Digital Commons Network^™

Performance Modeling For Cloud Platforms: A Data Driven Approach, Kewen Wang

Doctoral Dissertations

Software service providers are increasingly adopting cloud-based solutions to maximize resource utilization while minimizing operating cost. While performance predictability is becoming of paramount importance as the safety-critical nature of such systems continues to grow (e.g., IoT applications, infrastructure monitoring), however, large scale, high-degree of concurrency, and dynamic allocation of resources are making traditional performance modeling/tuning frameworks ill-suited that are not extendable. To address the aforementioned challenge, this thesis focuses on developing a data-driven performance modeling framework. Towards this objective, first, hierarchical performance models that can effectively capture and predict the execution time of a given job with high accuracy based …

Deep Learning Systems For Automated Lesion Detection, Segmentation, And Classification In Mammography, Dina Abdelhafiz

Doctoral Dissertations

Archival abstract submitted

Algorithms For Mitochondrial Genome Assembly And Haplogroup Assignment From Low-Coverage Whole-Genome Sequencing Data, Fahad Alqahtani

Doctoral Dissertations

Mitochondria are cellular organelles present with very rare exceptions in all eukaryotic cells. In most animals, the mitochondria have their own genome. The small size, high copy number, and the presence of both coding and regulatory regions that mutate at different rates make the mitochondrial genome an ideal genetic marker. Indeed, mitochondrial sequences have been used in applications ranging from maternal ancestry inference and tracing human migrations to forensic analysis. This thesis presents several novel bioinformatic tools enabling highly accurate mitochondrial genome reconstruction from low coverage from WGS data. First, we describe the Statistical Mitogenome Assembly with Repeats (SMART) pipeline …

Clinical Detection Of Diagnostic Biomarker Panels Using Microfluidic Electrochemical Immunoarrays, Abby Jones

Doctoral Dissertations

Cancer is a worldwide infliction. Cancer does not discriminate. Cancer does not care if you are young or old, rich or poor, disabled or in the prime of your life. It can be caused by genetics, environmental factors and/or lifestyle. The challenge, how do we diagnose and treat such a dynamic disease? Cancer detection is expensive, invasive, inaccurate and lacks sensitivity. New methods that rely on measurements of analytes in solution for detection and quantification are promising alternatives. Panels of protein biomarkers may aid in personalized diagnosis as protein levels in patients are often upregulated or down regulated in relation …

Correlative Framework Of Techniques For The Inspection, Evaluation, And Design Of Micro-Electronic Devices, Joseph Favata

Doctoral Dissertations

Trillions of micro- and nano-electronic devices are manufactured every year. They service countless electronic systems across a diverse range of applications ranging from civilian, military, and medical sectors. Examples of these devices include: packaged and board-mounted semiconductor devices such as ceramic capacitors, CPUs, GPUs, DSPs, etc., biomedical implantable electrochemical devices such as pacemakers, defibrillators, and neural stimulators, electromechanical sensors such as MEMS/NEMS accelerometers and positioning systems and many others. Though a diverse collection of devices, they are unified by their length scale. Particularly, with respect to the ever-present objectives of device miniaturization and performance improvement. Pressures to meet these objectives …

Model-Based Efficiency Analysis Of Power Plants With Carbon Footprint Constraints, Chen Chen

Doctoral Dissertations

Archival abstract submitted

Liquid Phase Adsorptive Desulfurization Using Modified Zeolites For Transportation And Fuel Cell Applications, Kevin Lee

Doctoral Dissertations

Sulfur removal from transportation fuels is essential for maintaining a pollutant-free environment and ensuring a healthy life. Adsorptive desulfurization using zeolites is an attractive desulfurization method, because of its low energy and cost requirements compared with conventional hydrodesulfurization. However, diffusion limitations within the micropores of zeolites can reduce their adsorption capacity, especially when refractory sulfur compounds are present. Moreover, the coexistence of aromatics in fuels exacerbates selective adsorption of sulfur. These challenges may be alleviated through careful physical and chemical modification, without compromising important structural properties of the zeolite. The introduction of mesoporosity via the surfactant-assisted method creates well-ordered mesopores …

