721 Matching Results

Search Results

Advanced search parameters have been applied.

Computational Complexity of Hopfield Networks

Description: There are three main results in this dissertation. They are PLS-completeness of discrete Hopfield network convergence with eight different restrictions, (degree 3, bipartite and degree 3, 8-neighbor mesh, dual of the knight's graph, hypercube, butterfly, cube-connected cycles and shuffle-exchange), exponential convergence behavior of discrete Hopfield network, and simulation of Turing machines by discrete Hopfield Network.
Date: August 1998
Creator: Tseng, Hung-Li
Partner: UNT Libraries

Neural Networks for Analysis of Top Quark Production

Description: Neural networks (NNs) provide a powerful and flexible tool for selecting a signal from a larger background. The D0 collaboration has used them extensively in studying t{anti t} decays. NNs were essential to the measurement of the t{anti t} production cross section in the all-jets channel (t{anti t} {yields} b {anti b}qqqq), and were also used in the measurement of the mass of the top quark in the lepton+jets channel (t{anti t} {yields} b{anti b}l{nu}q{anti q}). This paper will describe two new applications of neural networks to top quark analysis: the search for single top quark production, and an effort to increase the sensitivity in the dilepton channel t{anti t} {yields} b{anti b}e{anti {mu}}{nu}{anti {nu}} beyond that achieved in the published analysis.
Date: August 4, 1999
Creator: al., B. Abbott et
Partner: UNT Libraries Government Documents Department

Exploration of hierarchical leadership and connectivity in neural networks in vitro.

Description: Living neural networks are capable of processing information much faster than a modern computer, despite running at significantly lower clock speeds. Therefore, understanding the mechanisms neural networks utilize is an issue of substantial importance. Neuronal interaction dynamics were studied using histiotypic networks growing on microelectrode arrays in vitro. Hierarchical relationships were explored using bursting (when many neurons fire in a short time frame) dynamics, pairwise neuronal activation, and information theoretic measures. Together, these methods reveal that global network activity results from ignition by a small group of burst leader neurons, which form a primary circuit that is responsible for initiating most network-wide burst events. Phase delays between leaders and followers reveal information about the nature of the connection between the two. Physical distance from a burst leader appears to be an important factor in follower response dynamics. Information theory reveals that mutual information between neuronal pairs is also a function of physical distance. Activation relationships in developing networks were studied and plating density was found to play an important role in network connectivity development. These measures provide unique views of network connectivity and hierarchical relationship in vitro which should be included in biologically meaningful models of neural networks.
Date: December 2008
Creator: Ham, Michael I.
Partner: UNT Libraries

Distinguishing Sedimentary Depositional Environments on Mars Using In-Situ Luminescence Measurements and Neural Network Analysis.

Description: Although not recognized as a true geologic period on Mars, the term 'Martian Quaternary' emerged at the 2001 LPSC. The phrase helps focus attention on the fact that Mars does indeed exhibit dynamic surface processes andd geomorphic features active on time scales less than 10{sup 6} years. As on earth evolving landscapes can serve as important storehouses for paltoclimatic and paIeoenvironmental records. However, deciphering the 'Quaternary' history of Mars will require the development of 4 new set of tools that are uniquely suited to examining youthful sediments and geomorphic features.
Date: January 1, 2001
Creator: Lepper, K. E. (Kenneth E.) & Whitley, V. H. (Von H.)
Partner: UNT Libraries Government Documents Department

Concentration-dependent Effects of D-Methylphenidate on Frontal Cortex and Spinal Cord Networks in vitro

Description: Spontaneously active frontal cortex and spinal cord networks grown on microelectrode arrays were used to study effects of D-methylphenidate. These central nervous system tissues have relatively low concentrations of dopaminergic and noradrenergic neurons compared to the richly populated loci, yet exhibit similar neurophysiological responses to methylphenidate. The spontaneous spike activity of both tissues was inhibited in a concentration-dependent manner by serial additions of 1-500 µM methylphenidate. Methylphenidate is non-toxic as spike inhibition was recovered following washes. The average concentrations for 50% spike rate inhibition (IC50 ± SD) were 118 ± 52 (n= 6) and 57 ± 43 (n = 11) for frontal cortex and spinal cord networks, respectively. A 3 hour exposure of a network to 1 mM methylphenidate was nontoxic. The effective concentrations described in this study are within the therapeutic dosage range. Therefore, the platform may be used for further investigations of drug mechanisms.
Date: December 2004
Creator: Miller, Benjamin R.
Partner: UNT Libraries

