Physical Sciences and Mathematics Commons^™

Implementation Of A New Sigmoid Function In Backpropagation Neural Networks., Jeffrey A. Bonnell

Electronic Theses and Dissertations

This thesis presents the use of a new sigmoid activation function in backpropagation artificial neural networks (ANNs). ANNs using conventional activation functions may generalize poorly when trained on a set which includes quirky, mislabeled, unbalanced, or otherwise complicated data. This new activation function is an attempt to improve generalization and reduce overtraining on mislabeled or irrelevant data by restricting training when inputs to the hidden neurons are sufficiently small. This activation function includes a flattened, low-training region which grows or shrinks during back-propagation to ensure a desired proportion of inputs inside the low-training region. With a desired low-training proportion of …

A Case Study Of The Effects Of A Web Interface Redesign Based On Usability Guidelines., Paromita Bhattacharya

Electronic Theses and Dissertations

Usability guidelines for user interface (UI) design formulated by Nielsen and other authorities were used to redesign the web interface for a web based distance education course. The website's original UI, as judged by a group of web design professionals, violated standard guidelines for usability, readability, learnability, consistency, efficiency of use and flexibility. The original and redesigned UIs were tested using thinking aloud testing, a user satisfaction survey, and usability inspection. The results of user testing and heuristic evaluation suggest that the use of these guidelines in website redesign had a positive impact on user performance. A more conclusive result …

Improving Brain-Computer Interface Performance: Giving The P300 Speller Some Color., David B. Ryan

Electronic Theses and Dissertations

Individuals who suffer from severe motor disabilities face the possibility of the loss of speech. A Brain-Computer Interface (BCI) can provide a means for communication through non-muscular control. Current BCI systems use characters that flash from gray to white (GW), making adjacent character difficult to distinguish from the target. The current study implements two types of color stimulus (grey to color [GC] and color intensification [CI]) and I hypotheses that color stimuli will; (1) reduce distraction of nontargets (2) enhance target response (3) reduce eye strain. Online results (n=21) show that GC has increased information transfer rate over CI. Mean …

Social Media In Higher Education: Building Mutually Beneficial Student And Institutional Relationships Through Social Media., Megan L. Fuller

Electronic Theses and Dissertations

Social applications such as Facebook, YouTube, and Twitter have driven the public growth of Web 2.0. Universities and colleges are using social media to reach student prospects, keep contact with current students and alumni, and provide a mechanism for group collaboration and interaction in the classroom. Higher education institutions are influenced by current social media trends, and figuring out how to effectively interact with various constituencies within the social media environment can be challenging.

In this study, a group of higher education students were surveyed about their social media practices and preferences with a focus on education-related activities. The goal …

Modeling Human Group Behavior In Virtual Worlds, Fahad Shah

Electronic Theses and Dissertations

Virtual worlds and massively-multiplayer online games are rich sources of information about large-scale teams and groups, offering the tantalizing possibility of harvesting data about group formation, social networks, and network evolution. They provide new outlets for human social interaction that differ from both face-to-face interactions and non-physically-embodied social networking tools such as Facebook and Twitter. We aim to study group dynamics in these virtual worlds by collecting and analyzing public conversational patterns of users grouped in close physical proximity. To do this, we created a set of tools for monitoring, partitioning, and analyzing unstructured conversations between changing groups of participants …

Effective Task Transfer Through Indirect Encoding, Phillip Verbancsics

Electronic Theses and Dissertations

An important goal for machine learning is to transfer knowledge between tasks. For example, learning to play RoboCup Keepaway should contribute to learning the full game of RoboCup soccer. Often approaches to task transfer focus on transforming the original representation to fit the new task. Such representational transformations are necessary because the target task often requires new state information that was not included in the original representation. In RoboCup Keepaway, changing from the 3 vs. 2 variant of the task to 4 vs. 3 adds state information for each of the new players. In contrast, this dissertation explores the idea …

Techniques For Boosting The Performance In Content-Based Image Retrieval Systems, Ning Yu

Electronic Theses and Dissertations

Content-Based Image Retrieval has been an active research area for decades. In a CBIR system, one or more images are used as query to search for similar images. The similarity is measured on the low level features, such as color, shape, edge, texture. First, each image is processed and visual features are extract. Therefore each image becomes a point in the feature space. Then, if two images are close to each other in the feature space, they are considered similar. That is, the k nearest neighbors are considered the most similar images to the query image. In this K-Nearest Neighbor …

Harmony Oriented Architecture, Kyle A. Martin

Electronic Theses and Dissertations

This thesis presents Harmony Oriented Architecture: a novel architectural paradigm that applies the principles of Harmony Oriented Programming to the architecture of scalable and evolvable distributed systems. It is motivated by research on Ultra Large Scale systems that has revealed inherent limitations in human ability to design large-scale software systems that can only be overcome through radical alternatives to traditional object-oriented software engineering practice that simplifies the construction of highly scalable and evolvable system. HOP eschews encapsulation and information hiding, the core principles of objectoriented design, in favor of exposure and information sharing through a spatial abstraction. This helps to …

The Implications Of Virtual Environments In Digital Forensic Investigations, Farrah M. Patterson

Electronic Theses and Dissertations

This research paper discusses the role of virtual environments in digital forensic investigations. With virtual environments becoming more prevalent as an analysis tool in digital forensic investigations, it’s becoming more important for digital forensic investigators to understand the limitation and strengths of virtual machines. The study aims to expose limitations within commercial closed source virtual machines and open source virtual machines. The study provides a brief overview of history digital forensic investigations and virtual environments, and concludes with an experiment with four common open and closed source virtual machines; the effects of the virtual machines on the host machine as …

