Search

The problem addressed through this project is the decline in the retention of foreign students in U.S. institutions of higher learning due to the COVID-19 pandemic. In the 2020-2021 school year, international student enrollment declined by 15%, demonstrating the direct impact of the pandemic on U.S. institutions of higher learning (Svrluga, 2021). Moreover, the financial contribution of foreign students dropped by 27% to $28.4 billion in the same financial year, which was attributed to perceptions among international students that they were unwelcome, in addition to visa processing challenges. The purpose of the study was to examine factors influencing the retention of foreign learners in U.S. higher education institutions during the COVID-19 pandemic. The study aimed to identify the unique challenges international students faced coping with the COVID-19 pandemic while in the United States. The study further examined some key interventions higher education institutions should have implemented to ensure international students' experiences remained positive and existing challenges were resolved effectively. The study was guided by Spady's model, Tinto's theory, and the PICO framework (Considine et al., 2017; Nicoletti, 2019). The study was grounded on a systematic review of pertinent literature. The studies' relevance to the research questions was derived from journals, conference proceedings, electronic databases of U.S. colleges and universities, research reports from research organizations, and other electronic databases such as JSTOR, ProQuest, ScienceDirect, and Emerald. The PRISMA 2009 flow diagram guided the process of selecting studies. Results from quantitative studies were compared based on the methodologies adopted and results found out. Qualitative studies that were included were analyzed thematically. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]

Many-objective optimization is core to both artificial intelligence and data analytics as real-world problems commonly involve multiple objectives which are required to be optimized simultaneously. A large number of evolutionary algorithms have been developed to search for a set of Pareto optimal solutions for many-objective optimization problems. It is very rare that a many-objective evolutionary algorithm performs well in terms of both effectiveness and efficiency, two key evaluation criteria. Some algorithms may struggle to guide the solutions towards the Pareto front, e.g., Pareto-based algorithms, while other algorithms may have difficulty in diversifying the solutions evenly over the front on certain problems, e.g., decomposition-based algorithms. Furthermore, some effective algorithms may become very computationally expensive as the number of objectives increases, e.g., indicator-based algorithms. The aim of this thesis is to investigate how to make evolutionary algorithms perform well in terms of effectiveness and efficiency in many-objective optimization. After conducting a review of key concepts and the state of the art in the evolutionary many-objective optimization, this thesis shows how to improve the effectiveness of conventional Pareto-based algorithms on a challenging real-world problem in software engineering. This thesis then explores how to further enhance the effectiveness of leading many-objective evolutionary algorithms in general by extending the capability of a very popular and widely cited bi-goal evolution method. Last but not least, this thesis investigates how to strike a balance between effectiveness and efficiency of evolutionary algorithms when solving many-objective optimization problems. The work reported is based on either real-world or recognized synthetic datasets, and the proposed algorithms are compared and evaluated against leading algorithms in the field. The work does not only demonstrate ways of improving the effectiveness and efficiency of many-objective optimization algorithms but also led to promising areas for future research.

This thesis explores the ways in which schools accommodate the national vision of a multicultural society, specifically focusing on Indonesia's national vision of "Bhinneka Tunggal Ika" (Unity-in-Diversity). Employing a multiple-case-study research approach, it examines the ways head teachers, teachers, students, and parents from six state and private schools with different mixes of student ethnicity and religiosity promote this national vision in terms of educational practice. The research considers the perspectives of education stakeholders on the meanings of Unity-in-Diversity and their responses in promoting Unity-in-Diversity in schools and classrooms in contemporary Indonesia, drawing data from observations, interviews, photos, and documentary analysis. Using multicultural education as a conceptual lens, this research finds that promoting religious diversity is key to achieving Unity-in-Diversity in religious and democratic Indonesia, and schools generally have demonstrated some commitment to promoting inter-religious harmony as mandated. Yet, it argues that the discourses of diversity, particularly of religious diversity, continue to be spaces of complex interpretation, rather than settled policy positions. In schools, this process of interpretation is manifested in the school policies, school cultures, and teaching across subjects in classrooms. The strong narratives of religiosity and piety in the current education system, which are found to privilege the dominant religious culture(s), have ramifications in the accommodation of Unity-in-Diversity in practice. In response to the different schools' demographies, the nuanced accounts regarding the implementation of Unity-in-Diversity suggest the importance of considering identity and power dynamics, positionality, and contextuality in the school accommodation of the vision of diversity. In light of this research, this thesis also presents the usefulness and challenges of applying a theory of multicultural education, which has focused mainly on ethnic and racial diversity, in researching the issue of religious diversity. Based in the largest Muslim country in the world, this thesis offers a critical insight into the complex role of education in the development of an inclusive, religious society.