Neural Network Modeling of Weld Pool Shape in Pulsed-Laser Aluminum Welds

Description: A neural network model was developed to predict the weld pool shape for pulsed-laser aluminum welds. Several different network architectures were examined and the optimum architecture was identified. The neural network was then trained and, in spite of the small size of the training data set, the network accurately predicted the weld pool shape profiles. The neural network output was in the form of four weld pool shape parameters (depth, width, half-width, and area) and these were converted into predicted weld pool profiles with the use of the actual experimental poo1 profiles as templates. It was also shown that the neural network model could reliably predict the change from conduction-mode type shapes to keyhole-mode shapes.
Date: November 16, 1998
Creator: Iskander, Y.S.; Oblow, E.M. & Vitek, J.M.
Partner: UNT Libraries Government Documents Department

Neural Network method for Inverse Modeling of Material Deformation

Description: A method is described for inverse modeling of material deformation in applications of importance to the sheet metal forming industry. The method was developed in order to assess the feasibility of utilizing empirical data in the early stages of the design process as an alternative to conventional prototyping methods. Because properly prepared and employed artificial neural networks (ANN) were known to be capable of codifying and generalizing large bodies of empirical data, they were the natural choice for the application. The product of the work described here is a desktop ANN system that can produce in one pass an accurate die design for a user-specified part shape.
Date: July 10, 1999
Creator: Allen, J.D., Jr.; Ivezic, N.D. & Zacharia, T.
Partner: UNT Libraries Government Documents Department

Event identification from seismic/magnetic feature vectors: a comparative study

Description: The event identification problem plays a large role in the application of unattended ground sensors to the monitoring of borders and checkpoints. The choice of features and methods for classifying features affects how accurately these classifications are made. Finding features which reliably distinguish events of interest may require measurements based on separate physical phenomena. Classification methods include neural net versus fuzzy logic approaches, and within the neural category, different architectures and transfer functions for reaching decisions. This study examines ways of optimizing feature sets and surveys common techniques for classifying feature vectors corresponding to physical events. We apply each technique to samples of existing data, and compare discrimination attributes. Specifically, we calculate the confusion matrices for each technique applied to each sample dataset, and reduce them statistically to scalar scores. In addition, we gauge how the accuracy of each method is degraded by reducing the feature vector length by one element. Finally, we gather rough estimates of the relative cpu performance of the forward prediction algorithms.
Date: April 1, 1997
Creator: Wolford, J. K.
Partner: UNT Libraries Government Documents Department

Reservoir Characterization of Upper Devonian Gordon Sandstone, Jacksonburg, Stringtown Oil Field, Northwestern West Virginia

Description: This report gives results on use of a minipermeameter on cores to study very finescale trends in permeability, and use of neural networks to predict permeability in logged, uncored wells.
Date: May 21, 2002
Creator: Ameri, S.; Aminian, K.; Avary, K.L.; Bilgesu, H.I.; Hohn, M.E.; McDowell, R.R. et al.
Partner: UNT Libraries Government Documents Department

Pruning Neural Networks with Distribution Estimation Algorithms

Description: This paper describes the application of four evolutionary algorithms to the pruning of neural networks used in classification problems. Besides of a simple genetic algorithm (GA), the paper considers three distribution estimation algorithms (DEAs): a compact GA, an extended compact GA, and the Bayesian Optimization Algorithm. The objective is to determine if the DEAs present advantages over the simple GA in terms of accuracy or speed in this problem. The experiments used a feed forward neural network trained with standard back propagation and public-domain and artificial data sets. The pruned networks seemed to have better or equal accuracy than the original fully-connected networks. Only in a few cases, pruning resulted in less accurate networks. We found few differences in the accuracy of the networks pruned by the four EAs, but found important differences in the execution time. The results suggest that a simple GA with a small population might be the best algorithm for pruning networks on the data sets we tested.
Date: January 15, 2003
Creator: Cantu-Paz, E
Partner: UNT Libraries Government Documents Department