Replicative Bypass Of A Dna-Peptide Cross-Link And Oxidative Single And Tandem Dna Lesions In Human Cells, Spandana Naldiga

Doctoral Dissertations

DNA harbors the genetic information in all living organisms and plays a pivotal role in passing the genetic information to the next generation. All living cells are exposed to numerous exogenous and endogenous agents that damage the DNA to form different lesions or adducts in DNA. Fortuitously, diverse DNA repair mechanisms exist in the living cells, which repair these DNA adducts before the replication apparatus of the cell copies the DNA. However, frequently extensive DNA damage or slow repair may allow replication bypass of a DNA lesion. As a result, cells adapt a DNA damage tolerance mechanism, termed as translesion …

Using Strontium Coated Clay Nanoparticles For Bone Regeneration And Other Biomedical Applications, Anusha Elumalai

Doctoral Dissertations

The aim of this project is to coat halloysite nanotubes (HNTs) with strontium in an ecofriendly, simple, and non-expensive process. These particles, when doped in calcium phosphate cements (CPC), are predicted to increase the osteoconductive and antibacterial properties of three dimensional (3D) printed bone scaffolds.

The purpose of the 3D printed bone scaffolds is to assume the same function as the bones they are replacing but with several additional functionalities. These biomaterials will have the ability to be resorbed as new bone is formed. Due to inherent osteogenic factors and antibiotics released from doped HNTs during the reparative process, it …

Social Media Based Algorithmic Clinical Decision Support Learning From Behavioral Predispositions, Radhika V. Medury

Doctoral Dissertations

Behavioral disorders are disabilities characterized by an individual’s mood, thinking, and social interactions. The commonality of behavioral disorders amongst the United States population has increased in the last few years, with an estimated 50% of all Americans diagnosed with a behavioral disorder at some point in their lifetime. AttentionDeficit/Hyperactivity Disorder is one such behavioral disorder that is a severe public health concern because of its high prevalence, incurable nature, significant impact on domestic life, and peer relationships. Symptomatically, in theory, ADHD is characterized by inattention, hyperactivity, and impulsivity. Access to providers who can offer diagnosis and treat the disorder varies …

Nanoengineered Core-Shell Structures Using Tubile Halloysite Clay, Yusuf A. Darrat

Doctoral Dissertations

Halloysite nanotubes are a versatile nanomaterial that can be used in a wide variety of applications. They have a unique structure which could be described as a flat material that consists of silica on one side and alumina on the other; this structure is rolled up in a way naturally forming an internal 10-15 nm lumen and interlayer spacing. This could lead to many potential applications for example incorporating halloysite as a template material or as a support structure. They are an inexpensive clay material that is available in large quantities (thousands of tons), so they may be practically used …

Predicting Complex System Behavior Using Hybrid Modeling And Computational Intelligence, Vinayaka Gude

Doctoral Dissertations

“Modeling and prediction of complex systems is a challenging problem due to the sub-system interactions and dependencies. This research examines combining various computational intelligence algorithms and modeling techniques to provide insights into these complex processes and allow for better decision making. This hybrid methodology provided additional capabilities to analyze and predict the overall system behavior where a single model cannot be used to understand the complex problem. The systems analyzed here are flooding events and fetal health care. The impact of floods on road infrastructure is investigated using graph theory, agent-based traffic simulation, and Long Short-Term Memory deep learning to …

Human Behavior Understanding For Worker-Centered Intelligent Manufacturing, Wenjin Tao

Doctoral Dissertations

“In a worker-centered intelligent manufacturing system, sensing and understanding of the worker’s behavior are the primary tasks, which are essential for automatic performance evaluation & optimization, intelligent training & assistance, and human-robot collaboration. In this study, a worker-centered training & assistant system is proposed for intelligent manufacturing, which is featured with self-awareness and active-guidance. To understand the hand behavior, a method is proposed for complex hand gesture recognition using Convolutional Neural Networks (CNN) with multiview augmentation and inference fusion, from depth images captured by Microsoft Kinect. To sense and understand the worker in a more comprehensive way, a multi-modal approach …

