Search

Across the world, grandparents are increasingly becoming the primary caregivers of their grandchildren. China is no exception. The rapid growth of an ageing population, large scale rural-urban internal labour migration, and the large population of left-behind elders and children in rural China have made grandparenting in this area unprecedentedly prevalent. In 2016, there were 9.02 million left-behind children (whose parents were both internal migrants) in rural China, and nearly 90 percent of them were cared for by grandparents. Despite the prevalence of grandparenting in rural China, research on this topic is limited. It particularly lacks qualitative studies from grandparents' perspectives. Under these circumstances, this study aims to address this gap through a detailed qualitative study based on 38 in-depth interviews with grandparents from three different villages in rural China and ten interviews with key informants. The study explores the motivation for, patterns in, and impact of grandparenting on grandparents with themes captured from interview data and discusses the operation and function of grandparenting in the large "society-family-individual" system. The study argues that the prevalent grandparenting in rural China is deliberately led by the state and driven by an economic-growth orientation. This feature distinguishes today's grandparenting in rural China from the traditional Chinese grandparent-grandchild relationship and grandparenting in many Western countries. Grandparenting is profoundly influenced by and also has profound effects on the "society-family-individual" system. In this mutual impact process, it also demonstrates heterogeneity, as some grandfamilies and grandparents are more deeply influenced by modernity and the economic-growth orientation. This heterogeneity has an influence on both the grandparenting style and the grandparents' experience. In general, grandparenting benefits the functioning and development of society as a whole, but it has more profound and long-term negative effects on grandparents' subjective well-being. All in all, the study provides detailed pictures of rural grandparents' daily lives and grandparenting experiences. However, it is not restricted to the scope of individuals and families. Instead, it examines grandparenting from its causes to its impacts, discussing its operation and function in the more extensive "society-family-individual" system. It emphasises the relationship between grandparenting and the transformation of the entire system, as well as the mutual feedback between grandparenting and the society. By doing so, the study adds a comprehensive description and explanation of why and how grandparenting occurs, provides a tool to categorise and better study grandparents, offers an answer to the contentious question of grandparenting's impacts, and discusses grandparenting in the bigger picture to show grandparents' struggles and contributions.

The impact of nucleic acids on scientific and medical progress has been enormous. DNA is the blueprint for structure and function from the level of individual cells up to whole organisms, and various forms of RNA are all involved in the regulation of genetic information. The identification of specific nucleic acid sequences and/or their level of expression provide key information for the molecular identity and organism functional state. This is very useful in areas like biomedicine. Nucleic acids are also versatile structural materials at the nanoscale. The precise recognition pattern of the Watson-Crick base pairs makes them not only successful as genetic materials, but also capable of directing the assembly of highly structured materials with unique nanoscale features. Nucleic acids, either alone or in combination with other materials, have been used to create a number of nanoscale structures and devices that perform actively in an engineered environment. In this thesis, research towards nucleic acids as a fundamental tool for both diagnostic purpose and structural applications is described. The thesis mainly consists of two studies. The first study is focused on nucleic acid-based biosensing using fluorescent quantum dots (QDs) as donors in a fluorescence resonance energy transfer (FRET) assay for addressing the analytical needs for DNA or RNA detection. The optimization of protocols for synthesizing QD-DNA constructs and their applications in biosensing assays are discussed. A highly sensitive and specific microRNA (miRNA) assay was then developed by the integration of the QD-DNA constructs and an isothermal enzyme-mediated target recycling step, with a detection limit of 42 fM and excellent selectivity for miR-148 versus base-mismatched sequences and other miRNAs. This proposed method was successfully employed for detection of miR-21 using an alternative FRET pair, which was compared to qRT-PCR for the quantitative analysis of miR-21 in biological samples. The second study is focused on construction and characterization of nucleic acid-based hierarchical porous nanostructures and presents a concept for exploiting such constructs as scaffolds for enzyme immobilization and activity studies. Based on rolling circle replication (RCR), DNA or RNA structures with flower-shaped morphologies were synthesized by interactions between inorganic magnesium pyrophosphate (Mg2PPi) crystals and DNA or RNA strands in a time-dependent manner. Focusing on RNA-based structures, various characterization techniques were applied to understand the composition and structure of the RNA particles, and different methods were taken for immobilizing protein or enzymes onto RNA particle. As a proof of principle study, β-galactosidase (β-gal) and horseradish peroxidase (HRP) enzymes were coupled to the RNA particles, and both exhibited enhanced enzymatic activity and improved stability in comparison to free enzymes. This RNA-based biomaterial provides a model to develop a wide range of biocatalysts and offers the promise of potent protein loading and delivery system for biomedical applications.

