2021 UMKC Dissertations - Freely Available Online
Permanent URI for this collection
Browse
Recent Submissions
Item Second chance competitive autoencoders for understanding textual data(2021) Goudarzvand, Saria; Lee, Yugyung, 1960-Every day, an enormous amount of text data is produced. Sources of text data include news, social media, emails, text messages, medical reports, scientific publications, and fiction. To keep track of this data, there are categories, keywords, tags, or labels that are assigned to each text. Dimensionality reduction and topic modeling in Mining text data has received a lot of attention. Topic modeling is a statistical technique for revealing the underlying semantic structure in a large collection of documents. Applying conventional autoencoders on textual data often results in learning trivial and redundant representations due to high text dimensionality, sparsity, and following power-law word distribution. To address these challenges, we introduce three novel autoencoders, SCAT (Second Chance Autoencoder for Text), SSCAT (Similarity-based SCAT), and CSCAT (Coherent-based SCAT). Our autoencoders utilize competitive learning among the k winner neurons in the bottleneck layer, which become specialized in recognizing specific patterns, leading to learning more semantically meaningful representations of textual data. In addition, the SSCAT model presents a novel competition based on a similarity measurement to eliminate redundant features. Our experiments prove that SCAT, SSCAT, and CSCAT achieve high performance on several tasks, including classification, topic modeling, compared to LDA, k-Sparse, KATE, NVCTM, ZeroShotTM, and ProdLDA. Additionally, the proposed models are simpler and faster than the established approaches. This work contributes: (1) SCAT autoencoder utilizes the idea of k-competitive learning among the strongest and weakest, positive, and negative neurons in the bottleneck layer. The novelty stems from involving the weakest neurons in the competition process, which might hold meaningful representations but receive low activation values due to random initialization or being representative of rare words or topics. (2) SSCAT autoencoder presents the novel idea of a similarity-based criterion for selecting neurons that are eligible to enter the learning competition provided by the SCAT approach. This process prevents neurons from a high-similarity score to more than k/2 other neurons from entering the competition. We hypothesize that eliminating redundant features will result in better topic representation. (3)CSCAT autoencoder applies the coherent score for selecting the eligible neurons. In this approach, we eliminate neurons in which the highest features do not hold a high coherent score. (4) A thorough evaluation of our autoencoders compared to KATE, k-Sparse, LDA, and NVCTM. The evaluation includes topic modeling, topic coherence score, and document classification using the datasets: 20 Newsgroups, Wiki10+, and Reuters.Item An Approach For Scalable First-Order Rule Learning On Twitter Data(2021) Senapati, Monica; Rao, Praveen R.; Choi, Baek-YoungScalable Rule Learning (SRLearn) is a scalable divide-and-conquer approach with graph-based modeling of social media data, to scale up first-order rule learning through Markov Logic Networks on a commodity cluster on large scale Twitter data. SRLearn takes advantage of distributed systems to partition large-scale data into smaller but meaningful partitions based on user interaction and incorporates a gradient boosting approach with a tool called BoostSRL for first-order rule mining. We show how this scalable solution on first order predicates is more accurate and efficient than existing systems, such as ProbKB (a scalable system to construct probabilistic knowledge base) and XGBoost (extreme gradient boosting) on relational data.Item The Ordinance Project: Commemorating Kansas City's LGBTQ Landmark Legislation(2021) Williams, Austin Randall; Enriquez, Sandra I.; Davis, Donna M. (Donna Marie)This project documents the efforts of Kansas City activists, organizers, and politicians who successfully fought for the passage of a municipal nondiscrimination ordinance in the late 1980s and early 1990s. The ordinance outlawed discrimination in employment, housing, and public accommodations based upon a person’s sexual orientation or HIV status. Although this highly controversial piece of legislation initially failed to pass, the process of organizing around the possibility of such rights helped to form an engaged electorate within Kansas City’s LGBTQ communities. Many of the key individuals involved in this fight were lost during the AIDS epidemic—a disease which forced issues of LGBTQ discrimination to the forefront of the public’s attention. To commemorate the 25th anniversary of the passage of this civil rights ordinance, I set out to capture the oral histories of the activists, organizers, and politicians who are still with us. The efforts taken to commemorate this landmark achievement in Kansas City’s history resulted in a large body of original research and several academically rigorous projects designed to inform and engage the public through a wide variety of mediums: an oral history project, a feature-length documentary film, multiple community-based discussion groups, and a multimedia public history installation. I argue that these projects make a significant contribution to our knowledge of LGBTQ history, the history of the Midwest, the American AIDS epidemic, and several other connected subfields, such as the history of political activism and protest movements. While these projects are public facing, they are situated within scholarly debates and make a historiographical contribution to