eileen Kalmar (Committee Member); Tracy Kitchel (Committee Member) Subjects: Anatomy and Physiology

Doctor of Philosophy, Case Western Reserve University, 2019, EECS - Computer Engineering

With the onset of the big data era, designing efficient and secure machine learning frameworks to analyze large-scale data is in dire need. This dissertation considers two machine learning paradigms, the centralized learning scenario, where we study the secure outsourcing problem in cloud computing, and the distributed learning scenario, where we explore blockchain techniques to remove the untrusted central server to solve the security problems. In the centralized machine learning paradigm, inference using deep neural networks (DNNs) may be outsourced to the cloud due to its high computational cost, which, however, raises security concerns. Particularly, the data involved in DNNs can be highly sensitive, such as in medical, financial, commercial applications, and hence should be kept private. Besides, DNN models owned by research institutions or commercial companies are their valuable intellectual properties and can contain proprietary information, which should be protected as well. Moreover, an untrusted cloud service provider may return inaccurate and even erroneous computing results. To address above issues, we propose a secure outsourcing framework for deep neural network inference called SecureNets, which can preserve both a user's data privacy and his/her neural network model privacy, and also verify the computation results returned by the cloud. Specifically, we employ a secure matrix transformation scheme in SecureNets to avoid privacy leakage of the data and the model. Meanwhile, we propose a verification method that can efficiently verify the correctness of cloud computing results. Our simulation results on four- and five-layer deep neural networks demonstrate that SecureNets can reduce the processing runtime by up to 64%. Compared with CryptoNets, one of the previous schemes, SecureNets can increase the throughput by 104.45% while reducing the data transmission size by 69.78% per instance. We further improve the privacy level in SecureNets and implement (open full item for complete abstract)

Committee: Pan Li (Advisor); Loparo Kenneth (Committee Member); An Wang (Committee Member); Ayday Erman (Committee Member) Subjects: Computer Engineering

Doctor of Philosophy (PhD), Wright State University, 2016, Human Factors and Industrial/Organizational Psychology PhD

The purpose of this study was to examine key components of active learning. I hypothesized that feedback, accountability, and guided exploration were key components of active learning. I collected survey data from second year medical students (N = 103) in three different active learning interventions: peer instruction (PI), team-based learning (TBL), and problem-based learning (PBL), at six time points. My results did not consistently support my hypotheses. However, I observed a pattern of differences concerning feedback and accountability in the predicted direction in all three interventions. Feedback had a positive effect on professionalism in both PI and PBL, and accountability had positive effects on emotion control and professionalism in both PI and TBL. Also, I found results that raised issues related to each key component. Namely, that perceptions of feedback were influenced by the nature of questions, interactions between individuals, and the source of feedback. Furthermore, accountability was influenced by team membership and a proper measure of guided exploration needs to be developed. This study raised questions regarding which components of active learning affect important outcomes, and what issues affect key components of active learning.

Committee: Debra Steele-Johnson Ph.D. (Advisor); Dean Parmelee Ph.D. (Committee Member); Valerie Shalin Ph.D. (Committee Member); Nathan Bowling Ph.D. (Committee Member) Subjects: Psychology

Doctor of Philosophy, Miami University, 2013, Geology and Environmental Earth Science

This dissertation is an analogue for my own unique learning pathway in academia focused on projects in seismology and science education. My interests in seismology center on convergent boundaries, where friction between the downgoing and overriding plates causes stick-slip behavior and the eventual rapid rupture of the locked zone in a megathrust earthquakes. More recent seismic and geodetic measurements observe new types of stick-slip behavior that occur more slowly (hours to months) at the downdip edge of the locked zone. Broadly, these new behaviors are termed slow slip, which are often geodetically detected, but in some instances inferred through seismic detection of tectonic tremor or local earthquake swarms. My research focuses on tectonic tremor, for which I have developed an efficient detection technique to take advantage of its unique frequency content (2-5 Hz) to determine when the low amplitude, transient signal is active. I have applied this technique in two regions, Cascadia and southern Mexico, where it has performed similarly to other detection methods and often yielded previously unidentified detections. Moreover, in southern Mexico, a multi-year record of both seismic and geodetic data provide a first opportunity to investigate slow slip phenomena prior to the recent Mw 7.4 Ometepec earthquake. Recent observations and models suggest that slow slip can transfer stress updip to the locked zone, increasing the probability for a large earthquake. I search through seismic data and compile indications of slow slip, including geodetic and seismic signals that suggest the Ometepec earthquake was potentially triggered. My final project integrates my study of geohazards with my interests in science education, in order to train students as young scholars in an introductory geohazards course. In a traditional higher education classroom, lectures are a common way to convey information, yet cognitive research tells us students need to utilize greater and (open full item for complete abstract)

Committee: Michael Brudzinski (Advisor); William Hart (Committee Member); Brian Currie (Committee Chair); Dennis Keeler (Other); Heather DeShon (Other) Subjects: Educational Technology; Geophysics