In this study, ZnO, MgO, In2O3, SnO2, iron oxide nanowires and nanorods were successfully produced by using Pulsed Laser Deposition. It was observed that the parameters, such as gold nanodots, number of pulses on the targets, temperature, oxygen pressure and substrate, contribute to various results. Different metal catalyst or substrate results in different morphology. Controlling oxygen pressure and substrate temperature was found to be crucial for the ZnO, In2O3 and iron oxide nanowires growth using this technique. The morphology of ZnO nanowires is highly influenced by plasma plume. Increasing the shots on gold target would increase the width of MgO nanorods, but by increasing the number of gold pulses, the distribution of the sizes and dimensions of ZnO nanorods become smaller. However, in the case of In2O3, retaining the total number of gold shots between 5 shots and 25 shots is critical. The variation of the number of laser pulses on oxide target mainly affects the length of nanowires in the case of MgO. However, in regards to In2O3, it does affect both the width and length of the nanowires.

The development of a universal influenza vaccine is an ideal strategy to eliminate public health threats from influenza epidemics and pandemics. Influenza conserved structures are outstanding immunogens for such a vaccine. Layered protein nanoparticles composed of well-designed conserved influenza structures have improved immunogenicity with new physical and biochemical features. We constructed and purified structure-stabilized influenza matrix protein 2 ectodomain (M2e) and M2e-neuraminidase fusion proteins (M2e-NAs). We produced double-layered M2e-NA protein nanoparticles by chemical crosslinking M2e-NA onto M2e nanoparticle surfaces. Immunizations with these protein nanoparticles induced immune protection against influenza viruses with homologous and heterosubtypic NA in mice. Double-layered M2e-NA protein nanoparticles induced higher levels of humoral and cellular responses compared with their comprising protein mixture or M2e nanoparticles. In addition, the layered M2e-NA protein nanoparticle group demonstrated strong cytotoxic T cell responses. We also found that antibody responses contributed to heterosubtypic influenza viral immune protection. The protective immunity was long-lasting. To favorite rapid distribution and convenient administration of a universal influenza vaccine, we incorporated the FliC-adjuvanted M2e-NA nanoparticles into logistically simple microneedle patches (MNPs) for skin vaccination in mice. The nanoparticle MNPs significantly increased M2e and NA-specific antibody levels, the numbers of germinal center B cells, and IL-4 secreting splenocytes. MNPs incorporating the double-layered nanoparticles protected mice against homologous and heterosubtypic influenza viruses. Our results demonstrated that M2e-NA nanoparticles could be developed into a standalone or synergistic component of a universal influenza vaccine strategy. MNP skin vaccination can logistically simplify the vaccines for extensive distribution and administration.

With the advancement of next-generation sequencing technology, numerous disease/ phenotypic associations with the human microbiome and human genome are uncovered and revealed. In this dissertation, we take advantage of this and explore these association patterns using machine learning methods. We first design a deep learning method MDeep for microbiome-based prediction by considering both the taxon abundance and phylogenetic tree. MDeep models the taxonomic rank by the convolutional layers and captures the phylogenetic correlation on each taxonomic rank via the convolutional operation. Our simulations and real data analysis demonstrate that MDeep outperforms competing methods in both regression and binary classifications. In order to explore the diseases/ phenotypic associations with the human genome, we propose two machine learning frameworks. The first framework, WEVar, is a supervised learning framework by integrating the pre-computed scores from representative existing scoring methods, which will benefit from each individual method by automatically learning the relative contribution of each method and produce an ensemble score for the final prediction. Using simulation and real data studies, we show both context-free WEVar and context-dependent WEVar outperform the individual scoring methods on the state-of-the-art benchmark datasets. Furthermore, we find WEVar can prioritize experimentally validated non-coding variants in an LD block. The second framework, DeepMFIVar, is a deep multimodal learning framework for the functional interpretation of genetic variants. DeepMFIVar learns a predictive model linking DNA sequence context and clinical information to quantitative epigenetic signals. The mutation effect of the 210 million genetic variants are generated by the difference of the predicted epigenetic signal for the reference and for alternative alleles. The application to DNA methylation and histone modification demonstrates that DeepMFIVar can accurately predict locus-specific epigenetic signals using DNA sequence and clinical information, and it is also capable of prioritizing variants for downstream experiments.

