Search results

Title: Enhancing Explanation Generation in the CaJaDE system using Interactive User Feedback
Creator: Lee, Juseung
Date: 2022
Description: In today’s data-driven world, it is becoming increasingly difficult to interpret and understand query results after going through several...
Show moreIn today’s data-driven world, it is becoming increasingly difficult to interpret and understand query results after going through several manipulation steps, especially on a large database. There is a need for automated techniques that explain query results in a meaningful way. A recent study, CaJaDE(Context-Aware Join-Augmented Deep Explanations), presents a novel approach to generating explanations of query results including crucial contextual information. However, it becomes difficult to interpret explanations since the search space increases exponentially.In this thesis, we propose a new approach that introduces a user interaction model for a purpose of enhancing the generation of explanations in the CaJaDE system. We implemented a user interaction model that consists of three modules: User Selection, Recommendation Score, and User Rating. With these modules, our approach guides a user while exploring relevant join graphs, and lets them be involved in the decision-making process while generating join graphs. We demonstrate through performance experiments and user study that our approach is an effective method for users to understand explanations.
Show less

Title: Relations Between Inhibitory Control, Teacher Support, and Externalizing Behaviors in Elementary School Children
Creator: Kurian, Jennifer
Date: 2021
Description: The aim of this study was to examine the relation between child hot and cool inhibitory control (IC) at the beginning of the school year and...
Show moreThe aim of this study was to examine the relation between child hot and cool inhibitory control (IC) at the beginning of the school year and externalizing behaviors at the end of the year, and to determine if teacher support moderates this relation in early elementary school. Participants included a diverse sample of 138 children in grades 1 (n = 62) and 2 (n = 76), with a mean age of 7.2 years (SD = 10.1 months), about half of whom were male. Hot IC was assessed with the Puzzle Box Task and cool IC with the Happy-Sad Stroop Task. Teacher support was rated by independent observers using the Adapted Teaching Style Rating Scale. A composite teacher-report score based on ratings on subscales from two measures, the Strengths and Weaknesses of Attention Deficit Hyperactive Disorder Symptoms and Normal Behavior and the Strengths and Difficulties Questionnaire, was used to assess externalizing behavior at both time points. Results of hierarchical regression analyses revealed that, contrary to expectation, neither hot nor cool IC significantly predicted child externalizing behavior at the end of the school year. A moderation analysis also failed to show a significant moderating effect for teacher support. The only variable that significantly predicted externalizing behavior at the end of the year was externalizing behavior at the beginning of the year. There were significant concurrent associations between hot IC and externalizing behaviors at both the beginning and end of the school year. These findings suggest that externalizing behaviors in early elementary school are relatively stable. Thus, early and comprehensive intervention may be critical for implementing prevention strategies designed to increase self-regulation and thereby decrease externalizing behaviors after formal school entry.
Show less

Title: Growth kinetics of Salmonella enterica and Listeria monocytogenes during rehydration of dehydrated corn and subsequent storage
Creator: Mate, Madhuri
Date: 2022
Description: Dehydrated vegetables, including corn, are often used in restaurants and retail grocers. They do not support the growth of pathogens as their...
Show moreDehydrated vegetables, including corn, are often used in restaurants and retail grocers. They do not support the growth of pathogens as their moisture content is very low. After rehydration, these food products attain high water activity values suitable with neutral pH for the survival and proliferation of foodborne pathogens, including Salmonella enterica and Listeria monocytogenes. The purpose of the study was to examine the extent to which dehydrated corn supports the growth of S. enterica and L. monocytogenes during rehydration at 5 or 25°C water and following storage at 5, 10, and 25°C temperatures at 1, 3, 5 and 7 d intervals. Fresh corn was dehydrated at 60°C for 24 h. Dehydrated corn was inoculated with a 4-strain cocktail of either S. enterica or rifampicin-resistant L. monocytogenes, resulting in 4 log CFU/g, and held at ambient temperature for 24 h. This corn was then rehydrated using either 5 or 25°C water for 24 h. Throughout rehydration, corn samples were removed at intervals and enumerated. To enumerate S. enterica and L. monocytogenes, the samples were homogenized with BPB and BLEB respectively and cultivated on TSAYE with overlaid XLD or BHIARif200, respectively. Rehydrated corn was then stored at 5, 10, or 25°C and enumerated at intervals 1,3,5 and 7 d. Triplicate samples were assessed at each timepoint and three independent experiments were conducted for each rehydration water temperature. Growth rates were determined by DMFit and statistically analyzed using Student t-test. A p-value ≤0.05 was considered significant. Overall the growth rate of S. enterica was higher when rehydrated in 5°C water temperature and then stored at 25°C and was determined to be 0.61 ± 0.23 log CFU/g per d. This timepoint was also the shortest time required to increase by 1 log which was: 1.64 d, i.e. 39 h. For L. monocytogenes, the 25°C water rehydration showed the fastest growth rate when stored at 25°C. It took only 1.58 d or 37.8 h for 1 log increase in the population. After 5°C water rehydration of corn the highest populations of mesophilic bacteria and yeasts and molds were observed for 25°C storage ranging from 8.43 to 9.39 log CFU/g and 4.75 to 7.87 log CFU/g, respectively. After 25°C water rehydration, the highest population of mesophilic bacteria, 8.88 log CFU/g, was observed at 5°C storage at 5 d; yeasts and molds were 8.70 log CFU/g for 25°C storage on the same day. The results of this study determined that S. enterica and L.monocytogenes could survive and grow in dehydrated plant foods during rehydration and storage, highlighting the need for product assessments for these types of foods.
Show less

