A Regularizer Approach for RBF Networks Under the Concurrent Weight Failure Situation.

PubMed

Leung, Chi-Sing; Wan, Wai Yan; Feng, Ruibin

2017-06-01

Many existing results on fault-tolerant algorithms focus on the single fault source situation, where a trained network is affected by one kind of weight failure. In fact, a trained network may be affected by multiple kinds of weight failure. This paper first studies how the open weight fault and the multiplicative weight noise degrade the performance of radial basis function (RBF) networks. Afterward, we define the objective function for training fault-tolerant RBF networks. Based on the objective function, we then develop two learning algorithms, one batch mode and one online mode. Besides, the convergent conditions of our online algorithm are investigated. Finally, we develop a formula to estimate the test set error of faulty networks trained from our approach. This formula helps us to optimize some tuning parameters, such as RBF width.

SAGRAD: A Program for Neural Network Training with Simulated Annealing and the Conjugate Gradient Method.

PubMed

Bernal, Javier; Torres-Jimenez, Jose

2015-01-01

SAGRAD (Simulated Annealing GRADient), a Fortran 77 program for computing neural networks for classification using batch learning, is discussed. Neural network training in SAGRAD is based on a combination of simulated annealing and Møller's scaled conjugate gradient algorithm, the latter a variation of the traditional conjugate gradient method, better suited for the nonquadratic nature of neural networks. Different aspects of the implementation of the training process in SAGRAD are discussed, such as the efficient computation of gradients and multiplication of vectors by Hessian matrices that are required by Møller's algorithm; the (re)initialization of weights with simulated annealing required to (re)start Møller's algorithm the first time and each time thereafter that it shows insufficient progress in reaching a possibly local minimum; and the use of simulated annealing when Møller's algorithm, after possibly making considerable progress, becomes stuck at a local minimum or flat area of weight space. Outlines of the scaled conjugate gradient algorithm, the simulated annealing procedure and the training process used in SAGRAD are presented together with results from running SAGRAD on two examples of training data.

Training Feedforward Neural Networks Using Symbiotic Organisms Search Algorithm.

PubMed

Wu, Haizhou; Zhou, Yongquan; Luo, Qifang; Basset, Mohamed Abdel

2016-01-01

Symbiotic organisms search (SOS) is a new robust and powerful metaheuristic algorithm, which stimulates the symbiotic interaction strategies adopted by organisms to survive and propagate in the ecosystem. In the supervised learning area, it is a challenging task to present a satisfactory and efficient training algorithm for feedforward neural networks (FNNs). In this paper, SOS is employed as a new method for training FNNs. To investigate the performance of the aforementioned method, eight different datasets selected from the UCI machine learning repository are employed for experiment and the results are compared among seven metaheuristic algorithms. The results show that SOS performs better than other algorithms for training FNNs in terms of converging speed. It is also proven that an FNN trained by the method of SOS has better accuracy than most algorithms compared.

Development of Artificial Neural Network Model for Diesel Fuel Properties Prediction using Vibrational Spectroscopy.

PubMed

Bolanča, Tomislav; Marinović, Slavica; Ukić, Sime; Jukić, Ante; Rukavina, Vinko

2012-06-01

This paper describes development of artificial neural network models which can be used to correlate and predict diesel fuel properties from several FTIR-ATR absorbances and Raman intensities as input variables. Multilayer feed forward and radial basis function neural networks have been used to rapid and simultaneous prediction of cetane number, cetane index, density, viscosity, distillation temperatures at 10% (T10), 50% (T50) and 90% (T90) recovery, contents of total aromatics and polycyclic aromatic hydrocarbons of commercial diesel fuels. In this study two-phase training procedures for multilayer feed forward networks were applied. While first phase training algorithm was constantly the back propagation one, two second phase training algorithms were varied and compared, namely: conjugate gradient and quasi Newton. In case of radial basis function network, radial layer was trained using K-means radial assignment algorithm and three different radial spread algorithms: explicit, isotropic and K-nearest neighbour. The number of hidden layer neurons and experimental data points used for the training set have been optimized for both neural networks in order to insure good predictive ability by reducing unnecessary experimental work. This work shows that developed artificial neural network models can determine main properties of diesel fuels simultaneously based on a single and fast IR or Raman measurement.

Training Feedforward Neural Networks Using Symbiotic Organisms Search Algorithm

PubMed Central

Wu, Haizhou; Luo, Qifang

2016-01-01

Symbiotic organisms search (SOS) is a new robust and powerful metaheuristic algorithm, which stimulates the symbiotic interaction strategies adopted by organisms to survive and propagate in the ecosystem. In the supervised learning area, it is a challenging task to present a satisfactory and efficient training algorithm for feedforward neural networks (FNNs). In this paper, SOS is employed as a new method for training FNNs. To investigate the performance of the aforementioned method, eight different datasets selected from the UCI machine learning repository are employed for experiment and the results are compared among seven metaheuristic algorithms. The results show that SOS performs better than other algorithms for training FNNs in terms of converging speed. It is also proven that an FNN trained by the method of SOS has better accuracy than most algorithms compared. PMID:28105044

Training the Recurrent neural network by the Fuzzy Min-Max algorithm for fault prediction

DOE Office of Scientific and Technical Information (OSTI.GOV)

Zemouri, Ryad; Racoceanu, Daniel; Zerhouni, Noureddine

2009-03-05

In this paper, we present a training technique of a Recurrent Radial Basis Function neural network for fault prediction. We use the Fuzzy Min-Max technique to initialize the k-center of the RRBF neural network. The k-means algorithm is then applied to calculate the centers that minimize the mean square error of the prediction task. The performances of the k-means algorithm are then boosted by the Fuzzy Min-Max technique.

SAGRAD: A Program for Neural Network Training with Simulated Annealing and the Conjugate Gradient Method

PubMed Central

Bernal, Javier; Torres-Jimenez, Jose

2015-01-01

SAGRAD (Simulated Annealing GRADient), a Fortran 77 program for computing neural networks for classification using batch learning, is discussed. Neural network training in SAGRAD is based on a combination of simulated annealing and Møller’s scaled conjugate gradient algorithm, the latter a variation of the traditional conjugate gradient method, better suited for the nonquadratic nature of neural networks. Different aspects of the implementation of the training process in SAGRAD are discussed, such as the efficient computation of gradients and multiplication of vectors by Hessian matrices that are required by Møller’s algorithm; the (re)initialization of weights with simulated annealing required to (re)start Møller’s algorithm the first time and each time thereafter that it shows insufficient progress in reaching a possibly local minimum; and the use of simulated annealing when Møller’s algorithm, after possibly making considerable progress, becomes stuck at a local minimum or flat area of weight space. Outlines of the scaled conjugate gradient algorithm, the simulated annealing procedure and the training process used in SAGRAD are presented together with results from running SAGRAD on two examples of training data. PMID:26958442

Modification Of Learning Rate With Lvq Model Improvement In Learning Backpropagation

NASA Astrophysics Data System (ADS)

Tata Hardinata, Jaya; Zarlis, Muhammad; Budhiarti Nababan, Erna; Hartama, Dedy; Sembiring, Rahmat W.

2017-12-01

One type of artificial neural network is a backpropagation, This algorithm trained with the network architecture used during the training as well as providing the correct output to insert a similar but not the same with the architecture in use at training.The selection of appropriate parameters also affects the outcome, value of learning rate is one of the parameters which influence the process of training, Learning rate affects the speed of learning process on the network architecture.If the learning rate is set too large, then the algorithm will become unstable and otherwise the algorithm will converge in a very long period of time.So this study was made to determine the value of learning rate on the backpropagation algorithm. LVQ models of learning rate is one of the models used in the determination of the value of the learning rate of the algorithm LVQ.By modifying this LVQ model to be applied to the backpropagation algorithm. From the experimental results known to modify the learning rate LVQ models were applied to the backpropagation algorithm learning process becomes faster (epoch less).

Student beats the teacher: deep neural networks for lateral ventricles segmentation in brain MR

NASA Astrophysics Data System (ADS)

Ghafoorian, Mohsen; Teuwen, Jonas; Manniesing, Rashindra; Leeuw, Frank-Erik d.; van Ginneken, Bram; Karssemeijer, Nico; Platel, Bram

2018-03-01

Ventricular volume and its progression are known to be linked to several brain diseases such as dementia and schizophrenia. Therefore accurate measurement of ventricle volume is vital for longitudinal studies on these disorders, making automated ventricle segmentation algorithms desirable. In the past few years, deep neural networks have shown to outperform the classical models in many imaging domains. However, the success of deep networks is dependent on manually labeled data sets, which are expensive to acquire especially for higher dimensional data in the medical domain. In this work, we show that deep neural networks can be trained on muchcheaper-to-acquire pseudo-labels (e.g., generated by other automated less accurate methods) and still produce more accurate segmentations compared to the quality of the labels. To show this, we use noisy segmentation labels generated by a conventional region growing algorithm to train a deep network for lateral ventricle segmentation. Then on a large manually annotated test set, we show that the network significantly outperforms the conventional region growing algorithm which was used to produce the training labels for the network. Our experiments report a Dice Similarity Coefficient (DSC) of 0.874 for the trained network compared to 0.754 for the conventional region growing algorithm (p < 0.001).

Detection of pseudosinusoidal epileptic seizure segments in the neonatal EEG by cascading a rule-based algorithm with a neural network.

PubMed

Karayiannis, Nicolaos B; Mukherjee, Amit; Glover, John R; Ktonas, Periklis Y; Frost, James D; Hrachovy, Richard A; Mizrahi, Eli M

2006-04-01

This paper presents an approach to detect epileptic seizure segments in the neonatal electroencephalogram (EEG) by characterizing the spectral features of the EEG waveform using a rule-based algorithm cascaded with a neural network. A rule-based algorithm screens out short segments of pseudosinusoidal EEG patterns as epileptic based on features in the power spectrum. The output of the rule-based algorithm is used to train and compare the performance of conventional feedforward neural networks and quantum neural networks. The results indicate that the trained neural networks, cascaded with the rule-based algorithm, improved the performance of the rule-based algorithm acting by itself. The evaluation of the proposed cascaded scheme for the detection of pseudosinusoidal seizure segments reveals its potential as a building block of the automated seizure detection system under development.

A novel and generalized approach in the inversion of geoelectrical resistivity data using Artificial Neural Networks (ANN)

NASA Astrophysics Data System (ADS)

Raj, A. Stanley; Srinivas, Y.; Oliver, D. Hudson; Muthuraj, D.

2014-03-01

The non-linear apparent resistivity problem in the subsurface study of the earth takes into account the model parameters in terms of resistivity and thickness of individual subsurface layers using the trained synthetic data by means of Artificial Neural Networks (ANN). Here we used a single layer feed-forward neural network with fast back propagation learning algorithm. So on proper training of back propagation networks it tends to give the resistivity and thickness of the subsurface layer model of the field resistivity data with reference to the synthetic data trained in the appropriate network. During training, the weights and biases of the network are iteratively adjusted to make network performance function level more efficient. On adequate training, errors are minimized and the best result is obtained using the artificial neural networks. The network is trained with more number of VES data and this trained network is demonstrated by the field data. The accuracy of inversion depends upon the number of data trained. In this novel and specially designed algorithm, the interpretation of the vertical electrical sounding has been done successfully with the more accurate layer model.

Manifold absolute pressure estimation using neural network with hybrid training algorithm

PubMed Central

Selamat, Hazlina; Alimin, Ahmad Jais; Haniff, Mohamad Fadzli

2017-01-01

In a modern small gasoline engine fuel injection system, the load of the engine is estimated based on the measurement of the manifold absolute pressure (MAP) sensor, which took place in the intake manifold. This paper present a more economical approach on estimating the MAP by using only the measurements of the throttle position and engine speed, resulting in lower implementation cost. The estimation was done via two-stage multilayer feed-forward neural network by combining Levenberg-Marquardt (LM) algorithm, Bayesian Regularization (BR) algorithm and Particle Swarm Optimization (PSO) algorithm. Based on the results found in 20 runs, the second variant of the hybrid algorithm yields a better network performance than the first variant of hybrid algorithm, LM, LM with BR and PSO by estimating the MAP closely to the simulated MAP values. By using a valid experimental training data, the estimator network that trained with the second variant of the hybrid algorithm showed the best performance among other algorithms when used in an actual retrofit fuel injection system (RFIS). The performance of the estimator was also validated in steady-state and transient condition by showing a closer MAP estimation to the actual value. PMID:29190779

Antenna analysis using neural networks

NASA Technical Reports Server (NTRS)

Smith, William T.

1992-01-01

Conventional computing schemes have long been used to analyze problems in electromagnetics (EM). The vast majority of EM applications require computationally intensive algorithms involving numerical integration and solutions to large systems of equations. The feasibility of using neural network computing algorithms for antenna analysis is investigated. The ultimate goal is to use a trained neural network algorithm to reduce the computational demands of existing reflector surface error compensation techniques. Neural networks are computational algorithms based on neurobiological systems. Neural nets consist of massively parallel interconnected nonlinear computational elements. They are often employed in pattern recognition and image processing problems. Recently, neural network analysis has been applied in the electromagnetics area for the design of frequency selective surfaces and beam forming networks. The backpropagation training algorithm was employed to simulate classical antenna array synthesis techniques. The Woodward-Lawson (W-L) and Dolph-Chebyshev (D-C) array pattern synthesis techniques were used to train the neural network. The inputs to the network were samples of the desired synthesis pattern. The outputs are the array element excitations required to synthesize the desired pattern. Once trained, the network is used to simulate the W-L or D-C techniques. Various sector patterns and cosecant-type patterns (27 total) generated using W-L synthesis were used to train the network. Desired pattern samples were then fed to the neural network. The outputs of the network were the simulated W-L excitations. A 20 element linear array was used. There were 41 input pattern samples with 40 output excitations (20 real parts, 20 imaginary). A comparison between the simulated and actual W-L techniques is shown for a triangular-shaped pattern. Dolph-Chebyshev is a different class of synthesis technique in that D-C is used for side lobe control as opposed to pattern shaping. The interesting thing about D-C synthesis is that the side lobes have the same amplitude. Five-element arrays were used. Again, 41 pattern samples were used for the input. Nine actual D-C patterns ranging from -10 dB to -30 dB side lobe levels were used to train the network. A comparison between simulated and actual D-C techniques for a pattern with -22 dB side lobe level is shown. The goal for this research was to evaluate the performance of neural network computing with antennas. Future applications will employ the backpropagation training algorithm to drastically reduce the computational complexity involved in performing EM compensation for surface errors in large space reflector antennas.

Antenna analysis using neural networks

NASA Astrophysics Data System (ADS)

Smith, William T.

1992-09-01

Conventional computing schemes have long been used to analyze problems in electromagnetics (EM). The vast majority of EM applications require computationally intensive algorithms involving numerical integration and solutions to large systems of equations. The feasibility of using neural network computing algorithms for antenna analysis is investigated. The ultimate goal is to use a trained neural network algorithm to reduce the computational demands of existing reflector surface error compensation techniques. Neural networks are computational algorithms based on neurobiological systems. Neural nets consist of massively parallel interconnected nonlinear computational elements. They are often employed in pattern recognition and image processing problems. Recently, neural network analysis has been applied in the electromagnetics area for the design of frequency selective surfaces and beam forming networks. The backpropagation training algorithm was employed to simulate classical antenna array synthesis techniques. The Woodward-Lawson (W-L) and Dolph-Chebyshev (D-C) array pattern synthesis techniques were used to train the neural network. The inputs to the network were samples of the desired synthesis pattern. The outputs are the array element excitations required to synthesize the desired pattern. Once trained, the network is used to simulate the W-L or D-C techniques. Various sector patterns and cosecant-type patterns (27 total) generated using W-L synthesis were used to train the network. Desired pattern samples were then fed to the neural network. The outputs of the network were the simulated W-L excitations. A 20 element linear array was used. There were 41 input pattern samples with 40 output excitations (20 real parts, 20 imaginary).

Noise-enhanced convolutional neural networks.

PubMed

Audhkhasi, Kartik; Osoba, Osonde; Kosko, Bart

2016-06-01

Injecting carefully chosen noise can speed convergence in the backpropagation training of a convolutional neural network (CNN). The Noisy CNN algorithm speeds training on average because the backpropagation algorithm is a special case of the generalized expectation-maximization (EM) algorithm and because such carefully chosen noise always speeds up the EM algorithm on average. The CNN framework gives a practical way to learn and recognize images because backpropagation scales with training data. It has only linear time complexity in the number of training samples. The Noisy CNN algorithm finds a special separating hyperplane in the network's noise space. The hyperplane arises from the likelihood-based positivity condition that noise-boosts the EM algorithm. The hyperplane cuts through a uniform-noise hypercube or Gaussian ball in the noise space depending on the type of noise used. Noise chosen from above the hyperplane speeds training on average. Noise chosen from below slows it on average. The algorithm can inject noise anywhere in the multilayered network. Adding noise to the output neurons reduced the average per-iteration training-set cross entropy by 39% on a standard MNIST image test set of handwritten digits. It also reduced the average per-iteration training-set classification error by 47%. Adding noise to the hidden layers can also reduce these performance measures. The noise benefit is most pronounced for smaller data sets because the largest EM hill-climbing gains tend to occur in the first few iterations. This noise effect can assist random sampling from large data sets because it allows a smaller random sample to give the same or better performance than a noiseless sample gives. Copyright © 2015 Elsevier Ltd. All rights reserved.