The lack of mechanisation and modernisation adoption, unavoidable dependence on labour workforce, the lack of skills and experience, as well as the reliance on foreign workers, were amongst the challenges in improving the industrialised building system (IBS) in the Malaysian construction industry (MCI). Furthermore, the techniques and performance of the labourers in precast concrete for this study referred to hereon is Industrialiased Building System Precast Concrete Consruction (IBSPCC) are similar to those of the In-situ Concrete Construction (ICC), resulting in unnecessary repetition, which consequently creates non-value-added (NVA) activities that could affect productivity growth. To address the issues above, a significant improvement towards the IBSPCC, encompassing the production and the assembly work processes, may be achieved by adopting the lean approach, through which the value-added (VA) activities, NVA activities, and non-value added but required (NVAR) activities can be distinguished. Therefore, this research aimed to develop an IBSPCC work processes model, covering the reinforcement, precast moulding, concreting and assembling, using the lean approach as a strategy for work process improvement. To evaluate its effectiveness, each work process was compared against those of the ICC method, based on the duration in producing reinforced concrete (RC) structural frame (columns, beams and slabs), gang size, as well as the measure of their productivity level in terms of output per man-hour. Mixed methods study was undertaken, to provide comprehensive answer to the research question. Through this study, there were contradicted findings between questionnaires and the observation for both the IBSPCC and the ICC methods, in terms of the duration and gang size, for the reinforcement, formwork or precast moulding and concreting work processes. This also includes the findings obtained through the evaluation of productivity output per man-hour that indicates certain work processes of the IBSPCC has less output as compared to the ICC method. This confirms that some repetition of the ICC method occur in the IBSPCC work process. In other words, reliance on labour workforce and minimal machinery usage still persist. The only difference is in the location of producing the components, transportation, precast moulding and assembly work process. It was also found that a number of NVA activities are comparable with the seven classic categories such as overproduction, waiting, and unnecessary inventory, although the causes may be different. Surprisingly, a majority of the IBSPCC manufacturers have had an opposing view of overproduction as one of the NVA activities. To them, the more that they are able to produce, the better is the performance, which is in contrast with the lean approach. Apart from that, several NVA activities especially during assembly work process were uncontrollable due to procedures, working method, technological development or actions that need to be undertaken by others in the supply chain. This makes the mitigation of NVA activities more challenging. Due to the need for elimination or minimisation of NVA activities, the work process model has been developed, incorporating several strategies such as an optimisation, multi-tasking, standardisation, planning layout, just in time (JIT) and Kanban, which all have resulted in a simplified work process, less time, an improved labourers' working pattern and has a high potential to increase productivity. Although the case studies presented here cannot be generalised to all IBS classifications, the work process models could be a starting point to generate preliminary ideas on how the lean strategies can improve the IBSPCC work process. It is also hoped that this research will potentially increase awareness and knowledge regarding the lean approach amongst the IBSPCC stakeholders.