Title: RADIAL MAP ASSESSMENT APPROACH FOR DEEP LEARNING DENOISED CARDIAC MAGNETIC RESONANCE RECONSTRUCTION SHARPNESS
Creator: Mo, Fei
Date: 2021
Description: Deep Learning (DL) and Artificial Intelligence (AI) play important roles in the computer-aided medical diagnostics and precision medicine...
Show moreDeep Learning (DL) and Artificial Intelligence (AI) play important roles in the computer-aided medical diagnostics and precision medicine fields, capable of complementing human operators in disease diagnosis and treatment but optimizing and streamlining medical image display. While incredibly powerful, images produced via Deep Learning or Artificial Intelligence should be analyzed critically in order to be cognizant of how the algorithms are producing the new image and what the new imagine is. One such opportunity arose in the form of a unique collaborative project: the technical development of an image assessment tool that would analyze outputs between DL-based and non DL-based Magnetic Resonance Imaging reconstruction methods.More specifically, we examine the operator input dependence of the existing reference method in terms of accuracy and precision performance, and subsequently propose a new metric approach that preserves the heuristics of the intended quantification, overcomes operator dependence, and provides a relative comparative scoring approach that may normalize for angular dependence of examined images. In chapter 2 of this thesis, we provide a background description pertaining to the two imaging science principles that yielded our proposed method description and study design. First, if treated naively, the examined linear measurement approach exhibits potential bias with respect to the coordinate lattice space of the examined image. Second, the examined DL-based image reconstruction methods used in this thesis warrants an elaborate and explicit description of the measured noise and signal present in the reconstructed images. This specific reconstruction approach employs an iterative scheme with an embedded DL-based substep or filter to which we are blinded. In chapters 3 and 4 of this thesis, the imaging and DL-based image reconstruction experiments are described. These experiments employ cardiac MRI datasets from multiple clinical centers. We first outline the clinical and technical background for this approach, and then examine the quality of DL-based reconstructed image sharpness by two alternative methods: 1) by employing the gold-standard method that addresses the lattice point irregularity using a ‘re-gridding’ method, and 2) by applying our novel proposed method inspired by radial MRI k-space sampling, which exploits the mathematical properties of uniform radial sampling to yield the target voxel counts in the ‘gridded’ polar coordinate system. This new measure of voxel counts is shown to overcome the limitation due to the operator-dependence for the conventional approach. Furthermore, we propose this metric as a relative and comparative index between two alternative reconstruction methods from the same MRI k-space.
Show less

Title: DO GENERAL EDUCATION HIGH SCHOOL STUDENTS IN A BASIC PHYSICAL SCIENCE COURSE IMPROVE UPON ATTITUDES TOWARD SCIENCE LEARNING AND CONTENT MASTERY FOLLOWING VIRTUAL/REMOTE FLIPPED INSTRUCTION OR VIRTUAL/REMOTE NON – FLIPPED INQUIRY – BASED INSTRUCTION?
Creator: Martino, Robert S.
Date: 2022
Description: As we progress further into the 21st Century, high school science is being challenged on how to best deliver instruction to students. Teacher ...
Show moreAs we progress further into the 21st Century, high school science is being challenged on how to best deliver instruction to students. Teacher – centered instruction has long been de – emphasized in favor of inquiry – based instruction, although teacher – centered instruction still exists to a noticeable extent. Inquiry – based instruction, while more student – centered in its common practice, still involves the teacher as a guide during classroom direct instruction. Research has been ongoing to identify new and dynamic forms of science concept delivery that serve the needs of diversified science instruction (Keys & Bryan, 2001; Saldanha, 2007). Virtual instruction has become more commonplace, and it was fully implemented during this study. It has become incumbent upon science education researchers to explore and identify the most effective means of virtual instruction, means that are student – centered, engaging, interesting, and that both improve student science content understanding and attitudes toward science. Flipped instruction is a more recently – incorporated form of student – centered instruction that has students experiencing classroom routines at home and homework routines in class, and that is why this instruction is referred to as being “flipped.” Hunley (2016) examined teacher and student perception of flipped instruction in a science classroom, while Howell (2013) explored it in a ninth – grade physical science honors classroom. At the onset of this study, relatively few studies were available about this newer form of instruction within high school science instruction, no studies were available that involved high school general education physical science courses, and certainly no studies were available that compared virtual/flipped and non – flipped general education physical science instruction at the onset of this study. This study researched the effect of virtually – implemented flipped instruction on high school students’ understanding and attitude toward science. Instruction was completely virtual/remote (online), and at home, for all students in this study. In investigating the effect of this type of instruction, this study examined student academic performance and attitudes (and intentions and beliefs) toward science in two units of a high school Integrated Chemistry and Physics (Physical Science) course. Sixty – six students from Southlake High School, a midwestern U.S. high school, took part in the study. Sixty – four of those students took the unit assessments. Half of the students (test group) were instructed via virtual/remote flipped instruction and the other half (control group) were instructed via virtual/remote non – flipped, inquiry – based instruction during the first unit. During the second unit, the test group students who were instructed via virtual/remote flipped instruction switched with the control group and were instructed via virtual/remote non-flipped inquiry – based instruction, while the control group students who were instructed via virtual/remote non-flipped instruction were instructed via virtual/remote flipped instruction. The students in both groups were surveyed three times, using the Behaviors, Related Attitudes, and Intentions Toward Science (BRAINS) (Summers, 2016) instrument student questionnaire and survey for their attitudes (and beliefs and intentions) toward science (once prior to the first unit, once after the first unit, and once following the second unit). Student test results and survey responses were then analyzed to identify which instructional style was more effective for student learning and whether student attitudes (and intentions, and beliefs) favored one instructional style over the other. Student science attitudes (and beliefs and intentions) and academic performance were evaluated throughout the study. There was an increase in control group student science attitudes (and beliefs and intentions), from the pre – study survey to the post – unit 1 survey following their receipt of non – flipped virtual/remote instruction in the first unit. There was a lower increase in test group student science attitudes (and beliefs and intentions), from lower pre – study attitudes (compared with the control group) following the test group’s receipt of flipped virtual/remote instruction in the first unit,. Following the second unit, both the control group and test group again showed increases in attitude (and beliefs and intentions) compared with the pre – study survey results, with the control group again showing greater increases than the study group. Student academic performance favored the control group as it outperformed the test group in both the first unit and the second unit, even when the test group received the virtually – delivered flipped instruction in the first unit. The findings of the study showed that virtually implemented flipped instruction resulted in no advantage for the test group in terms of greater improvement in attitudes (or beliefs or intentions) toward science and no advantage for the test group in terms of learning science content in general education Integrated Chemistry and Physics (Physical Science). These results indicate that this form of teaching may not be effective in improving general education Physical Science student learning and student attitudes (and beliefs and intentions) toward science. Therefore, the use of virtually implemented flipped instruction in this general education science course will need to be further studied to determine its effect on student learning and student attitudes (or even beliefs and intentions) toward science.
Show less

