Downloads & Free Reading Options - Results

Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. This study will first make an introduction to the field of artificial neural networks, then describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be constructed and used to implement the operation of the net. Its ability will be demonstrated in differentiating between different patterns and even correcting a noisy pattern and recognizing it. The Hopfield network, the Hamming network and the Carpenter/Grossberg network will be individually utilized in developing an algorithm for pattern recognition and classification. The maximum- likelihood sequence estimation function will be mapped onto a neural network structure. The application of this structure computations for data detection in digital communications receivers will be described. A computer simulation program will be constructed and used to show that neural networks offer attractive implementation alternatives for MLSE. Keywords: Neural networks; Hopfield net; Hamming net; Carpenter/Grossberg net; Pattern recognition.

Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. This study will first make an introduction to the field of artificial neural networks, then describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be constructed and used to implement the operation of the net. Its ability will be demonstrated in differentiating between different patterns and even correcting a noisy pattern and recognizing it. The Hopfield network, the Hamming network and the Carpenter/Grossberg network will be individually utilized in developing an algorithm for pattern recognition and classification. The maximum- likelihood sequence estimation function will be mapped onto a neural network structure. The application of this structure computations for data detection in digital communications receivers will be described. A computer simulation program will be constructed and used to show that neural networks offer attractive implementation alternatives for MLSE. Keywords: Neural networks; Hopfield net; Hamming net; Carpenter/Grossberg net; Pattern recognition.

Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. This study will first make an introduction to the field of artificial neural networks, then describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be constructed and used to implement the operation of the net. Its ability will be demonstrated in differentiating between different patterns and even correcting a noisy pattern and recognizing it. The Hopfield network, the Hamming network and the Carpenter/Grossberg network will be individually utilized in developing an algorithm for pattern recognition and classification. The maximum- likelihood sequence estimation function will be mapped onto a neural network structure. The application of this structure computations for data detection in digital communications receivers will be described. A computer simulation program will be constructed and used to show that neural networks offer attractive implementation alternatives for MLSE. Keywords: Neural networks; Hopfield net; Hamming net; Carpenter/Grossberg net; Pattern recognition.

Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. This study will first make an introduction to the field of artificial neural networks, then describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be constructed and used to implement the operation of the net. Its ability will be demonstrated in differentiating between different patterns and even correcting a noisy pattern and recognizing it. The Hopfield network, the Hamming network and the Carpenter/Grossberg network will be individually utilized in developing an algorithm for pattern recognition and classification. The maximum- likelihood sequence estimation function will be mapped onto a neural network structure. The application of this structure computations for data detection in digital communications receivers will be described. A computer simulation program will be constructed and used to show that neural networks offer attractive implementation alternatives for MLSE. Keywords: Neural networks; Hopfield net; Hamming net; Carpenter/Grossberg net; Pattern recognition.

Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. This study will first make an introduction to the field of artificial neural networks, then describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be constructed and used to implement the operation of the net. Its ability will be demonstrated in differentiating between different patterns and even correcting a noisy pattern and recognizing it. The Hopfield network, the Hamming network and the Carpenter/Grossberg network will be individually utilized in developing an algorithm for pattern recognition and classification. The maximum- likelihood sequence estimation function will be mapped onto a neural network structure. The application of this structure computations for data detection in digital communications receivers will be described. A computer simulation program will be constructed and used to show that neural networks offer attractive implementation alternatives for MLSE. Keywords: Neural networks; Hopfield net; Hamming net; Carpenter/Grossberg net; Pattern recognition.

Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. This study will first make an introduction to the field of artificial neural networks, then describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be constructed and used to implement the operation of the net. Its ability will be demonstrated in differentiating between different patterns and even correcting a noisy pattern and recognizing it. The Hopfield network, the Hamming network and the Carpenter/Grossberg network will be individually utilized in developing an algorithm for pattern recognition and classification. The maximum- likelihood sequence estimation function will be mapped onto a neural network structure. The application of this structure computations for data detection in digital communications receivers will be described. A computer simulation program will be constructed and used to show that neural networks offer attractive implementation alternatives for MLSE. Keywords: Neural networks; Hopfield net; Hamming net; Carpenter/Grossberg net; Pattern recognition.