Children with Autism Spectrum Disorder (ASD) behave in different ways from other children as well as face enormous problems in their daily life. Previous research has found that physical treatment not only improves their body functions, but also benefits their mental development. Play therapy is a very effective way to help children with ASD since playing is the way children express themselves. However, after field research and benchmarking, I found that there are few toys helping them with their developmental coordination, which is necessary especially for children with ASD. In this paper, I have designed a series of devices for children with different levels of ASD to play with their different body parts to improve bimanual coordination, limb coordination and mental development. I developed and tested various colors, sizes, shapes, materials of mock-ups with children with ASD and tried to apply different kinds of feedbacks to the device. Specifically, bright colors are more attractive to children with ASD, but too bright colors should be avoided since some children with ASD have hypersensitivity with colors. Toys with a big size catch more of their attention than those with a small size despite of colors. Besides, combining materials with different softness could benefit children with ASD in therapy. Auditory elements have little attraction to children with ASD with normal sight unless it is very loud. By contrast, visual feedbacks such as light and shape change attract children with ASD more. I applied those findings to my design Bubbo-a toy designed for play therapy, which children with ASD could use to crawl on the floor or just touch and squeeze.

Whey proteins lactoferrin (Lf) and beta-lactoglobulin (β-Lg) from deer milk were isolated by Fast Protein Liquid Chromatography. Amino acids of Lf from deer milk were sequenced by LC-MS for the first time. The antibacterial activities of Lf, β-Lg, and their peptides were investigated in this research. Deer Lf slowed bacterial growth. Peptide Lfcin which was derived from deer Lf exhibited strong antibacterial activity against E.coli ATCC 25922 and L. acidophilus ATCC 4356. Deer β-Lg and its hydrolysates had strong antibacterial activity against S. aureus ATCC 25923.

The first chapter of my dissertation explores the roles “quants” or technically trained workers play at hedge funds. I quantify the impact of hiring “quants” on hedge fund strategy and risk taking by exploiting H-1B lottery results and a policy shock that significantly reduced the future supply of foreign high-skilled labor in the U.S. I find that the H-1B visa program allows hedge funds to pursue strategies that are more quantitative (e.g., hedging, systematic trading, etc.) as opposed to fundamentals-based (e.g., event driven). Although there is evidence of substitution between fund strategies, I also find that hedge funds overall become more diversified, both in terms of regional focus and fund investment style.Traditional theories on corporate financial policy mostly focus on firm-specific determinants and assume that capital structure choices are made independently of the actions of peer firms. The second chapter of my dissertation introduces a theoretical model that embodies the dynamic of how peer behavior affects a firm’s optimal choice of debt level. Specifically, peer effect is captured by a measure of how much the liquidation value of assets are adversely impacted if peer firms choose to liquidate simultaneously, the effect of which is more prominent for industries with highly specialized assets – hence low asset redeployability. The theoretical model predicts that a firm’s optional debt level is lower when the liquidation prices of its assets are expected to be more severely impacted by the concurrent liquidation of peer firms. Moreover, in those situations where peer firms’ cash flows are more closely correlated, the optional debt level will be higher. Individual firms’ attempts to maximize their own utility lead to an industry-wide over leverage, further exacerbating the risk of general crises in an already highly correlated industry.The third chapter of my dissertation explores the potential of using deep learning models to enhance equity trading strategies such as momentum and reversal trading. A deep feed forward neural network model (DFN) is fed with a training data set (1965-2000) that uses the rolling Z-scored cumulative returns of various horizons as predicative variables. With the model parameters generated by the training set, the model is applied to the validation set (2000-2016) to generate a signal for each stock predicting its likelihood of outperforming the cross-sectional medium. Based on this model-predicted signal, I have constructed an long-short investment portfolio out of the validation data set. The deep learning portfolios yield an annualized return of over 20 percent and generate significantly large alphas over commonly used factor models.