Title: Sensemaking for Power Asymmetries in Anti-Oppressive Design Practice
Creator: Meharry, Jessica J
Date: 2022
Description: Within professional design practice in capitalist market contexts, the goals of user-centered and human-centered design methodologies is to...
Show moreWithin professional design practice in capitalist market contexts, the goals of user-centered and human-centered design methodologies is to make algorithmically-based technologies understandable for users, satisfy customer needs and desires, and thereby increase corporate profitability. However, there is growing concern that the computational methods, data management, and business models that drive these technologies are leading to global asymmetries of knowledge, information, and power. The asymmetries of power generated by these designed interactions can be considered the kind of wicked problem that design seeks to address. Yet the dominant goals and methods of professional design practice limit their ability to design ethically within market contexts. These methodologies fail to adequately consider systemic context and power relations, potential for bias in algorithmic computation, and specific forms of systemic oppression. These gaps then lead to inadequate design solutions. This study explores these gaps in design methodologies that could be transferable to a range of professional (and non-professional) practices by looking at potential new levers within familiar design methods and their effectiveness as facilitating problem reframing towards equitable solutions. This dissertation advances knowledge in design by exploring how professional designers can better understand how to use sensemaking processes for salience of power asymmetries, algorithmic materiality, and systemic oppression. It proposes an anti-oppressive design framework that is rooted in a critically-informed design praxis. These orientations rethink and recreate design knowledge by helping professional designers shift the market-focused paradigm for which they are designing.
Show less

Title: PER – AND POLYFLUOROALKYL SUBSTANCES FATE AND TRANSPORT IN SEDIMENTS, SAND, AND ADSORBENT MEDIA
Creator: Manwatkar, Prashik
Date: 2022
Description: Perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) are two important organic chemicals of the per- and polyfuoroalkyl...
Show morePerfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) are two important organic chemicals of the per- and polyfuoroalkyl substances (PFAS) group that have contaminated land, water, and the air since 1950. The continuous release of PFAS from the surface of land into water is not easy to forecast and an appropriate treatment method needs to be economically viable since there are currently around 42,000 suspect industrial and municipal sites in the United States. For a true reproduction of real-world pollution patterns, we constructed polypropylene tanks, performed laboratory-based experiments, and analyzed the samples using EPA method 533. In this study, we examined the fate and transport of long- and short-chain PFAS, including PFOA, PFOS, and perfluorobutanesulfonic acid (PFBS), from sediments, adsorbent media, and sands under overlaying water tanks. Granular activated carbon (GAC), biochar (BC), and Fluorosorb® (FS) were also added between the contaminated sediments and the sand layer in order to observe capping effectiveness. As one of the best ways to treat contaminated sediments on a large scale, adsorbent beds may reduce contaminants migration and support the degradation of contaminants. We found that all three chemicals were able to pass through the adsorbent layers of 3-4 inches from 4-5 inches of contaminated sediments and reach the top surface of the beds (25-30 inches). In the top 5-7 inches, PFBS concentration varied from 0.28 ppb to 0.78 ppb for all adsorbent tanks for 7 days. Whereas the bottom contaminated sediments concentrations of PFBS were 8518 ppb to 9481 ppb. We also observed the concentrations at top ports increased by 0.59 ppb to 2.31 ppb in 21 days, and ultimately, 0.58 ppb to 7.07 ppb in 69 days. While PFOA and PFOS found different metabolites in all layers, they provided noticeably lesser concentrations in contaminated sediments compared to PFBS. Further, the results of this study can be useful for validating the contaminant transport model predictions by identifying linear or nonlinear sorption equilibrium processes and diffusion-dispersion processes in sediment, sand, and various adsorbent media.
Show less

Title: Comparing the effects of an adjunct brief action planning intervention to standard treatment in a heterogeneous sample of chronic pain patients
Creator: Mikrut, Cassandra Leona
Date: 2022
Description: Objectives: Behavioral treatments for chronic pain have been associated with positive outcomes, but they are often time consuming in nature....
Show moreObjectives: Behavioral treatments for chronic pain have been associated with positive outcomes, but they are often time consuming in nature. The aim of the present study was to investigate the effectiveness of a brief behavioral treatment for chronic pain and compare Brief Action Planning used in conjunction with treatment as usual (BAP + TAU) to TAU, on changes in pain severity, pain interference, pain self-efficacy, quality of life, and anxiety and depression in a heterogeneous sample of chronic pain patients. Methods: A total of 172 participants were recruited from an urban pain clinic. Eighty-five participants were quasi-randomly assigned to the BAP + TAU group and 87 participants were quasi-randomly assigned to the TAU control group. After completing T1 measures, two iterations of the BAP protocol were delivered to the intervention group by a trained clinician over the phone, with two weeks in between iterations. The TAU group received check-in calls, collecting brief mood and pain scores, to control for clinician contact. All participants completed T2 measures following the last phone call. Validated measures were used at T1 and T2 to examine participant outcomes. Results: Two-way repeated measures analysis of variance (ANOVA) tests were used to test the primary hypotheses that there would be a Group x Time interaction, on pain severity, pain interference, pain self-efficacy, quality of life (QOL), and anxiety and depression, such that participants assigned to the BAP + TAU group would endorse improved scores from T1 to T2, while TAU participants would not. Results showed a significant Group x Time interaction on pain severity and anxiety and depression. However, there was not a significant Group x Time interaction on pain interference, pain self-efficacy, or QOL. Discussion: These findings provide preliminary support for the effectiveness of BAP, as an adjunctive treatment to TAU, when provided by a trained clinician, as a treatment for reducing pain severity and anxiety and depression, in a heterogeneous chronic pain population. These results advance the current BAP literature, providing preliminary support for using BAP with individuals with a wide variety of chronic pain diagnoses.
Show less