EdD, University of Cincinnati, 2002, Education : Curriculum and Instruction

The purpose of this qualitative case study was to examine the influence of interaction on active learning, learning outcomes, and community bonding in an online technology course. Participants in the study were 65 students and four instructors of undergraduate computer courses at a two-year suburban branch campus of a large urban, midwestern university. The courses met three times for an orientation and two testing sessions. Online interaction occurred via email and the virtual classroom (chats) and discussion boards of an online instructional software called Blackboard®. Qualitative and quantitative data were collected from face-to-face class observations, synchronous chat observations and transcripts, learner-to-instructor emails, instructor-to-learner emails, discussion board messages posted by the participants, semi-structured interviews, semi-structured focus groups, and course documents. Many participants were technology majors. The online instructors required a demonstration of online technology knowledge by the end of the first week of the course. Therefore, this research offered a unique opportunity to focus on the active learning, learning outcomes, and community bonding without the online technology barriers faced by many online students. The results suggested that the synchronous virtual classroom chats had the most influence on active learning, learning outcomes, and community bonding in these online technology courses. Furthermore, the virtual classroom student participation positively correlated with test grades in three of the four online classes with one class exhibiting statistical significance.

Committee: Dr. Kenneth Martin (Advisor) Subjects:

Master of Computer Science, Miami University, 2010, Computer Science and Systems Analysis

Software reverse engineering is the process of extracting knowledge from a software system and then creating high-level abstractions to communicate that knowledge. This is vital to supporting long-term maintenance of the system. One such abstraction, or view, is to split the classes of the system into two sets -domain concept classes and peripheral classes. That is, the classes that relate to the domain of the system, and those classes that just help with the operation and functioning of the system. Supervised machine learning is a technique that can be used to label domain concept classes and peripheral classes given a training set. However, manually creating a training set is inefficient. The goal of the research is to present a method and tool to semi-automate the creation of a training set for using supervised learning to classify domain concept classes and peripheral classes in a software system.

Committee: Gerald Gannod PhD (Advisor); James Kiper PhD (Committee Member); Michael Zmuda PhD (Committee Member) Subjects: Computer Science

Doctor of Philosophy, University of Toledo, 2012, Judith Herb College of Education

Ohio needs to increase the number of college-educated citizens to improve the state's economy and to remain competitive in the global economy. The Ohio Board of Regents challenged colleges to develop better methods of retaining students to increase graduation rates. Learning community curriculum models have been used to restructure the curriculum, student learning, and the classroom environments to improve student persistence and academic achievement. However, there is limited research on how learning communities are used in business programs in community colleges. This multi-case study examines how three learning community curriculum models have changed the learning environment and the findings provide a comprehensive, contextually rich description of each learning community based on an insider's perspective. A document analysis, three classroom observation, and interviews with 34 participants were included in the multi-case study. The findings show how learning communities facilitate student involvement and social and academic integration, and describe best practices of learning communities in community colleges.

Committee: David Meabon PhD (Committee Chair); Debra Gentry PhD (Committee Member); Mary Edwards PhD (Committee Member); Ron Abrams PhD (Committee Member); David Hyslop PhD (Committee Member) Subjects: Adult Education

Master of Science, The Ohio State University, 2024, Civil Engineering

Extreme weather events continually challenge the integrity of transmission lines, necessitating robust predictive models for grid resilience. This thesis presents two comprehensive data-driven approaches to assess and predict wind-induced transmission line interruptions, leveraging advanced machine learning techniques and innovative data sampling methods. In the first phase, predictive models that establish reliable associations between weather, environmental factors, and line characteristics to forecast interruptions are developed. A novel event-based sampling approach is introduced to address the challenges of unbalanced datasets, improving model accuracy by identifying input subspaces that are most challenging for classification. Among various machine learning models tested, XGBoost demonstrated the highest predictive accuracy of 93.5\%. Sensitivity analyses using Shapley values underscored the importance of wind gust, line length, and mean sea pressure in interruption events. These models offer valuable insights for real-time outage mitigation and long-term planning for line reinforcement and resource allocation during extreme weather events. The second phase enhances the best-performing model from the first phase by addressing class imbalance through novel GAN model created by combining advanced synthetic data generation using Conditional Tabular GAN (CTGAN) augmented with InfoGAN methodologies. Additionally, to make use of the large synthetic data created generated by the model, an adapted Query by Committee (QBC) active learning framework is proposed. The method strategically selects the most informative data points from the large pool of real and synthetic data points, optimizing model performance. When compaing our developed model with a model trained using conventional methods, our model showed a substantial increase of 5\% in testing accuracy for typical testing datasets and an even bigger increase of 17\% when presented with a more challenging test (open full item for complete abstract)

Committee: Abdollah Shafieezadeh (Advisor); Jieun Hur (Committee Member); Chen Chen (Committee Member) Subjects: Civil Engineering; Computer Science

Doctor of Philosophy, The Ohio State University, 2022, Computer Science and Engineering