Brain-Inspired Constructive Learning Algorithms with Evolutionally Additive Nonlinear Neurons

NASA Astrophysics Data System (ADS)

Fang, Le-Heng; Lin, Wei; Luo, Qiang

In this article, inspired partially by the physiological evidence of brain’s growth and development, we developed a new type of constructive learning algorithm with evolutionally additive nonlinear neurons. The new algorithms have remarkable ability in effective regression and accurate classification. In particular, the algorithms are able to sustain a certain reduction of the loss function when the dynamics of the trained network are bogged down in the vicinity of the local minima. The algorithm augments the neural network by adding only a few connections as well as neurons whose activation functions are nonlinear, nonmonotonic, and self-adapted to the dynamics of the loss functions. Indeed, we analytically demonstrate the reduction dynamics of the algorithm for different problems, and further modify the algorithms so as to obtain an improved generalization capability for the augmented neural networks. Finally, through comparing with the classical algorithm and architecture for neural network construction, we show that our constructive learning algorithms as well as their modified versions have better performances, such as faster training speed and smaller network size, on several representative benchmark datasets including the MNIST dataset for handwriting digits.

Convergence and objective functions of some fault/noise-injection-based online learning algorithms for RBF networks.

PubMed

Ho, Kevin I-J; Leung, Chi-Sing; Sum, John

2010-06-01

In the last two decades, many online fault/noise injection algorithms have been developed to attain a fault tolerant neural network. However, not much theoretical works related to their convergence and objective functions have been reported. This paper studies six common fault/noise-injection-based online learning algorithms for radial basis function (RBF) networks, namely 1) injecting additive input noise, 2) injecting additive/multiplicative weight noise, 3) injecting multiplicative node noise, 4) injecting multiweight fault (random disconnection of weights), 5) injecting multinode fault during training, and 6) weight decay with injecting multinode fault. Based on the Gladyshev theorem, we show that the convergence of these six online algorithms is almost sure. Moreover, their true objective functions being minimized are derived. For injecting additive input noise during training, the objective function is identical to that of the Tikhonov regularizer approach. For injecting additive/multiplicative weight noise during training, the objective function is the simple mean square training error. Thus, injecting additive/multiplicative weight noise during training cannot improve the fault tolerance of an RBF network. Similar to injective additive input noise, the objective functions of other fault/noise-injection-based online algorithms contain a mean square error term and a specialized regularization term.

Neural network for image compression

NASA Astrophysics Data System (ADS)

Panchanathan, Sethuraman; Yeap, Tet H.; Pilache, B.

1992-09-01

In this paper, we propose a new scheme for image compression using neural networks. Image data compression deals with minimization of the amount of data required to represent an image while maintaining an acceptable quality. Several image compression techniques have been developed in recent years. We note that the coding performance of these techniques may be improved by employing adaptivity. Over the last few years neural network has emerged as an effective tool for solving a wide range of problems involving adaptivity and learning. A multilayer feed-forward neural network trained using the backward error propagation algorithm is used in many applications. However, this model is not suitable for image compression because of its poor coding performance. Recently, a self-organizing feature map (SOFM) algorithm has been proposed which yields a good coding performance. However, this algorithm requires a long training time because the network starts with random initial weights. In this paper we have used the backward error propagation algorithm (BEP) to quickly obtain the initial weights which are then used to speedup the training time required by the SOFM algorithm. The proposed approach (BEP-SOFM) combines the advantages of the two techniques and, hence, achieves a good coding performance in a shorter training time. Our simulation results demonstrate the potential gains using the proposed technique.

New recursive-least-squares algorithms for nonlinear active control of sound and vibration using neural networks.

PubMed

Bouchard, M

2001-01-01

In recent years, a few articles describing the use of neural networks for nonlinear active control of sound and vibration were published. Using a control structure with two multilayer feedforward neural networks (one as a nonlinear controller and one as a nonlinear plant model), steepest descent algorithms based on two distinct gradient approaches were introduced for the training of the controller network. The two gradient approaches were sometimes called the filtered-x approach and the adjoint approach. Some recursive-least-squares algorithms were also introduced, using the adjoint approach. In this paper, an heuristic procedure is introduced for the development of recursive-least-squares algorithms based on the filtered-x and the adjoint gradient approaches. This leads to the development of new recursive-least-squares algorithms for the training of the controller neural network in the two networks structure. These new algorithms produce a better convergence performance than previously published algorithms. Differences in the performance of algorithms using the filtered-x and the adjoint gradient approaches are discussed in the paper. The computational load of the algorithms discussed in the paper is evaluated for multichannel systems of nonlinear active control. Simulation results are presented to compare the convergence performance of the algorithms, showing the convergence gain provided by the new algorithms.

Predicting Positive and Negative Relationships in Large Social Networks.

PubMed

Wang, Guan-Nan; Gao, Hui; Chen, Lian; Mensah, Dennis N A; Fu, Yan

2015-01-01

In a social network, users hold and express positive and negative attitudes (e.g. support/opposition) towards other users. Those attitudes exhibit some kind of binary relationships among the users, which play an important role in social network analysis. However, some of those binary relationships are likely to be latent as the scale of social network increases. The essence of predicting latent binary relationships have recently began to draw researchers' attention. In this paper, we propose a machine learning algorithm for predicting positive and negative relationships in social networks inspired by structural balance theory and social status theory. More specifically, we show that when two users in the network have fewer common neighbors, the prediction accuracy of the relationship between them deteriorates. Accordingly, in the training phase, we propose a segment-based training framework to divide the training data into two subsets according to the number of common neighbors between users, and build a prediction model for each subset based on support vector machine (SVM). Moreover, to deal with large-scale social network data, we employ a sampling strategy that selects small amount of training data while maintaining high accuracy of prediction. We compare our algorithm with traditional algorithms and adaptive boosting of them. Experimental results of typical data sets show that our algorithm can deal with large social networks and consistently outperforms other methods.

Neural network-based run-to-run controller using exposure and resist thickness adjustment

NASA Astrophysics Data System (ADS)

Geary, Shane; Barry, Ronan

2003-06-01

This paper describes the development of a run-to-run control algorithm using a feedforward neural network, trained using the backpropagation training method. The algorithm is used to predict the critical dimension of the next lot using previous lot information. It is compared to a common prediction algorithm - the exponentially weighted moving average (EWMA) and is shown to give superior prediction performance in simulations. The manufacturing implementation of the final neural network showed significantly improved process capability when compared to the case where no run-to-run control was utilised.

Bringing Interpretability and Visualization with Artificial Neural Networks

ERIC Educational Resources Information Center

Gritsenko, Andrey

2017-01-01

Extreme Learning Machine (ELM) is a training algorithm for Single-Layer Feed-forward Neural Network (SLFN). The difference in theory of ELM from other training algorithms is in the existence of explicitly-given solution due to the immutability of initialed weights. In practice, ELMs achieve performance similar to that of other state-of-the-art…

Algorithm for Training a Recurrent Multilayer Perceptron

NASA Technical Reports Server (NTRS)

Parlos, Alexander G.; Rais, Omar T.; Menon, Sunil K.; Atiya, Amir F.

2004-01-01

An improved algorithm has been devised for training a recurrent multilayer perceptron (RMLP) for optimal performance in predicting the behavior of a complex, dynamic, and noisy system multiple time steps into the future. [An RMLP is a computational neural network with self-feedback and cross-talk (both delayed by one time step) among neurons in hidden layers]. Like other neural-network-training algorithms, this algorithm adjusts network biases and synaptic-connection weights according to a gradient-descent rule. The distinguishing feature of this algorithm is a combination of global feedback (the use of predictions as well as the current output value in computing the gradient at each time step) and recursiveness. The recursive aspect of the algorithm lies in the inclusion of the gradient of predictions at each time step with respect to the predictions at the preceding time step; this recursion enables the RMLP to learn the dynamics. It has been conjectured that carrying the recursion to even earlier time steps would enable the RMLP to represent a noisier, more complex system.

CONEDEP: COnvolutional Neural network based Earthquake DEtection and Phase Picking

NASA Astrophysics Data System (ADS)

Zhou, Y.; Huang, Y.; Yue, H.; Zhou, S.; An, S.; Yun, N.

2017-12-01

We developed an automatic local earthquake detection and phase picking algorithm based on Fully Convolutional Neural network (FCN). The FCN algorithm detects and segments certain features (phases) in 3 component seismograms to realize efficient picking. We use STA/LTA algorithm and template matching algorithm to construct the training set from seismograms recorded 1 month before and after the Wenchuan earthquake. Precise P and S phases are identified and labeled to construct the training set. Noise data are produced by combining back-ground noise and artificial synthetic noise to form the equivalent scale of noise set as the signal set. Training is performed on GPUs to achieve efficient convergence. Our algorithm has significantly improved performance in terms of the detection rate and precision in comparison with STA/LTA and template matching algorithms.

Railway obstacle detection algorithm using neural network

NASA Astrophysics Data System (ADS)

Yu, Mingyang; Yang, Peng; Wei, Sen

2018-05-01

Aiming at the difficulty of detection of obstacle in outdoor railway scene, a data-oriented method based on neural network to obtain image objects is proposed. First, we mark objects of images(such as people, trains, animals) acquired on the Internet. and then use the residual learning units to build Fast R-CNN framework. Then, the neural network is trained to get the target image characteristics by using stochastic gradient descent algorithm. Finally, a well-trained model is used to identify an outdoor railway image. if it includes trains and other objects, it will issue an alert. Experiments show that the correct rate of warning reached 94.85%.

Efficient Training of Supervised Spiking Neural Network via Accurate Synaptic-Efficiency Adjustment Method.

PubMed

Xie, Xiurui; Qu, Hong; Yi, Zhang; Kurths, Jurgen

2017-06-01

The spiking neural network (SNN) is the third generation of neural networks and performs remarkably well in cognitive tasks, such as pattern recognition. The temporal neural encode mechanism found in biological hippocampus enables SNN to possess more powerful computation capability than networks with other encoding schemes. However, this temporal encoding approach requires neurons to process information serially on time, which reduces learning efficiency significantly. To keep the powerful computation capability of the temporal encoding mechanism and to overcome its low efficiency in the training of SNNs, a new training algorithm, the accurate synaptic-efficiency adjustment method is proposed in this paper. Inspired by the selective attention mechanism of the primate visual system, our algorithm selects only the target spike time as attention areas, and ignores voltage states of the untarget ones, resulting in a significant reduction of training time. Besides, our algorithm employs a cost function based on the voltage difference between the potential of the output neuron and the firing threshold of the SNN, instead of the traditional precise firing time distance. A normalized spike-timing-dependent-plasticity learning window is applied to assigning this error to different synapses for instructing their training. Comprehensive simulations are conducted to investigate the learning properties of our algorithm, with input neurons emitting both single spike and multiple spikes. Simulation results indicate that our algorithm possesses higher learning performance than the existing other methods and achieves the state-of-the-art efficiency in the training of SNN.

An Efficient Supervised Training Algorithm for Multilayer Spiking Neural Networks

PubMed Central

Xie, Xiurui; Qu, Hong; Liu, Guisong; Zhang, Malu; Kurths, Jürgen

2016-01-01

The spiking neural networks (SNNs) are the third generation of neural networks and perform remarkably well in cognitive tasks such as pattern recognition. The spike emitting and information processing mechanisms found in biological cognitive systems motivate the application of the hierarchical structure and temporal encoding mechanism in spiking neural networks, which have exhibited strong computational capability. However, the hierarchical structure and temporal encoding approach require neurons to process information serially in space and time respectively, which reduce the training efficiency significantly. For training the hierarchical SNNs, most existing methods are based on the traditional back-propagation algorithm, inheriting its drawbacks of the gradient diffusion and the sensitivity on parameters. To keep the powerful computation capability of the hierarchical structure and temporal encoding mechanism, but to overcome the low efficiency of the existing algorithms, a new training algorithm, the Normalized Spiking Error Back Propagation (NSEBP) is proposed in this paper. In the feedforward calculation, the output spike times are calculated by solving the quadratic function in the spike response model instead of detecting postsynaptic voltage states at all time points in traditional algorithms. Besides, in the feedback weight modification, the computational error is propagated to previous layers by the presynaptic spike jitter instead of the gradient decent rule, which realizes the layer-wised training. Furthermore, our algorithm investigates the mathematical relation between the weight variation and voltage error change, which makes the normalization in the weight modification applicable. Adopting these strategies, our algorithm outperforms the traditional SNN multi-layer algorithms in terms of learning efficiency and parameter sensitivity, that are also demonstrated by the comprehensive experimental results in this paper. PMID:27044001

Synthesis of recurrent neural networks for dynamical system simulation.

PubMed

Trischler, Adam P; D'Eleuterio, Gabriele M T

2016-08-01

We review several of the most widely used techniques for training recurrent neural networks to approximate dynamical systems, then describe a novel algorithm for this task. The algorithm is based on an earlier theoretical result that guarantees the quality of the network approximation. We show that a feedforward neural network can be trained on the vector-field representation of a given dynamical system using backpropagation, then recast it as a recurrent network that replicates the original system's dynamics. After detailing this algorithm and its relation to earlier approaches, we present numerical examples that demonstrate its capabilities. One of the distinguishing features of our approach is that both the original dynamical systems and the recurrent networks that simulate them operate in continuous time. Copyright © 2016 Elsevier Ltd. All rights reserved.

A novel constructive-optimizer neural network for the traveling salesman problem.

PubMed

Saadatmand-Tarzjan, Mahdi; Khademi, Morteza; Akbarzadeh-T, Mohammad-R; Moghaddam, Hamid Abrishami

2007-08-01

In this paper, a novel constructive-optimizer neural network (CONN) is proposed for the traveling salesman problem (TSP). CONN uses a feedback structure similar to Hopfield-type neural networks and a competitive training algorithm similar to the Kohonen-type self-organizing maps (K-SOMs). Consequently, CONN is composed of a constructive part, which grows the tour and an optimizer part to optimize it. In the training algorithm, an initial tour is created first and introduced to CONN. Then, it is trained in the constructive phase for adding a number of cities to the tour. Next, the training algorithm switches to the optimizer phase for optimizing the current tour by displacing the tour cities. After convergence in this phase, the training algorithm switches to the constructive phase anew and is continued until all cities are added to the tour. Furthermore, we investigate a relationship between the number of TSP cities and the number of cities to be added in each constructive phase. CONN was tested on nine sets of benchmark TSPs from TSPLIB to demonstrate its performance and efficiency. It performed better than several typical Neural networks (NNs), including KNIES_TSP_Local, KNIES_TSP_Global, Budinich's SOM, Co-Adaptive Net, and multivalued Hopfield network as wall as computationally comparable variants of the simulated annealing algorithm, in terms of both CPU time and accuracy. Furthermore, CONN converged considerably faster than expanding SOM and evolved integrated SOM and generated shorter tours compared to KNIES_DECOMPOSE. Although CONN is not yet comparable in terms of accuracy with some sophisticated computationally intensive algorithms, it converges significantly faster than they do. Generally speaking, CONN provides the best compromise between CPU time and accuracy among currently reported NNs for TSP.

A decentralized training algorithm for Echo State Networks in distributed big data applications.

PubMed

Scardapane, Simone; Wang, Dianhui; Panella, Massimo

2016-06-01

The current big data deluge requires innovative solutions for performing efficient inference on large, heterogeneous amounts of information. Apart from the known challenges deriving from high volume and velocity, real-world big data applications may impose additional technological constraints, including the need for a fully decentralized training architecture. While several alternatives exist for training feed-forward neural networks in such a distributed setting, less attention has been devoted to the case of decentralized training of recurrent neural networks (RNNs). In this paper, we propose such an algorithm for a class of RNNs known as Echo State Networks. The algorithm is based on the well-known Alternating Direction Method of Multipliers optimization procedure. It is formulated only in terms of local exchanges between neighboring agents, without reliance on a coordinating node. Additionally, it does not require the communication of training patterns, which is a crucial component in realistic big data implementations. Experimental results on large scale artificial datasets show that it compares favorably with a fully centralized implementation, in terms of speed, efficiency and generalization accuracy. Copyright © 2015 Elsevier Ltd. All rights reserved.

T-wave end detection using neural networks and Support Vector Machines.

PubMed

Suárez-León, Alexander Alexeis; Varon, Carolina; Willems, Rik; Van Huffel, Sabine; Vázquez-Seisdedos, Carlos Román

2018-05-01

In this paper we propose a new approach for detecting the end of the T-wave in the electrocardiogram (ECG) using Neural Networks and Support Vector Machines. Both, Multilayer Perceptron (MLP) neural networks and Fixed-Size Least-Squares Support Vector Machines (FS-LSSVM) were used as regression algorithms to determine the end of the T-wave. Different strategies for selecting the training set such as random selection, k-means, robust clustering and maximum quadratic (Rényi) entropy were evaluated. Individual parameters were tuned for each method during training and the results are given for the evaluation set. A comparison between MLP and FS-LSSVM approaches was performed. Finally, a fair comparison of the FS-LSSVM method with other state-of-the-art algorithms for detecting the end of the T-wave was included. The experimental results show that FS-LSSVM approaches are more suitable as regression algorithms than MLP neural networks. Despite the small training sets used, the FS-LSSVM methods outperformed the state-of-the-art techniques. FS-LSSVM can be successfully used as a T-wave end detection algorithm in ECG even with small training set sizes. Copyright © 2018 Elsevier Ltd. All rights reserved.