Deep learning has in recent years come to dominate the previously separate fields of research in machine learning, computer vision, natural language understanding and speech recognition. Despite breakthroughs in training deep networks, there remains a lack of understanding of both the optimization and structure of deep networks. The approach advocated by many researchers in the field has been to train monolithic networks with excess complexity, and strong regularization --- an approach that leaves much to desire in efficiency. Instead we propose that carefully designing networks in consideration of our prior knowledge of the task and learned representation can improve the memory and compute efficiency of state-of-the art networks, and even improve generalization --- what we propose to denote as structural priors. We present two such novel structural priors for convolutional neural networks, and evaluate them in state-of-the-art image classification CNN architectures. The first of these methods proposes to exploit our knowledge of the low-rank nature of most filters learned for natural images by structuring a deep network to learn a collection of mostly small, low-rank, filters. The second addresses the filter/channel extents of convolutional filters, by learning filters with limited channel extents. The size of these channel-wise basis filters increases with the depth of the model, giving a novel sparse connection structure that resembles a tree root. Both methods are found to improve the generalization of these architectures while also decreasing the size and increasing the efficiency of their training and test-time computation. Finally, we present work towards conditional computation in deep neural networks, moving towards a method of automatically learning structural priors in deep networks. We propose a new discriminative learning model, conditional networks, that jointly exploit the accurate representation learning capabilities of deep neural networks with the efficient conditional computation of decision trees. Conditional networks yield smaller models, and offer test-time flexibility in the trade-off of computation vs. accuracy.

The state of conservation practice and the fate of historic buildings and cities reflect the critical issue of the value of heritage in Malaysia. The idea of safeguarding heritage has been far behind the agenda of the government until recent years. This has resulted in physical and cultural heritage loss. The consciousness of the importance of heritage has triggered the implementation of the National Heritage Act 2005 which focuses on safeguarding the nation's heritage. The implementation of the act, however, has been late when compared to the poor state of built heritage in Malaysia. This Base Line research focuses on how the conservation of heritage can lead to the regeneration of historic cities. The study seeks to identify the important characteristics of historic cities which give them identity and to examine the relationship between the conservation of historic buildings and the regeneration of historic cities, particularly in Malaysia, in order to understand the elements which should be taken into consideration in regeneration or redevelopment schemes carried out within historic cities. The scope of the research covers both theoretical and practical aspects of conservation practice and regeneration initiatives of historic cities, specifically relevant to Malaysia, taking into consideration social, cultural, political, and economic aspects and stakeholders' aspirations. This thesis explains the relationship between conservation of historic buildings and regeneration of historic cities by looking at the examples of regeneration schemes in the UK and development proposals by local authorities in Malaysia. The results from the study give an understanding of the scenario in the historic areas in Melaka and George Town, to assess whether the conservation proposals carried out have improved the area and have helped in regenerating the economy of the area, thus, to providing a baseline to demonstrate that the conservation of historic buildings can lead to the regeneration of historic areas in Malaysia.

Electrochemical amination of acetone with methylamine as nitrogen source, the electric double layer on Cu electrode was studied in the conditions of electrochemical amination of acetone with methylamine as Nitrogen source. Therefore, double layer capacitance is used as a key descriptor to investigate the surface pre-configuration or preference of adsorption. It was found that in alkaline water solution, acetone (hydrophobic) and methylamine (hydrophilic) will result in different local dielectric property, which cause increase/decrease in Helmholtz capacitance. Besides, it was observed that the working Cu electrode is corroded using mixtures of acetone and methylamine even under reductive potential conditions (-0.75V vs RHE). We employed Grand Canonical density functional theory to understand this dynamic process under potential from a microscopic perspective. We show that amine ligands in solution directly adsorb on the electrode, coordinate with the metal center, drive the rearrangement of the copper surface by extracting Cu atoms in low coordination positions, finally forming a detached Cu-tetramine di-cation complex in solution, even at negative potential conditions.