Temporal Connectionist Expert Systems Using a Temporal Backpropagation Algorithm

Description: Representing time has been considered a general problem for artificial intelligence research for many years. More recently, the question of representing time has become increasingly important in representing human decision making process through connectionist expert systems. Because most human behaviors unfold over time, any attempt to represent expert performance, without considering its temporal nature, can often lead to incorrect results. A temporal feedforward neural network model that can be applied to a number of neural network application areas, including connectionist expert systems, has been introduced. The neural network model has a multi-layer structure, i.e. the number of layers is not limited. Also, the model has the flexibility of defining output nodes in any layer. This is especially important for connectionist expert system applications. A temporal backpropagation algorithm which supports the model has been developed. The model along with the temporal backpropagation algorithm makes it extremely practical to define any artificial neural network application. Also, an approach that can be followed to decrease the memory space used by weight matrix has been introduced. The algorithm was tested using a medical connectionist expert system to show how best we describe not only the disease but also the entire course of the disease. The system, first, was trained using a pattern that was encoded from the expert system knowledge base rules. Following then, series of experiments were carried out using the temporal model and the temporal backpropagation algorithm. The first series of experiments was done to determine if the training process worked as predicted. In the second series of experiments, the weight matrix in the trained system was defined as a function of time intervals before presenting the system with the learned patterns. The result of the two experiments indicate that both approaches produce correct results. The only difference between the two results ...
Date: December 1993
Creator: Civelek, Ferda N. (Ferda Nur)
Partner: UNT Libraries

Advanced Signal Analysis for Forensic Applications of Ground Penetrating Radar

Description: Ground penetrating radar (GPR) systems have traditionally been used to image subsurface objects. The main focus of this paper is to evaluate an advanced signal analysis technique. Instead of compiling spatial data for the analysis, this technique conducts object recognition procedures based on spectral statistics. The identification feature of an object type is formed from the training vectors by a singular-value decomposition procedure. To illustrate its capability, this procedure is applied to experimental data and compared to the performance of the neural-network approach.
Date: June 2004
Creator: Koppenjan, Steven; Streeton, Matthew; Lee, Hua; Lee, Michael & Ono, Sashi
Partner: UNT Libraries Government Documents Department

Identifying generalities in data sets using periodic Hopfield networks : initial status report.

Description: We present a novel class of dynamic neural networks that is capable of learning, in an unsupervised manner, attractors that correspond to generalities in a data set. Upon presentation of a test stimulus, the networks follow a sequence of attractors that correspond to subsets of increasing size or generality in the original data set. The networks, inspired by those of the insect antennal lobe, build upon a modified Hopfield network in which nodes are periodically suppressed, global inhibition is gradually strengthened, and the weight of input neurons is gradually decreased relative to recurrent connections. This allows the networks to converge on a Hopfield network's equilibrium within each suppression cycle, and to switch between attractors in between cycles. The fast mutually reinforcing excitatory connections that dominate dynamics within cycles ensures the robust error-tolerant behavior that characterizes Hopfield networks. The cyclic inhibition releases the network from what would otherwise be stable equilibriums or attractors. Increasing global inhibition and decreasing dependence on the input leads successive attractors to differ, and to display increasing generality. As the network is faced with stronger inhibition, only neurons connected with stronger mutually excitatory connections will remain on; successive attractors will consist of sets of neurons that are more strongly correlated, and will tend to select increasingly generic characteristics of the data. Using artificial data, we were able to identify configurations of the network that appeared to produce a sequence of increasingly general results. The next logical steps are to apply these networks to suitable real-world data that can be characterized by a hierarchy of increasing generality and observe the network's performance. This report describes the work, data, and results, the current understanding of the results, and how the work could be continued. The code, data, and preliminary results are included and are available as an archive.
Date: December 1, 2004
Creator: Link, Hamilton E. & Backer, Alejandro
Partner: UNT Libraries Government Documents Department