Photoacoustic image reconstruction via deep learning

NASA Astrophysics Data System (ADS)

Antholzer, Stephan; Haltmeier, Markus; Nuster, Robert; Schwab, Johannes

2018-02-01

Applying standard algorithms to sparse data problems in photoacoustic tomography (PAT) yields low-quality images containing severe under-sampling artifacts. To some extent, these artifacts can be reduced by iterative image reconstruction algorithms which allow to include prior knowledge such as smoothness, total variation (TV) or sparsity constraints. These algorithms tend to be time consuming as the forward and adjoint problems have to be solved repeatedly. Further, iterative algorithms have additional drawbacks. For example, the reconstruction quality strongly depends on a-priori model assumptions about the objects to be recovered, which are often not strictly satisfied in practical applications. To overcome these issues, in this paper, we develop direct and efficient reconstruction algorithms based on deep learning. As opposed to iterative algorithms, we apply a convolutional neural network, whose parameters are trained before the reconstruction process based on a set of training data. For actual image reconstruction, a single evaluation of the trained network yields the desired result. Our presented numerical results (using two different network architectures) demonstrate that the proposed deep learning approach reconstructs images with a quality comparable to state of the art iterative reconstruction methods.

A hybrid linear/nonlinear training algorithm for feedforward neural networks.

PubMed

McLoone, S; Brown, M D; Irwin, G; Lightbody, A

1998-01-01

This paper presents a new hybrid optimization strategy for training feedforward neural networks. The algorithm combines gradient-based optimization of nonlinear weights with singular value decomposition (SVD) computation of linear weights in one integrated routine. It is described for the multilayer perceptron (MLP) and radial basis function (RBF) networks and then extended to the local model network (LMN), a new feedforward structure in which a global nonlinear model is constructed from a set of locally valid submodels. Simulation results are presented demonstrating the superiority of the new hybrid training scheme compared to second-order gradient methods. It is particularly effective for the LMN architecture where the linear to nonlinear parameter ratio is large.

Matching algorithm of missile tail flame based on back-propagation neural network

NASA Astrophysics Data System (ADS)

Huang, Da; Huang, Shucai; Tang, Yidong; Zhao, Wei; Cao, Wenhuan

2018-02-01

This work presents a spectral matching algorithm of missile plume detection that based on neural network. The radiation value of the characteristic spectrum of the missile tail flame is taken as the input of the network. The network's structure including the number of nodes and layers is determined according to the number of characteristic spectral bands and missile types. We can get the network weight matrixes and threshold vectors through training the network using training samples, and we can determine the performance of the network through testing the network using the test samples. A small amount of data cause the network has the advantages of simple structure and practicality. Network structure composed of weight matrix and threshold vector can complete task of spectrum matching without large database support. Network can achieve real-time requirements with a small quantity of data. Experiment results show that the algorithm has the ability to match the precise spectrum and strong robustness.

System identification of an unmanned quadcopter system using MRAN neural

NASA Astrophysics Data System (ADS)

Pairan, M. F.; Shamsudin, S. S.

2017-12-01

This project presents the performance analysis of the radial basis function neural network (RBF) trained with Minimal Resource Allocating Network (MRAN) algorithm for real-time identification of quadcopter. MRAN’s performance is compared with the RBF with Constant Trace algorithm for 2500 input-output pair data sampling. MRAN utilizes adding and pruning hidden neuron strategy to obtain optimum RBF structure, increase prediction accuracy and reduce training time. The results indicate that MRAN algorithm produces fast training time and more accurate prediction compared with standard RBF. The model proposed in this paper is capable of identifying and modelling a nonlinear representation of the quadcopter flight dynamics.

Evaluation of a parallel implementation of the learning portion of the backward error propagation neural network: experiments in artifact identification.

PubMed Central

Sittig, D. F.; Orr, J. A.

1991-01-01

Various methods have been proposed in an attempt to solve problems in artifact and/or alarm identification including expert systems, statistical signal processing techniques, and artificial neural networks (ANN). ANNs consist of a large number of simple processing units connected by weighted links. To develop truly robust ANNs, investigators are required to train their networks on huge training data sets, requiring enormous computing power. We implemented a parallel version of the backward error propagation neural network training algorithm in the widely portable parallel programming language C-Linda. A maximum speedup of 4.06 was obtained with six processors. This speedup represents a reduction in total run-time from approximately 6.4 hours to 1.5 hours. We conclude that use of the master-worker model of parallel computation is an excellent method for obtaining speedups in the backward error propagation neural network training algorithm. PMID:1807607

Two algorithms for neural-network design and training with application to channel equalization.

PubMed

Sweatman, C Z; Mulgrew, B; Gibson, G J

1998-01-01

We describe two algorithms for designing and training neural-network classifiers. The first, the linear programming slab algorithm (LPSA), is motivated by the problem of reconstructing digital signals corrupted by passage through a dispersive channel and by additive noise. It constructs a multilayer perceptron (MLP) to separate two disjoint sets by using linear programming methods to identify network parameters. The second, the perceptron learning slab algorithm (PLSA), avoids the computational costs of linear programming by using an error-correction approach to identify parameters. Both algorithms operate in highly constrained parameter spaces and are able to exploit symmetry in the classification problem. Using these algorithms, we develop a number of procedures for the adaptive equalization of a complex linear 4-quadrature amplitude modulation (QAM) channel, and compare their performance in a simulation study. Results are given for both stationary and time-varying channels, the latter based on the COST 207 GSM propagation model.

Logistics Distribution Center Location Evaluation Based on Genetic Algorithm and Fuzzy Neural Network

NASA Astrophysics Data System (ADS)

Shao, Yuxiang; Chen, Qing; Wei, Zhenhua

Logistics distribution center location evaluation is a dynamic, fuzzy, open and complicated nonlinear system, which makes it difficult to evaluate the distribution center location by the traditional analysis method. The paper proposes a distribution center location evaluation system which uses the fuzzy neural network combined with the genetic algorithm. In this model, the neural network is adopted to construct the fuzzy system. By using the genetic algorithm, the parameters of the neural network are optimized and trained so as to improve the fuzzy system’s abilities of self-study and self-adaptation. At last, the sampled data are trained and tested by Matlab software. The simulation results indicate that the proposed identification model has very small errors.

Interacting neural networks.

PubMed

Metzler, R; Kinzel, W; Kanter, I

2000-08-01

Several scenarios of interacting neural networks which are trained either in an identical or in a competitive way are solved analytically. In the case of identical training each perceptron receives the output of its neighbor. The symmetry of the stationary state as well as the sensitivity to the used training algorithm are investigated. Two competitive perceptrons trained on mutually exclusive learning aims and a perceptron which is trained on the opposite of its own output are examined analytically. An ensemble of competitive perceptrons is used as decision-making algorithms in a model of a closed market (El Farol Bar problem or the Minority Game. In this game, a set of agents who have to make a binary decision is considered.); each network is trained on the history of minority decisions. This ensemble of perceptrons relaxes to a stationary state whose performance can be better than random.

Interacting neural networks

NASA Astrophysics Data System (ADS)

Metzler, R.; Kinzel, W.; Kanter, I.

2000-08-01

Several scenarios of interacting neural networks which are trained either in an identical or in a competitive way are solved analytically. In the case of identical training each perceptron receives the output of its neighbor. The symmetry of the stationary state as well as the sensitivity to the used training algorithm are investigated. Two competitive perceptrons trained on mutually exclusive learning aims and a perceptron which is trained on the opposite of its own output are examined analytically. An ensemble of competitive perceptrons is used as decision-making algorithms in a model of a closed market (El Farol Bar problem or the Minority Game. In this game, a set of agents who have to make a binary decision is considered.); each network is trained on the history of minority decisions. This ensemble of perceptrons relaxes to a stationary state whose performance can be better than random.

On the improvement of neural cryptography using erroneous transmitted information with error prediction.

PubMed

Allam, Ahmed M; Abbas, Hazem M

2010-12-01

Neural cryptography deals with the problem of "key exchange" between two neural networks using the mutual learning concept. The two networks exchange their outputs (in bits) and the key between the two communicating parties is eventually represented in the final learned weights, when the two networks are said to be synchronized. Security of neural synchronization is put at risk if an attacker is capable of synchronizing with any of the two parties during the training process. Therefore, diminishing the probability of such a threat improves the reliability of exchanging the output bits through a public channel. The synchronization with feedback algorithm is one of the existing algorithms that enhances the security of neural cryptography. This paper proposes three new algorithms to enhance the mutual learning process. They mainly depend on disrupting the attacker confidence in the exchanged outputs and input patterns during training. The first algorithm is called "Do not Trust My Partner" (DTMP), which relies on one party sending erroneous output bits, with the other party being capable of predicting and correcting this error. The second algorithm is called "Synchronization with Common Secret Feedback" (SCSFB), where inputs are kept partially secret and the attacker has to train its network on input patterns that are different from the training sets used by the communicating parties. The third algorithm is a hybrid technique combining the features of the DTMP and SCSFB. The proposed approaches are shown to outperform the synchronization with feedback algorithm in the time needed for the parties to synchronize.

On-line training of recurrent neural networks with continuous topology adaptation.

PubMed

Obradovic, D

1996-01-01

This paper presents an online procedure for training dynamic neural networks with input-output recurrences whose topology is continuously adjusted to the complexity of the target system dynamics. This is accomplished by changing the number of the elements of the network hidden layer whenever the existing topology cannot capture the dynamics presented by the new data. The training mechanism is based on the suitably altered extended Kalman filter (EKF) algorithm which is simultaneously used for the network parameter adjustment and for its state estimation. The network consists of a single hidden layer with Gaussian radial basis functions (GRBF), and a linear output layer. The choice of the GRBF is induced by the requirements of the online learning. The latter implies the network architecture which permits only local influence of the new data point in order not to forget the previously learned dynamics. The continuous topology adaptation is implemented in our algorithm to avoid memory and computational problems of using a regular grid of GRBF'S which covers the network input space. Furthermore, we show that the resulting parameter increase can be handled "smoothly" without interfering with the already acquired information. If the target system dynamics are changing over time, we show that a suitable forgetting factor can be used to "unlearn" the no longer-relevant dynamics. The quality of the recurrent network training algorithm is demonstrated on the identification of nonlinear dynamic systems.

Reduced kernel recursive least squares algorithm for aero-engine degradation prediction

NASA Astrophysics Data System (ADS)

Zhou, Haowen; Huang, Jinquan; Lu, Feng

2017-10-01

Kernel adaptive filters (KAFs) generate a linear growing radial basis function (RBF) network with the number of training samples, thereby lacking sparseness. To deal with this drawback, traditional sparsification techniques select a subset of original training data based on a certain criterion to train the network and discard the redundant data directly. Although these methods curb the growth of the network effectively, it should be noted that information conveyed by these redundant samples is omitted, which may lead to accuracy degradation. In this paper, we present a novel online sparsification method which requires much less training time without sacrificing the accuracy performance. Specifically, a reduced kernel recursive least squares (RKRLS) algorithm is developed based on the reduced technique and the linear independency. Unlike conventional methods, our novel methodology employs these redundant data to update the coefficients of the existing network. Due to the effective utilization of the redundant data, the novel algorithm achieves a better accuracy performance, although the network size is significantly reduced. Experiments on time series prediction and online regression demonstrate that RKRLS algorithm requires much less computational consumption and maintains the satisfactory accuracy performance. Finally, we propose an enhanced multi-sensor prognostic model based on RKRLS and Hidden Markov Model (HMM) for remaining useful life (RUL) estimation. A case study in a turbofan degradation dataset is performed to evaluate the performance of the novel prognostic approach.

Practical training framework for fitting a function and its derivatives.

PubMed

Pukrittayakamee, Arjpolson; Hagan, Martin; Raff, Lionel; Bukkapatnam, Satish T S; Komanduri, Ranga

2011-06-01

This paper describes a practical framework for using multilayer feedforward neural networks to simultaneously fit both a function and its first derivatives. This framework involves two steps. The first step is to train the network to optimize a performance index, which includes both the error in fitting the function and the error in fitting the derivatives. The second step is to prune the network by removing neurons that cause overfitting and then to retrain it. This paper describes two novel types of overfitting that are only observed when simultaneously fitting both a function and its first derivatives. A new pruning algorithm is proposed to eliminate these types of overfitting. Experimental results show that the pruning algorithm successfully eliminates the overfitting and produces the smoothest responses and the best generalization among all the training algorithms that we have tested.

A conjugate gradients/trust regions algorithms for training multilayer perceptrons for nonlinear mapping

NASA Technical Reports Server (NTRS)

Madyastha, Raghavendra K.; Aazhang, Behnaam; Henson, Troy F.; Huxhold, Wendy L.

1992-01-01

This paper addresses the issue of applying a globally convergent optimization algorithm to the training of multilayer perceptrons, a class of Artificial Neural Networks. The multilayer perceptrons are trained towards the solution of two highly nonlinear problems: (1) signal detection in a multi-user communication network, and (2) solving the inverse kinematics for a robotic manipulator. The research is motivated by the fact that a multilayer perceptron is theoretically capable of approximating any nonlinear function to within a specified accuracy. The algorithm that has been employed in this study combines the merits of two well known optimization algorithms, the Conjugate Gradients and the Trust Regions Algorithms. The performance is compared to a widely used algorithm, the Backpropagation Algorithm, that is basically a gradient-based algorithm, and hence, slow in converging. The performances of the two algorithms are compared with the convergence rate. Furthermore, in the case of the signal detection problem, performances are also benchmarked by the decision boundaries drawn as well as the probability of error obtained in either case.

Implementations of back propagation algorithm in ecosystems applications

NASA Astrophysics Data System (ADS)

Ali, Khalda F.; Sulaiman, Riza; Elamir, Amir Mohamed

2015-05-01

Artificial Neural Networks (ANNs) have been applied to an increasing number of real world problems of considerable complexity. Their most important advantage is in solving problems which are too complex for conventional technologies, that do not have an algorithmic solutions or their algorithmic Solutions is too complex to be found. In general, because of their abstraction from the biological brain, ANNs are developed from concept that evolved in the late twentieth century neuro-physiological experiments on the cells of the human brain to overcome the perceived inadequacies with conventional ecological data analysis methods. ANNs have gained increasing attention in ecosystems applications, because of ANN's capacity to detect patterns in data through non-linear relationships, this characteristic confers them a superior predictive ability. In this research, ANNs is applied in an ecological system analysis. The neural networks use the well known Back Propagation (BP) Algorithm with the Delta Rule for adaptation of the system. The Back Propagation (BP) training Algorithm is an effective analytical method for adaptation of the ecosystems applications, the main reason because of their capacity to detect patterns in data through non-linear relationships. This characteristic confers them a superior predicting ability. The BP algorithm uses supervised learning, which means that we provide the algorithm with examples of the inputs and outputs we want the network to compute, and then the error is calculated. The idea of the back propagation algorithm is to reduce this error, until the ANNs learns the training data. The training begins with random weights, and the goal is to adjust them so that the error will be minimal. This research evaluated the use of artificial neural networks (ANNs) techniques in an ecological system analysis and modeling. The experimental results from this research demonstrate that an artificial neural network system can be trained to act as an expert ecosystem analyzer for many applications in ecological fields. The pilot ecosystem analyzer shows promising ability for generalization and requires further tuning and refinement of the basis neural network system for optimal performance.

Artificial Neural Network Approach in Laboratory Test Reporting:  Learning Algorithms.

PubMed

Demirci, Ferhat; Akan, Pinar; Kume, Tuncay; Sisman, Ali Riza; Erbayraktar, Zubeyde; Sevinc, Suleyman

2016-08-01

In the field of laboratory medicine, minimizing errors and establishing standardization is only possible by predefined processes. The aim of this study was to build an experimental decision algorithm model open to improvement that would efficiently and rapidly evaluate the results of biochemical tests with critical values by evaluating multiple factors concurrently. The experimental model was built by Weka software (Weka, Waikato, New Zealand) based on the artificial neural network method. Data were received from Dokuz Eylül University Central Laboratory. "Training sets" were developed for our experimental model to teach the evaluation criteria. After training the system, "test sets" developed for different conditions were used to statistically assess the validity of the model. After developing the decision algorithm with three iterations of training, no result was verified that was refused by the laboratory specialist. The sensitivity of the model was 91% and specificity was 100%. The estimated κ score was 0.950. This is the first study based on an artificial neural network to build an experimental assessment and decision algorithm model. By integrating our trained algorithm model into a laboratory information system, it may be possible to reduce employees' workload without compromising patient safety. © American Society for Clinical Pathology, 2016. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

A neural network approach for image reconstruction in electron magnetic resonance tomography.

PubMed

Durairaj, D Christopher; Krishna, Murali C; Murugesan, Ramachandran

2007-10-01

An object-oriented, artificial neural network (ANN) based, application system for reconstruction of two-dimensional spatial images in electron magnetic resonance (EMR) tomography is presented. The standard back propagation algorithm is utilized to train a three-layer sigmoidal feed-forward, supervised, ANN to perform the image reconstruction. The network learns the relationship between the 'ideal' images that are reconstructed using filtered back projection (FBP) technique and the corresponding projection data (sinograms). The input layer of the network is provided with a training set that contains projection data from various phantoms as well as in vivo objects, acquired from an EMR imager. Twenty five different network configurations are investigated to test the ability of the generalization of the network. The trained ANN then reconstructs two-dimensional temporal spatial images that present the distribution of free radicals in biological systems. Image reconstruction by the trained neural network shows better time complexity than the conventional iterative reconstruction algorithms such as multiplicative algebraic reconstruction technique (MART). The network is further explored for image reconstruction from 'noisy' EMR data and the results show better performance than the FBP method. The network is also tested for its ability to reconstruct from limited-angle EMR data set.

Training Recurrent Neural Networks With the Levenberg-Marquardt Algorithm for Optimal Control of a Grid-Connected Converter.

