Physical Sciences and Mathematics Commons^™

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 …

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 …

Adaptive Feature Engineering Modeling For Ultrasound Image Classification For Decision Support, Hatwib Mugasa

Doctoral Dissertations

Ultrasonography is considered a relatively safe option for the diagnosis of benign and malignant cancer lesions due to the low-energy sound waves used. However, the visual interpretation of the ultrasound images is time-consuming and usually has high false alerts due to speckle noise. Improved methods of collection image-based data have been proposed to reduce noise in the images; however, this has proved not to solve the problem due to the complex nature of images and the exponential growth of biomedical datasets. Secondly, the target class in real-world biomedical datasets, that is the focus of interest of a biopsy, is usually …

Feature Space Modeling For Accurate And Efficient Learning From Non-Stationary Data, Ayesha Akter

Doctoral Dissertations

A non-stationary dataset is one whose statistical properties such as the mean, variance, correlation, probability distribution, etc. change over a specific interval of time. On the contrary, a stationary dataset is one whose statistical properties remain constant over time. Apart from the volatile statistical properties, non-stationary data poses other challenges such as time and memory management due to the limitation of computational resources mostly caused by the recent advancements in data collection technologies which generate a variety of data at an alarming pace and volume. Additionally, when the collected data is complex, managing data complexity, emerging from its dimensionality and …

Multidimensional Feature Engineering For Post-Translational Modification Prediction Problems, Norman Mapes Jr.

Doctoral Dissertations

Protein sequence data has been produced at an astounding speed. This creates an opportunity to characterize these proteins for the treatment of illness. A crucial characterization of proteins is their post translational modifications (PTM). There are 20 amino acids coded by DNA after coding (translation) nearly every protein is modified at an amino acid level. We focus on three specific PTMs. First is the bonding formed between two cysteine amino acids, thus introducing a loop to the straight chain of a protein. Second, we predict which cysteines can generally be modified (oxidized). Finally, we predict which lysine amino acids are …

Spatiotemporal Subspace Feature Tracking By Mining Discriminatory Characteristics, Richard D. Appiah

Doctoral Dissertations

Recent advancements in data collection technologies have made it possible to collect heterogeneous data at complex levels of abstraction, and at an alarming pace and volume. Data mining, and most recently data science seek to discover hidden patterns and insights from these data by employing a variety of knowledge discovery techniques. At the core of these techniques is the selection and use of features, variables or properties upon which the data were acquired to facilitate effective data modeling. Selecting relevant features in data modeling is critical to ensure an overall model accuracy and optimal predictive performance of future effects. The …

Full Simulation For The Qweak Experiment At 1.16 And 0.877 Gev And Their Impact On Extracting The Pv Asymmetry In The N→Δ A Transition, Hend Abdullah Nuhait

Doctoral Dissertations

The Qweak project is seeking to find new physics beyond the Standard Model. It is aimed to measure the weak charge of the proton, which has never been measured, at 4% precision at low momentum transfer. The experiment is performed by scattering electrons from protons and exploiting parity violation in the weak interaction at low four-momentum transfer.

In this experiment, two measurements were considered: which are elastic and inelastic. The elastic is to measure the proton's weak charge. In addition, the inelastic asymmetry measurement, which will extract the low energy constant dΔ. That measurement works in the neutral current …

Motion-Capture-Based Hand Gesture Recognition For Computing And Control, Andrew Gardner

Doctoral Dissertations

This dissertation focuses on the study and development of algorithms that enable the analysis and recognition of hand gestures in a motion capture environment. Central to this work is the study of unlabeled point sets in a more abstract sense. Evaluations of proposed methods focus on examining their generalization to users not encountered during system training.

In an initial exploratory study, we compare various classification algorithms based upon multiple interpretations and feature transformations of point sets, including those based upon aggregate features (e.g. mean) and a pseudo-rasterization of the capture space. We find aggregate feature classifiers to be balanced across …

Anxiolytic Effects Of Propranolol And Diphenoxylate On Mice And Automated Stretch-Attend Posture Analysis, Kevin Scott Holly

Doctoral Dissertations

The prevention of social anxiety, performance anxiety, and social phobia via the combination of two generic drugs, diphenoxylate HC1 (opioid) plus atropine sulfate (anticholinergic) and propranolol HCl (beta blocker) was evaluated in mice through behavioral studies. A patent published on a September 8, 2011 by Benjamin D. Holly, US 2011/0218215 Al, prompted the research. The drug combination of diphenoxylate and atropine plus propranolol could be an immediate treatment for patients suffering from acute phobic and social anxiety disorders. Demonstrating the anxiolytic effects of the treatment on mice would validate a mouse model for neuroscientist to be used to detect the …