Title: EMBEDDING RELATIONSHIPS: THE INDIRECT EFFECTS OF WORK RELATIONSHIPS ON TURNOVER INTENT
Creator: McDonald, Jordan C.
Date: 2022
Description: With the onset of the “Great Resignation” following the onset of the COVID-19 pandemic, employees are quitting jobs at unprecedented levels....
Show moreWith the onset of the “Great Resignation” following the onset of the COVID-19 pandemic, employees are quitting jobs at unprecedented levels. Although the traditional model of turnover (Mobley, 1977; Mobley, Griffeth, Hand, & Meglino, 1979) links job attitudes and turnover intentions as key determinants in understanding the turnover process, there is a growing recognition of the importance of studying contextual variables, namely social relations, in expanding our understanding of employee turnover and retention. Job embeddedness (Mitchell et al., 2001) and social capital theories (Granovetter, 1973; Burt, 1992; Lin, 1982) implicate employees’ social networks as additional factors worth investigating in understanding employee turnover. The aim of the current study was to study an expanded model of turnover by examining whether different types of social relationships at work differentially related to work experiences and attitudes that, in turn, related to turnover intentions. The current research leveraged an ego-centric method to collect information on employees’ social networks at work along with work experience and attitudinal constructs. The results of the study found that expressive relationship networks (i.e., friendship networks) had a positive, significant effect on employees’ job embeddedness, with an indication of a marginal indirect effect with organizational commitment. Surprisingly, employees’ instrumental networks were not significantly related to any work experience or attitudinal factors. There was no support for the hypothesized indirect effects linking social networks, work experiences and attitudes, and turnover intentions. Practical implications and directions for future research are discussed.
Show less

Title: AN IMPROVED VALIDATED METHOD FOR THE DETERMINATION OF SHORT-CHAIN FATTY ACIDS IN HUMAN FECAL SAMPLES BY GC-FID
Creator: Freeman, Morganne M
Date: 2022
Description: Short-chain fatty acids (SCFAs) are metabolites produced by the gut microbiota through the fermentation of non-digestible carbohydrates....
Show moreShort-chain fatty acids (SCFAs) are metabolites produced by the gut microbiota through the fermentation of non-digestible carbohydrates. Recent studies suggest that gut microbiota composition, diet and metabolic status play an important role in the production of SCFAs. Current methods for the analysis of SCFAs are complex and inconsistent between research studies. The primary objective of this study was to develop a simplified method for standardized SCFA analysis in human fecal samples by gas chromatography with flame ionization detection (GC-FID). A secondary objective was to apply the method to fecal samples from a previous randomized, crossover clinical trial comparing participants with pre-diabetes mellitus and insulin resistance (IR-group, n=20) to a metabolically healthy reference group (R-group, n=9) after daily consumption of a red raspberry smoothie (RRB, 1 cup fresh-weight equivalent) with or without fructo-oligosaccharide (RRB + FOS, 1 cup RRB + 8g FOS) over a 4-week intervention period. Extraction parameters, including solvent selection and water content of the sample, were investigated before finalizing the method. Freeze-dried fecal samples (0.5 g) were suspended in 5 mL of milli-Q water, vortexed and centrifuged at 3,214 x g for 10 minutes. The supernatant was transferred to a clean tube, acidified with 5.0 M HCl and centrifuged again at 12,857 x g for 5 minutes. The resulting supernatant was transferred to a GC vial for analysis by GC-FID. Linear regression data for standards at concentrations 5-2000 ppm ranged from 0.99994-0.99998. Limit of detection (LOD) ranged from 0.02-0.23 µg/mL. Limit of quantification (LOQ) ranged from 0.08-0.78 µg/mL. The validated method was then applied to fecal samples collected from a previously conducted study. Nine SCFAs were identified and quantified (acetic, propionic, iso-butyric, butyric, iso-valeric, valeric, 4-methyl valeric, hexanoic and heptanoic acids). Statistical analysis (Student’s t-test, ANCOVA) was performed on PC-SAS 9.4 (SAS Institute). Acetic acid was significantly lower in the IR-group compared to the R-group before starting intervention (baseline, Week 0, IR v R-group, p=0.014). Intervention analysis comparing RRB to RRB + FOS at 4 weeks (WK4) showed a significant difference in 4-methyl valeric acid (p = 0.040) in the R-group. Trends of decreased SCFA content after 4-weeks of RRB and RRB + FOS compared to baseline were observed in both groups, though changes were not significantly different between dietary interventions at 4 weeks (p>0.05). Metabolic status and dietary intervention are discussed in relation to their impact on SCFA content in fecal samples and mechanisms of biological use as a metabolite. Limitations of the study include sample size and using only feces and not other biological samples for SCFAs analysis, which may be considered for future research.
Show less