PubMed

Fu, Xingang; Li, Shuhui; Fairbank, Michael; Wunsch, Donald C; Alonso, Eduardo

2015-09-01

This paper investigates how to train a recurrent neural network (RNN) using the Levenberg-Marquardt (LM) algorithm as well as how to implement optimal control of a grid-connected converter (GCC) using an RNN. To successfully and efficiently train an RNN using the LM algorithm, a new forward accumulation through time (FATT) algorithm is proposed to calculate the Jacobian matrix required by the LM algorithm. This paper explores how to incorporate FATT into the LM algorithm. The results show that the combination of the LM and FATT algorithms trains RNNs better than the conventional backpropagation through time algorithm. This paper presents an analytical study on the optimal control of GCCs, including theoretically ideal optimal and suboptimal controllers. To overcome the inapplicability of the optimal GCC controller under practical conditions, a new RNN controller with an improved input structure is proposed to approximate the ideal optimal controller. The performance of an ideal optimal controller and a well-trained RNN controller was compared in close to real-life power converter switching environments, demonstrating that the proposed RNN controller can achieve close to ideal optimal control performance even under low sampling rate conditions. The excellent performance of the proposed RNN controller under challenging and distorted system conditions further indicates the feasibility of using an RNN to approximate optimal control in practical applications.

Three learning phases for radial-basis-function networks.

PubMed

Schwenker, F; Kestler, H A; Palm, G

2001-05-01

In this paper, learning algorithms for radial basis function (RBF) networks are discussed. Whereas multilayer perceptrons (MLP) are typically trained with backpropagation algorithms, starting the training procedure with a random initialization of the MLP's parameters, an RBF network may be trained in many different ways. We categorize these RBF training methods into one-, two-, and three-phase learning schemes. Two-phase RBF learning is a very common learning scheme. The two layers of an RBF network are learnt separately; first the RBF layer is trained, including the adaptation of centers and scaling parameters, and then the weights of the output layer are adapted. RBF centers may be trained by clustering, vector quantization and classification tree algorithms, and the output layer by supervised learning (through gradient descent or pseudo inverse solution). Results from numerical experiments of RBF classifiers trained by two-phase learning are presented in three completely different pattern recognition applications: (a) the classification of 3D visual objects; (b) the recognition hand-written digits (2D objects); and (c) the categorization of high-resolution electrocardiograms given as a time series (ID objects) and as a set of features extracted from these time series. In these applications, it can be observed that the performance of RBF classifiers trained with two-phase learning can be improved through a third backpropagation-like training phase of the RBF network, adapting the whole set of parameters (RBF centers, scaling parameters, and output layer weights) simultaneously. This, we call three-phase learning in RBF networks. A practical advantage of two- and three-phase learning in RBF networks is the possibility to use unlabeled training data for the first training phase. Support vector (SV) learning in RBF networks is a different learning approach. SV learning can be considered, in this context of learning, as a special type of one-phase learning, where only the output layer weights of the RBF network are calculated, and the RBF centers are restricted to be a subset of the training data. Numerical experiments with several classifier schemes including k-nearest-neighbor, learning vector quantization and RBF classifiers trained through two-phase, three-phase and support vector learning are given. The performance of the RBF classifiers trained through SV learning and three-phase learning are superior to the results of two-phase learning, but SV learning often leads to complex network structures, since the number of support vectors is not a small fraction of the total number of data points.

Research on particle swarm optimization algorithm based on optimal movement probability

NASA Astrophysics Data System (ADS)

Ma, Jianhong; Zhang, Han; He, Baofeng

2017-01-01

The particle swarm optimization algorithm to improve the control precision, and has great application value training neural network and fuzzy system control fields etc.The traditional particle swarm algorithm is used for the training of feed forward neural networks,the search efficiency is low, and easy to fall into local convergence.An improved particle swarm optimization algorithm is proposed based on error back propagation gradient descent. Particle swarm optimization for Solving Least Squares Problems to meme group, the particles in the fitness ranking, optimization problem of the overall consideration, the error back propagation gradient descent training BP neural network, particle to update the velocity and position according to their individual optimal and global optimization, make the particles more to the social optimal learning and less to its optimal learning, it can avoid the particles fall into local optimum, by using gradient information can accelerate the PSO local search ability, improve the multi beam particle swarm depth zero less trajectory information search efficiency, the realization of improved particle swarm optimization algorithm. Simulation results show that the algorithm in the initial stage of rapid convergence to the global optimal solution can be near to the global optimal solution and keep close to the trend, the algorithm has faster convergence speed and search performance in the same running time, it can improve the convergence speed of the algorithm, especially the later search efficiency.

The DSFPN, a new neural network for optical character recognition.

PubMed

Morns, L P; Dlay, S S

1999-01-01

A new type of neural network for recognition tasks is presented in this paper. The network, called the dynamic supervised forward-propagation network (DSFPN), is based on the forward only version of the counterpropagation network (CPN). The DSFPN, trains using a supervised algorithm and can grow dynamically during training, allowing subclasses in the training data to be learnt in an unsupervised manner. It is shown to train in times comparable to the CPN while giving better classification accuracies than the popular backpropagation network. Both Fourier descriptors and wavelet descriptors are used for image preprocessing and the wavelets are proven to give a far better performance.

Accurate modeling of switched reluctance machine based on hybrid trained WNN

DOE Office of Scientific and Technical Information (OSTI.GOV)

Song, Shoujun, E-mail: sunnyway@nwpu.edu.cn; Ge, Lefei; Ma, Shaojie

2014-04-15

According to the strong nonlinear electromagnetic characteristics of switched reluctance machine (SRM), a novel accurate modeling method is proposed based on hybrid trained wavelet neural network (WNN) which combines improved genetic algorithm (GA) with gradient descent (GD) method to train the network. In the novel method, WNN is trained by GD method based on the initial weights obtained per improved GA optimization, and the global parallel searching capability of stochastic algorithm and local convergence speed of deterministic algorithm are combined to enhance the training accuracy, stability and speed. Based on the measured electromagnetic characteristics of a 3-phase 12/8-pole SRM, themore » nonlinear simulation model is built by hybrid trained WNN in Matlab. The phase current and mechanical characteristics from simulation under different working conditions meet well with those from experiments, which indicates the accuracy of the model for dynamic and static performance evaluation of SRM and verifies the effectiveness of the proposed modeling method.« less

Biological engineering applications of feedforward neural networks designed and parameterized by genetic algorithms.

PubMed

Ferentinos, Konstantinos P

2005-09-01

Two neural network (NN) applications in the field of biological engineering are developed, designed and parameterized by an evolutionary method based on the evolutionary process of genetic algorithms. The developed systems are a fault detection NN model and a predictive modeling NN system. An indirect or 'weak specification' representation was used for the encoding of NN topologies and training parameters into genes of the genetic algorithm (GA). Some a priori knowledge of the demands in network topology for specific application cases is required by this approach, so that the infinite search space of the problem is limited to some reasonable degree. Both one-hidden-layer and two-hidden-layer network architectures were explored by the GA. Except for the network architecture, each gene of the GA also encoded the type of activation functions in both hidden and output nodes of the NN and the type of minimization algorithm that was used by the backpropagation algorithm for the training of the NN. Both models achieved satisfactory performance, while the GA system proved to be a powerful tool that can successfully replace the problematic trial-and-error approach that is usually used for these tasks.

Combining two open source tools for neural computation (BioPatRec and Netlab) improves movement classification for prosthetic control.

PubMed

Prahm, Cosima; Eckstein, Korbinian; Ortiz-Catalan, Max; Dorffner, Georg; Kaniusas, Eugenijus; Aszmann, Oskar C

2016-08-31

Controlling a myoelectric prosthesis for upper limbs is increasingly challenging for the user as more electrodes and joints become available. Motion classification based on pattern recognition with a multi-electrode array allows multiple joints to be controlled simultaneously. Previous pattern recognition studies are difficult to compare, because individual research groups use their own data sets. To resolve this shortcoming and to facilitate comparisons, open access data sets were analysed using components of BioPatRec and Netlab pattern recognition models. Performances of the artificial neural networks, linear models, and training program components were compared. Evaluation took place within the BioPatRec environment, a Matlab-based open source platform that provides feature extraction, processing and motion classification algorithms for prosthetic control. The algorithms were applied to myoelectric signals for individual and simultaneous classification of movements, with the aim of finding the best performing algorithm and network model. Evaluation criteria included classification accuracy and training time. Results in both the linear and the artificial neural network models demonstrated that Netlab's implementation using scaled conjugate training algorithm reached significantly higher accuracies than BioPatRec. It is concluded that the best movement classification performance would be achieved through integrating Netlab training algorithms in the BioPatRec environment so that future prosthesis training can be shortened and control made more reliable. Netlab was therefore included into the newest release of BioPatRec (v4.0).

Gross domestic product estimation based on electricity utilization by artificial neural network

NASA Astrophysics Data System (ADS)

Stevanović, Mirjana; Vujičić, Slađana; Gajić, Aleksandar M.

2018-01-01

The main goal of the paper was to estimate gross domestic product (GDP) based on electricity estimation by artificial neural network (ANN). The electricity utilization was analyzed based on different sources like renewable, coal and nuclear sources. The ANN network was trained with two training algorithms namely extreme learning method and back-propagation algorithm in order to produce the best prediction results of the GDP. According to the results it can be concluded that the ANN model with extreme learning method could produce the acceptable prediction of the GDP based on the electricity utilization.

An application of artificial neural networks to experimental data approximation

NASA Technical Reports Server (NTRS)

Meade, Andrew J., Jr.

1993-01-01

As an initial step in the evaluation of networks, a feedforward architecture is trained to approximate experimental data by the backpropagation algorithm. Several drawbacks were detected and an alternative learning algorithm was then developed to partially address the drawbacks. This noniterative algorithm has a number of advantages over the backpropagation method and is easily implemented on existing hardware.

Predicting the survival of diabetes using neural network

NASA Astrophysics Data System (ADS)

Mamuda, Mamman; Sathasivam, Saratha

2017-08-01

Data mining techniques at the present time are used in predicting diseases of health care industries. Neural Network is one among the prevailing method in data mining techniques of an intelligent field for predicting diseases in health care industries. This paper presents a study on the prediction of the survival of diabetes diseases using different learning algorithms from the supervised learning algorithms of neural network. Three learning algorithms are considered in this study: (i) The levenberg-marquardt learning algorithm (ii) The Bayesian regulation learning algorithm and (iii) The scaled conjugate gradient learning algorithm. The network is trained using the Pima Indian Diabetes Dataset with the help of MATLAB R2014(a) software. The performance of each algorithm is further discussed through regression analysis. The prediction accuracy of the best algorithm is further computed to validate the accurate prediction

Application of Two-Dimensional AWE Algorithm in Training Multi-Dimensional Neural Network Model

DTIC Science & Technology

2003-07-01

hybrid scheme . the general neural network method (Table 3.1). The training process of the software- ACKNOWLEDGMENT "Neuralmodeler" is shown in Fig. 3.2...engineering. Artificial neural networks (ANNs) have emerged Training a neural network model is the key of as a powerful technique for modeling general neural...coefficients am, the derivatives method of moments (MoM). The variables in the of matrix I have to be generated . A closed form model are frequency

Training trajectories by continuous recurrent multilayer networks.

PubMed

Leistritz, L; Galicki, M; Witte, H; Kochs, E

2002-01-01

This paper addresses the problem of training trajectories by means of continuous recurrent neural networks whose feedforward parts are multilayer perceptrons. Such networks can approximate a general nonlinear dynamic system with arbitrary accuracy. The learning process is transformed into an optimal control framework where the weights are the controls to be determined. A training algorithm based upon a variational formulation of Pontryagin's maximum principle is proposed for such networks. Computer examples demonstrating the efficiency of the given approach are also presented.

Pruning Neural Networks with Distribution Estimation Algorithms

DOE Office of Scientific and Technical Information (OSTI.GOV)

Cantu-Paz, E

2003-01-15

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 themore » 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.« less

Relabeling exchange method (REM) for learning in neural networks

NASA Astrophysics Data System (ADS)

Wu, Wen; Mammone, Richard J.

1994-02-01

The supervised training of neural networks require the use of output labels which are usually arbitrarily assigned. In this paper it is shown that there is a significant difference in the rms error of learning when `optimal' label assignment schemes are used. We have investigated two efficient random search algorithms to solve the relabeling problem: the simulated annealing and the genetic algorithm. However, we found them to be computationally expensive. Therefore we shall introduce a new heuristic algorithm called the Relabeling Exchange Method (REM) which is computationally more attractive and produces optimal performance. REM has been used to organize the optimal structure for multi-layered perceptrons and neural tree networks. The method is a general one and can be implemented as a modification to standard training algorithms. The motivation of the new relabeling strategy is based on the present interpretation of dyslexia as an encoding problem.

Impact of Linearity and Write Noise of Analog Resistive Memory Devices in a Neural Algorithm Accelerator

DOE Office of Scientific and Technical Information (OSTI.GOV)

Jacobs-Gedrim, Robin B.; Agarwal, Sapan; Knisely, Kathrine E.

Resistive memory (ReRAM) shows promise for use as an analog synapse element in energy-efficient neural network algorithm accelerators. A particularly important application is the training of neural networks, as this is the most computationally-intensive procedure in using a neural algorithm. However, training a network with analog ReRAM synapses can significantly reduce the accuracy at the algorithm level. In order to assess this degradation, analog properties of ReRAM devices were measured and hand-written digit recognition accuracy was modeled for the training using backpropagation. Bipolar filamentary devices utilizing three material systems were measured and compared: one oxygen vacancy system, Ta-TaO x, andmore » two conducting metallization systems, Cu-SiO 2, and Ag/chalcogenide. Analog properties and conductance ranges of the devices are optimized by measuring the response to varying voltage pulse characteristics. Key analog device properties which degrade the accuracy are update linearity and write noise. Write noise may improve as a function of device manufacturing maturity, but write nonlinearity appears relatively consistent among the different device material systems and is found to be the most significant factor affecting accuracy. As a result, this suggests that new materials and/or fundamentally different resistive switching mechanisms may be required to improve device linearity and achieve higher algorithm training accuracy.« less

Impact of Linearity and Write Noise of Analog Resistive Memory Devices in a Neural Algorithm Accelerator

DOE PAGES

Jacobs-Gedrim, Robin B.; Agarwal, Sapan; Knisely, Kathrine E.; ...

2017-12-01

Resistive memory (ReRAM) shows promise for use as an analog synapse element in energy-efficient neural network algorithm accelerators. A particularly important application is the training of neural networks, as this is the most computationally-intensive procedure in using a neural algorithm. However, training a network with analog ReRAM synapses can significantly reduce the accuracy at the algorithm level. In order to assess this degradation, analog properties of ReRAM devices were measured and hand-written digit recognition accuracy was modeled for the training using backpropagation. Bipolar filamentary devices utilizing three material systems were measured and compared: one oxygen vacancy system, Ta-TaO x, andmore » two conducting metallization systems, Cu-SiO 2, and Ag/chalcogenide. Analog properties and conductance ranges of the devices are optimized by measuring the response to varying voltage pulse characteristics. Key analog device properties which degrade the accuracy are update linearity and write noise. Write noise may improve as a function of device manufacturing maturity, but write nonlinearity appears relatively consistent among the different device material systems and is found to be the most significant factor affecting accuracy. As a result, this suggests that new materials and/or fundamentally different resistive switching mechanisms may be required to improve device linearity and achieve higher algorithm training accuracy.« less

Condition monitoring of 3G cellular networks through competitive neural models.

PubMed

Barreto, Guilherme A; Mota, João C M; Souza, Luis G M; Frota, Rewbenio A; Aguayo, Leonardo

2005-09-01

We develop an unsupervised approach to condition monitoring of cellular networks using competitive neural algorithms. Training is carried out with state vectors representing the normal functioning of a simulated CDMA2000 network. Once training is completed, global and local normality profiles (NPs) are built from the distribution of quantization errors of the training state vectors and their components, respectively. The global NP is used to evaluate the overall condition of the cellular system. If abnormal behavior is detected, local NPs are used in a component-wise fashion to find abnormal state variables. Anomaly detection tests are performed via percentile-based confidence intervals computed over the global and local NPs. We compared the performance of four competitive algorithms [winner-take-all (WTA), frequency-sensitive competitive learning (FSCL), self-organizing map (SOM), and neural-gas algorithm (NGA)] and the results suggest that the joint use of global and local NPs is more efficient and more robust than current single-threshold methods.

Vector Quantization Algorithm Based on Associative Memories

NASA Astrophysics Data System (ADS)

Guzmán, Enrique; Pogrebnyak, Oleksiy; Yáñez, Cornelio; Manrique, Pablo

This paper presents a vector quantization algorithm for image compression based on extended associative memories. The proposed algorithm is divided in two stages. First, an associative network is generated applying the learning phase of the extended associative memories between a codebook generated by the LBG algorithm and a training set. This associative network is named EAM-codebook and represents a new codebook which is used in the next stage. The EAM-codebook establishes a relation between training set and the LBG codebook. Second, the vector quantization process is performed by means of the recalling stage of EAM using as associative memory the EAM-codebook. This process generates a set of the class indices to which each input vector belongs. With respect to the LBG algorithm, the main advantages offered by the proposed algorithm is high processing speed and low demand of resources (system memory); results of image compression and quality are presented.

Comparison of crisp and fuzzy character networks in handwritten word recognition

NASA Technical Reports Server (NTRS)

Gader, Paul; Mohamed, Magdi; Chiang, Jung-Hsien

1992-01-01

Experiments involving handwritten word recognition on words taken from images of handwritten address blocks from the United States Postal Service mailstream are described. The word recognition algorithm relies on the use of neural networks at the character level. The neural networks are trained using crisp and fuzzy desired outputs. The fuzzy outputs were defined using a fuzzy k-nearest neighbor algorithm. The crisp networks slightly outperformed the fuzzy networks at the character level but the fuzzy networks outperformed the crisp networks at the word level.