Direct And Inverse Scattering Problems For Domains With Multiple Corners, Jiang Yihong

Doctoral Dissertations

Direct and inverse scattering problems have wide applications in geographical exploration, radar, sonar, medical imaging and non-destructive testing. In many applications, the obstacles are not smooth. Corner singularity challenges the design of a forward solver. Also, the nonlinearity and ill-posedness of the inverse problem challenge the design of an efficient, robust and accurate imaging method.

This dissertation presents numerical methods for solving the direct and inverse scattering problems for domains with multiple corners. The acoustic wave is sent from the transducers, scattered by obstacles and received by the transducers. This forms the response matrix data. The goal for the direct …

Intrusion Detection System Of Industrial Control Networks Using Network Telemetry, Stanislav Ponomarev

Doctoral Dissertations

Industrial Control Systems (ICSs) are designed, implemented, and deployed in most major spheres of production, business, and entertainment. ICSs are commonly split into two subsystems - Programmable Logic Controllers (PLCs) and Supervisory Control And Data Acquisition (SCADA) systems - to achieve high safety, allow engineers to observe states of an ICS, and perform various configuration updates. Before wide adoption of the Internet, ICSs used "air-gap" security measures, where the ICS network was isolated from other networks, including the Internet, by a physical disconnect [1]. This level of security allowed ICS protocol designers to concentrate on the availability and safety of …

Sensitivity Of Mixed Models To Computational Algorithms Of Time Series Data, Gunaime Nevine

Doctoral Dissertations

Statistical analysis is influenced by implementation of the algorithms used to execute the computations associated with various statistical techniques. Over many years; very important criteria for model comparison has been studied and examined, and two algorithms on a single dataset have been performed numerous times. The goal of this research is not comparing two or more models on one dataset, but comparing models with numerical algorithms that have been used to solve them on the same dataset.

In this research, different models have been broadly applied in modeling and their contrasting which are affected by the numerical algorithms in different …

Dual Channel-Based Network Traffic Authentication, David Irakiza

Doctoral Dissertations

In a local network or the Internet in general, data that is transmitted between two computers (also known as network traffic or simply, traffic) in that network is usually classified as being of a malicious or of a benign nature by a traffic authentication system employing databases of previously observed malicious or benign traffic signatures, i.e., blacklists or whitelists, respectively. These lists typically consist of either the destinations (i.e., IP addresses or domain names) to which traffic is being sent or the statistical properties of the traffic, e.g., packet size, rate of connection establishment, etc. The drawback with the list-based …

Topology Dependence Of Ppm-Based Internet Protocol Traceback Schemes, Ankunda R. Kiremire

Doctoral Dissertations

Multiple schemes that utilize probabilistic packet marking (PPM) have been proposed to deal with Distributed Denial of Service (DDoS) attacks by reconstructing their attack graphs and identifying the attack sources.

In the first part of this dissertation, we present our contribution to the family of PPM-based schemes for Internet Protocol (IP) traceback. Our proposed approach, Prediction-Based Scheme (PBS), consists of marking and traceback algorithms that reduce scheme convergence times by dealing with the problems of data loss and incomplete attack graphs exhibited by previous PPM-based schemes.

Compared to previous PPM-based schemes, the PBS marking algorithm ensures that traceback is possible …

Modeling Profile-Attribute Disclosure In Online Social Networks From A Game Theoretic Perspective, Jundong Chen

Doctoral Dissertations

Privacy settings are a crucial part of any online social network as users are confronted with determining which and how many profile attributes to disclose. Revealing more attributes increases users chances of finding friends and yet leaves users more vulnerable to dangers such as identity theft. In this dissertation, we consider the problem of finding the optimal strategy for the disclosure of user attributes in social networks from a game-theoretic perspective.