Title: Intelligent Job Scheduling on High Performance Computing Systems
Creator: Fan, Yuping
Date: 2021
Description: Job scheduler is a crucial component in high-performance computing (HPC) systems. It sorts and allocates jobs according to site policies and...
Show moreJob scheduler is a crucial component in high-performance computing (HPC) systems. It sorts and allocates jobs according to site policies and resource availability. It plays an important role in the efficient use of system resources and users satisfaction. Existing HPC job schedulers typically leverage simple heuristics to schedule jobs. However, the rapid growth in system infrastructure and the introduction of diverse workloads pose serious challenges to the traditional heuristic approaches. First, the current approaches concentrate on CPU footprint and ignore the performance of other resources. Second, the scheduling policies are manually designed and only consider some isolated job information, such as job size and runtime estimate. Such a manual design process prevents the schedulers from making informative decisions by extracting the abundant environment information (i.e., system and queue information). Moreover, they can hardly adapt to workload changes, leading to degraded scheduling performance. These challenges call for a new job scheduling framework that can extract useful information from diverse workloads and the increasingly complicated system environment, and finally make well-informed scheduling decisions in real time.In this work, we propose an intelligent HPC job scheduling framework to address these emerging challenges. Our research takes advantage of advanced machine learning and optimization methods to extract useful workload- and system-specific information and to further educate the framework to make efficient scheduling decisions under various system configurations and diverse workloads. The framework contains four major efforts. First, we focus on providing more accurate job runtime estimations. Estimated job runtime is one of the most important factors affecting scheduling decisions. However, user provided runtime estimates are highly inaccurate and existing solutions are prone to underestimation which causes jobs to be killed. We leverage and enhance a machine learning method called Tobit model to improve the accuracy of job runtime estimates at the same time reduce underestimation rate. More importantly, using TRIP’s improved job runtime estimates boosts scheduling performance by up to 45%. Second, we conduct research on multi-resource scheduling. HPC systems are undergoing significant changes in recent years. New hardware devices, such as GPU and burst buffer, have been integrated into production HPC systems, which significantly expands the schedulable resources. Unfortunately, the current production schedulers allocate jobs solely based on CPU footprint, which severely hurts system performance. In our work, we propose a framework taking all scalable resources into consideration by transforming this problem into multi-objective optimization (MOO) problem and rapid solving it via genetic algorithm. Next, we leverage reinforcement learning (RL) to automatically learn efficient workload- and system-specific scheduling policies. Existing HPC schedulers either use generalized and simple heuristics or optimization methods that ignore workload and system characteristics. To overcome this issue, we design a new scheduling agent DRAS to automatically learn efficient scheduling policies. DRAS leverages the advance in deep reinforcement learning and incorporates the key features of HPC scheduling in the form of a hierarchical neural network structure. We develop a three-phase training process to help DRAS effectively learn the scheduling environment (i.e., the system and its workloads) and to rapidly converge to an optimal policy. Finally, we explore the problem of scheduling mixed workloads, i.e., rigid, malleable and on-demand workloads, on a single HPC system. Traditionally, rigid jobs are the main tenants of HPC systems. In recent years, malleable applications, i.e., jobs that can change sizes before and during execution, are emerging on HPC systems. In addition, dedicated clusters were the main platforms to run on-demand jobs, i.e., jobs needed to be completed in the shortest time possible. As the sizes of on-demand jobs are growing, HPC systems become more cost-efficient platforms for on-demand jobs. However, existing studies do not consider the problem of scheduling all three types of workloads. In our work, we propose six mechanisms, which combine checkpointing, shrink, expansion techniques, to schedule the mixed workloads on one HPC system.
Show less

Title: Developing Novel Optimization Algorithms Applied To Building Energy Performance and Indoor Air Quality
Creator: Faramarzi, Afshin
Date: 2021
Description: Residential and commercial buildings account for 23% of global energy use. In the United States, space heating, cooling, and lighting energy...
Show moreResidential and commercial buildings account for 23% of global energy use. In the United States, space heating, cooling, and lighting energy use accounts for 38%, 9%, and 7% of building energy consumption, which results in 54% of the total energy consumption of the building. Energy efficiency improvements in buildings require consideration of optimal design, operation, and control of building components (e.g., mechanical and envelope systems). We can address this task by taking advantage of computational optimization methods throughout the design, operation, and control processes.Non-gradient metaheuristic optimization methods known as metaheuristics are some of the most popular and widely used optimization methods in Building Performance Optimization (BPO) problems. Conventional metaheuristics usually have simple mathematical models with low rate of convergence. On the other hand, high-performance metaheuristic optimizers are efficient and usually have a fast rate of convergence, but their mathematical models are hard to understand and implement. As such, researchers are usually not inclined to employ them in solving their problems. To this end, we aimed at developing optimization algorithms which borrow simplicity from conventional methods and efficiency from high-performance optimizers to solve problems fast and efficiently while being welcomed by users from throughout the world. Therefore, the overarching objective of this work is defined to first develop novel optimization algorithms which are simple in mathematical models and still efficient in solving optimization benchmark problems and then apply the methods to building energy performance and indoor air quality (IAQ) problems. In the first objective of this work, which is the development phase, two continuous optimization methods and one binary optimizer are developed and are separately described in three different tasks. The first method called Equilibrium Optimizer (EO) is a simple method inspired by the mass balance equation in a control volume. The second optimization method called Marine Predators Algorithm (MPA) is a more complicated method compared to EO and is inspired by widespread foraging strategies between marine predators in the ocean ecosystem. Finally, the third method is the binary version of an already developed equilibrium optimizer called Binary Equilibrium Optimizer (BEO). The second objective of the dissertation is the application phase which focuses on the application of the developed methods and other widely used methods in research and industry for solving the almost new BPO and IAQ problems. The results showed that the developed methods were able to either reach more energy-efficient solutions compared to the other methods or to show a considerably faster rate of convergence compared to other methods in the problems in which the optimal solutions are similarly obtained by different methods.
Show less