Acoustic echo cancellation (AEC) and active noise control (ANC) have attracted increasing attention in research and industrial applications over the past few decades. Conventionally, AEC and ANC are addressed using methods that are based on adaptive signal processing with the least mean square algorithm as the foundation. They are linear systems and do not perform satisfactorily in the presence of nonlinear distortions. However, nonlinear distortions are inevitable in applications of AEC and ANC due to the limited quality of electronic devices such as amplifiers and loudspeakers. Considering the capacity of deep learning in modeling complex nonlinear relationships, we propose deep learning approaches to address AEC and ANC problems in this dissertation. Different from traditional signal processing methods, we formulate AEC as deep learning based speech separation. The proposed approach, called deep AEC, suppresses echo and noise by separating the near-end speech from a microphone signal with the accessible far-end signal as additional information. Our study of deep AEC starts with magnitude-domain estimation, and a recurrent neural network with bidirectional long short-term memory (BLSTM) is trained to estimate a spectral magnitude mask (SMM) from the microphone and far-end signals. Later, a convolutional recurrent network (CRN) is utilized for complex spectral mapping and results in better speech quality. In addition, we explore combining deep learning based and traditional AEC algorithms to further improve AEC performance. Although deep AEC produces significant improvements over traditional AEC methods, there exists a tradeoff between echo suppression and near-end speech quality. To address this, we propose a neural cascade architecture to leverage the advantages of magnitude-domain and complex-domain estimation. The proposed cascade architecture consists of two modules. A CRN is employed in the first module for complex spectral mapping. The output is then (open full item for complete abstract)

Committee: DeLiang Wang (Advisor); Wei-Lun Chao (Committee Member); Eric Fosler-Lussier (Committee Member) Subjects: Acoustics; Computer Engineering; Computer Science

Doctor of Nursing Practice , Case Western Reserve University, 2022, School of Nursing

Clinical judgment (CJ) is widely considered an essential nursing skill, yet many new graduate nurses (NGNs) lack the CJ skills needed to safely care for patients. Transformation is needed in clinical nursing education to improve the preparation of NGNs for the delivery of safe and effective patient care. This study aims to fill a gap in the literature and provide evidence on the effectiveness of teaching methods in clinical nursing education. The purpose of this study was to examine the effect of active learning strategies (ALS) on the CJ of nursing students and learners' perspectives regarding CJ development. A quasi-experimental, two-group, longitudinal study was conducted using a convenience sample of 92 senior BSN students from a midwestern state university school of nursing. Approximately one-half of students participated in standard post-conference activities (N=42) with the other half (N=50) participating in ALS with structured faculty debriefing. Data was collected using the Lasater Clinical Judgment Rubric (LCJR) by students for self-evaluation and by faculty for assessment of student performance at three times during the semester: pre-intervention (week 3), midterm (week 8), and post-intervention (week 14). Total LCJR scores were used to reflect overall CJ with subscale scores for noticing, interpreting, responding, and reflecting. Analysis included descriptive statistics; independent samples t-tests; repeated measures ANOVA, and a two-way mixed ANOVA. Both groups demonstrated improvement in LCJR total and subscale scores over time (p < .001), with control group faculty rating students higher than intervention group faculty at baseline, midterm, and post-intervention (p < .05). Students self-assessment scores for both groups were similar at baseline (p > .05), but control group students began scoring themselves higher at weeks 8 and 14 (p < .05). There was no interaction effect between time and LCJR Total scores for intervention and control gro (open full item for complete abstract)

Committee: Deborah Lindell (Advisor); Amy Weaver (Committee Member); Molly Jackson (Committee Member) Subjects: Education; Nursing

Doctor of Philosophy, The Ohio State University, 2020, Industrial and Systems Engineering

The use of augmented reality and gaming technologies to enhance active learning in simulation environments has generated a great deal of interest. These simulated environments are low risk and have minimal real-world consequences for erroneous actions or delayed diagnosis. From a human factors perspective, augmented reality enhances the ability of trainees to perform sensemaking of subtle symptoms to accurately diagnose and treat a patient. From an operations research perspective, game theory provides the ability to find equilibria for dyadic bimatrix adversarial games which allows offline prediction of what an opponent will likely do, and the ability to perform mechanism design. In applying some of these concepts to the context of gaming environments, we can simulate scenarios with adversarial elements. From a methodological perspective, there are many possible options within a game and an individual decision maker is not in complete control of all the factor settings that influence outcomes. Mathematical estimations can be made for the situation where multiple decision makers select options and receive rewards that depend on the selections made by all players. Findings are presented from a study that was conducted in a simulation-based environment with 42 recruited medical students. For the stimuli, a video of an augmented reality trauma patient was `painted' on a table. The patient became increasingly sick over four consecutive stages of a case. The three patient cases were gunshot wound, superheated airway, and tension pneumothorax. The experimental condition prompted participants for an articulation of their mental model between stages and provided expert coaching via an audio-taped verbal presentation on the cues and mechanism of injury for the case. The mental model prompts included questions about asking for applicable cues and information driving a diagnosis, treatment goals, interventions, and predictions for case progression with and without treatm (open full item for complete abstract)

Committee: Theodore Allen PhD (Advisor); Emily Patterson PhD (Advisor); Cathy Xia PhD (Committee Member) Subjects: Industrial Engineering