We model the privacy settings' dynamics of social networks with three game-theoretic approaches. In a two-user game, each user selects an ideal number of attributes to disclose to each …

A Knowledge Discovery Approach For The Detection Of Power Grid State Variable Attacks, Nathan Wallace

Doctoral Dissertations

As the level of sophistication in power system technologies increases, the amount of system state parameters being recorded also increases. This data not only provides an opportunity for monitoring and diagnostics of a power system, but it also creates an environment wherein security can be maintained. Being able to extract relevant information from this pool of data is one of the key challenges still yet to be obtained in the smart grid. The potential exists for the creation of innovative power grid cybersecurity applications, which harness the information gained from advanced analytics. Such analytics can be based on the extraction …

Numerical Solutions For Problems With Complex Physics In Complex Geometry, Yifan Wang

Doctoral Dissertations

In this dissertation, two high order accurate numerical methods, Spectral Element Method (SEM) and Discontinuous Galerkin method (DG), are discussed and investigated. The advantages of both methods and their applicable areas are studied. Particular problems in complex geometry with complex physics are investigated and their high order accurate numerical solutions obtained by using either SEM or DG are presented. Furthermore, the Smoothed Particle Hydrodynamics (SPH) (a mesh-free weighted interpolation method) is implemented on graphics processing unit (GPU). Some numerical simulations of the complex flow with a free surface are presented and discussed to show the advantages of SPH method in …

Performance Modeling And Optimization Techniques For Heterogeneous Computing, Supada Laosooksathit

Doctoral Dissertations

Since Graphics Processing Units (CPUs) have increasingly gained popularity amoung non-graphic and computational applications, known as General-Purpose computation on GPU (GPGPU), CPUs have been deployed in many clusters, including the world's fastest supercomputer. However, to make the most efficiency from a GPU system, one should consider both performance and reliability of the system.

This dissertation makes four major contributions. First, the two-level checkpoint/restart protocol that aims to reduce the checkpoint and recovery costs with a latency hiding strategy in a system between a CPU (Central Processing Unit) and a GPU is proposed. The experimental results and analysis reveals some benefits, …

Vulnerability Analysis Of Cyber-Behavioral Biometric Authentication, Abdul Serwadda

Doctoral Dissertations

Research on cyber-behavioral biometric authentication has traditionally assumed naïve (or zero-effort) impostors who make no attempt to generate sophisticated forgeries of biometric samples. Given the plethora of adversarial technologies on the Internet, it is questionable as to whether the zero-effort threat model provides a realistic estimate of how these authentication systems would perform in the wake of adversity. To better evaluate the efficiency of these authentication systems, there is need for research on algorithmic attacks which simulate the state-of-the-art threats.

To tackle this problem, we took the case of keystroke and touch-based authentication and developed a new family of algorithmic …

Predicting Threat Potential Using Cyber Sensors, Mark Anthony Thompson

Doctoral Dissertations

The proliferation of the Internet has created a culture of a connected society dependent upon technology for communication and information sharing needs. In this dissertation, we hypothesize that attackers are increasingly using electronic resources that are capable of leaving a digital footprint, such as social media services, e-mail, text messages, blogs, and websites for the communication, planning, and coordination of attacks. In its current form, however, traffic analysis is primarily concerned with using communications volume to extract intelligence information, but largely ignores the content of communications transmissions that is needed to meet the security challenges and demands of continually emerging …

Nonlinear Granger Causality And Its Application In Decoding Of Human Reaching Intentions, Mengting Liu

Doctoral Dissertations

Multi-electrode recording is a key technology that allows the brain mechanisms of decision making, cognition, and their breakdown in diseases to be studied from a network perspective. As the hypotheses concerning the role of neural interactions in cognitive paradigms become increasingly more elaborate, the ability to evaluate the direction of neural interactions in neural networks holds the key to distinguishing their functional significance.

Granger Causality (GC) is used to detect the directional influence of signals between multiple locations. To extract the nonlinear directional flow, GC was completed through a nonlinear predictive approach using radial basis functions (RBF). Furthermore, to obtain …

Using Power-Law Properties Of Social Groups For Cloud Defense And Community Detection, Justin L. Rice

Doctoral Dissertations

The power-law distribution can be used to describe various aspects of social group behavior. For mussels, sociobiological research has shown that the Lévy walk best describes their self-organizing movement strategy. A mussel's step length is drawn from a power-law distribution, and its direction is drawn from a uniform distribution. In the area of social networks, theories such as preferential attachment seek to explain why the degree distribution tends to be scale-free. The aim of this dissertation is to glean insight from these works to help solve problems in two domains: cloud computing systems and community detection.

Privacy and security are …

Snoop-Forge-Replay Attack On Continuous Verification With Keystrokes, Khandaker Abir Rahman

Doctoral Dissertations

We present a new attack called the snoop-forge-replay attack on the keystroke-based continuous verification systems. We performed the attacks on two levels – 1) feature-level and 2) sample-level.