Technology scaling brings about the need for computationally efficient methods for circuit analysis, optimization, and synthesis. Convex optimization is a special class of mathematical optimization problems which minimize convex functions over convex sets. For general optimization problems, finding a local minimum is often easier than finding a global optimal. The inherent convexity ensures that a local minimum must be a global minimum so that convex optimization problems can be solved efficiently and reliably. These techniques are widely used in the EDA community because not only many problems can inherently be modeled as convex optimization problems, but also convex approximations work well even for non-convex problems. In particular, this dissertation explores several problems in VLSI design by applying recent convex optimization techniques. This dissertation first addresses the equation-based analog synthesis which finds design parameters that achieve optimal performance with a given circuit topology. The problem of analog synthesis in deep submicron technology is that convex modeling of circuit performance functions is not accurate enough while non-convex models are generally hard to solve. This dissertation proposes two techniques to solving this problem by leveraging the recent advances in semidefinite programming problem (SDP) relaxation of polynomial optimization. The first technique addresses the computational intractability of SDP relaxation in polynomial optimization formulation. Modeling analog performance functions via general polynomials allows for highly accurate fitting of SPICE data due to their inherent non-convexity and non-linearity. However, the computational complexity of solving the corresponding SDP relaxation scales exponentially with the degree of a polynomial. This dissertation studies fitting high-degree but sparse polynomials with special structure to enable efficient subsequent optimization phase. Results demonstrate that by handling higher-degree polynomials, the fitting error can be drastically reduced, which translates to a significant increase in the rate of constraint satisfaction. The second technique addresses the main challenge of using geometric programming (GP): the mismatch between what GP can accurately fit and the behavior of true analog performance functions. This approach proposes fitting device models as high-order monomials, defined as the exponential functions of polynomials in the logarithmic variables. The introduction of high-order monomials allows significant improvements in model accuracy. Using SDP-relaxations, the proposed strategy is able to obtain efficient near-optimal global solutions. Next, this dissertation focuses on computational improvements in power grid reduction. Analysis and simulation of power grids for large integrated circuits is computationally expensive due to the tremendous number of elements in power grids. The goal of power grid reduction is to produce a sparse approximation of the original grid with far fewer elements, that approximately preserves the voltage-current relations, so that the simulation time of the reduced model is greatly reduced. This dissertation models the power grid using the admittance matrix, known as the graph Laplacian, and proposes a convex optimization formulation of the Laplacian matrix approximation. In particular, this approach adopts the L1-norm regularization to enforce sparsity in the resulting power grid and controls the approximation quality only for currents within realistic ranges. This improves the previous methods which are overly conservative in terms of degree of sparsity. Next, this dissertation considers energy-efficiency in hardware implementations of machine learning applications using approximate computing. Unlike existing work at transistor level, gate level, or micro-architecture level, this dissertation exploits the inherent error-resilience of machine learning applications at the algorithmic level using the random projection technique, also known as sketching. Sketching of large matrices into their lower-dimensional representations can significantly reduce the cost of computation and storage and save the energy. This dissertation develops a highly energy-efficient hardware implementation of random projection by taking the hardware unique cost, e.g. latency, area and energy, into consideration. In particular, this dissertation demonstrates that randomness inherent in the projection matrices can be exploited to drastically reduce operand bit-width to achieve greater savings in the cost of hardware implementation. Finally, this dissertation develops a lossy compression algorithm for saving the expensive one-time-programmable memory in the context of SRAM physical unclonable functions (PUF). PUFs are ``silicon fingerprints'' that generate entropy from manufacturing process variation and are used in secure key generation and device authentication. SRAM PUFs generate unique randomness using the power-up values from the mismatch between cross-coupled inverters. However, the major challenge of using SRAM PUFs lies in the reliability of responses. Error-correction-codes (ECC) are required to be used in order to guarantee reliable reconstruction. Code syndrome of ECC needs to be stored, as helper data, in non-volatile memories. In particular, for original silicon authentication, helper data needs to be stored in one-time-programmable memory, which is limited and expensive. The proposed lossy compression algorithm allows a significant reduction of helper data size (HDS) for reliability mask by treating a fraction of reliable bits as unreliable. Based on the observation that cell reliability information is correlated across temperature, the proposed scheme utilizes the stochastic concentration theory to obtain an accurate confidence level for average bit-error-rate, which permits HDS reduction using only room-temperature characterization.