Title: The Feasibility of Honeycomb Structure to Enhance Daylighting and Energy Performance for High-Rise Buildings
Creator: Geng, Camelia Mina
Date: 2022
Description: The world population is increasing at a fast rate and the projection is that there will be more than 12 billion people by the year 2050. It is...
Show moreThe world population is increasing at a fast rate and the projection is that there will be more than 12 billion people by the year 2050. It is also expected that at least 70% of the population will reside and work in urban areas (mostly cities) in some sort of high-rise building. At the same time, the climate is rapidly changing to increase the effects of man-made global warming. Conceivably, energy conservation, daylighting performance, thermal comfort and environmentally friendly high-rise buildings are necessary to facilitate sustainable working and living environments. The roles of the architects and planners are paramount at this critical era of history of mankind; for one thing they are responsible for the planning and design of sustainable high-rise buildings.Recently, there has been significant research to connect a branch of Biophilia design, which is Biomorphic architecture. This has developed a wonderful design approach, termed the Biomorphic idea. This focuses on the enhancement of the physical and psychological connection with nature, to acquire more natural light and the outside connection targeting energy saving. More and more, high-rise buildings are being designed following Biomorphic approaches. As such, these buildings are defined as sustainable and primarily, because they are energy efficient and, and in many cases tend to minimize the use of fossil fuels while promoting the use of renewable and clean energy sources. As such, a honeycomb structure approach successfully applies to high-rise building design. The intend of this research document is to simulate Biomorphic honeycomb structure which is the hexagonal rotation ring structure including 32 stories in18 different hexagon high-rise building configurations, to develop true daylighting and energy. performance. This is achieved by the using Grasshopper-Climate Studio simulation tool and multiple fuzzy mathematics for decision making. This document will provide a comparison of daylighting including sDA, ASE, sDG and the illuminance results from these 3 series of the 18 models configuring different honeycomb structures of high-rise buildings. The results prove that the hexagon honeycomb structure for high-rise building is feasibility and targets green buildings standards such as LEED V4.1 The success of the method depends on developing multiple criteria of Poisson ratio and Gaussian curvature within the hexagon structure to create different honeycomb facades and rotation of the ring for office high-rise building which is also a qualitative nature of the Biomorphic design parameters.
Show less

Title: Two Essays on Cryptocurrency Markets
Creator: Fan, Lei
Date: 2022
Description: Understanding the dependence relationships among cryptocurrencies and equity markets is of interest to both academics and researchers. This...
Show moreUnderstanding the dependence relationships among cryptocurrencies and equity markets is of interest to both academics and researchers. This dissertation is comprised of two essays to add to this understanding. In the first essay, I investigate the interdependencies among the level of informational efficiency of four cryptocurrencies. I examine the correlations between the market efficiencies of cryptocurrencies using the rolling window method. I find that the correlations between those levels of market efficiencies are time-varying and influenced by the market condition and external events. I extend the study by employing Granger causality tests to analyze the causal relationships among these levels of market efficiency. I find that the Granger causalities among the levels of the cryptocurrency market efficiencies are time-varying and impacted by the level of the market efficiencies. In the second essay, I investigate the pairwise dependencies and causalities between the returns of the cryptocurrencies and six equity market indices. I examine the pairwise dependencies between the returns of cryptocurrencies and those of the equity indices by using the DCC-GARCH framework. I find the dynamic conditional correlations between the cryptocurrencies and equity indices are time-varying and generally weak. Furthermore, I study the causal relationship between cryptocurrencies and equity indices by employing the rolling Granger causality test. I find that the Granger causalities between cryptocurrencies and equity indices are time-varying, and more unidirectional Granger causalities are found from cryptocurrencies to equity indices. In addition, I examine the impact of cryptocurrency returns on the correlations between the equity market indices, and likewise, the impact of equity market returns on the correlations between the cryptocurrencies. I find that the cryptocurrency price fluctuations have minimal impact on the correlations between equity indices. Moreover, the dynamic conditional correlation between cryptocurrencies is unaffected by equity price innovations except for some extreme events. These findings could have implications for understanding the relationships among cryptocurrencies and equity markets and for investors wishing to incorporate these relationships in their portfolio choices.
Show less

Title: PLAYER MOTIVATION AND TRAINING EFFECTIVENESS: INSIGHTS FROM A STRUCTURAL MODEL OF GAME-BASED LEARNING
Creator: Gandara, Daniel A.
Date: 2022
Description: Digital game-based learning (DGBL) delivers training through video games. Practitioners are using DGBL in attempts to increase motivation,...
Show moreDigital game-based learning (DGBL) delivers training through video games. Practitioners are using DGBL in attempts to increase motivation, promote learning, and increase transfer in training. Theory and models of DGBL aim to explain how motivation is created to yield these benefits, and studies have compared DGBL to traditional methods, yet the tenets of these theories remain largely unexamined. The present study tested the process-outcome link of Garris et al.’s (2002) input-process-outcome model, examined the effect of positive and negative user judgments on behavior and learning, and expanded the model to include trainee reactions and adaptive transfer. Participants (N = 254) learned about identifying misinformation online by playing Fake It to Make It, a social-impact game that teaches core critical thinking skills. Autoregressive cross-lagged (ARCL) panel analysis was used to analyze and compare models to test the hypothesized relationships among judgments and behavior scores across six game levels in predicting six learning outcomes, including adaptive transfer tasks evaluating online sources. Findings indicated that each judgment was predicted by its own lagged judgment and lagged behavior. Additionally, positive user judgments predicted reactions, post-training self-efficacy, and motivation to transfer, while frustration inhibited declarative knowledge. Results also demonstrated that behavior and declarative knowledge predicted performance on the adaptive transfer tasks. Research recommendations and practice implications are discussed relative to using games to deliver training with emphasis on motivational properties and targeted outcomes.
Show less