Ray Collection Bounding Volume Hierarchy, Kris Krishna Rivera

Electronic Theses and Dissertations

This thesis presents Ray Collection BVH, an improvement over a current day Ray Tracing acceleration structure to both build and perform the steps necessary to efficiently render dynamic scenes. Bounding Volume Hierarchy (BVH) is a commonly used acceleration structure, which aides in rendering complex scenes in 3D space using Ray Tracing by breaking the scene of triangles into a simple hierarchical structure. The algorithm this thesis explores was developed in an attempt at accelerating the process of both constructing this structure, and also using it to render these complex scenes more efficiently. The idea of using "ray collection" as a …

Virtual Router Approach For Wireless Ad Hoc Networks, Ai Hua Ho

Electronic Theses and Dissertations

Wireless networks have become increasingly popular in recent years. There are two variations of mobile wireless networks: infrastructure mobile networks and infrastructureless mobile networks. The latter are also known as mobile ad hoc network (MANET). MANETs have no fixed routers. Instead, mobile nodes function as relay nodes or routers, which discover and maintain communication connections between source nodes and destination nodes for various data transmission sessions. In other words, an MANET is a self-organizing multi-hop wireless network in which all nodes within a given geographical area participate in the routing and data forwarding process. Such networks are scalable and self-healing. …

Minimax And Maximin Fitting Of Geometric Objects To Sets Of Points, Yan B. Mayster

Electronic Theses and Dissertations

This thesis addresses several problems in the facility location sub-area of computational geometry. Let S be a set of n points in the plane. We derive algorithms for approximating S by a step function curve of size k < n, i.e., by an x-monotone orthogonal polyline ℜ with k < n horizontal segments. We use the vertical distance to measure the quality of the approximation, i.e., the maximum distance from a point in S to the horizontal segment directly above or below it. We consider two types of problems: min-ε, where the goal is to minimize the error for a …

Memory Access Patterns For Cellular Automata Using Gpgpus, James Michael Balasalle

Electronic Theses and Dissertations

Today's graphical processing units have hundreds of individual processing cores that can be used for general purpose computation of mathematical and scientific problems. Due to their hardware architecture, these devices are especially effective when solving problems that exhibit a high degree of spatial locality. Cellular automata use small, local neighborhoods to determine successive states of individual elements and therefore, provide an excellent opportunity for the application of general purpose GPU computing. However, the GPU presents a challenging environment because it lacks many of the features of traditional CPUs, such as automatic, on-chip caching of data. To fully realize the potential …

Simulation, Application, And Resilience Of An Organic Neuromorphic Architecture, Made With Organic Bistable Devices And Organic Field Effect Transistors, Robert A. Nawrocki

Electronic Theses and Dissertations

This thesis presents work done simulating a type of organic neuromorphic architecture, modeled after Artificial Neural Network, and termed Synthetic Neural Network, or SNN. The first major contribution of this thesis is development of a single-transistor-single-organic-bistable-device-per-input circuit that approximates behavior of an artificial neuron. The efficacy of this design is validated by comparing the behavior of a single synthetic neuron to that of an artificial neuron as well as two examples involving a network of synthetic neurons. The analysis utilizes electrical characteristics of polymer electronic elements, namely Organic Bistable Device and Organic Field Effect Transistor, created in the laboratory at …

The Power Of Quantum Walk Insights, Implementation, And Applications, Chen Fu Chiang

Electronic Theses and Dissertations

In this thesis, I investigate quantum walks in quantum computing from three aspects: the insights, the implementation, and the applications. Quantum walks are the quantum analogue of classical random walks. For the insights of quantum walks, I list and explain the required components for quantizing a classical random walk into a quantum walk. The components are, for instance, Markov chains, quantum phase estimation, and quantum spectrum theorem. I then demonstrate how the product of two reflections in the walk operator provides a quadratic speed-up, in comparison to the classical counterpart. For the implementation of quantum walks, I show the construction …

Research In High Performance And Low Power Computer Systems For Data-Intensive Environment, Pengju Shang

Electronic Theses and Dissertations

According to the data affinity, DAFA re-organizes data to maximize the parallelism of the affinitive data, and also subjective to the overall load balance. This enables DAFA to realize the maximum number of map tasks with data-locality. Besides the system performance, power consumption is another important concern of current computer systems. In the U.S. alone, the energy used by servers which could be saved comes to 3.17 million tons of carbon dioxide, or 580,678 cars {Kar09}. However, the goals of high performance and low energy consumption are at odds with each other. An ideal power management strategy should be able …

The Extended Kalman-Consensus Filter, Andrew Pellett

Electronic Theses and Dissertations

An algorithm called the extended Kalman-Consensus filter is developed as an extension of the Kalman-Consensus filter to the non-linear case. The extended Kalman- Consensus filter is a technique for estimating the state of a non-linear process disturbed by noise using multiple observations from a distributed set of sense nodes. All sense nodes attempt to estimate the same state by determining how their observations affect that state, and by communicating with neighbor nodes. The algorithm is designed to be more accurate through measurement diversity, scalable to a large number of nodes, and robust against loss of nodes during operation. Simulations are …

Large Margin Random Forests On Mixed Type Data, Sheng Liu

Electronic Theses and Dissertations

Incorporating various sources of biological information is important for biological discovery. For example, genes have a multi-view representation. They can be represented by features such as sequence length and physical-chemical properties. They can also be represented by pairwise similarities, gene expression levels, and phylogenetics position. Hence, the types vary from numerical features to categorical features. An efficient way of learning from observations with a multi-view representation of mixed type of data is thus important. We propose a large margin random forests classification approach based on random forests proximity. Random forests accommodate mixed data types naturally. Large margin classifiers are obtained …