Text mining is involved in a broad scope of applications in diverse domains that mainly, but not exclusively, serve political, commercial, medical and academic needs. Along with the rapid development of the Internet technology in recent thirty years and the advent of online social media and network in a decade, text data is obliged to entail features of online social data streams, for example, the explosive growth, the constantly changing content and the huge volume. As a result, text mining is no longer merely oriented to textual content itself, but requires consideration of surroundings and combining theories and techniques of stream processing and social network analysis, which give birth to a wide range of applications used for understanding thoughts spread over the world , such as sentiment analysis, mass surveillance and market prediction. Automatically discovering sequences of words that represent appropriate themes in a collection of documents, topic detection closely associated with document clustering and classification. These two tasks play integral roles in revealing deep insight into the text content in the whole text mining framework. However, most existing detection techniques cannot adapt to the dynamic social context. This shows bottlenecks of detecting performance and deficiencies of topic models. In this thesis, we take aim at text data stream, investigating novel techniques and solutions for robust text mining to tackle arising challenges associated with the online social context by incorporating methodologies of stream processing, topic detection and document clustering and classification. In particular, we have advanced the state-of-theart by making the following contributions: 1. A Multi-Window based Ensemble Learning (MWEL) framework is proposed for imbalanced streaming data that comprehensively improves the classification performance. MWEL ensures that the ensemble classifier is maintained up to date and adaptive to the evolving data distribution by applying a multi-window monitoring mechanism and efficient updating strategy. 2. A semi-supervised learning method is proposed to detect latent topics from news streams and the corresponding social context with a constraint propagation scheme to adequately exploit the hidden geometrical structure as supervised information in given data space. A collective learning algorithm is proposed to integrate the textual content into the social context. A locally weighted scheme is afterwards proposed to seek an improvement of the algorithm stability. 3. A Robust Hierarchical Ensemble (RHE) framework is introduced to enhance the robustness of the topic model. It, on the one hand, reduces repercussions caused by outliers and noises, and on the other overcomes inherent defects of text data. RHE adapts to the changing distribution of text stream by constructing a flexible document hierarchy which can be dynamically adjusted. A discussion of how to extract the most valuable social context is conducted with experiments for the purpose of removing some noises from the surroundings and efficiency of the proposed.

This dissertation examines the determinants of competition and consumer access in the health care market, and supply- and demand-side determinants of health care use under the Affordable Care Act (ACA). The first essay studies insurer entry into the federally-facilitated health insurance market under the ACA. Motivated by the fact that insurers’ service areas can be subsets of rating areas, and the substantial variation in plan composition within a rating area, I explore variations in the type of plans offered and insurers’ decisions to enter a rating area. I find that availability of medical providers, population size, and metropolitan status are important in insurers’ decisions to enter a rating area. Medical cost affects the entry of restricted network plans. The second essay examines how supply-side incentives affect treatment choice for depression. Using claims data from Florida’s Medicaid program, I find large variations in initiating antidepressant treatment among newly diagnosed patients with three plan types: Fee-for-Service (FFS), Primary Care Case Management (PCCMs) and Accountable Care Organizations (ACOs). Compared to FFS, PCCMs and ACOs are more likely to provide antidepressant but no office-based care. I use the control function approach to mitigate the self-selection bias and find that ACOs tend to use lower cost medication options. Despite the use of low-cost alternatives for ACOs, no differences are found in subsequent psychiatric hospitalization or emergency room visits among plans. Different provider contractual relationships may partially explain treatment choice differences. The third essay investigates whether the ACA policy of free preventive services affects utilization of preventive care. I use variation in commercially-insured enrollees to examine the demand and supply prices of four preventive services. Despite an average 53 percentage point decrease in demand prices for these services, the actual service use only increased by 17 percent from 2007 to 2011, possibly due to little or no change in prices paid to providers. Using risk adjustment tools to predict and control for patient underlying health status, I find similar changes in demand prices and rates of service use across six health plan types, consistent with preventive visits being provider rather than consumer choices.