Bond Behavior Of Advanced Fiber Reinforced Composite-Concrete Joints, Xingxing Zou

Doctoral Dissertations

“Externally bonding advanced composite materials to concrete structures is an effective way to improve their strength, ductility, and durability. The interfacial bond behavior is fundamental to understand the overall structural performance of concrete structures strengthened with advanced composite materials. This study includes a comprehensive investigation of the bond behavior of composite-concrete joints with different fiber reinforced composite types. First, a direct approach to determine the bond-slip relationship for fiber reinforced cementitious matrix (FRCM)-concrete joints based on fiber strain measurements was proposed. Then, an analytical solution to predict the full-range response of FRCM-concrete joints was derived by assuming a trilinear bond-slip …

Development Of Lightweight Materials By Meso- And Microstructure Control, Myranda Shea Spratt

Doctoral Dissertations

”In this work, two lightweight structures – lattice structures and metal matrix syntactic foams (MMSF) – were studied. Honeycomb lattices were manufactured by powder bed selective laser melting (SLM) from 304L stainless steel. The wall thicknesses of these structures ranged from 0.2 to 0.5 mm. Surface roughness was the primary cause of dimensional mismatch between the expected and as-built structures with an average wall thickness increase of 0.12 mm. The strength of the honeycombs increased with increasing wall thickness. A feature of the SLM microstructure, the melt pool boundary, was also studied as a part of this work. 3D models …

Poisoning Attacks On Learning-Based Keystroke Authentication And A Residue Feature Based Defense, Zibo Wang

Doctoral Dissertations

Behavioral biometrics, such as keystroke dynamics, are characterized by relatively large variation in the input samples as compared to physiological biometrics such as fingerprints and iris. Recent advances in machine learning have resulted in behaviorbased pattern learning methods that obviate the effects of variation by mapping the variable behavior patterns to a unique identity with high accuracy. However, it has also exposed the learning systems to attacks that use updating mechanisms in learning by injecting imposter samples to deliberately drift the data to impostors’ patterns. Using the principles of adversarial drift, we develop a class of poisoning attacks, named Frog-Boiling …

A Framework Of Multi-Dimensional And Multi-Scale Modeling With Applications, Zilong Li

Doctoral Dissertations

In this dissertation, a framework for multi-dimensional and multi-scale modeling is proposed. The essential idea is based on oriented space curves, which can be represented as a 3D slender object or 1D step parameters. SMILES and Masks provide functionalities that extend slender objects into branched and other objects. We treat the conversion between 1D, 2D, 3D, and 4D representations as data unification. A mathematical analysis of different methods applied to helices (a special type of space curves) is also provided. Computational implementation utilizes Model-ViewController design principles to integrate data unification with graphical visualizations to create a dashboard. Applications of multi-dimensional …

Acoustic Confinement And Characterization Of A Microwave Plasma, Seth Lee Pree '09

Doctoral Dissertations

High amplitude acoustic fields are used to confine, characterize, and manipulate collisional plasmas with temperatures of a few thousand Kelvin. This dissertation describes the theory, experimental techniques, and apparatus necessary both to generate high amplitude sound in a few thousand Kelvin plasma and to use that sound field to manipulate the plasma within a resonant acoustic cavity. The acoustic field in a spherically symmetric oscillating plasma has been measured to have a Mach number of .03, which is sufficient to cause acoustic radiation pressure effects to confine the plasma to the center of its container. This field also generates convection …

Isospin Dependence Of Nucleon-Nucleon Short-Range Correlations In Inclusive Scattering With Tritium And Helium-3, Shujie Li

Doctoral Dissertations

The nucleon-nucleon (NN) potential has a strong repulsive core. When a two-nucleon (sub)system falls into this range, they will interact strongly at short distance, and move away from each other with momenta above the Fermi level. This is called the NN Short-range Correlations (SRCs).