(1) Feature-level attack targets specific keystroke-based continuous verification method or system. In feature-level attacks, we performed a series of experiments using keystroke data from 50 users who typed approximately 1200 to 2300 keystrokes of free text during three different periods. The experiments consisted of two parts. In the first part, we conducted zero-effort verification experiments with two verifiers ("R" and "S") and obtained Equal Error Rates (EERs) between 10% and 15% under …

Adaptive Grid Based Localized Learning For Multidimensional Data, Sheetal Saini

Doctoral Dissertations

Rapid advances in data-rich domains of science, technology, and business has amplified the computational challenges of "Big Data" synthesis necessary to slow the widening gap between the rate at which the data is being collected and analyzed for knowledge. This has led to the renewed need for efficient and accurate algorithms, framework, and algorithmic mechanisms essential for knowledge discovery, especially in the domains of clustering, classification, dimensionality reduction, feature ranking, and feature selection. However, data mining algorithms are frequently challenged by the sparseness due to the high dimensionality of the datasets in such domains which is particularly detrimental to the …

Failure Prediction For High-Performance Computing Systems, Narate Taerat

Doctoral Dissertations

The failure rate in high-performance computing (HPC) systems continues to escalate as the number of components in these systems increases. This affects the scalability and the performance of parallel applications in large-scale HPC systems. Fault tolerance (FT) mechanisms help mitigating the impact of failures on parallel applications. However, utilizing such mechanisms requires additional overhead. Besides, the overuse of FT mechanisms results in unnecessarily large overhead in the parallel applications. Knowing when and where failures will occur can greatly reduce the excessive overhead. As such, failure prediction is critical in order to effectively utilize FT mechanisms. In addition, it also helps …

Near-Optimal Scheduling And Decision-Making Models For Reactive And Proactive Fault Tolerance Mechanisms, Nichamon Naksinehaboon

Doctoral Dissertations

As High Performance Computing (HPC) systems increase in size to fulfill computational power demand, the chance of failure occurrences dramatically increases, resulting in potentially large amounts of lost computing time. Fault Tolerance (FT) mechanisms aim to mitigate the impact of failure occurrences to the running applications. However, the overhead of FT mechanisms increases proportionally to the HPC systems' size. Therefore, challenges arise in handling the expensive overhead of FT mechanisms while minimizing the large amount of lost computing time due to failure occurrences.

In this dissertation, a near-optimal scheduling model is built to determine when to invoke a hybrid checkpoint …

Improving The Accuracy Of The Generalized Fdtd-Q Scheme For Solving The Linear Time-Dependent Schrödinger Equation, James John Elliot Iii

Doctoral Dissertations

This dissertation improves the accuracy of the Generalized Finite Difference Time Domain (FDTD) scheme by determining a differential operator that is capable of achieving reasonable accuracy when used to obtain even-order derivatives up to order fourteen. The Generalized FDTD scheme is an explicit, scheme used to solve the time-dependent Schrödinger equation, and being an explicit scheme, it must utilize a carefully devised ratio of the temporal step to the spatial step to maintain numerical stability. This ratio is called the mesh ratio, and the Generalized FDTD scheme allows this ratio to be significantly relaxed. As the mesh ratio increases, the …

Shape Reconstruction And Classification Using The Response Matrix, Wei Wang

Doctoral Dissertations

This dissertation presents a novel method for the inverse scattering problem for extended target. The acoustic or electromagnetic wave is scattered by the target and received by all the transducers around the target. The scattered field on all the transducers forms the response matrix which contains the information of the geometry of the target. The objective of the inverse scattering problem is to reconstruct the shape of the scatter using the Response Matrix.

There are two types of numerical methods for solving the inverse problem: the direct imaging method and the iterative method. Two direct imaging methods, MUSIC method and …

A Modeling And Simulation Framework For Electrokinetic Nanoparticle Treatment, James Phillips

Doctoral Dissertations

The focus of this research is to model and provide a simulation framework for the packing of differently sized spheres within a hard boundary. The novel contributions of this dissertation are the cylinders of influence (COI) method and sectoring method implementations. The impetus for this research stems from modeling electrokinetic nanoparticle (EN) treatment, which packs concrete pores with differently sized nanoparticles. We show an improved speed of the simulation compared to previously published results of EN treatment simulation while obtaining similar porosity reduction results. We mainly focused on readily, commercially available particle sizes of 2 nm and 20 nm particles, …