Title: Child Temperament, Attachment, and Loneliness: The Mediating Effects of Social Competence
Creator: Evans, Lindsey M
Date: 2021
Description: Chronic loneliness is a risk factor associated with adverse psychological, physical, and academic outcomes. Converging evidence suggests that...
Show moreChronic loneliness is a risk factor associated with adverse psychological, physical, and academic outcomes. Converging evidence suggests that young children experience and can reliably report on their own loneliness. Due to the significant negative sequalae associated with childhood loneliness, it is critically important to examine risk factors for child loneliness. The aims of this study were two-fold: (a) to examine if temperament (i.e., negative affect, effortful control, and inhibitory control) and attachment security assessed at 4 years of age predict loneliness at age 6; and (b) to determine if social competence at age 5 mediates the relation between temperament and attachment security at age 4 and loneliness at age 6. Participants included a diverse sample of 796 4-year old children, about half of whom were male. At age 4, temperament was assessed with the Rothbart Child Behavior Questionnaire and three inhibitory control tasks, and attachment security was assessed with the Attachment Q-Sort. At age 5, the Social Skills Rating Scale was used to assess social competence, and, at age 6, loneliness was assessed with the Loneliness and Social Dissatisfaction Questionnaire. Results of hierarchical regression analyses indicated that lower levels of effortful control and inhibitory control at age 4 significantly predicted higher levels of loneliness at age 6. Also, lower levels of negative affect and higher levels of effortful control and attachment security at age 4 significantly predicted higher levels of social competence at age 5. However, social competence at age 5 did not predict loneliness at age 6. There was no evidence that social competence at age 5 mediated the relation between age 4 temperament, attachment security and age 6 loneliness. These findings reveal that early self-regulation is associated with later child-reported loneliness and that intervention for children who struggle with cognitive regulation may be effective in decreasing risk for later loneliness.
Show less

Title: COMPUTATIONAL FLUID DYNAMICS SIMULATION OF CARBON CAPTURE UNIT USING AN AMINE-BASED SOLID SORBENT
Creator: Esmaeili Rad, Farnaz
Date: 2021
Description: Carbon capture and sequestration (CCS) is one of the key technologies to reduce the emission of carbon dioxide, including that from exiting...
Show moreCarbon capture and sequestration (CCS) is one of the key technologies to reduce the emission of carbon dioxide, including that from exiting flue gas of fossil fuel-fired power plants. The goal of this project is the development of a computational fluid dynamics (CFD) model to predict the extent of CO2 capture in a circulating fluidized bed carbon capture unit using novel amine-based solid sorbents.In this study, first the hydrodynamics of the carbonation section of the carbon capture unit was investigated. Then, the performance of the amine-based solid sorbents toward capturing carbon dioxide from flue gas and the extent of CO2 adsorption in the carbonation section were studied. At the second stage of the study, the regeneration of the sorbents and desorption of carbon dioxide from carbonated solid sorbents in the regeneration section of the carbon capture unit was investigated. At the third stage of the study, the hydrodynamics of the entire loop of the integrated carbonation and regeneration sections were simulated. Two-dimensional non-reactive CFD simulations of the entire loop, including the carbonator, regenerator, and two loop-seal fluidized beds, were performed to study the details of the solid circulation in the system in a stable operational condition. At the fourth stage of the study, the effect of the carbonated solids’ residence time in the regeneration section was investigated by extending the regenerator fluidized bed height and adding to the volume of the system. Heated surfaces, which resembled heating coils in the regenerator cylinder, were also added to the system to investigate the effect of the temperature. The heated surface of the immersed coils in the bed provided sufficient energy for the endothermic regeneration reaction to keep the temperature of the bed at the desired temperature. Finally, the verified models of the carbonation section, the regenerations section, and non-reactive simulation of the CFB loop were used to simulate the entire circulating fluidized bed carbon capture unit, with an integrated carbonator and regenerator system using amine-based solid sorbents. The extent of CO2 capture in the carbonation section and desorption of carbon dioxide in the regeneration section were predicted. Our study showed the potential of continuous carbon capture by amine-based solid sorbents through the circulating fluidized bed CO2 capture unit.
Show less

Title: Improvement and Validation of Multiyear Auroral Analysis to Categorize Scintillation Event Layer
Creator: English, Breanna R.
Date: 2022
Description: Ionospheric irregularities scintillate electromagnetic waves, such as Global Positioning System (GPS) signals, as they pass through the...
Show moreIonospheric irregularities scintillate electromagnetic waves, such as Global Positioning System (GPS) signals, as they pass through the ionosphere, especially in auroral zones. A previous method was developed to determine which layer of the ionosphere these scintillation events occurred in by analyzing optical all sky images (ASI). The results of determining the ionospheric scattering layer using the ratio of 630 nm (red) intensity to 428 nm (blue) intensity were compared to a radar-based method of determining the scintillation layer, and it was found that the results disagreed. In this work, the ASI method is critically analyzed to identify possible errors or sensitivities in the original method that might resolve the discrepancy. This is done by improving and validating the nighttime auroral cloud detection method by comparing to National Oceanic and Atmospheric Administration (NOAA) satellite cloud data. Then a sensitivity analysis is performed on the ASI method to determine which parameters of the method the results are sensitive to. The keogram cloud detection method is improved by automating the selection of the keogram time points that are used to calculate a flat-field gain correction, and by calculating the flat field gain for each year rather than calculatingit once and using it for all years of the study. Keogram cloud detection using the coefficient of variation is verified by comparing the keogram results to true sky conditions based on NOAA cloud mask data, and using detection theory to determine the optimal coefficient of variation threshold. We find that the ideal keogram threshold was 0.37 producing a disagreement rate of 22.4%. The ASI image analysis criteria tested are: the ASI azimuth and elevation mapping files, the magnetic zenith limit, the number of pixels of the ASI that are being analyzed, the duration of the scintillation event that is analyzed, and the red-to-blue ratio threshold. It is found that only changing the red-to-blue ratio threshold has a significant effect on the ASI method, with the red-to-blue ratio that minimizes the number of misattributed layers found to be 1.43.
Show less