Neural Network Approach to Locating Cryptography in Object Code

Description: Finding and identifying cryptography is a growing concern in the malware analysis community. In this paper, artificial neural networks are used to classify functional blocks from a disassembled program as being either cryptography related or not. The resulting system, referred to as NNLC (Neural Net for Locating Cryptography) is presented and results of applying this system to various libraries are described.
Date: September 1, 2009
Creator: Wright, Jason L. & Manic, Milos
Partner: UNT Libraries Government Documents Department

Freeform Cursive Handwriting Recognition Using a Clustered Neural Network

Description: Optical character recognition (OCR) software has advanced greatly in recent years. Machine-printed text can be scanned and converted to searchable text with word accuracy rates around 98%. Reasonably neat hand-printed text can be recognized with about 85% word accuracy. However, cursive handwriting still remains a challenge, with state-of-the-art performance still around 75%. Algorithms based on hidden Markov models have been only moderately successful, while recurrent neural networks have delivered the best results to date. This thesis explored the feasibility of using a special type of feedforward neural network to convert freeform cursive handwriting to searchable text. The hidden nodes in this network were grouped into clusters, with each cluster being trained to recognize a unique character bigram. The network was trained on writing samples that were pre-segmented and annotated. Post-processing was facilitated in part by using the network to identify overlapping bigrams that were then linked together to form words and sentences. With dictionary assisted post-processing, the network achieved word accuracy of 66.5% on a small, proprietary corpus. The contributions in this thesis are threefold: 1) the novel clustered architecture of the feed-forward neural network, 2) the development of an expanded set of observers combining image masks, modifiers, and feature characterizations, and 3) the use of overlapping bigrams as the textual working unit to assist in context analysis and reconstruction.
Date: August 2015
Creator: Bristow, Kelly H.
Partner: UNT Libraries

Identifying Energy-Efficient Concurrency Levels using Machine Learning

Description: Multicore microprocessors have been largely motivated by the diminishing returns in performance and the increased power consumption of single-threaded ILP microprocessors. With the industry already shifting from multicore to many-core microprocessors, software developers must extract more thread-level parallelism from applications. Unfortunately, low power-efficiency and diminishing returns in performance remain major obstacles with many cores. Poor interaction between software and hardware, and bottlenecks in shared hardware structures often prevent scaling to many cores, even in applications where a high degree of parallelism is potentially available. In some cases, throwing additional cores at a problem may actually harm performance and increase power consumption. Better use of otherwise limitedly beneficial cores by software components such as hypervisors and operating systems can improve system-wide performance and reliability, even in cases where power consumption is not a main concern. In response to these observations, we evaluate an approach to throttle concurrency in parallel programs dynamically. We throttle concurrency to levels with higher predicted efficiency from both performance and energy standpoints, and we do so via machine learning, specifically artificial neural networks (ANNs). One advantage of using ANNs over similar techniques previously explored is that the training phase is greatly simplified, thereby reducing the burden on the end user. Using machine learning in the context of concurrency throttling is novel. We show that ANNs are effective for identifying energy-efficient concurrency levels in multithreaded scientific applications, and we do so using physical experimentation on a state-of-the-art quad-core Xeon platform.
Date: July 23, 2007
Creator: Curtis-Maury, M; Singh, K; Blagojevic, F; Nikolopoulos, D S; de Supinski, B R; Schulz, M et al.
Partner: UNT Libraries Government Documents Department

An Approach to Performance Prediction for Parallel Applications

Description: Accurately modeling and predicting performance for large-scale applications becomes increasingly difficult as system complexity scales dramatically. Analytic predictive models are useful, but are difficult to construct, usually limited in scope, and often fail to capture subtle interactions between architecture and software. In contrast, we employ multilayer neural networks trained on input data from executions on the target platform. This approach is useful for predicting many aspects of performance, and it captures full system complexity. Our models are developed automatically from the training input set, avoiding the difficult and potentially error-prone process required to develop analytic models. This study focuses on the high-performance, parallel application SMG2000, a much studied code whose variations in execution times are still not well understood. Our model predicts performance on two large-scale parallel platforms within 5%-7% error across a large, multi-dimensional parameter space.
Date: May 17, 2005
Creator: Ipek, E; de Supinski, B R; Schulz, M & McKee, S A
Partner: UNT Libraries Government Documents Department

Achieving New Source Performance Standards (NSPS) Emission Standards Through Integration of Low-NOx Burners with an Optimization Plan for Boiler Combustion