Previous experiments reported a neutron-proton pair (isosinglet) dominance in high-momentum nucleons. In inclusive electron scattering, this np dominance will cause a scaling behavior of cross sections at 1.4

Tuning Surface Properties Of Titanium Dioxide For Carbon Monoxide Oxidation Over Copper-Titanium Dioxide, Guoqiang Cao

Doctoral Dissertations

Besides being widely used as a probe reaction, carbon monoxide (CO) oxidation has been playing a crucial role in multiple industrial applications, including automotive emission elimination, chemical production, and abatement of toxic waste. Copper (Cu) based catalysts have been extensively investigated for CO oxidation, because of its low cost and comparable activity to precious metal-based catalysts. Meanwhile, the surface properties, such as the oxidation state of copper species and oxygen species, depend on the synthesis process. Surface properties would eventually impact the catalytic performance. Therefore, it is desirable to build the property-activity relationship to understand factors that can affect the …

Computational Studies Of Small Molecule Transport And Binding, Yong Liu

Doctoral Dissertations

In this thesis, I have studied two classes of systems using molecular dynamics (MD) simulations as a primary approach. The first class of systems is responsible for transport of small molecules. Specically, I have investigated diffusion pathways of inhibitory gas CO within the enzyme FeFe-hydrogenase as well as water transport in articial water channels. The second class of systems are biological proteins and their interactions with ligands. I have focused on Regulators of G-protein Signaling (RGS) proteins and their binding to thiadiazolidinone (TDZD) based small-molecule inhibitors as well as phosphosdiesterase 6 (PDE6) binding with cyclic guanosine monophosphate (cGMP).

Specically, I …

Manipulating Single Polymer Molecules For Applications In Nanomaterials, Ruiwen Chen

Doctoral Dissertations

Polymeric nanoparticles have been utilized in an increasing number of fields over the past two decades due to their unique properties such as design flexibility and good biocompatibility. Despite various techniques available to produce polymer nanoparticles, the preparation of small nanoparticles with customized functions in the sub 20 nm dimension remains challenging. Inspired by the self-organizing behavior of natural biomacromolecules, a class of single-chain nanoparticles (SCNP) are synthesized featuring biomimicry and ultrafine size. These nanoparticles are prepared from self-folding of polymer precursors bearing reactive pendant groups by intramolecular cross-linking reactions. A variety of cross-linking chemistries are available including covalent, dynamic …

Experimental And Numerical Failure Analysis Of Deep Steel Column Sections, Shokoufeh Zargar Zargar Shoushtari

Doctoral Dissertations

The availability of reliable numerical models is essential to reduce the uncertainties present in the prediction of structural behavior. Experimental studies allow the calibration and development of numerical models capable of characterizing the realistic behavior of structural elements and components until the limit state of collapse is approached. Exterior columns in perimeter steel moment-resisting frame structures that are exposed to strong earthquakes experience bending moment demands with high levels of axial load due to overturning. Deep wide flange sections can be used as exterior columns to increase the lateral stiffness of moment frames without significantly increasing the overall weight of …

Future Hydrological Failure Probability Of Dams In New England Under Land Use And Climate Change, Iman Hosseini-Shakib

Doctoral Dissertations

Floods lead to the overtopping of dams which is the main cause of dam failures and can result in significant loss of lives and property. This study investigates how the hydrological failure probability of dams in New England may change with future changes in climate and land use. Non-stationarity of future precipitation caused by the anthropogenic climate change and altered watershed concentration times caused by anthropogenic alterations such as urbanization, industrialization or deforestation can impact the mechanisms of runoff production and transfer. This can potentially change the frequency, magnitude, or duration of floods. Therefore, due to different flood patterns and …

A Broad Iceland Plume Associated With Two Phase Transitions At The 660 Km Discontinuity: Constraints From Receiver Functions, Dan Wang

