Physical Sciences and Mathematics Commons^™

Finding Functional Gene Relationships Using The Semantic Gene Organizer (Sgo), Kevin Erich Heinrich

Masters Theses

Understanding functional gene relationships is a major challenge in bioninformatics and computational biology. Currently, many approaches extract gene relationships via term co-occurrence models from the biomedical literature. Unfortunately, however, many genes that are experimentally identified to be related have not been previously studied together. As a result, many automated models fail to help researchers understand the nature of the relationships. In this work, the particular schema used tomine genomic data is called Latent
Semantic Indexing (LSI). LSI performs a singular-value decomposition (SVD) to produce a low-rank approximation of the data set. Effectively, it allows queries to be interpreted in a …

Software Reconfigurability For Heterogeneous Robot Cooperation, Maureen Chandra

Masters Theses

Previous work in multi-robot cooperation has aimed at gaining autonomy and fault tolerance in the robot team. Most attempt to accomplish this by dynamically assigning roles or tasks to the robots and pre-designing the solution for heterogeneous robot teams with known sensing capabilities. However, pre-designed solutions fail when changes occur in the robot team composition or in the available environmental sensors at run-time. Automated solution design is thus needed to accomplish autonomy and fault tolerance in multi-agent systems. Very little work has been done in automating robot solutions at run-time due to the difficulty of adapting to many unexpected events …

Itera- Tive Reconstruction Framework For High-Resolution X-Ray Ct Data, Thomas Matthew Benson

Masters Theses

Small animal medical imaging has become an important tool for researchers as it allows noninvasively screening animal models for pathologies as well as monitoring dis- ease progression and therapy response. Currently, clinical CT scanners typically use a Filtered Backprojection (FBP) based method for image reconstruction. This algorithm is fast and generally produces acceptable results, but has several drawbacks. Firstly, it is based upon line integrals, which do not accurately describe the process of X-ray attenuation. Secondly, noise in the projection data is not properly modeled with FBP. On the other hand, iterative algorithms allow the integration of more complicated sys- …

3d Visualization Modules For Chemical Engineering – A Web-Based Approach Using Java And Opengl, Sharad Anant Gupta

Masters Theses

The main objective of this work is to implement web-based educational modules for chemical engineering students. Phase behavior is a topic with which the students seem to struggle with, particularly for mixtures, where a 2-D representation of the phase diagram falls far short of the understanding a 3-D model can provide. Using the platform-independence of Java and the graphics capability of OpenGL, three phase diagram Java applets have been developed. Users can view these web-based 3D applets by installing a plug-in. These modules provide users with an ability to rotate the 3D models, slice through them, zoom into them and …

Isosurface Extraction In The Visualization Toolkit Using The Extrema Skeleton Algorithm, Subha Parvathy Mahaadevan

Masters Theses

Generating isosurfaces is a very useful technique in data visualization for understanding the distribution of scalar data. Often, when the size of the data set is really large, as in the case with data produced by medical imaging applications, engineering simulations or geographic information systems applications, the use of traditional methods like marching cubes makes repeated generation of isosurfaces a very time consuming task. This thesis investigated the use of the Extrema Skeleton algorithm to speed up repeated isosurface generation in the visualization package, Visualization Toolkit (VTK). The objective was to reduce the number of non-isosurface cells visited to generate …

Integrated Visualization Of Diffusion Tensor And Functional Mri, Nathaniel Richard Fout

Masters Theses

To understand the morphology, structure, and function of the human brain and the underlying relationships therein has long been a goal of mankind. Technologies are constantly emerging and evolving in an effort to realize this goal, with each new development potentially providing another piece of the puzzle. Among those technologies are advanced imaging modalities such as Diffusion Tensor MRI (DT-MRI) [2] and Functional MRI (fMRI) [8], whose purpose is to provide deeper insight into the neural network connecting functional units of the cerebral cortex. More specifically, DT-MRI captures a description of the fibrous structures (such as nerve fibers) in the …

A Coarse-Grain Parallel Implementation Of The Block Tridiagonal Divide And Conquer Algorithm For Symmetric Eigenproblems., Robert M. Day

Masters Theses

Cuppen’s divide and conquer technique for symmetric tridiagonal eigenproblems, along with Gu and Eisenstat’s modification for improvement of the eigenvector computation, has yielded a stable, efficient, and widely-used algorithm. This algorithm has now been extended to a larger class of matrices, namely symmetric block tridiagonal eigenproblems. The Block Tridiagonal Divide and Conquer algorithm has shown several characteristics that make it suitable for a number of applications, such as the Self-Consistent-Field procedure in quantum chemistry.