Title: Algorithms for Discrete Data in Statistics and Operations Research
Creator: Schwartz, William K.
Date: 2021
Description: This thesis develops mathematical background for the design of algorithms for discrete-data problems, two in statistics and one in operations...
Show moreThis thesis develops mathematical background for the design of algorithms for discrete-data problems, two in statistics and one in operations research. Chapter 1 gives some background on what chapters 2 to 4 have in common. It also defines some basic terminology that the other chapters use.Chapter 2 offers a general approach to modeling longitudinal network data, including exponential random graph models (ERGMs), that vary according to certain discrete-time Markov chains (The abstract of chapter 2 borrows heavily from the abstract of Schwartz et al., 2021). It connects conditional and Markovian exponential families, permutation- uniform Markov chains, various (temporal) ERGMs, and statistical considerations such as dyadic independence and exchangeability. Markovian exponential families are explored in depth to prove that they and only they have exponential family finite sample distributions with the same parameter as that of the transition probabilities. Many new statistical and algebraic properties of permutation-uniform Markov chains are derived. We introduce exponential random ?-multigraph models, motivated by our result on replacing ? observations of a permutation-uniform Markov chain of graphs with a single observation of a corresponding multigraph. Our approach simplifies analysis of some network and autoregressive models from the literature. Removing models’ temporal dependence but not interpretability permitted us to offer closed-form expressions for maximum likelihood estimators that previously did not have closed-form expression available. Chapter 3 designs novel, exact, conditional tests of statistical goodness-of-fit for mixed membership stochastic block models (MMSBMs) of networks, both directed and undirected. The tests employ a ?²-like statistic from which we define p-values for the general null hypothesis that the observed network’s distribution is in the MMSBM as well as for the simple null hypothesis that the distribution is in the MMSBM with specified parameters. For both tests the alternative hypothesis is that the distribution is unconstrained, and they both assume we have observed the block assignments. As exact tests that avoid asymptotic arguments, they are suitable for both small and large networks. Further we provide and analyze a Monte Carlo algorithm to compute the p-value for the simple null hypothesis. In addition to our rigorous results, simulations demonstrate the validity of the test and the convergence of the algorithm. As a conditional test, it requires the algorithm sample the fiber of a sufficient statistic. In contrast to the Markov chain Monte Carlo samplers common in the literature, our algorithm is an exact simulation, so it is faster, more accurate, and easier to implement. Computing the p-value for the general null hypothesis remains an open problem because it depends on an intractable optimization problem. We discuss the two schools of thought evident in the literature on how to deal with such problems, and we recommend a future research program to bridge the gap those two schools. Chapter 4 investigates an auctioneer’s revenue maximization problem in combinatorial auctions. In combinatorial auctions bidders express demand for discrete packages of multiple units of multiple, indivisible goods. The auctioneer’s NP-complete winner determination problem (WDP) is to fit these packages together within the available supply to maximize the bids’ sum. To shorten the path practitioners traverse from from legalese auction rules to computer code, we offer a new wdp formalism to reflect how government auctioneers sell billions of dollars of radio-spectrum licenses in combinatorial auctions today. It models common tie-breaking rules by maximizing a sum of bid vectors lexicographically. After a novel pre-solving technique based on package bids’ marginal values, we develop an algorithm for the WDP. In developing the algorithm’s branch-and-bound part adapted to lexicographic maximization, we discover a partial explanation of why classical WDP has been successful in using the linear programming relaxation: it equals the Lagrangian dual. We adapt the relaxation to lexicographic maximization. The algorithm’s dynamic-programming part retrieves already computed partial solutions from a novel data structure suited specifically to our WDP formalism. Finally we show that the data structure can “warm start” a popular algorithm for solving for opportunity-cost prices.
Show less

Title: TOPICDP – ENSURING DIFFERENTIAL PRIVACY FOR TOPIC MINING
Creator: Sharma, Jayashree
Date: 2021
Description: Topic mining enables applications to recognize patterns and draw insights from text data, which can be used for applications such as sentiment...
Show moreTopic mining enables applications to recognize patterns and draw insights from text data, which can be used for applications such as sentiment analysis, building of recommender systems and classifiers. The text data can be a set of documents or emails or product feedback and reviews. Each document is analysed using probabilistic models and statistical analysis to discover patterns that reflects underlying topics.TopicDP is a differentially private topic mining technique, which injects well-calibrated Gaussian noise into the matrix output of the topic mining model generated from LDA algorithm. This method ensures differential privacy and good utility of the topic mining model. We derive smooth sensitivity for the Gaussian mechanism via sensitivity sampling, which resses the major challenges of high sensitivity in case of topic mining for differential privacy. Furthermore, we theoretically prove the differential privacy guarantee and utility error bounds of TopicDP. Finally, we conduct extensive experiments on two real-word text datasets (Enron email and Amazon Product Reviews), and the experimental results demonstrate that TopicDP can generate better privacy preserving performance for topic mining as compared against other state-of-the-art differential privacy mechanisms.
Show less

repository.iit

Search the repository

Pages

Pages

Enabled Filters

Refine Results

Type

Date

Academic status of creator

Department

Subject

Creator

Part of