Membership generation using multilayer neural network

NASA Technical Reports Server (NTRS)

Kim, Jaeseok

1992-01-01

There has been intensive research in neural network applications to pattern recognition problems. Particularly, the back-propagation network has attracted many researchers because of its outstanding performance in pattern recognition applications. In this section, we describe a new method to generate membership functions from training data using a multilayer neural network. The basic idea behind the approach is as follows. The output values of a sigmoid activation function of a neuron bear remarkable resemblance to membership values. Therefore, we can regard the sigmoid activation values as the membership values in fuzzy set theory. Thus, in order to generate class membership values, we first train a suitable multilayer network using a training algorithm such as the back-propagation algorithm. After the training procedure converges, the resulting network can be treated as a membership generation network, where the inputs are feature values and the outputs are membership values in the different classes. This method allows fairly complex membership functions to be generated because the network is highly nonlinear in general. Also, it is to be noted that the membership functions are generated from a classification point of view. For pattern recognition applications, this is highly desirable, although the membership values may not be indicative of the degree of typicality of a feature value in a particular class.

Using an Extended Kalman Filter Learning Algorithm for Feed-Forward Neural Networks to Describe Tracer Correlations

NASA Technical Reports Server (NTRS)

Lary, David J.; Mussa, Yussuf

2004-01-01

In this study a new extended Kalman filter (EKF) learning algorithm for feed-forward neural networks (FFN) is used. With the EKF approach, the training of the FFN can be seen as state estimation for a non-linear stationary process. The EKF method gives excellent convergence performances provided that there is enough computer core memory and that the machine precision is high. Neural networks are ideally suited to describe the spatial and temporal dependence of tracer-tracer correlations. The neural network performs well even in regions where the correlations are less compact and normally a family of correlation curves would be required. For example, the CH4-N2O correlation can be well described using a neural network trained with the latitude, pressure, time of year, and CH4 volume mixing ratio (v.m.r.). The neural network was able to reproduce the CH4-N2O correlation with a correlation coefficient between simulated and training values of 0.9997. The neural network Fortran code used is available for download.

Markov Chain Monte Carlo Bayesian Learning for Neural Networks

NASA Technical Reports Server (NTRS)

Goodrich, Michael S.

2011-01-01

Conventional training methods for neural networks involve starting al a random location in the solution space of the network weights, navigating an error hyper surface to reach a minimum, and sometime stochastic based techniques (e.g., genetic algorithms) to avoid entrapment in a local minimum. It is further typically necessary to preprocess the data (e.g., normalization) to keep the training algorithm on course. Conversely, Bayesian based learning is an epistemological approach concerned with formally updating the plausibility of competing candidate hypotheses thereby obtaining a posterior distribution for the network weights conditioned on the available data and a prior distribution. In this paper, we developed a powerful methodology for estimating the full residual uncertainty in network weights and therefore network predictions by using a modified Jeffery's prior combined with a Metropolis Markov Chain Monte Carlo method.

Can surgical simulation be used to train detection and classification of neural networks?

PubMed

Zisimopoulos, Odysseas; Flouty, Evangello; Stacey, Mark; Muscroft, Sam; Giataganas, Petros; Nehme, Jean; Chow, Andre; Stoyanov, Danail

2017-10-01

Computer-assisted interventions (CAI) aim to increase the effectiveness, precision and repeatability of procedures to improve surgical outcomes. The presence and motion of surgical tools is a key information input for CAI surgical phase recognition algorithms. Vision-based tool detection and recognition approaches are an attractive solution and can be designed to take advantage of the powerful deep learning paradigm that is rapidly advancing image recognition and classification. The challenge for such algorithms is the availability and quality of labelled data used for training. In this Letter, surgical simulation is used to train tool detection and segmentation based on deep convolutional neural networks and generative adversarial networks. The authors experiment with two network architectures for image segmentation in tool classes commonly encountered during cataract surgery. A commercially-available simulator is used to create a simulated cataract dataset for training models prior to performing transfer learning on real surgical data. To the best of authors' knowledge, this is the first attempt to train deep learning models for surgical instrument detection on simulated data while demonstrating promising results to generalise on real data. Results indicate that simulated data does have some potential for training advanced classification methods for CAI systems.

Application of the clinical matrix to the diagnosis of leukemia

NASA Astrophysics Data System (ADS)

Pakkala, Sampath Y.; Lin, Frank C.

1992-07-01

A system for diagnosing leukemia subtypes has been formulated using neural networks. The statistical data of the symptoms collected by hematologists is fed into a single training set using a neural network, where the network is trained by using fast backpropagation algorithm, which when done can help the general practitioners for making diagnoses on the basis of signs and symptoms alone.

On-line node fault injection training algorithm for MLP networks: objective function and convergence analysis.

PubMed

Sum, John Pui-Fai; Leung, Chi-Sing; Ho, Kevin I-J

2012-02-01

Improving fault tolerance of a neural network has been studied for more than two decades. Various training algorithms have been proposed in sequel. The on-line node fault injection-based algorithm is one of these algorithms, in which hidden nodes randomly output zeros during training. While the idea is simple, theoretical analyses on this algorithm are far from complete. This paper presents its objective function and the convergence proof. We consider three cases for multilayer perceptrons (MLPs). They are: (1) MLPs with single linear output node; (2) MLPs with multiple linear output nodes; and (3) MLPs with single sigmoid output node. For the convergence proof, we show that the algorithm converges with probability one. For the objective function, we show that the corresponding objective functions of cases (1) and (2) are of the same form. They both consist of a mean square errors term, a regularizer term, and a weight decay term. For case (3), the objective function is slight different from that of cases (1) and (2). With the objective functions derived, we can compare the similarities and differences among various algorithms and various cases.

Scalable learning method for feedforward neural networks using minimal-enclosing-ball approximation.

PubMed

Wang, Jun; Deng, Zhaohong; Luo, Xiaoqing; Jiang, Yizhang; Wang, Shitong

2016-06-01

Training feedforward neural networks (FNNs) is one of the most critical issues in FNNs studies. However, most FNNs training methods cannot be directly applied for very large datasets because they have high computational and space complexity. In order to tackle this problem, the CCMEB (Center-Constrained Minimum Enclosing Ball) problem in hidden feature space of FNN is discussed and a novel learning algorithm called HFSR-GCVM (hidden-feature-space regression using generalized core vector machine) is developed accordingly. In HFSR-GCVM, a novel learning criterion using L2-norm penalty-based ε-insensitive function is formulated and the parameters in the hidden nodes are generated randomly independent of the training sets. Moreover, the learning of parameters in its output layer is proved equivalent to a special CCMEB problem in FNN hidden feature space. As most CCMEB approximation based machine learning algorithms, the proposed HFSR-GCVM training algorithm has the following merits: The maximal training time of the HFSR-GCVM training is linear with the size of training datasets and the maximal space consumption is independent of the size of training datasets. The experiments on regression tasks confirm the above conclusions. Copyright © 2016 Elsevier Ltd. All rights reserved.

LiteNet: Lightweight Neural Network for Detecting Arrhythmias at Resource-Constrained Mobile Devices.

PubMed

He, Ziyang; Zhang, Xiaoqing; Cao, Yangjie; Liu, Zhi; Zhang, Bo; Wang, Xiaoyan

2018-04-17

By running applications and services closer to the user, edge processing provides many advantages, such as short response time and reduced network traffic. Deep-learning based algorithms provide significantly better performances than traditional algorithms in many fields but demand more resources, such as higher computational power and more memory. Hence, designing deep learning algorithms that are more suitable for resource-constrained mobile devices is vital. In this paper, we build a lightweight neural network, termed LiteNet which uses a deep learning algorithm design to diagnose arrhythmias, as an example to show how we design deep learning schemes for resource-constrained mobile devices. Compare to other deep learning models with an equivalent accuracy, LiteNet has several advantages. It requires less memory, incurs lower computational cost, and is more feasible for deployment on resource-constrained mobile devices. It can be trained faster than other neural network algorithms and requires less communication across different processing units during distributed training. It uses filters of heterogeneous size in a convolutional layer, which contributes to the generation of various feature maps. The algorithm was tested using the MIT-BIH electrocardiogram (ECG) arrhythmia database; the results showed that LiteNet outperforms comparable schemes in diagnosing arrhythmias, and in its feasibility for use at the mobile devices.

LiteNet: Lightweight Neural Network for Detecting Arrhythmias at Resource-Constrained Mobile Devices

PubMed Central

Zhang, Xiaoqing; Cao, Yangjie; Liu, Zhi; Zhang, Bo; Wang, Xiaoyan

2018-01-01

By running applications and services closer to the user, edge processing provides many advantages, such as short response time and reduced network traffic. Deep-learning based algorithms provide significantly better performances than traditional algorithms in many fields but demand more resources, such as higher computational power and more memory. Hence, designing deep learning algorithms that are more suitable for resource-constrained mobile devices is vital. In this paper, we build a lightweight neural network, termed LiteNet which uses a deep learning algorithm design to diagnose arrhythmias, as an example to show how we design deep learning schemes for resource-constrained mobile devices. Compare to other deep learning models with an equivalent accuracy, LiteNet has several advantages. It requires less memory, incurs lower computational cost, and is more feasible for deployment on resource-constrained mobile devices. It can be trained faster than other neural network algorithms and requires less communication across different processing units during distributed training. It uses filters of heterogeneous size in a convolutional layer, which contributes to the generation of various feature maps. The algorithm was tested using the MIT-BIH electrocardiogram (ECG) arrhythmia database; the results showed that LiteNet outperforms comparable schemes in diagnosing arrhythmias, and in its feasibility for use at the mobile devices. PMID:29673171

Accelerated Training for Large Feedforward Neural Networks

NASA Technical Reports Server (NTRS)

Stepniewski, Slawomir W.; Jorgensen, Charles C.

1998-01-01

In this paper we introduce a new training algorithm, the scaled variable metric (SVM) method. Our approach attempts to increase the convergence rate of the modified variable metric method. It is also combined with the RBackprop algorithm, which computes the product of the matrix of second derivatives (Hessian) with an arbitrary vector. The RBackprop method allows us to avoid computationally expensive, direct line searches. In addition, it can be utilized in the new, 'predictive' updating technique of the inverse Hessian approximation. We have used directional slope testing to adjust the step size and found that this strategy works exceptionally well in conjunction with the Rbackprop algorithm. Some supplementary, but nevertheless important enhancements to the basic training scheme such as improved setting of a scaling factor for the variable metric update and computationally more efficient procedure for updating the inverse Hessian approximation are presented as well. We summarize by comparing the SVM method with four first- and second- order optimization algorithms including a very effective implementation of the Levenberg-Marquardt method. Our tests indicate promising computational speed gains of the new training technique, particularly for large feedforward networks, i.e., for problems where the training process may be the most laborious.

A comparative study of breast cancer diagnosis based on neural network ensemble via improved training algorithms.

PubMed

Azami, Hamed; Escudero, Javier

2015-08-01

Breast cancer is one of the most common types of cancer in women all over the world. Early diagnosis of this kind of cancer can significantly increase the chances of long-term survival. Since diagnosis of breast cancer is a complex problem, neural network (NN) approaches have been used as a promising solution. Considering the low speed of the back-propagation (BP) algorithm to train a feed-forward NN, we consider a number of improved NN trainings for the Wisconsin breast cancer dataset: BP with momentum, BP with adaptive learning rate, BP with adaptive learning rate and momentum, Polak-Ribikre conjugate gradient algorithm (CGA), Fletcher-Reeves CGA, Powell-Beale CGA, scaled CGA, resilient BP (RBP), one-step secant and quasi-Newton methods. An NN ensemble, which is a learning paradigm to combine a number of NN outputs, is used to improve the accuracy of the classification task. Results demonstrate that NN ensemble-based classification methods have better performance than NN-based algorithms. The highest overall average accuracy is 97.68% obtained by NN ensemble trained by RBP for 50%-50% training-test evaluation method.

Development of automated system based on neural network algorithm for detecting defects on molds installed on casting machines

NASA Astrophysics Data System (ADS)

Bazhin, V. Yu; Danilov, I. V.; Petrov, P. A.

2018-05-01

During the casting of light alloys and ligatures based on aluminum and magnesium, problems of the qualitative distribution of the metal and its crystallization in the mold arise. To monitor the defects of molds on the casting conveyor, a camera with a resolution of 780 x 580 pixels and a shooting rate of 75 frames per second was selected. Images of molds from casting machines were used as input data for neural network algorithm. On the preparation of a digital database and its analytical evaluation stage, the architecture of the convolutional neural network was chosen for the algorithm. The information flow from the local controller is transferred to the OPC server and then to the SCADA system of foundry. After the training, accuracy of neural network defect recognition was about 95.1% on a validation split. After the training, weight coefficients of the neural network were used on testing split and algorithm had identical accuracy with validation images. The proposed technical solutions make it possible to increase the efficiency of the automated process control system in the foundry by expanding the digital database.

Comparative Analysis of Neural Network Training Methods in Real-time Radiotherapy.

PubMed

Nouri, S; Hosseini Pooya, S M; Soltani Nabipour, J

2017-03-01

The motions of body and tumor in some regions such as chest during radiotherapy treatments are one of the major concerns protecting normal tissues against high doses. By using real-time radiotherapy technique, it is possible to increase the accuracy of delivered dose to the tumor region by means of tracing markers on the body of patients. This study evaluates the accuracy of some artificial intelligence methods including neural network and those of combination with genetic algorithm as well as particle swarm optimization (PSO) estimating tumor positions in real-time radiotherapy. One hundred recorded signals of three external markers were used as input data. The signals from 3 markers thorough 10 breathing cycles of a patient treated via a cyber-knife for a lung tumor were used as data input. Then, neural network method and its combination with genetic or PSO algorithms were applied determining the tumor locations using MATLAB© software program. The accuracies were obtained 0.8%, 12% and 14% in neural network, genetic and particle swarm optimization algorithms, respectively. The internal target volume (ITV) should be determined based on the applied neural network algorithm on training steps.

On the fusion of tuning parameters of fuzzy rules and neural network

NASA Astrophysics Data System (ADS)

Mamuda, Mamman; Sathasivam, Saratha

2017-08-01

Learning fuzzy rule-based system with neural network can lead to a precise valuable empathy of several problems. Fuzzy logic offers a simple way to reach at a definite conclusion based upon its vague, ambiguous, imprecise, noisy or missing input information. Conventional learning algorithm for tuning parameters of fuzzy rules using training input-output data usually end in a weak firing state, this certainly powers the fuzzy rule and makes it insecure for a multiple-input fuzzy system. In this paper, we introduce a new learning algorithm for tuning the parameters of the fuzzy rules alongside with radial basis function neural network (RBFNN) in training input-output data based on the gradient descent method. By the new learning algorithm, the problem of weak firing using the conventional method was addressed. We illustrated the efficiency of our new learning algorithm by means of numerical examples. MATLAB R2014(a) software was used in simulating our result The result shows that the new learning method has the best advantage of training the fuzzy rules without tempering with the fuzzy rule table which allowed a membership function of the rule to be used more than one time in the fuzzy rule base.

Deep Marginalized Sparse Denoising Auto-Encoder for Image Denoising

NASA Astrophysics Data System (ADS)

Ma, Hongqiang; Ma, Shiping; Xu, Yuelei; Zhu, Mingming

2018-01-01

Stacked Sparse Denoising Auto-Encoder (SSDA) has been successfully applied to image denoising. As a deep network, the SSDA network with powerful data feature learning ability is superior to the traditional image denoising algorithms. However, the algorithm has high computational complexity and slow convergence rate in the training. To address this limitation, we present a method of image denoising based on Deep Marginalized Sparse Denoising Auto-Encoder (DMSDA). The loss function of Sparse Denoising Auto-Encoder is marginalized so that it satisfies both sparseness and marginality. The experimental results show that the proposed algorithm can not only outperform SSDA in the convergence speed and training time, but also has better denoising performance than the current excellent denoising algorithms, including both the subjective and objective evaluation of image denoising.

Geometrical features assessment of liver's tumor with application of artificial neural network evolved by imperialist competitive algorithm.

PubMed

Keshavarz, M; Mojra, A

2015-05-01

Geometrical features of a cancerous tumor embedded in biological soft tissue, including tumor size and depth, are a necessity in the follow-up procedure and making suitable therapeutic decisions. In this paper, a new socio-politically motivated global search strategy which is called imperialist competitive algorithm (ICA) is implemented to train a feed forward neural network (FFNN) to estimate the tumor's geometrical characteristics (FFNNICA). First, a viscoelastic model of liver tissue is constructed by using a series of in vitro uniaxial and relaxation test data. Then, 163 samples of the tissue including a tumor with different depths and diameters are generated by making use of PYTHON programming to link the ABAQUS and MATLAB together. Next, the samples are divided into 123 samples as training dataset and 40 samples as testing dataset. Training inputs of the network are mechanical parameters extracted from palpation of the tissue through a developing noninvasive technology called artificial tactile sensing (ATS). Last, to evaluate the FFNNICA performance, outputs of the network including tumor's depth and diameter are compared with desired values for both training and testing datasets. Deviations of the outputs from desired values are calculated by a regression analysis. Statistical analysis is also performed by measuring Root Mean Square Error (RMSE) and Efficiency (E). RMSE in diameter and depth estimations are 0.50 mm and 1.49, respectively, for the testing dataset. Results affirm that the proposed optimization algorithm for training neural network can be useful to characterize soft tissue tumors accurately by employing an artificial palpation approach. Copyright © 2015 John Wiley & Sons, Ltd.