Infrared spectroscopy is an important technique for measuring airborne chemicals, for pollution monitoring and to warn of toxic compound releases. Infrared spectroscopy provides both detection and identification of airborne components. Computer-assisted classification tools, including pattern recognition and artificial neural network techniques, have been applied to a collection of infrared spectra of organophosphorus compounds, and these have successfully discriminated commercial pesticide compounds from military nerve agents, precursors, and hydrolysis products. Infrared spectra for previous tests came from a commercial infrared library, with permission, from military laboratories, and from defense contractors. In order to further test such classification tools, additional infrared spectra from the NIST gas-phase infrared library were added to the data set. These additional spectra probed the tendency of the trained classifiers to misidentify unrelated spectra into the trained classes. Infrared spectra used in this effort were gathered from a variety of sources. Different instrument operators collected them at a number of locations, in a variety of spectral data collection designs, and they were delivered in a variety of digital formats. The spectra were treated mathematically to remove artifacts from their collection. Preprocessing techniques used included Fisher weighting and principal component analysis. Classifications were made using the k-nearest neighbor classifier, feed forward neural networks, trained with a variety of techniques, and radial basis function networks. The results from these classification techniques will be reported and compared.

Infrared spectroscopy is an important technique for measuring airborne chemicals, for pollution monitoring and to warn of toxic compound releases. Infrared spectroscopy provides both detection and identification of airborne components. Computer-assisted classification tools, including pattern recognition and artificial neural network techniques, have been applied to a collection of infrared spectra of organophosphorus compounds, and these have successfully discriminated commercial pesticide compounds from military nerve agents, precursors, and hydrolysis products. Infrared spectra for previous tests came from a commercial infrared library, with permission, from military laboratories, and from defense contractors. In order to further test such classification tools, additional infrared spectra from the NIST gas-phase infrared library were added to the data set. These additional spectra probed the tendency of the trained classifiers to misidentify unrelated spectra into the trained classes. Infrared spectra used in this effort were gathered from a variety of sources. Different instrument operators collected them at a number of locations, in a variety of spectral data collection designs, and they were delivered in a variety of digital formats. The spectra were treated mathematically to remove artifacts from their collection. Preprocessing techniques used included Fisher weighting and principal component analysis. Classifications were made using the k-nearest neighbor classifier, feed forward neural networks, trained with a variety of techniques, and radial basis function networks. The results from these classification techniques will be reported and compared.

Infrared spectroscopy is an important technique for measuring airborne chemicals, for pollution monitoring and to warn of toxic compound releases. Infrared spectroscopy provides both detection and identification of airborne components. Computer-assisted classification tools, including pattern recognition and artificial neural network techniques, have been applied to a collection of infrared spectra of organophosphorus compounds, and these have successfully discriminated commercial pesticide compounds from military nerve agents, precursors, and hydrolysis products. Infrared spectra for previous tests came from a commercial infrared library, with permission, from military laboratories, and from defense contractors. In order to further test such classification tools, additional infrared spectra from the NIST gas-phase infrared library were added to the data set. These additional spectra probed the tendency of the trained classifiers to misidentify unrelated spectra into the trained classes. Infrared spectra used in this effort were gathered from a variety of sources. Different instrument operators collected them at a number of locations, in a variety of spectral data collection designs, and they were delivered in a variety of digital formats. The spectra were treated mathematically to remove artifacts from their collection. Preprocessing techniques used included Fisher weighting and principal component analysis. Classifications were made using the k-nearest neighbor classifier, feed forward neural networks, trained with a variety of techniques, and radial basis function networks. The results from these classification techniques will be reported and compared.

Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. In this study, we are first going to make an introduction to the field of artificial neural networks, then we are going to describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be constructed and used to implement the operation of the net. Its ability will be demonstrated in differentiating between different patterns and even correcting a noisy pattern and recognizing it. The Hopfield network, the Hamming network and the Carpenter / Grossberg network will be individually utilized in developing an algorithm for pattern recognition and classification. The maximum-likelihood sequence estimation function will be mapped onto a neural network structure. The application of this structure computations for data detection in digital communications receivers will be described. A computer simulation program will be constructed and used to show that neural networks offer attractive implementation alternatives for MLSE.

This three-part discussion provides the necessary background in understanding the make-up and operation of the generalized biological neuron and its role in memory and pattern recognition in the brain. There are many specific types of neurons which have adapted to perform specialized operations as part of the central nervous system. Motor neurons deal with the operation of the muscles; optical neurons deal with the operations of the eyes. Although there are differences in physical make-up, their basic operation is similiar and is known as the generalized biological neuron. The physical structure and operation of this neuron will be examined. A brief discussion will follow on how the inter-connection and paralleled structure of neural networks in the optical system and cortex of the brain are able to send, store, and recall information dealing with pattern recognition. This will be followed by a brief discussion of the first mathematical model developed of the generalized biological model. Its make-up is the building block for the vast majority of neural network models which followed. Keywords: Computerized simulation.

In Phase I of this project, Physical Optics Corporation (POC) accomplished the goal of the original proposal which was to develop and optimize a unique neural network (NN) algorithm that performs rapid spectral signal processing and identification. POC's NN algorithm was tested with extremely noisy Raman spectra from Lawrence Livermore National Laboratory and experimentally showed at least ten times better sensitivity and reliability than conventional spectral signal processing methods. POC built a portable demonstration system with POC's NN and successfully demonstrated real-time spectral signature identification operations. POC proposed, for Phase II implementation, a holographic optical neural network (HONN) system that is capable of rapid hyperspectral imaging through an acoustic-optic tunable filter (AOTF), real-time spectral feature identification, and mapping. The success of the Phase II project will make automatic and rapid hyperspectral image analysis and feature location possible. Neural networks, Holographic optical spectral feature identification, Portable smart spectrometer, Hyperspectral image processing.

For application to chemical sensor arrays, the ideal pattern recognition is accurate, fast, simple to train, robust to outliers, has low memory requirements, and has the ability to produce a measure of classification certainty. In this work, four data sets representing typical chemical sensor array data were used to compare seven pattern recognition algorithms nearest neighbor, Mahalanobis linear discriminant analysis, Bayesian linear discriminant analysis, SIMCA, back propagation neural networks, probabilistic neural networks (PNN), and learning vector quantization (LVQ) for their ability to meet the criteria. LVQ and PNN exhibited high classification accuracy and met many of the qualitative criteria for an ideal algorithm. Based on these results, a new algorithm (LVQ-PNN) that incorporates the best features of PNN and LVQ was developed. The LVQ-PNN algorithm was further improved by the addition of a faster training procedure. It was then compared with the other seven algorithms. The LVQ-PNN method achieved excellent classification performance. A general procedure for selecting the optimal rejection threshold for a PNN based algorithm using Monte Carlo simulations also was demonstrated. This outlier rejection strategy was implemented for an LVQ-PNN classifier and found consistently to reject ambiguous patterns.

For application to chemical sensor arrays, the ideal pattern recognition is accurate, fast, simple to train, robust to outliers, has low memory requirements, and has the ability to produce a measure of classification certainty. In this work, four data sets representing typical chemical sensor array data were used to compare seven pattern recognition algorithms nearest neighbor, Mahalanobis linear discriminant analysis, Bayesian linear discriminant analysis, SIMCA, back propagation neural networks, probabilistic neural networks (PNN), and learning vector quantization (LVQ) for their ability to meet the criteria. LVQ and PNN exhibited high classification accuracy and met many of the qualitative criteria for an ideal algorithm. Based on these results, a new algorithm (LVQ-PNN) that incorporates the best features of PNN and LVQ was developed. The LVQ-PNN algorithm was further improved by the addition of a faster training procedure. It was then compared with the other seven algorithms. The LVQ-PNN method achieved excellent classification performance. A general procedure for selecting the optimal rejection threshold for a PNN based algorithm using Monte Carlo simulations also was demonstrated. This outlier rejection strategy was implemented for an LVQ-PNN classifier and found consistently to reject ambiguous patterns.