Medical image segmentation is the procedure to delineate anatomical structures and other regions of interest in various image modalities. While crucial and often a prerequisite step for other analysis tasks, accurate automatic segmentation is difficult to obtain, especially for three dimensional (3D) data. Recently, deep learning techniques have revolutionized many domains of artificial intelligence (AI) including image search, speech recognition and 2D/3D natural image/video segmentation. When it comes to 3D image segmentation, however, the majority of deep learning solutions either treat 3D volumes as stacked 2D slices, overlooking the adjacent information between slices, or directly perform 3D convolutional operations with isotropic kernels that are inconsistent with the anisotropic dimensions in 3D medical data. Neural networks based on 3D convolutions tend to be computationally costly, as well as require much more training data to account for the increased number of parameters that need to be tuned. The scarcity of annotated data in medical imaging also adds up the difficulty.To remedy the aforementioned drawbacks of existing solutions, we propose two works for 3D volume segmentation in this dissertation. The first work is multi-view ensemble convolutional neural network (CNN) framework in which multiple decision maps generated along different 2D views are integrated. The second work is a novel end-to-end deep learning architecture that combines CNN and Recurrent Neural Network (RNN) to better leverage the dimensional anisotropism in 3D medical data. Our model is designed with the aim to take advantage of CNNs remarkable power in capturing multi-scale 2D features, while rely on multi-view ensemble learning or inter-slice sequential learning to ensure certain level of output consistency through inter-slice contextual constraints. Experiments conducted on hippocampus magnetic resonance imaging (MRI) data for both work demonstrate that the multi-view solution and the joint CNN-RNN model achieve significant improvements over single-view approaches, and outperform many state-of-the-art solutions in hippocampus segmentation. Our work also show better results when comparing with 3D CNN method. In addition, we further validate our proposed work on other neuroimage segmentation task, i.e., multiple-class segmentation for brain tumor (glioma) using pre-operative multi-modal MRI scans. Also, experimental results demonstrate that our proposed solutions can effectively improve the accuracy and consistency of the tumor segmentation, and show a very comparative performance when compared to the state-of-the-art solutions.

Hangout! is a mobile social-activity app, encouraging users to connect with friends and family. Users can specify their preferred leisure activity, whether that may be camping or caving, kayaking or surfing, running or rock climbing—whatever type of fun they are looking for. Combining a GPS location and a user’s history of outside recreational experiences, the app provides news about the area in real time, along with any emergency notifications for issuing caution. As the name implies, Hangout! is an application celebrating the recreational places you frequent and the people you socialize with. High-quality imagery and video helps bolster a high aesthetic.

A chemical-proteomic platform to monitor cysteine sensitivity to transnitrosation Yani Zhou Dissertation advisor: Dr. Eranthie Weerapana Abstract S-nitrosation has emerged as a ubiquitous endogenous protein posttranslational modification that significantly impacts cellular protein function through a variety of mechanisms. Despite the advent of chemical and proteomic methods to study S-nitrosation, the subset of cellular cysteine residues that show uniquely high reactivity to endogenous transnitrosation donors is poorly characterized. To further these existing global studies, a cysteine-reactivity profiling strategy was applied herein to rank ~600 cysteine residues by sensitivity to S-nitrosoglutathione. These proteomic studies revealed several previously uncharacterized sites of S-nitrosation, including Cys58 in HADH2. Further characterization revealed that HADH2 catalytic activity is allosterically regulated by S-nitrosation, and this modification occurs in cells at (patho)physiological levels of nitrosative stress. Functional role of Cys58 and its regulation by S-nitrosation facilitated the identification of RB-21-CA as a potential covalent Cys58 inhibitor. Global analysis of GSNO, S-nitroso-Coenzyme A and Thioredoxin-C73-SNO transnitrosation identified 756 cysteines with different sensitivity to each of three SNO donors. Systematic evaluation on transnitrosation selectivity revealed that specific interaction of transnitrosation donor with its protein target is a key component governing the selective transnitrosation of a specific cysteine residue. Together, these studies illustrated the potential of cysteine-reactivity profiling strategy for evaluating the substrate specificity of transnitrosation donors and enable the identification of previously uncharacterized, functionally relevant sites of S-nitrosation. Another cysteine oxoform, S-glutathionylation, is the disulfide formation of a protein cysteine residue with glutathione. Although glucose starvation is known to induce redox-disturbance, global and individual protein S-glutathionylation in response to glucose metabolism or mitochondrial activity remains largely unknown. By using a clickable glutathione approach, which forms clickable glutathione by the use of a mutant of glutathione synthetase, we found that protein S-glutathionylation is readily induced in response to glucose starvation when mitochondrial reactive oxygen species are elevated in cells, and glucose is the major determinant for inducing reversible glutathionylation. Application of a proteomic mass spectrometry platform identified over 1,300 S-glutathionylated. Confirmation of S-glutathionylation for selected proteins by in gel analysis further validated the mass spectrometry results, and highlights the dynamic change of S-glutathionylation on an individual protein level. In order to expand on the understanding of the functional role of the cysteine residues in biological systems, we evaluated a panel of 1,3,5-triazine- and 4-aminopiperidine-based cysteine-reactive small-molecules on two proteins, apoptosis signal-regulating kinase 1 (ASK1) and peroxiredoxin 1 (Prdx1), between which a intermolecular disulfide forms and results in the activation of mitogen-activated protein kinase pathway. In-gel fluorescence revealed that RB-11-CA and SMC-1, both of which contain an n-octyl group as a diversity element, showed greatest selectivity and potency to ASK1 and Prdx1, respectively. Further mass spectrometry analysis identified cysteine 225 of ASK1 and cysteine 173 of Prdx1 are the sites of covalent probe modification.