Doctoral Dissertations

"In spite of the fact that Iceland is frequently regarded as the archetypal example of mantle plumes, the existence, depth extent, origin, dimension and excess temperature of the hypothesized plume remain enigmatic and hotly debated. The controversy mostly originates from the limited vertical resolution of seismic tomography techniques and the associated uncertainty in the depth and lateral extents of the lower wavespeed anomaly. Here we utilize a robust receiver-function-based technique to image the topography of the 410 and 660 km discontinuities bordering the mantle transition zone beneath Iceland and surrounding oceanic regions, and construct thermal and seismic wavespeed models of …

Seismological Investigations Of Lithospheric Structure Beneath The East African Rift System And The Kalahari Craton, Tuo Wang

Doctoral Dissertations

"In spite of numerous geophysical studies have been conducted in East African Rift System and Southern Africa, the structure and evolution of the crust and upper mantle are still controversial. In this study, we applied multiple seismological techniques to investigate the structure in the crust and uppermost mantle beneath the East African Rift System and Kalahari Craton in southern Africa. First, the technique of ambient noise tomography was applied to image shear wave velocity model from surface to the depth of 40 km beneath the Malawi and Luangwa rift zones. For the first time, a high-resolution 3-D shear wave velocity …

Development Of A System Architecture For The Prediction Of Student Success Using Machine Learning Techniques, Tatiana A. Cardona

Doctoral Dissertations

“ The goals of higher education have evolved through time based on the impact that technology development and industry have on productivity. Nowadays, jobs demand increased technical skills, and the supply of prepared personnel to assume those jobs is insufficient. The system of higher education needs to evaluate their practices to realize the potential of cultivating an educated and technically skilled workforce. Currently, completion rates at universities are too low to accomplish the aim of closing the workforce gap. Recent reports indicate that 40 percent of freshman at four-year public colleges will not graduate, and rates of completion are even …

Effect Of Turbulence Model Closure Coefficient Uncertainty On Scramjet Flow Field Analysis, Martin Albert Di Stefano

Doctoral Dissertations

“The numerical modeling of supersonic combustion ramjet (scramjet) engine flow paths plays an important role in the design of hypersonic air-breathing propulsion systems. Due to the complexity of the flow field physics and limited experimental data, numerical models possess uncertainties which should be addressed to improve the prediction accuracy of the simulations. In this study, the effect of turbulence model closure coefficient uncertainty on the Reynolds-averaged Navier-Stokes solution of a scramjet isolator and scramjet strut fuel injector flow field is investigated with an uncertainty quantification and sensitivity analysis study for commonly used turbulence models. Turbulence models considered in this work …

Improved Attenuation And Crosstalk Modeling Techniques For High-Speed Channels, Shaohui Yong

Doctoral Dissertations

”As digital systems are moving in the direction of faster data transmission rate and higher density of circuits, the problem of the far-end crosstalk (FEXT) and frequency-dependent attenuation are becoming the major factors that limit signal integrity performance. This research is focusing on providing several more comprehensive and accurate modeling approaches for striplines on fabricated printed circuit board (PCB). By characterizing the dielectric permittivity of prepreg and core, dielectric loss tangent, and copper foil surface roughness using measurement data, a better agreement between the stripline model and measurement is achieved. First, a method is proposed to extract dielectric loss tangent …

A Framework For A Successful Additive Repair System, Todd E. Sparks

Doctoral Dissertations

“The goal of this research is to generate a revolutionary improvement to the usability and usefulness of additive repair technology by integrating a set of tools into a seamless work flow. Insufficient automation in the current repair process is a huge hurdle in achieving cost-effective, reliable repairs. Many opportunities have been missed due to inconsistency, quality issues and lack of robustness and flexibility. The present work addresses deficiencies in preparatory steps such as 2D and 3D geometry processing, parameter estimation, and path planning as well as on-machine execution of the path plan. The bulk of the effort is focused on …