Correcting Evaluation Bias of Relational Classifiers with Network Cross Validation

DTIC Science & Technology

2010-01-01

classi- fication algorithms: simple random resampling (RRS), equal-instance random resampling (ERS), and network cross-validation ( NCV ). The first two... NCV procedure that eliminates overlap between test sets altogether. The procedure samples for k disjoint test sets that will be used for evaluation...propLabeled ∗ S) nodes from train Pool in f erenceSet =network − trainSet F = F ∪ < trainSet, test Set, in f erenceSet > end for output: F NCV addresses

Improving Maritime Domain Awareness Using Neural Networks for Target of Interest Classification

DTIC Science & Technology

2015-03-01

spreading SCG scaled conjugate gradient xv THIS PAGE INTENTIONALLY LEFT BLANK xvi EXECUTIVE SUMMARY The research detailed in this thesis is a...algorithms were explored for training the neural networks: resilient backpropagation (RP) and scaled conjugate gradient backpropagation ( SCG ). The...results of the neural network training performance are presented using mean squared error convergence plots. In all implementations, the SCG learning

Inverse kinematics problem in robotics using neural networks

NASA Technical Reports Server (NTRS)

Choi, Benjamin B.; Lawrence, Charles

1992-01-01

In this paper, Multilayer Feedforward Networks are applied to the robot inverse kinematic problem. The networks are trained with endeffector position and joint angles. After training, performance is measured by having the network generate joint angles for arbitrary endeffector trajectories. A 3-degree-of-freedom (DOF) spatial manipulator is used for the study. It is found that neural networks provide a simple and effective way to both model the manipulator inverse kinematics and circumvent the problems associated with algorithmic solution methods.

Applying FastSLAM to Articulated Rovers

NASA Astrophysics Data System (ADS)

Hewitt, Robert Alexander

This thesis presents the navigation algorithms designed for use on Kapvik, a 30 kg planetary micro-rover built for the Canadian Space Agency; the simulations used to test the algorithm; and novel techniques for terrain classification using Kapvik's LIDAR (Light Detection And Ranging) sensor. Kapvik implements a six-wheeled, skid-steered, rocker-bogie mobility system. This warrants a more complicated kinematic model for navigation than a typical 4-wheel differential drive system. The design of a 3D navigation algorithm is presented that includes nonlinear Kalman filtering and Simultaneous Localization and Mapping (SLAM). A neural network for terrain classification is used to improve navigation performance. Simulation is used to train the neural network and validate the navigation algorithms. Real world tests of the terrain classification algorithm validate the use of simulation for training and the improvement to SLAM through the reduction of extraneous LIDAR measurements in each scan.

Distributed computing methodology for training neural networks in an image-guided diagnostic application.

PubMed

Plagianakos, V P; Magoulas, G D; Vrahatis, M N

2006-03-01

Distributed computing is a process through which a set of computers connected by a network is used collectively to solve a single problem. In this paper, we propose a distributed computing methodology for training neural networks for the detection of lesions in colonoscopy. Our approach is based on partitioning the training set across multiple processors using a parallel virtual machine. In this way, interconnected computers of varied architectures can be used for the distributed evaluation of the error function and gradient values, and, thus, training neural networks utilizing various learning methods. The proposed methodology has large granularity and low synchronization, and has been implemented and tested. Our results indicate that the parallel virtual machine implementation of the training algorithms developed leads to considerable speedup, especially when large network architectures and training sets are used.

Elements of an algorithm for optimizing a parameter-structural neural network

NASA Astrophysics Data System (ADS)

Mrówczyńska, Maria

2016-06-01

The field of processing information provided by measurement results is one of the most important components of geodetic technologies. The dynamic development of this field improves classic algorithms for numerical calculations in the aspect of analytical solutions that are difficult to achieve. Algorithms based on artificial intelligence in the form of artificial neural networks, including the topology of connections between neurons have become an important instrument connected to the problem of processing and modelling processes. This concept results from the integration of neural networks and parameter optimization methods and makes it possible to avoid the necessity to arbitrarily define the structure of a network. This kind of extension of the training process is exemplified by the algorithm called the Group Method of Data Handling (GMDH), which belongs to the class of evolutionary algorithms. The article presents a GMDH type network, used for modelling deformations of the geometrical axis of a steel chimney during its operation.

Android malware detection based on evolutionary super-network

NASA Astrophysics Data System (ADS)

Yan, Haisheng; Peng, Lingling

2018-04-01

In the paper, an android malware detection method based on evolutionary super-network is proposed in order to improve the precision of android malware detection. Chi square statistics method is used for selecting characteristics on the basis of analyzing android authority. Boolean weighting is utilized for calculating characteristic weight. Processed characteristic vector is regarded as the system training set and test set; hyper edge alternative strategy is used for training super-network classification model, thereby classifying test set characteristic vectors, and it is compared with traditional classification algorithm. The results show that the detection method proposed in the paper is close to or better than traditional classification algorithm. The proposed method belongs to an effective Android malware detection means.

Spatial-Temporal Reasoning Applications of Computational Intelligence in the Game of Go and Computer Networks

DTIC Science & Technology

2012-01-01

dimensionality, Tesauro used a backpropagation- based , three-layer neural network and implemented the outcome from a self-play game as the reinforcement signal...a school of fish, flock of birds, and colony of ants. Our literature review reveals that no one has used PSO to train the neural network ...trained with a variant of PSO called cellular PSO (CPSO). CSRN is a supervised learning neural network (SLNN). The proposed algorithm for the

Two papers on feed-forward networks

NASA Technical Reports Server (NTRS)

Buntine, Wray L.; Weigend, Andreas S.

1991-01-01

Connectionist feed-forward networks, trained with back-propagation, can be used both for nonlinear regression and for (discrete one-of-C) classification, depending on the form of training. This report contains two papers on feed-forward networks. The papers can be read independently. They are intended for the theoretically-aware practitioner or algorithm-designer; however, they also contain a review and comparison of several learning theories so they provide a perspective for the theoretician. The first paper works through Bayesian methods to complement back-propagation in the training of feed-forward networks. The second paper addresses a problem raised by the first: how to efficiently calculate second derivatives on feed-forward networks.

Fast, Simple and Accurate Handwritten Digit Classification by Training Shallow Neural Network Classifiers with the ‘Extreme Learning Machine’ Algorithm

PubMed Central

McDonnell, Mark D.; Tissera, Migel D.; Vladusich, Tony; van Schaik, André; Tapson, Jonathan

2015-01-01

Recent advances in training deep (multi-layer) architectures have inspired a renaissance in neural network use. For example, deep convolutional networks are becoming the default option for difficult tasks on large datasets, such as image and speech recognition. However, here we show that error rates below 1% on the MNIST handwritten digit benchmark can be replicated with shallow non-convolutional neural networks. This is achieved by training such networks using the ‘Extreme Learning Machine’ (ELM) approach, which also enables a very rapid training time (∼ 10 minutes). Adding distortions, as is common practise for MNIST, reduces error rates even further. Our methods are also shown to be capable of achieving less than 5.5% error rates on the NORB image database. To achieve these results, we introduce several enhancements to the standard ELM algorithm, which individually and in combination can significantly improve performance. The main innovation is to ensure each hidden-unit operates only on a randomly sized and positioned patch of each image. This form of random ‘receptive field’ sampling of the input ensures the input weight matrix is sparse, with about 90% of weights equal to zero. Furthermore, combining our methods with a small number of iterations of a single-batch backpropagation method can significantly reduce the number of hidden-units required to achieve a particular performance. Our close to state-of-the-art results for MNIST and NORB suggest that the ease of use and accuracy of the ELM algorithm for designing a single-hidden-layer neural network classifier should cause it to be given greater consideration either as a standalone method for simpler problems, or as the final classification stage in deep neural networks applied to more difficult problems. PMID:26262687

Analysis Resilient Algorithm on Artificial Neural Network Backpropagation

NASA Astrophysics Data System (ADS)

Saputra, Widodo; Tulus; Zarlis, Muhammad; Widia Sembiring, Rahmat; Hartama, Dedy

2017-12-01

Prediction required by decision makers to anticipate future planning. Artificial Neural Network (ANN) Backpropagation is one of method. This method however still has weakness, for long training time. This is a reason to improve a method to accelerate the training. One of Artificial Neural Network (ANN) Backpropagation method is a resilient method. Resilient method of changing weights and bias network with direct adaptation process of weighting based on local gradient information from every learning iteration. Predicting data result of Istanbul Stock Exchange training getting better. Mean Square Error (MSE) value is getting smaller and increasing accuracy.

Evaluation of multilayer perceptron algorithms for an analysis of network flow data

NASA Astrophysics Data System (ADS)

Bieniasz, Jedrzej; Rawski, Mariusz; Skowron, Krzysztof; Trzepiński, Mateusz

2016-09-01

The volume of exchanged information through IP networks is larger than ever and still growing. It creates a space for both benign and malicious activities. The second one raises awareness on security network devices, as well as network infrastructure and a system as a whole. One of the basic tools to prevent cyber attacks is Network Instrusion Detection System (NIDS). NIDS could be realized as a signature-based detector or an anomaly-based one. In the last few years the emphasis has been placed on the latter type, because of the possibility of applying smart and intelligent solutions. An ideal NIDS of next generation should be composed of self-learning algorithms that could react on known and unknown malicious network activities respectively. In this paper we evaluated a machine learning approach for detection of anomalies in IP network data represented as NetFlow records. We considered Multilayer Perceptron (MLP) as the classifier and we used two types of learning algorithms - Backpropagation (BP) and Particle Swarm Optimization (PSO). This paper includes a comprehensive survey on determining the most optimal MLP learning algorithm for the classification problem in application to network flow data. The performance, training time and convergence of BP and PSO methods were compared. The results show that PSO algorithm implemented by the authors outperformed other solutions if accuracy of classifications is considered. The major disadvantage of PSO is training time, which could be not acceptable for larger data sets or in real network applications. At the end we compared some key findings with the results from the other papers to show that in all cases results from this study outperformed them.

Nonlinear inversion of electrical resistivity imaging using pruning Bayesian neural networks

NASA Astrophysics Data System (ADS)

Jiang, Fei-Bo; Dai, Qian-Wei; Dong, Li

2016-06-01

Conventional artificial neural networks used to solve electrical resistivity imaging (ERI) inversion problem suffer from overfitting and local minima. To solve these problems, we propose to use a pruning Bayesian neural network (PBNN) nonlinear inversion method and a sample design method based on the K-medoids clustering algorithm. In the sample design method, the training samples of the neural network are designed according to the prior information provided by the K-medoids clustering results; thus, the training process of the neural network is well guided. The proposed PBNN, based on Bayesian regularization, is used to select the hidden layer structure by assessing the effect of each hidden neuron to the inversion results. Then, the hyperparameter α k , which is based on the generalized mean, is chosen to guide the pruning process according to the prior distribution of the training samples under the small-sample condition. The proposed algorithm is more efficient than other common adaptive regularization methods in geophysics. The inversion of synthetic data and field data suggests that the proposed method suppresses the noise in the neural network training stage and enhances the generalization. The inversion results with the proposed method are better than those of the BPNN, RBFNN, and RRBFNN inversion methods as well as the conventional least squares inversion.

A Structure-Adaptive Hybrid RBF-BP Classifier with an Optimized Learning Strategy

PubMed Central

Wen, Hui; Xie, Weixin; Pei, Jihong

2016-01-01

This paper presents a structure-adaptive hybrid RBF-BP (SAHRBF-BP) classifier with an optimized learning strategy. SAHRBF-BP is composed of a structure-adaptive RBF network and a BP network of cascade, where the number of RBF hidden nodes is adjusted adaptively according to the distribution of sample space, the adaptive RBF network is used for nonlinear kernel mapping and the BP network is used for nonlinear classification. The optimized learning strategy is as follows: firstly, a potential function is introduced into training sample space to adaptively determine the number of initial RBF hidden nodes and node parameters, and a form of heterogeneous samples repulsive force is designed to further optimize each generated RBF hidden node parameters, the optimized structure-adaptive RBF network is used for adaptively nonlinear mapping the sample space; then, according to the number of adaptively generated RBF hidden nodes, the number of subsequent BP input nodes can be determined, and the overall SAHRBF-BP classifier is built up; finally, different training sample sets are used to train the BP network parameters in SAHRBF-BP. Compared with other algorithms applied to different data sets, experiments show the superiority of SAHRBF-BP. Especially on most low dimensional and large number of data sets, the classification performance of SAHRBF-BP outperforms other training SLFNs algorithms. PMID:27792737

Bearing Fault Diagnosis under Variable Speed Using Convolutional Neural Networks and the Stochastic Diagonal Levenberg-Marquardt Algorithm

PubMed Central

Tra, Viet; Kim, Jaeyoung; Kim, Jong-Myon

2017-01-01

This paper presents a novel method for diagnosing incipient bearing defects under variable operating speeds using convolutional neural networks (CNNs) trained via the stochastic diagonal Levenberg-Marquardt (S-DLM) algorithm. The CNNs utilize the spectral energy maps (SEMs) of the acoustic emission (AE) signals as inputs and automatically learn the optimal features, which yield the best discriminative models for diagnosing incipient bearing defects under variable operating speeds. The SEMs are two-dimensional maps that show the distribution of energy across different bands of the AE spectrum. It is hypothesized that the variation of a bearing’s speed would not alter the overall shape of the AE spectrum rather, it may only scale and translate it. Thus, at different speeds, the same defect would yield SEMs that are scaled and shifted versions of each other. This hypothesis is confirmed by the experimental results, where CNNs trained using the S-DLM algorithm yield significantly better diagnostic performance under variable operating speeds compared to existing methods. In this work, the performance of different training algorithms is also evaluated to select the best training algorithm for the CNNs. The proposed method is used to diagnose both single and compound defects at six different operating speeds. PMID:29211025

Neural network explanation using inversion.

PubMed

Saad, Emad W; Wunsch, Donald C

2007-01-01

An important drawback of many artificial neural networks (ANN) is their lack of explanation capability [Andrews, R., Diederich, J., & Tickle, A. B. (1996). A survey and critique of techniques for extracting rules from trained artificial neural networks. Knowledge-Based Systems, 8, 373-389]. This paper starts with a survey of algorithms which attempt to explain the ANN output. We then present HYPINV, a new explanation algorithm which relies on network inversion; i.e. calculating the ANN input which produces a desired output. HYPINV is a pedagogical algorithm, that extracts rules, in the form of hyperplanes. It is able to generate rules with arbitrarily desired fidelity, maintaining a fidelity-complexity tradeoff. To our knowledge, HYPINV is the only pedagogical rule extraction method, which extracts hyperplane rules from continuous or binary attribute neural networks. Different network inversion techniques, involving gradient descent as well as an evolutionary algorithm, are presented. An information theoretic treatment of rule extraction is presented. HYPINV is applied to example synthetic problems, to a real aerospace problem, and compared with similar algorithms using benchmark problems.

A Distributed Learning Method for ℓ1-Regularized Kernel Machine over Wireless Sensor Networks

PubMed Central

Ji, Xinrong; Hou, Cuiqin; Hou, Yibin; Gao, Fang; Wang, Shulong

2016-01-01

In wireless sensor networks, centralized learning methods have very high communication costs and energy consumption. These are caused by the need to transmit scattered training examples from various sensor nodes to the central fusion center where a classifier or a regression machine is trained. To reduce the communication cost, a distributed learning method for a kernel machine that incorporates ℓ1 norm regularization (ℓ1-regularized) is investigated, and a novel distributed learning algorithm for the ℓ1-regularized kernel minimum mean squared error (KMSE) machine is proposed. The proposed algorithm relies on in-network processing and a collaboration that transmits the sparse model only between single-hop neighboring nodes. This paper evaluates the proposed algorithm with respect to the prediction accuracy, the sparse rate of model, the communication cost and the number of iterations on synthetic and real datasets. The simulation results show that the proposed algorithm can obtain approximately the same prediction accuracy as that obtained by the batch learning method. Moreover, it is significantly superior in terms of the sparse rate of model and communication cost, and it can converge with fewer iterations. Finally, an experiment conducted on a wireless sensor network (WSN) test platform further shows the advantages of the proposed algorithm with respect to communication cost. PMID:27376298

Application of dynamic recurrent neural networks in nonlinear system identification

NASA Astrophysics Data System (ADS)

Du, Yun; Wu, Xueli; Sun, Huiqin; Zhang, Suying; Tian, Qiang

2006-11-01

An adaptive identification method of simple dynamic recurrent neural network (SRNN) for nonlinear dynamic systems is presented in this paper. This method based on the theory that by using the inner-states feed-back of dynamic network to describe the nonlinear kinetic characteristics of system can reflect the dynamic characteristics more directly, deduces the recursive prediction error (RPE) learning algorithm of SRNN, and improves the algorithm by studying topological structure on recursion layer without the weight values. The simulation results indicate that this kind of neural network can be used in real-time control, due to its less weight values, simpler learning algorithm, higher identification speed, and higher precision of model. It solves the problems of intricate in training algorithm and slow rate in convergence caused by the complicate topological structure in usual dynamic recurrent neural network.

Big Data: A Parallel Particle Swarm Optimization-Back-Propagation Neural Network Algorithm Based on MapReduce.

PubMed

Cao, Jianfang; Cui, Hongyan; Shi, Hao; Jiao, Lijuan

2016-01-01

A back-propagation (BP) neural network can solve complicated random nonlinear mapping problems; therefore, it can be applied to a wide range of problems. However, as the sample size increases, the time required to train BP neural networks becomes lengthy. Moreover, the classification accuracy decreases as well. To improve the classification accuracy and runtime efficiency of the BP neural network algorithm, we proposed a parallel design and realization method for a particle swarm optimization (PSO)-optimized BP neural network based on MapReduce on the Hadoop platform using both the PSO algorithm and a parallel design. The PSO algorithm was used to optimize the BP neural network's initial weights and thresholds and improve the accuracy of the classification algorithm. The MapReduce parallel programming model was utilized to achieve parallel processing of the BP algorithm, thereby solving the problems of hardware and communication overhead when the BP neural network addresses big data. Datasets on 5 different scales were constructed using the scene image library from the SUN Database. The classification accuracy of the parallel PSO-BP neural network algorithm is approximately 92%, and the system efficiency is approximately 0.85, which presents obvious advantages when processing big data. The algorithm proposed in this study demonstrated both higher classification accuracy and improved time efficiency, which represents a significant improvement obtained from applying parallel processing to an intelligent algorithm on big data.