This paper focuses on graph layouts with constraints using force-directed simulations. Existing graph drawings with constraints include placement of a particular vertex or a group of vertices at a specified location, constraining placement of vertices and edges to specified rows and/or columns. We propose an alternative way of specifying constraints by allowing the user to interactively draw a boundary wherein the graph layout will be constrained. Such boundary constraints may be saved and applied to other graphs as well. In addition, the boundary may be of different topology such as a donut shape, or figure-eight shape, etc. We model these boundaries as a set of additional forces that contribute to the forces acting on graph vertices. Because our proposed approach is force-directed, it can take advantage of optimizations of other force-directed graph layout algorithms. Furthermore, one can utilize the knowledge of the size of the graph to be visualized and the size of the interior of the boundary region to scale the forces appropriately to achieve a uniform distribution of vertices. We tested this idea on several data sets and different boundary constraints.

Bacterial endotoxins (lipopolysaccharides (LPS)) are strong elicitors of the human immune system by interacting with serum and membrane proteins such as lipopolysaccharide-binding protein (LBP) and CD14 with high specificity. At LPS concentrations as low as 0.3 ng/ml, such interactions may lead to severe pathophysiological effects, including sepsis and septic shock. One approach to inhibit an uncontrolled inflammatory reaction is the use of appropriate polycationic and amphiphilic antimicrobial peptides, here called synthetic anti-LPS peptides (SALPs). We designed various SALP structures and investigated their ability to inhibit LPS-induced cytokine secretion in vitro, their protective effect in a mouse model of sepsis, and their cytotoxicity in physiological human cells. Using a variety of biophysical techniques, we investigated selected SALPs with considerable differences in their biological responses to characterize and understand the mechanism of LPS inactivation by SALPs. Our investigations show that neutralization of LPS by peptides is associated with a fluidization of the LPS acyl chains, a strong exothermic Coulomb interaction between the two compounds, and a drastic change of the LPS aggregate type from cubic into multilamellar, with an increase in the aggregate sizes, inhibiting the binding of LBP and other mammalian proteins to the endotoxin. At the same time, peptide binding to phospholipids of human origin (e.g., phosphatidylcholine) does not cause essential structural changes, such as changes in membrane fluidity and bilayer structure. The absence of cytotoxicity is explained by the high specificity of the interaction of the peptides with LPS.

This research explores the politics of editorship in the Kompas and Republika dailies in covering conflict between Muslims and Christians in Maluku. It is attempting to answer why the two most prominent Indonesian newspapers cover the same event differently. As a Catholic-affiliated newspaper, Kompas defended the Christian interests; and as a Muslim newspaper, Republika efended the Muslim interests. Two approaches are employed here to understand the two newspapers’ difference in reporting. The first approach is philosophical framework underpinning the theory of interpretation, known as hermeneutics, and the second one is the political economy which shaped the two newspapers in their current characteristics. Any text or event is open for interpretation, and an interpretation depends on the interpreter’s worldview and attitudes. This factor has made Kompas’s and Republika’s difference in reporting. The political economy, or specifically, the power behind the press contributed to the difference in policies of the two newspapers’ editorship. In the wider context of Indonesia, the repressive New Order regime shaped the “infant” characteristics of Indonesian journalism. Transition from authoritarian regime to a democratic one is also reflected in the press life. The Indonesian press now should reconsider its role along with the democratic reform. Previously repressed to be the mouthpiece of the government, today the Indonesian press becomes a critical watchdog which is crucial in shaping the new Indonesia. The changing role of the Indonesian press now has its momentum, but it depends on the press workers to determine the press role for the future.