Description: The objective of this project was to demonstrate the use of an Integrated Combustion Optimization System to achieve NO{sub X} emission levels in the range of 0.15 to 0.22 lb/MMBtu while simultaneously enabling increased power output. The project plan consisted of the integration of low-NO{sub X} burners and advanced overfire air technology with various process measurement and control devices on the Holcomb Station Unit 1 boiler. The plan included the use of sophisticated neural networks or other artificial intelligence technologies and complex software to optimize several operating parameters, including NO{sub X} emissions, boiler efficiency, and CO emissions. The program was set up in three phases. In Phase I, the boiler was equipped with sensors that can be used to monitor furnace conditions and coal flow to permit improvements in boiler operation. In Phase II, the boiler was equipped with burner modifications designed to reduce NO{sub X} emissions and automated coal flow dampers to permit on-line fuel balancing. In Phase III, the boiler was to be equipped with an overfire air system to permit deep reductions in NO{sub X} emissions. Integration of the overfire air system with the improvements made in Phases I and II would permit optimization of boiler performance, output, and emissions. This report summarizes the overall results from Phases I and II of the project. A significant amount of data was collected from the combustion sensors, coal flow monitoring equipment, and other existing boiler instrumentation to monitor performance of the burner modifications and the coal flow balancing equipment.
Date: December 31, 2006
Creator: Penrod, Wayne
Partner: UNT Libraries Government Documents Department

Photometric Redshifts for the Dark Energy Survey and VISTA and Implications for Large Scale Structure

Description: We conduct a detailed analysis of the photometric redshift requirements for the proposed Dark Energy Survey (DES) using two sets of mock galaxy simulations and an artificial neural network code-ANNz. In particular, we examine how optical photometry in the DES grizY bands can be complemented with near infra-red photometry from the planned VISTA Hemisphere Survey (VHS) in the JHK{sub s} bands in order to improve the photometric redshift estimate by a factor of two at z > 1. We draw attention to the effects of galaxy formation scenarios such as reddening on the photo-z estimate and using our neural network code, calculate A{sub v} for these reddened galaxies. We also look at the impact of using different training sets when calculating photometric redshifts. In particular, we find that using the ongoing DEEP2 and VVDS-Deep spectroscopic surveys to calibrate photometric redshifts for DES, will prove effective. However we need to be aware of uncertainties in the photometric redshift bias that arise when using different training sets as these will translate into errors in the dark energy equation of state parameter, w. Furthermore, we show that the neural network error estimate on the photometric redshift may be used to remove outliers from our samples before any kind of cosmological analysis, in particular for large-scale structure experiments. By removing all galaxies with a 1{sigma} photo-z scatter greater than 0.1 from our DES+VHS sample, we can constrain the galaxy power spectrum out to a redshift of 2 and reduce the fractional error on this power spectrum by {approx}15-20% compared to using the entire catalogue.
Date: November 1, 2007
Creator: Banerji, Manda; Abdalla, Filipe B.; Lahav, Ofer; London, /University Coll.; Lin, Huan & /Fermilab
Partner: UNT Libraries Government Documents Department

Use of Microarray Test Data for Toxicogenomic Prediction-Multi-Intelligent Systems for Toxicogenomic Applications (MISTA)

Description: The YAHSGS LLC and Oak Ridge National Laboratory established a CRADA to develop a computational neural network and wavelets software to facilitate providing national needs for toxicity prediction and overcome the voracious drain of resources (money and time) being directed to the development of pharmaceutical agents. The research project was supported through a STTR Phase I task by NIEHS in 2004. The research deploys state-of-the-art computational neural networks and wavelets to make toxicity prediction on three independent bases: (1) quantitative structure-activity relationships, (2) microarray data, and (3) Massively Parallel Signature Sequencing technology. Upon completion of Phase I, a prototype software Multi-Intelligent System for Toxicogenomic and Applications (MISTA) was developed, the utility's feasibility was demonstrated, and a Phase II proposal was jointly prepared and submitted to NIEHS for funding evaluation. The goals and objectives of the program have been achieved.
Date: September 12, 2005
Creator: Wasson, J.S. & Lu, P.-Y.
Partner: UNT Libraries Government Documents Department