For application to chemical sensor arrays, the ideal pattern recognition is accurate, fast, simple to train, robust to outliers, has low memory requirements, and has the ability to produce a measure of classification certainty. In this work, four data sets representing typical chemical sensor array data were used to compare seven pattern recognition algorithms nearest neighbor, Mahalanobis linear discriminant analysis, Bayesian linear discriminant analysis, SIMCA, back propagation neural networks, probabilistic neural networks (PNN), and learning vector quantization (LVQ) for their ability to meet the criteria. LVQ and PNN exhibited high classification accuracy and met many of the qualitative criteria for an ideal algorithm. Based on these results, a new algorithm (LVQ-PNN) that incorporates the best features of PNN and LVQ was developed. The LVQ-PNN algorithm was further improved by the addition of a faster training procedure. It was then compared with the other seven algorithms. The LVQ-PNN method achieved excellent classification performance. A general procedure for selecting the optimal rejection threshold for a PNN based algorithm using Monte Carlo simulations also was demonstrated. This outlier rejection strategy was implemented for an LVQ-PNN classifier and found consistently to reject ambiguous patterns.

No Abstract Available

No Abstract Available

Thesis advisor, Tri T. Ha

Thesis advisor, Tri T. Ha

The Final Proceedings for DYNN'96 International workshop on Dynamics of Neural Networks and Application to, 12 March 1996 - 13 March 1996. The articles contain information on dynamics of chaotic non-linear systems and their uses in image processing and pattern recognition, and practical application of dynamic neural networks for VLSI implementation for image processing and pattern recognition.

The Final Proceedings for DYNN'96 International workshop on Dynamics of Neural Networks and Application to, 12 March 1996 - 13 March 1996. The articles contain information on dynamics of chaotic non-linear systems and their uses in image processing and pattern recognition, and practical application of dynamic neural networks for VLSI implementation for image processing and pattern recognition.

Includes bibliographical references and index

 ............................

The application of artificial neural networks (ANNs) for the estimation of HI gas mass fraction (\fgas) is investigated, based on a sample of 13,674 galaxies in the Sloan Digital Sky Survey (SDSS) with HI detections or upper limits from the Arecibo Legacy Fast Arecibo L-band Feed Array (ALFALFA). We show that, for an example set of fixed input parameters ($g-r$ colour and $i$-band surface brightness), a multidimensional quadratic model yields \fgas\ scaling relations with a smaller scatter (0.22 dex) than traditional linear fits (0.32 dex), demonstrating that non-linear methods can lead to an improved performance over traditional approaches. A more extensive ANN analysis is performed using 15 galaxy parameters that capture variation in stellar mass, internal structure, environment and star formation. Of the 15 parameters investigated, we find that $g-r$ colour, followed by stellar mass surface density, bulge fraction and specific star formation rate have the best connection with \fgas. By combining two control parameters, that indicate how well a given galaxy in SDSS is represented by the ALFALFA training set (\pr) and the scatter in the training procedure (\sigf), we develop a strategy for quantifying which SDSS galaxies our ANN can be adequately applied to, and the associated errors in the \fgas\ estimation. In contrast to previous works, our \fgas\ estimation has no systematic trend with galactic parameters such as M$_{\star}$, $g-r$ and SFR. We present a catalog of \fgas\ estimates for more than half a million galaxies in the SDSS, of which $\sim$ 150,000 galaxies have a secure selection parameter with average scatter in the \fgas\ estimation of 0.22 dex.