The formation of the Himalayas and surrounding mountain ranges is one of the most significant tectonic events of the Cenozoic, with far reaching impacts on both regional and global climate. The Pamir represents the indented westward continuation of the northern margin of the Tibetan Plateau making them a key study area to examine internal deformation and intracontinental subduction processes. Yet the tectonics of the Pamir Salient is poorly understood. Much debate is ongoing, regarding the extent of indentation and mechanisms driving uplift of the Pamir. This project seeks to provide constraints on the timing of Pamir evolution and examine the impact of Pamir formation on the regional climate using a variety of techniques to date and interpret observed variations in terms of provenance, tectonism and climate, the results of which are presented here. At ~41 Ma the Tarim Sea retreated from the Aertashi region. Continental deposition ensued, dominated by the western Kunlun to the south. By 24-22 Ma, basinward expansion of the NW margin of Tibet coincides with the initiation of Pamir indentation. Activation of strike slip faulting, growth strata and exhumation along the Kashgar-Yecheng Transfer System is evidenced in the Aertashi region which may be linked to a pull-apart basin. By ~18.5 Ma, the onset of northward thrusting of the Kunlun hinterland and southward migrating foreland of the Tian Shan positions the Aertashi region in the depocenter of the Tarim Basin. By ~15 Ma, the Pamir Salient had propagated sufficiently far north, that the north Pamir, and by 11 Ma the Central Pamir were proximal to Aertashi. The Eocene – mid Miocene climate in the Tarim Basin was highly seasonal with evidence of increased aridity. Aridification was interrupted by a short period of increased humidity at ~17.6-15 Ma which can be linked to the mid-Miocene climate optimum, or a facies change. Limited comparative studies make climatic conclusions tentative.

This thesis presents the first detailed multiproxy provenance study of the modern Nile catchment, together with Oligocene-Recent sediments of the offshore Nile cone, and demonstrates for the first time that the palaeodrainage of the Nile has remained relatively stable over the last 30 Ma. Detrital mineral and bulk-rock analyses of modern river, wadi and bedrock samples demonstrate the role of basement lithology, sedimentary recycling, and modern geomorphological features in controlling the composition of sediment reaching the Nile cone. The bulk provenance signature of the modern Nile is dominated by the input of basic detritus from the Cenozoic Ethiopian Large Igneous Province. Detrital zircon signatures are dominated by variations in bedrock lithology, and the availability of easily-eroded Phanerozoic cover sediments. These represent an important source of detritus to the river and are characterised by the presence of strongly negative εHf populations at c. 600 and 1000 Ma, and significant populations of pre-Neoproterozoic grains. Young zircon grains derived from Ethiopian Cenozoic volcanic rocks are also recognised in samples from the Blue Nile, Atbara and Nile trunk. The Sr, Nd and Hf signature of Oligocene Nile cone sediments can only be explained if there is significant input of mafic material being received from the hinterland. The Ethiopian Large Igneous Province represents the only possible source for this detritus, indicating that the Nile has been connected to the Ethiopian Highlands since this time. Zircons with U-Pb ages and hafnium isotope signatures compatible with the Ethiopian LIP have been supplied to the Nile cone since at least the Miocene (16 Ma). Such zircons were not observed in Oligocene 2 delta sediments, but Sr, Nd and Hf isotope data for Oligocene delta muds as old as 31 Ma require significant mafic input. The combined U-Pb and Hf-isotope zircon dataset for Nile catchment and cone sediments represents a significant archive that documents the evolution of the North African crust, in particular highlighting phases in the development of the Arabian Nubian Shield and amalgamation of Gondwana in late Neoproterozoic times.