Outlier detection is a fundamental task that is used in numerous data analytic applications. It tackles the problem of identifying rare or atypical points that widely diverge from the general behavior or model of the data. The process of detecting outliers and subsequently using them for data analysis relies on the underlying application. For example, outlier detection can be employed as a preprocessing step to clean the data set from erroneous measurements and noisy data points. On the other hand, it can also be used to isolate suspicious or interesting patterns in the data. Examples include fraud detection, customer relationship management, network intrusion, clinical diagnosis, and biological data analysis. Although many successful algorithms have been developed for outlier detection, several challenges have haunted researchers and practitioners for decades. The first one is limited algorithm scalability. Due to the fast evolution of World Wide Web, the collected data can easily reach terabyte- or even petabyte- scale. Most existing approaches, ranging from statistical methods to geometric methods, and from density-based approaches to information theory based approaches, suffer from limited scalability and do not work well on large scale data. The second one is to detect outliers in the irregular, dynamic semi-structured data such as trees and graphs. There have been some research on finding outliers from the graphs. What are the definitions for meaningful outliers in the graph context? How can we detect them accurately and efficiently? The third challenge is to build a unified and modular detection system which provides researchers a complete toolbox for outlier detection tasks. Our research aims at designing the next-generation outlier detection algorithms thattackle the above three challenges. To achieve better scalability, we have done an extensive empirical study on different optimization techniques for distance-based outlier detection. Also, we proposed an ranking scheme driven by the Locality Sensitive Hashing (LSH), which finds all outliers by only visiting a small portion of the data (10%). Find similar points of each point, or all pair similarity search, is the key operation for many distance-based, density-based and cluster-based outliers. We optimized this fundamental kernel in metric space on MapReduce platform, and scaled the algorithm to hundreds of machines and solved the inadequate memory issue. For semi-structured outlier detection, we first designed a clustering-based algorithm, and a generic clustering algorithm for sets/multisets, trees and graphs. We also studied a concrete detection application on the semi-structured knowledge base, and found more than one million anomalies. Finally, we integrated our work seamlessly into a detection framework , which accepts different types of data. Users also enjoy the freedom of choosing and comparing different algorithms.

Active Galactic Nuclei (AGNs) are the most mystic objects in the universe. They are usually very far away from our Galaxy, which means that they are ancient objects. They are also luminous and have unique features in their spectra. Studying AGNs helps understanding the early universe and the evolution of galaxies. This Dissertation aims to research the structure of AGNs and the cooling function in the AGNs environment. I first investigate what optical/ultraviolet spectroscopic features would be produced by Broad-line Region (BLR) clouds crossing our line of sight to the accretion disk, the source of the optical/UV continuum. This research, prompted by recent X-ray observations, suggests that single cloud has little effect on the optical/UV spectrum. However, an ensemble of clouds produces a strong distinctive feature between the Lyman limit and Lyα. The extent of these features indicates the line-of-sight covering factor of clouds and may explain the ubiquitous AGN spectral break around 1100Å. I next study, considering the physical parameters of AGNs, how the gas cooling function changes at high temperature (T > 104 K) over a wide range of density (nH < 1012 cm−3) and metallicity (Z < 30Z⊙). I find that both density and metallicity change the ionization status of the gas. I provide numerical cooling functions by describing the total cooling as a sum of four parts: that due to H&He, the heavy elements, electron-electron bremsstrahlung, and grains. Finally, I also provide a function giving the electron fraction, which can be used to convert the cooling function into a cooling rate. Last, I extend the cooling-function study to the seldom-explored low-temperature range (T < 104 K). For primordial gas, gas lacking elements heavier than B, I find that radiative attachment and Compton recoil are important cooling processes when the gas kinetic temperature is lower than the temperature of the cosmic microwave background. I also find that collisional de-excitation of HD and H2 is not important above 1000K unlike claims of previous studies. For the dust-free solar case, we identify water as the dominant coolant in high density-environments. We also analyze the parameter ranges where metal, metal molecules, or all molecules, dominate the total cooling. We provide the density, above which the metal or metal molecules become the dominant coolants, as a function of temperature and metallicity. For the ISM case, with dust and depleted abundances, we find that dust does not directly cool the gas. Rather, dust modifies he cooling by affecting the chemistral balance. Similar to the high-temperature case, I also provide numerical cooling data.