This thesis discusses the steps taken to implement a coarse-grain parallel version of the Block Tridiagonal Divide and Conquer algorithm, suitable for a parallel supercomputer or …

Sparta: A Graphical User Interface For Malicious Mobile Code Fingerprint-Ing., Victor P. De Cerqueira

Masters Theses

This thesis introduces and describes SPARTA (for Stochastic Profiling Application for the Rendering of Trees and Automata), a graphical user interface used as a front end to a collection of tools written in C that collectively convert a log of registry system calls performed by an application into binary descriptions of PSTs (for Probabilistic Suffix Trees) and PSAs (for Probabilistic Suffix Automata), which are models used to represent application behavior on Windows-based systems. SPARTA works by rendering these binary descriptions into graphical form, showcasing a variety of features intended to make the user interaction with PSTs and PSAs informative and …

Thermal Evolution Of Planetesimals And Protoplanets In The Terrestrial Planet Region: Code Optimization And Implementation On A Distributed Grid Using Netsolve, Amitabha Ghosh

Masters Theses

A code for asteroidal heat transfer and growth is optimized for performance. The Gauss elimination routine for the solver is replaced by a sparse matrix routine. Finite element matrix assembly operations are rewritten to reduce operations involving 3D arrays to 1D. Advantage is taken of the sparse matrix structure of finite element matrices in reducing 2D arrays to 1D. The number of vector touches are reduced to the extent possible, by carrying over statements from one iteration to the next. The number of do loops are reduced by merging several do loops into one. The optimization reduced the CPU time …

Neural Supernets: Structuring Artificial Neural Networks According To Cluster Analysis, Simon Lucas Winberg

Masters Theses

A typical feed forward neural network relies solely on its training algorithm, such as backprop or quickprop, to determine suitable weight values for an architecture chosen by the human operator. The architecture itself is typically a fully connected structuring of neurons and synapses where each hidden neuron is connected to every neuron in the next layer. Such architecture does not reflect the structure of the data used to train it. Similarly, in the case where random initial weight values are used, these initial weights are also unlikely to relate to the training set. Thus the job of the training algorithm …

On A Grid-Based Interface To A Special-Purpose Hardware Cluster, Jeanne Marie Lehrter

Masters Theses

Grid computing is an important and useful tool that enables resource sharing within a computational community. In the grid computing model, each node of the grid is a computational resource that may be used by applications residing on other nodes of the grid or on machines outside of the grid. To make grid computing more accessible to all users, the Innovative Computing Lab on campus at the University of Tennessee has produced a software package called NetSolve. NetSolve is a client-server application that defines interfaces for accessing and utilizing the software and hardware resources of the nodes in the grid …

Design And Implementation Of The L-Bone And Logistical Tools, Edward Scott Atchley

Masters Theses

The purpose of this paper is to outline the design criteria and implementation of the Logistical Backbone (L-Bone) and the Logistical Tools. These tools, along with IBP and the exNode Library, allow storage to be used as a network resource. These are components of the Network Storage Stack, a design by the Logistical Computing and Internetworking Lab at the University of Tennessee. Having storage as a network resource enables users to do many things that are either difficult or not possible today, such as moving and sharing very large files across administrative domains, improving performance through caching and improving fault-tolerance …

Design And Simulation Of Self-Organizing Microbial Cellular Automata, Michael Patrick Campfield

Masters Theses

This paper discusses the design and implementation of cellular automata based on the alteration of genetic sequences in bacteria. The work is composed of five chapters covering the problem, the system’s design, the software simulation of the system and future issues on the problem. The section covering the problem explores the reasons for this work as well as issues that this work solves. The section covering system design details the modified genetic sequences and the algorithm that these sequences implement. The simulation section describes the layout of an experiment along with the test cases experimented on. Finally, the future work …

Adaptive Simulated Annealing: An Alternative Approach For The Error Minimization Of Neural Networks, Yuxing Sun

Masters Theses

This work introduces an alternative algorithm, simulated annealing, to minimize the prediction error from neural networks that traditionally use back-propagation methods. The simulated annealing algorithm stochastically samples the parameter space formed by weights of the neural network until a minimal error is found.

Three problems were investigated in this work: the radiator problem, the spiral problem, and the time series prediction problem. Each of them was examined using the same neural network architecture, i.e., a 2-layer network, and trained by both back-propagation and simulated annealing.

The simulated annealing algorithm consumes longer computation time in searching for the global minimum than …

Initial Feasible Solutions Of Three Dimensional Transportation Problems, Mamoru Aiga

Masters Theses

The purpose of this paper is to investigate methods to obtain initial feasible solutions of three dimensional transportation problems. Schell's procedure was tested on various randomly generated problems, and it was determined that this algorithm did not always yield an initial feasible solution. Thus a modified Schell procedure was developed.

Computer programs were written to compare the modified Schell procedure with Phase I of Simplex method. It was concluded, from cases tested, that the modified Schell procedure requires much less computing time and generally gives a feasible solution closer to the optimum solution.

Computer Implemented Synthesis Of Multiple Level Combinational Networks, Alfredo Caldwell

Masters Theses

The purpose of this investigation has been to implement, in the form of computer programs, two algorithms which are used in the synthesis of multiple level combinational networks. The algorithms implemented were devised by Professor Paul E. Wood, Jr., of M.I.T., and by Professor Eugene L. Lawler, of The University of Michigan. In the course of the investigation a more efficient way of implementing the two algorithms was discovered. The combined version of the algorithms takes advantage of the best features of the original algorithms.

In the synthesis of multiple level combinational networks minimal complexity is high desirable. The whole …