Neural network approach to the inverse problem of the crack-depth determination from ultrasonic backscattering data

NASA Astrophysics Data System (ADS)

Takadoya, M.; Notake, M.; Kitahara, M.; Achenbach, J. D.; Guo, Q. C.; Peterson, M. L.

A neural network approach has been developed to determine the depth of a surface breaking crack in a steel plate from ultrasonic backscattering data. The network is trained by the use of a feedforward three-layered network together with a back-propagation algorithm for error corrections. Synthetic data are employed for network training. The signal used for crack isonification is a mode converted 45 deg transverse wave. The plate with a surface breaking crack is immersed in water, and the crack is insonified from the opposite uncracked side of the plate. A numerical analysis of the backscattered field is carried out based on the elastic wave theory by the use of the boundary element method. The numerical analysis provides synthetic data for the training of the network. The training data have been calculated for cracks with specific increments of the experimental data which are different from the training data.

Container-code recognition system based on computer vision and deep neural networks

NASA Astrophysics Data System (ADS)

Liu, Yi; Li, Tianjian; Jiang, Li; Liang, Xiaoyao

2018-04-01

Automatic container-code recognition system becomes a crucial requirement for ship transportation industry in recent years. In this paper, an automatic container-code recognition system based on computer vision and deep neural networks is proposed. The system consists of two modules, detection module and recognition module. The detection module applies both algorithms based on computer vision and neural networks, and generates a better detection result through combination to avoid the drawbacks of the two methods. The combined detection results are also collected for online training of the neural networks. The recognition module exploits both character segmentation and end-to-end recognition, and outputs the recognition result which passes the verification. When the recognition module generates false recognition, the result will be corrected and collected for online training of the end-to-end recognition sub-module. By combining several algorithms, the system is able to deal with more situations, and the online training mechanism can improve the performance of the neural networks at runtime. The proposed system is able to achieve 93% of overall recognition accuracy.

Geometrical structure of Neural Networks: Geodesics, Jeffrey's Prior and Hyper-ribbons

NASA Astrophysics Data System (ADS)

Hayden, Lorien; Alemi, Alex; Sethna, James

2014-03-01

Neural networks are learning algorithms which are employed in a host of Machine Learning problems including speech recognition, object classification and data mining. In practice, neural networks learn a low dimensional representation of high dimensional data and define a model manifold which is an embedding of this low dimensional structure in the higher dimensional space. In this work, we explore the geometrical structure of a neural network model manifold. A Stacked Denoising Autoencoder and a Deep Belief Network are trained on handwritten digits from the MNIST database. Construction of geodesics along the surface and of slices taken from the high dimensional manifolds reveal a hierarchy of widths corresponding to a hyper-ribbon structure. This property indicates that neural networks fall into the class of sloppy models, in which certain parameter combinations dominate the behavior. Employing this information could prove valuable in designing both neural network architectures and training algorithms. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No . DGE-1144153.

Privacy-preserving backpropagation neural network learning.

PubMed

Chen, Tingting; Zhong, Sheng

2009-10-01

With the development of distributed computing environment , many learning problems now have to deal with distributed input data. To enhance cooperations in learning, it is important to address the privacy concern of each data holder by extending the privacy preservation notion to original learning algorithms. In this paper, we focus on preserving the privacy in an important learning model, multilayer neural networks. We present a privacy-preserving two-party distributed algorithm of backpropagation which allows a neural network to be trained without requiring either party to reveal her data to the other. We provide complete correctness and security analysis of our algorithms. The effectiveness of our algorithms is verified by experiments on various real world data sets.

Learning polynomial feedforward neural networks by genetic programming and backpropagation.

PubMed

Nikolaev, N Y; Iba, H

2003-01-01

This paper presents an approach to learning polynomial feedforward neural networks (PFNNs). The approach suggests, first, finding the polynomial network structure by means of a population-based search technique relying on the genetic programming paradigm, and second, further adjustment of the best discovered network weights by an especially derived backpropagation algorithm for higher order networks with polynomial activation functions. These two stages of the PFNN learning process enable us to identify networks with good training as well as generalization performance. Empirical results show that this approach finds PFNN which outperform considerably some previous constructive polynomial network algorithms on processing benchmark time series.

Speech enhancement based on neural networks improves speech intelligibility in noise for cochlear implant users.

PubMed

Goehring, Tobias; Bolner, Federico; Monaghan, Jessica J M; van Dijk, Bas; Zarowski, Andrzej; Bleeck, Stefan

2017-02-01

Speech understanding in noisy environments is still one of the major challenges for cochlear implant (CI) users in everyday life. We evaluated a speech enhancement algorithm based on neural networks (NNSE) for improving speech intelligibility in noise for CI users. The algorithm decomposes the noisy speech signal into time-frequency units, extracts a set of auditory-inspired features and feeds them to the neural network to produce an estimation of which frequency channels contain more perceptually important information (higher signal-to-noise ratio, SNR). This estimate is used to attenuate noise-dominated and retain speech-dominated CI channels for electrical stimulation, as in traditional n-of-m CI coding strategies. The proposed algorithm was evaluated by measuring the speech-in-noise performance of 14 CI users using three types of background noise. Two NNSE algorithms were compared: a speaker-dependent algorithm, that was trained on the target speaker used for testing, and a speaker-independent algorithm, that was trained on different speakers. Significant improvements in the intelligibility of speech in stationary and fluctuating noises were found relative to the unprocessed condition for the speaker-dependent algorithm in all noise types and for the speaker-independent algorithm in 2 out of 3 noise types. The NNSE algorithms used noise-specific neural networks that generalized to novel segments of the same noise type and worked over a range of SNRs. The proposed algorithm has the potential to improve the intelligibility of speech in noise for CI users while meeting the requirements of low computational complexity and processing delay for application in CI devices. Copyright © 2016 The Authors. Published by Elsevier B.V. All rights reserved.

Improving Pattern Recognition and Neural Network Algorithms with Applications to Solar Panel Energy Optimization

NASA Astrophysics Data System (ADS)

Zamora Ramos, Ernesto

Artificial Intelligence is a big part of automation and with today's technological advances, artificial intelligence has taken great strides towards positioning itself as the technology of the future to control, enhance and perfect automation. Computer vision includes pattern recognition and classification and machine learning. Computer vision is at the core of decision making and it is a vast and fruitful branch of artificial intelligence. In this work, we expose novel algorithms and techniques built upon existing technologies to improve pattern recognition and neural network training, initially motivated by a multidisciplinary effort to build a robot that helps maintain and optimize solar panel energy production. Our contributions detail an improved non-linear pre-processing technique to enhance poorly illuminated images based on modifications to the standard histogram equalization for an image. While the original motivation was to improve nocturnal navigation, the results have applications in surveillance, search and rescue, medical imaging enhancing, and many others. We created a vision system for precise camera distance positioning motivated to correctly locate the robot for capture of solar panel images for classification. The classification algorithm marks solar panels as clean or dirty for later processing. Our algorithm extends past image classification and, based on historical and experimental data, it identifies the optimal moment in which to perform maintenance on marked solar panels as to minimize the energy and profit loss. In order to improve upon the classification algorithm, we delved into feedforward neural networks because of their recent advancements, proven universal approximation and classification capabilities, and excellent recognition rates. We explore state-of-the-art neural network training techniques offering pointers and insights, culminating on the implementation of a complete library with support for modern deep learning architectures, multilayer percepterons and convolutional neural networks. Our research with neural networks has encountered a great deal of difficulties regarding hyperparameter estimation for good training convergence rate and accuracy. Most hyperparameters, including architecture, learning rate, regularization, trainable parameters (or weights) initialization, and so on, are chosen via a trial and error process with some educated guesses. However, we developed the first quantitative method to compare weight initialization strategies, a critical hyperparameter choice during training, to estimate among a group of candidate strategies which would make the network converge to the highest classification accuracy faster with high probability. Our method provides a quick, objective measure to compare initialization strategies to select the best possible among them beforehand without having to complete multiple training sessions for each candidate strategy to compare final results.

Network-centric decision architecture for financial or 1/f data models

NASA Astrophysics Data System (ADS)

Jaenisch, Holger M.; Handley, James W.; Massey, Stoney; Case, Carl T.; Songy, Claude G.

2002-12-01

This paper presents a decision architecture algorithm for training neural equation based networks to make autonomous multi-goal oriented, multi-class decisions. These architectures make decisions based on their individual goals and draw from the same network centric feature set. Traditionally, these architectures are comprised of neural networks that offer marginal performance due to lack of convergence of the training set. We present an approach for autonomously extracting sample points as I/O exemplars for generation of multi-branch, multi-node decision architectures populated by adaptively derived neural equations. To test the robustness of this architecture, open source data sets in the form of financial time series were used, requiring a three-class decision space analogous to the lethal, non-lethal, and clutter discrimination problem. This algorithm and the results of its application are presented here.

Nutrient Stress Detection in Corn Using Neural Networks and AVIRIS Hyperspectral Imagery

NASA Technical Reports Server (NTRS)

Estep, Lee

2001-01-01

AVIRIS image cube data has been processed for the detection of nutrient stress in corn by both known, ratio-type algorithms and by trained neural networks. The USDA Shelton, NE, ARS Variable Rate Nitrogen Application (VRAT) experimental farm was the site used in the study. Upon application of ANOVA and Dunnett multiple comparsion tests on the outcome of both the neural network processing and the ratio-type algorithm results, it was found that the neural network methodology provides a better overall capability to separate nutrient stressed crops from in-field controls.

Edge detection for optical synthetic aperture based on deep neural network

NASA Astrophysics Data System (ADS)

Tan, Wenjie; Hui, Mei; Liu, Ming; Kong, Lingqin; Dong, Liquan; Zhao, Yuejin

2017-09-01

Synthetic aperture optics systems can meet the demands of the next-generation space telescopes being lighter, larger and foldable. However, the boundaries of segmented aperture systems are much more complex than that of the whole aperture. More edge regions mean more imaging edge pixels, which are often mixed and discretized. In order to achieve high-resolution imaging, it is necessary to identify the gaps between the sub-apertures and the edges of the projected fringes. In this work, we introduced the algorithm of Deep Neural Network into the edge detection of optical synthetic aperture imaging. According to the detection needs, we constructed image sets by experiments and simulations. Based on MatConvNet, a toolbox of MATLAB, we ran the neural network, trained it on training image set and tested its performance on validation set. The training was stopped when the test error on validation set stopped declining. As an input image is given, each intra-neighbor area around the pixel is taken into the network, and scanned pixel by pixel with the trained multi-hidden layers. The network outputs make a judgment on whether the center of the input block is on edge of fringes. We experimented with various pre-processing and post-processing techniques to reveal their influence on edge detection performance. Compared with the traditional algorithms or their improvements, our method makes decision on a much larger intra-neighbor, and is more global and comprehensive. Experiments on more than 2,000 images are also given to prove that our method outperforms classical algorithms in optical images-based edge detection.

Classification of E-Nose Aroma Data of Four Fruit Types by ABC-Based Neural Network

PubMed Central

Adak, M. Fatih; Yumusak, Nejat

2016-01-01

Electronic nose technology is used in many areas, and frequently in the beverage industry for classification and quality-control purposes. In this study, four different aroma data (strawberry, lemon, cherry, and melon) were obtained using a MOSES II electronic nose for the purpose of fruit classification. To improve the performance of the classification, the training phase of the neural network with two hidden layers was optimized using artificial bee colony algorithm (ABC), which is known to be successful in exploration. Test data were given to two different neural networks, each of which were trained separately with backpropagation (BP) and ABC, and average test performances were measured as 60% for the artificial neural network trained with BP and 76.39% for the artificial neural network trained with ABC. Training and test phases were repeated 30 times to obtain these average performance measurements. This level of performance shows that the artificial neural network trained with ABC is successful in classifying aroma data. PMID:26927124

Classification of E-Nose Aroma Data of Four Fruit Types by ABC-Based Neural Network.

PubMed

Adak, M Fatih; Yumusak, Nejat

2016-02-27

Electronic nose technology is used in many areas, and frequently in the beverage industry for classification and quality-control purposes. In this study, four different aroma data (strawberry, lemon, cherry, and melon) were obtained using a MOSES II electronic nose for the purpose of fruit classification. To improve the performance of the classification, the training phase of the neural network with two hidden layers was optimized using artificial bee colony algorithm (ABC), which is known to be successful in exploration. Test data were given to two different neural networks, each of which were trained separately with backpropagation (BP) and ABC, and average test performances were measured as 60% for the artificial neural network trained with BP and 76.39% for the artificial neural network trained with ABC. Training and test phases were repeated 30 times to obtain these average performance measurements. This level of performance shows that the artificial neural network trained with ABC is successful in classifying aroma data.

WS-BP: An efficient wolf search based back-propagation algorithm

NASA Astrophysics Data System (ADS)

Nawi, Nazri Mohd; Rehman, M. Z.; Khan, Abdullah

2015-05-01

Wolf Search (WS) is a heuristic based optimization algorithm. Inspired by the preying and survival capabilities of the wolves, this algorithm is highly capable to search large spaces in the candidate solutions. This paper investigates the use of WS algorithm in combination with back-propagation neural network (BPNN) algorithm to overcome the local minima problem and to improve convergence in gradient descent. The performance of the proposed Wolf Search based Back-Propagation (WS-BP) algorithm is compared with Artificial Bee Colony Back-Propagation (ABC-BP), Bat Based Back-Propagation (Bat-BP), and conventional BPNN algorithms. Specifically, OR and XOR datasets are used for training the network. The simulation results show that the WS-BP algorithm effectively avoids the local minima and converge to global minima.

The application of neural networks to the SSME startup transient

NASA Technical Reports Server (NTRS)

Meyer, Claudia M.; Maul, William A.

1991-01-01

Feedforward neural networks were used to model three parameters during the Space Shuttle Main Engine startup transient. The three parameters were the main combustion chamber pressure, a controlled parameter, the high pressure oxidizer turbine discharge temperature, a redlined parameter, and the high pressure fuel pump discharge pressure, a failure-indicating performance parameter. Network inputs consisted of time windows of data from engine measurements that correlated highly to the modeled parameter. A standard backpropagation algorithm was used to train the feedforward networks on two nominal firings. Each trained network was validated with four additional nominal firings. For all three parameters, the neural networks were able to accurately predict the data in the validation sets as well as the training set.

Automated target recognition and tracking using an optical pattern recognition neural network

NASA Technical Reports Server (NTRS)

Chao, Tien-Hsin

1991-01-01

The on-going development of an automatic target recognition and tracking system at the Jet Propulsion Laboratory is presented. This system is an optical pattern recognition neural network (OPRNN) that is an integration of an innovative optical parallel processor and a feature extraction based neural net training algorithm. The parallel optical processor provides high speed and vast parallelism as well as full shift invariance. The neural network algorithm enables simultaneous discrimination of multiple noisy targets in spite of their scales, rotations, perspectives, and various deformations. This fully developed OPRNN system can be effectively utilized for the automated spacecraft recognition and tracking that will lead to success in the Automated Rendezvous and Capture (AR&C) of the unmanned Cargo Transfer Vehicle (CTV). One of the most powerful optical parallel processors for automatic target recognition is the multichannel correlator. With the inherent advantages of parallel processing capability and shift invariance, multiple objects can be simultaneously recognized and tracked using this multichannel correlator. This target tracking capability can be greatly enhanced by utilizing a powerful feature extraction based neural network training algorithm such as the neocognitron. The OPRNN, currently under investigation at JPL, is constructed with an optical multichannel correlator where holographic filters have been prepared using the neocognitron training algorithm. The computation speed of the neocognitron-type OPRNN is up to 10(exp 14) analog connections/sec that enabling the OPRNN to outperform its state-of-the-art electronics counterpart by at least two orders of magnitude.

A hybrid intelligent algorithm for portfolio selection problem with fuzzy returns

NASA Astrophysics Data System (ADS)

Li, Xiang; Zhang, Yang; Wong, Hau-San; Qin, Zhongfeng

2009-11-01

Portfolio selection theory with fuzzy returns has been well developed and widely applied. Within the framework of credibility theory, several fuzzy portfolio selection models have been proposed such as mean-variance model, entropy optimization model, chance constrained programming model and so on. In order to solve these nonlinear optimization models, a hybrid intelligent algorithm is designed by integrating simulated annealing algorithm, neural network and fuzzy simulation techniques, where the neural network is used to approximate the expected value and variance for fuzzy returns and the fuzzy simulation is used to generate the training data for neural network. Since these models are used to be solved by genetic algorithm, some comparisons between the hybrid intelligent algorithm and genetic algorithm are given in terms of numerical examples, which imply that the hybrid intelligent algorithm is robust and more effective. In particular, it reduces the running time significantly for large size problems.

Satellite image analysis using neural networks

NASA Technical Reports Server (NTRS)

Sheldon, Roger A.

1990-01-01

The tremendous backlog of unanalyzed satellite data necessitates the development of improved methods for data cataloging and analysis. Ford Aerospace has developed an image analysis system, SIANN (Satellite Image Analysis using Neural Networks) that integrates the technologies necessary to satisfy NASA's science data analysis requirements for the next generation of satellites. SIANN will enable scientists to train a neural network to recognize image data containing scenes of interest and then rapidly search data archives for all such images. The approach combines conventional image processing technology with recent advances in neural networks to provide improved classification capabilities. SIANN allows users to proceed through a four step process of image classification: filtering and enhancement, creation of neural network training data via application of feature extraction algorithms, configuring and training a neural network model, and classification of images by application of the trained neural network. A prototype experimentation testbed was completed and applied to climatological data.