the field. To demonstrate the academic rigor that was considered during the construction and implementation of these projects, an essay is included in this portfolio that will place my research and interpretations into the greater historiographical context of LGBTQ civil rights measures and the American HIV/AIDS epidemic. Furthermore, this paper will detail how I dealt with the unique challenges posed by conducting oral histories, incorporating a wide variety of experiences into a cohesive narrative, and delivering this narrative to the public through multiple forms of media and events intended to convey and elicit emotion without sacrificing critical treatment and analysis of the historical evidence and the implications for current and future social justice issues.Item A study of carbon and iron charged point defects in gallium nitride: electronic structure implications for high-power photoconductive solid state switch applications(2021) Ghazwani, Mofareh Ahmed; Rulis, Paul Michael, 1976-There is growing demand for high-performance electronics in high volt- age, high current, and high frequency efficiency requirements that current materials (e.g., Si) are not fulfilling. Within the last few years, the rate of development of Si power electron- ics has slowed as the MOSFET silicon power asymptotically approached its theoretical limits. Gallium natride (GaN) grown on top of a silicon substrate could displace silicon across a significant portion of the power management market. Doping elements in bulk GaN may influence and enhance its prop- erties. Carbon doping of GaN is potentially efficient and useful material for photo-conductive solid-state switches (PCSSs), also called photo-conductive semiconductor switches. However, to make effective use of the rich capa- bilities of device-scale engineering design tools (e.g., Technology Computer- Aided Design (TCAD)) it is necessary to know a variety of material de- pendent parameters for which experimental results have not been obtained. Therefore, the ability to determine those parameters via ab initio calcula- tions is essential, especially when the material contains some type of defect or dopant. To overcomes this dilemma, we proposed a simulation methodology to ex- tract the needed parameters form atomistic ab initio calculation of bulk (un- doped) GaN, carbon-doped GaN, and iron-doped GaN. The proposed method chain was successfully produced the required parameters including electronic structure, polarization properties, phonon calculation, and mechanical and spectroscopic properties for GaN, C-doped GaN, and Fe-GaN crystals. The parameter values were subsequently used in a TCAD tool to compute trans- port properties and breakdown voltage of GaN, C-doped GaN, and Fe-GaN. Result shows that all material properties such as mechanical, optical, polar- ization, transport properties, and the breakdown transport properties and breakdown voltage changed due to the presence of dopants. The comparison of breakdown voltage models for C-doped and Fe-doped GaN channel layers revealed that Fe-doped GaN has a greater breakdown voltage. To produce a more accurate simulation of GaN HEMT, it is necessary to take into account the parameters of a genuine model with their actual values rather than rely- ing on a generic dopant. Key Words: PCSS, GaN, C-doped GaN, Fe-doped GaN, point defect, electronic structure, polarization properties, Piezoelectric constant, phonon calculation, mechanical and optical properties, transport properties, XANES/ELNES spectrscopy, device Simulation (TCAD), Multiscal Methods, OLCAO, break- down voltage, electron velocity, mobility, scattering rate.Item Unfolding Untapped Stories: A Narrative Inquiry of Teachers' Experiences of Working With Students Who Have Faced Trauma or Traumatic Events.(2021) Ladhawala, Ami; Barger, RitaThe purpose of this narrative inquiry is to understand the teachers’ experiences of working with students who have faced trauma or traumatic events. The number of children facing trauma exposures to violence, crime, and abuse in the classroom has increased. The insufficient knowledge among educators about the effect of trauma on students lays the foundation to improve training related to trauma. The following questions are addressed by this narrative study: (1) How do teachers address students’ challenging behavior which is manifested because of trauma or traumatic events? (2) How do teachers describe the feelings of success and failure of working with students who have faced trauma or traumatic events? (3) What stories do teachers talk about experiencing secondary trauma due to working with students who have faced trauma or traumatic events? The study collected data using surveys, interviews, and journal entries. Surveys and open-ended interviews provided the information of teachers’ experiences, which was used to construct a narrative profile for each participant. One hundred seventeen survey links were sent to teachers, out of which 76 surveys were sent back. The surveys revealed 22 participants who were interested in face-to-face interviews. Twelve of these participants were selected using the maximum variation sampling method, including gender, ethnicity, teaching experience, and area of specialization. The overall findings from this study suggest that with the growing number of Adverse Childhood Experiences, trauma-informed care needs to be woven into the curriculum of teacher programs. Teachers today encounter students who have faced trauma, which puts them at the receiving end of secondary trauma. Hence, mental health needs to be prioritized for the teachers. The findings from this study might be informative for new and experienced teachers, administrators, and teacher preparation institutions.