Abstract: Trans-11 vaccenic acid (VA) is the predominant trans fatty acid in dairy fat and is the major precursor to endogenous synthesis of cis9,trans11-conjugated linoleic acid (CLA) in humans and animals. Epidemiological studies have shown the positive association between trans fat intake and incidence of coronary heart disease. Nevertheless, CLA, categorized as a group of trans fatty acids, has been shown to possess anti-carcinogenic, hypolipidemic and anti-diabetic benefits in several animal models as well as certain human populations, possibly via activating peroxisome proliferator-activated receptor (PPAR) related metabolic pathways. The subsequent effort in enriching CLA in dairy products (e.g. butter) has led to a concomitant increase in VA, whose bioactivity and health implications were not fully appreciated. Interestingly, VA is the major natural trans fat found in the diet. Therefore, the objectives of this thesis were to assess the effect of dietary supplementation of synthetic VA on lipid metabolism especially during conditions of dyslipidemia and metabolic syndrome, and to delineate the intestinal and hepatic metabolic pathways potentially modulated by VA. The JCR:LA-cp rat model, when homozygous for the cp trait (cp/cp), develop leptin receptor deficiency which leads to symptoms of metabolic syndrome and pre-diabetes including obesity, insulin resistance, hepatic steatosis, hypertriglyceridemia and exacerbated production of hepatic very low-density lipoproteins and intestinal chylomicrons (CM). Gas chromatography analysis on nascent lymph shows that VA was effectively absorbed into the intestine. In addition, VA from natural source (i.e. beef fat) showed higher intestinal bioavailability compared to synthetic VA. Dietary supplementation of 1.0% (w/w) synthetic VA to JCR:LA-cp rats (but not lean healthy controls) demonstrated a profound reduction in plasma triglyceride, total cholesterol, low-density lipoprotein-cholesterol, non-esterified fatty acid and haptoglobin concentrations (51%, p

In IP over WDM networks, IP links are realized by subscribing to lightpath services from a WDM network. Survivability is a key issue in designing such networks. There are two kinds of Lightpath services available: protected and unprotected. The research in [3] mainly considers the cases in which the only constraint is that the IP network remains connected after any single fiber-link failure. The problem was formulated into a mixed integer linear programming (MILP) problem. This thesis considers the additional requirement of a minimum survivable network bandwidth, and incorporates this as a criterion into the network design. First a new MILP formulation is proposed which satisfies the minimum survivable network bandwidth requirements. In addition, the computational requirement for the MILP is measured in simulations. Simulations show that the MILP has reasonable computation costs for small to moderate size networks. These costs increase with the increase of survivable bandwidth requirements.

This is a metatheoretical work in literacy and deafness that synthesizes the various models and theories in mainstream literacy and analyzes their potential contributions to understanding and improving the literacy skills of students who are deaf or hard of hearing. Detailed instructional suggestions are provided for practice and further research as well as for metatheorizing in the field. Using Ritzer’s paradigm approach as the metatheorizing tool, this study confirms the persistent pattern in the field of literacy and deafness: many students who are deaf or hard of hearing do not read and write as well as their hearing counterparts. Furthermore, there has been no consensus on the reasons for these reading difficulties or on the appropriate instructional approaches to improve the reading levels of students who are deaf or hard of hearing. By discussing the common ground in the goal of literacy education, which is proposed in this study as literate thought, namely, the ability to access and utilize captured information, this study attempts to connect literate thought to a broad concept of literacy which includes all the information that could be captured. Thus, instead of focusing on only print literacy, the reconceptualized literacy validates print literacy, performance literacy, and caption literacy as roughly equal vehicles for the development of literate thought. Such an understanding of literacy will provide significant contributions to the field of education, especially for traditionally marginalized populations who cannot benefit sufficiently from the conventional reading world. This study provides two general directions for resolving the predicament of the limited literacy skills of students who are deaf or hard of hearing: 1) finding the source underlying the apparent difficulties and trying to settle the problem, and 2) seeking the ends of being literate and trying to find an alternative means to parallel the route of script literacy. Additionally, this study goes beyond literacy education and provides implications for students with disabilities with respect to accessing general educational curricula such as science and social studies. Accordingly, instructional suggestions in the areas of Visual Phonics, performance literacy, and the integration of performance literacy and inquiry-based science are presented.