Generalised Transfer Functions of Neural Networks

NASA Astrophysics Data System (ADS)

Fung, C. F.; Billings, S. A.; Zhang, H.

1997-11-01

When artificial neural networks are used to model non-linear dynamical systems, the system structure which can be extremely useful for analysis and design, is buried within the network architecture. In this paper, explicit expressions for the frequency response or generalised transfer functions of both feedforward and recurrent neural networks are derived in terms of the network weights. The derivation of the algorithm is established on the basis of the Taylor series expansion of the activation functions used in a particular neural network. This leads to a representation which is equivalent to the non-linear recursive polynomial model and enables the derivation of the transfer functions to be based on the harmonic expansion method. By mapping the neural network into the frequency domain information about the structure of the underlying non-linear system can be recovered. Numerical examples are included to demonstrate the application of the new algorithm. These examples show that the frequency response functions appear to be highly sensitive to the network topology and training, and that the time domain properties fail to reveal deficiencies in the trained network structure.

SKYNET: an efficient and robust neural network training tool for machine learning in astronomy

NASA Astrophysics Data System (ADS)

Graff, Philip; Feroz, Farhan; Hobson, Michael P.; Lasenby, Anthony

2014-06-01

We present the first public release of our generic neural network training algorithm, called SKYNET. This efficient and robust machine learning tool is able to train large and deep feed-forward neural networks, including autoencoders, for use in a wide range of supervised and unsupervised learning applications, such as regression, classification, density estimation, clustering and dimensionality reduction. SKYNET uses a `pre-training' method to obtain a set of network parameters that has empirically been shown to be close to a good solution, followed by further optimization using a regularized variant of Newton's method, where the level of regularization is determined and adjusted automatically; the latter uses second-order derivative information to improve convergence, but without the need to evaluate or store the full Hessian matrix, by using a fast approximate method to calculate Hessian-vector products. This combination of methods allows for the training of complicated networks that are difficult to optimize using standard backpropagation techniques. SKYNET employs convergence criteria that naturally prevent overfitting, and also includes a fast algorithm for estimating the accuracy of network outputs. The utility and flexibility of SKYNET are demonstrated by application to a number of toy problems, and to astronomical problems focusing on the recovery of structure from blurred and noisy images, the identification of gamma-ray bursters, and the compression and denoising of galaxy images. The SKYNET software, which is implemented in standard ANSI C and fully parallelized using MPI, is available at http://www.mrao.cam.ac.uk/software/skynet/.

Predicting coronary artery disease using different artificial neural network models.

PubMed

Colak, M Cengiz; Colak, Cemil; Kocatürk, Hasan; Sağiroğlu, Seref; Barutçu, Irfan

2008-08-01

Eight different learning algorithms used for creating artificial neural network (ANN) models and the different ANN models in the prediction of coronary artery disease (CAD) are introduced. This work was carried out as a retrospective case-control study. Overall, 124 consecutive patients who had been diagnosed with CAD by coronary angiography (at least 1 coronary stenosis > 50% in major epicardial arteries) were enrolled in the work. Angiographically, the 113 people (group 2) with normal coronary arteries were taken as control subjects. Multi-layered perceptrons ANN architecture were applied. The ANN models trained with different learning algorithms were performed in 237 records, divided into training (n=171) and testing (n=66) data sets. The performance of prediction was evaluated by sensitivity, specificity and accuracy values based on standard definitions. The results have demonstrated that ANN models trained with eight different learning algorithms are promising because of high (greater than 71%) sensitivity, specificity and accuracy values in the prediction of CAD. Accuracy, sensitivity and specificity values varied between 83.63%-100%, 86.46%-100% and 74.67%-100% for training, respectively. For testing, the values were more than 71% for sensitivity, 76% for specificity and 81% for accuracy. It may be proposed that the use of different learning algorithms other than backpropagation and larger sample sizes can improve the performance of prediction. The proposed ANN models trained with these learning algorithms could be used a promising approach for predicting CAD without the need for invasive diagnostic methods and could help in the prognostic clinical decision.

An improved multi-domain convolution tracking algorithm

NASA Astrophysics Data System (ADS)

Sun, Xin; Wang, Haiying; Zeng, Yingsen

2018-04-01

Along with the wide application of the Deep Learning in the field of Computer vision, Deep learning has become a mainstream direction in the field of object tracking. The tracking algorithm in this paper is based on the improved multidomain convolution neural network, and the VOT video set is pre-trained on the network by multi-domain training strategy. In the process of online tracking, the network evaluates candidate targets sampled from vicinity of the prediction target in the previous with Gaussian distribution, and the candidate target with the highest score is recognized as the prediction target of this frame. The Bounding Box Regression model is introduced to make the prediction target closer to the ground-truths target box of the test set. Grouping-update strategy is involved to extract and select useful update samples in each frame, which can effectively prevent over fitting. And adapt to changes in both target and environment. To improve the speed of the algorithm while maintaining the performance, the number of candidate target succeed in adjusting dynamically with the help of Self-adaption parameter Strategy. Finally, the algorithm is tested by OTB set, compared with other high-performance tracking algorithms, and the plot of success rate and the accuracy are drawn. which illustrates outstanding performance of the tracking algorithm in this paper.

Neural Network and Response Surface Methodology for Rocket Engine Component Optimization

NASA Technical Reports Server (NTRS)

Vaidyanathan, Rajkumar; Papita, Nilay; Shyy, Wei; Tucker, P. Kevin; Griffin, Lisa W.; Haftka, Raphael; Fitz-Coy, Norman; McConnaughey, Helen (Technical Monitor)

2000-01-01

The goal of this work is to compare the performance of response surface methodology (RSM) and two types of neural networks (NN) to aid preliminary design of two rocket engine components. A data set of 45 training points and 20 test points obtained from a semi-empirical model based on three design variables is used for a shear coaxial injector element. Data for supersonic turbine design is based on six design variables, 76 training, data and 18 test data obtained from simplified aerodynamic analysis. Several RS and NN are first constructed using the training data. The test data are then employed to select the best RS or NN. Quadratic and cubic response surfaces. radial basis neural network (RBNN) and back-propagation neural network (BPNN) are compared. Two-layered RBNN are generated using two different training algorithms, namely solverbe and solverb. A two layered BPNN is generated with Tan-Sigmoid transfer function. Various issues related to the training of the neural networks are addressed including number of neurons, error goals, spread constants and the accuracy of different models in representing the design space. A search for the optimum design is carried out using a standard gradient-based optimization algorithm over the response surfaces represented by the polynomials and trained neural networks. Usually a cubic polynominal performs better than the quadratic polynomial but exceptions have been noticed. Among the NN choices, the RBNN designed using solverb yields more consistent performance for both engine components considered. The training of RBNN is easier as it requires linear regression. This coupled with the consistency in performance promise the possibility of it being used as an optimization strategy for engineering design problems.

Solar radiation and precipitable water modeling for Turkey using artificial neural networks

NASA Astrophysics Data System (ADS)

Şenkal, Ozan

2015-08-01

Artificial neural network (ANN) method was applied for modeling and prediction of mean precipitable water and solar radiation in a given location and given date (month), given altitude, temperature, pressure and humidity in Turkey (26-45ºE and 36-42ºN) during the period of 2000-2002. Resilient Propagation (RP) learning algorithms and logistic sigmoid transfer function were used in the network. To train the network, meteorological measurements taken by the Turkish State Meteorological Service (TSMS) and Wyoming University for the period from 2000 to 2002 from five stations distributed in Turkey were used as training data. Data from years (2000 and 2001) were used for training, while the year 2002 was used for testing and validating the model. The RP algorithm were first used for determination of the precipitable water and subsequently, computation of the solar radiation, in these stations Root Mean Square Error (RMSE) between the estimated and measured values for monthly mean daily sum for precipitable water and solar radiation values have been found as 0.0062 gr/cm2 and 0.0603 MJ/m2 (training cities), 0.5652 gr/cm2 and 3.2810 MJ/m2 (testing cities), respectively.

Vision-based mobile robot navigation through deep convolutional neural networks and end-to-end learning

NASA Astrophysics Data System (ADS)

Zhang, Yachu; Zhao, Yuejin; Liu, Ming; Dong, Liquan; Kong, Lingqin; Liu, Lingling

2017-09-01

In contrast to humans, who use only visual information for navigation, many mobile robots use laser scanners and ultrasonic sensors along with vision cameras to navigate. This work proposes a vision-based robot control algorithm based on deep convolutional neural networks. We create a large 15-layer convolutional neural network learning system and achieve the advanced recognition performance. Our system is trained from end to end to map raw input images to direction in supervised mode. The images of data sets are collected in a wide variety of weather conditions and lighting conditions. Besides, the data sets are augmented by adding Gaussian noise and Salt-and-pepper noise to avoid overfitting. The algorithm is verified by two experiments, which are line tracking and obstacle avoidance. The line tracking experiment is proceeded in order to track the desired path which is composed of straight and curved lines. The goal of obstacle avoidance experiment is to avoid the obstacles indoor. Finally, we get 3.29% error rate on the training set and 5.1% error rate on the test set in the line tracking experiment, 1.8% error rate on the training set and less than 5% error rate on the test set in the obstacle avoidance experiment. During the actual test, the robot can follow the runway centerline outdoor and avoid the obstacle in the room accurately. The result confirms the effectiveness of the algorithm and our improvement in the network structure and train parameters

Prediction of air pollutant concentration based on sparse response back-propagation training feedforward neural networks.

PubMed

Ding, Weifu; Zhang, Jiangshe; Leung, Yee

2016-10-01

In this paper, we predict air pollutant concentration using a feedforward artificial neural network inspired by the mechanism of the human brain as a useful alternative to traditional statistical modeling techniques. The neural network is trained based on sparse response back-propagation in which only a small number of neurons respond to the specified stimulus simultaneously and provide a high convergence rate for the trained network, in addition to low energy consumption and greater generalization. Our method is evaluated on Hong Kong air monitoring station data and corresponding meteorological variables for which five air quality parameters were gathered at four monitoring stations in Hong Kong over 4 years (2012-2015). Our results show that our training method has more advantages in terms of the precision of the prediction, effectiveness, and generalization of traditional linear regression algorithms when compared with a feedforward artificial neural network trained using traditional back-propagation.

A review and analysis of neural networks for classification of remotely sensed multispectral imagery

NASA Technical Reports Server (NTRS)

Paola, Justin D.; Schowengerdt, Robert A.

1993-01-01

A literature survey and analysis of the use of neural networks for the classification of remotely sensed multispectral imagery is presented. As part of a brief mathematical review, the backpropagation algorithm, which is the most common method of training multi-layer networks, is discussed with an emphasis on its application to pattern recognition. The analysis is divided into five aspects of neural network classification: (1) input data preprocessing, structure, and encoding; (2) output encoding and extraction of classes; (3) network architecture, (4) training algorithms; and (5) comparisons to conventional classifiers. The advantages of the neural network method over traditional classifiers are its non-parametric nature, arbitrary decision boundary capabilities, easy adaptation to different types of data and input structures, fuzzy output values that can enhance classification, and good generalization for use with multiple images. The disadvantages of the method are slow training time, inconsistent results due to random initial weights, and the requirement of obscure initialization values (e.g., learning rate and hidden layer size). Possible techniques for ameliorating these problems are discussed. It is concluded that, although the neural network method has several unique capabilities, it will become a useful tool in remote sensing only if it is made faster, more predictable, and easier to use.

Multi-modality image fusion based on enhanced fuzzy radial basis function neural networks.

PubMed

Chao, Zhen; Kim, Dohyeon; Kim, Hee-Joung

2018-04-01

In clinical applications, single modality images do not provide sufficient diagnostic information. Therefore, it is necessary to combine the advantages or complementarities of different modalities of images. Recently, neural network technique was applied to medical image fusion by many researchers, but there are still many deficiencies. In this study, we propose a novel fusion method to combine multi-modality medical images based on the enhanced fuzzy radial basis function neural network (Fuzzy-RBFNN), which includes five layers: input, fuzzy partition, front combination, inference, and output. Moreover, we propose a hybrid of the gravitational search algorithm (GSA) and error back propagation algorithm (EBPA) to train the network to update the parameters of the network. Two different patterns of images are used as inputs of the neural network, and the output is the fused image. A comparison with the conventional fusion methods and another neural network method through subjective observation and objective evaluation indexes reveals that the proposed method effectively synthesized the information of input images and achieved better results. Meanwhile, we also trained the network by using the EBPA and GSA, individually. The results reveal that the EBPGSA not only outperformed both EBPA and GSA, but also trained the neural network more accurately by analyzing the same evaluation indexes. Copyright © 2018 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.

Research on Abnormal Detection Based on Improved Combination of K - means and SVDD

NASA Astrophysics Data System (ADS)

Hao, Xiaohong; Zhang, Xiaofeng

2018-01-01

In order to improve the efficiency of network intrusion detection and reduce the false alarm rate, this paper proposes an anomaly detection algorithm based on improved K-means and SVDD. The algorithm first uses the improved K-means algorithm to cluster the training samples of each class, so that each class is independent and compact in class; Then, according to the training samples, the SVDD algorithm is used to construct the minimum superspheres. The subordinate relationship of the samples is determined by calculating the distance of the minimum superspheres constructed by SVDD. If the test sample is less than the center of the hypersphere, the test sample belongs to this class, otherwise it does not belong to this class, after several comparisons, the final test of the effective detection of the test sample.In this paper, we use KDD CUP99 data set to simulate the proposed anomaly detection algorithm. The results show that the algorithm has high detection rate and low false alarm rate, which is an effective network security protection method.

Automated isotope identification algorithm using artificial neural networks

DOE Office of Scientific and Technical Information (OSTI.GOV)

Kamuda, Mark; Stinnett, Jacob; Sullivan, Clair

There is a need to develop an algorithm that can determine the relative activities of radio-isotopes in a large dataset of low-resolution gamma-ray spectra that contains a mixture of many radio-isotopes. Low-resolution gamma-ray spectra that contain mixtures of radio-isotopes often exhibit feature over-lap, requiring algorithms that can analyze these features when overlap occurs. While machine learning and pattern recognition algorithms have shown promise for the problem of radio-isotope identification, their ability to identify and quantify mixtures of radio-isotopes has not been studied. Because machine learning algorithms use abstract features of the spectrum, such as the shape of overlapping peaks andmore » Compton continuum, they are a natural choice for analyzing radio-isotope mixtures. An artificial neural network (ANN) has be trained to calculate the relative activities of 32 radio-isotopes in a spectrum. Furthermore, the ANN is trained with simulated gamma-ray spectra, allowing easy expansion of the library of target radio-isotopes. In this paper we present our initial algorithms based on an ANN and evaluate them against a series measured and simulated spectra.« less

Automated isotope identification algorithm using artificial neural networks

DOE PAGES

Kamuda, Mark; Stinnett, Jacob; Sullivan, Clair

2017-04-12

There is a need to develop an algorithm that can determine the relative activities of radio-isotopes in a large dataset of low-resolution gamma-ray spectra that contains a mixture of many radio-isotopes. Low-resolution gamma-ray spectra that contain mixtures of radio-isotopes often exhibit feature over-lap, requiring algorithms that can analyze these features when overlap occurs. While machine learning and pattern recognition algorithms have shown promise for the problem of radio-isotope identification, their ability to identify and quantify mixtures of radio-isotopes has not been studied. Because machine learning algorithms use abstract features of the spectrum, such as the shape of overlapping peaks andmore » Compton continuum, they are a natural choice for analyzing radio-isotope mixtures. An artificial neural network (ANN) has be trained to calculate the relative activities of 32 radio-isotopes in a spectrum. Furthermore, the ANN is trained with simulated gamma-ray spectra, allowing easy expansion of the library of target radio-isotopes. In this paper we present our initial algorithms based on an ANN and evaluate them against a series measured and simulated spectra.« less

Recursive Bayesian recurrent neural networks for time-series modeling.

PubMed

Mirikitani, Derrick T; Nikolaev, Nikolay

2010-02-01

This paper develops a probabilistic approach to recursive second-order training of recurrent neural networks (RNNs) for improved time-series modeling. A general recursive Bayesian Levenberg-Marquardt algorithm is derived to sequentially update the weights and the covariance (Hessian) matrix. The main strengths of the approach are a principled handling of the regularization hyperparameters that leads to better generalization, and stable numerical performance. The framework involves the adaptation of a noise hyperparameter and local weight prior hyperparameters, which represent the noise in the data and the uncertainties in the model parameters. Experimental investigations using artificial and real-world data sets show that RNNs equipped with the proposed approach outperform standard real-time recurrent learning and extended Kalman training algorithms for recurrent networks, as well as other contemporary nonlinear neural models, on time-series modeling.

Hacia la predicción del Número R de Wolf de manchas solares utilizando Redes Neuronales con retardos temporales

NASA Astrophysics Data System (ADS)

Francile, C.; Luoni, M. L.

We present a prediction of the time series of the Wolf number R of sunspots using "time lagged feed forward neural networks". We use two types of networks: the focused and distributed ones which were trained with the back propagation of errors algorithm and the temporal back propagation algorithm respectively. As inputs to neural networks we use the time series of the number R averaged annually and monthly with the method IR5. As data sets for training and test we choose certain intervals of the time series similar to other works, in order to compare the results. Finally we discuss the topology of the networks used, the number of delays used, the number of neurons per layer, the number of hidden layers and the results in the prediction of the series between one and six steps ahead. FULL TEXT IN SPANISH

Improved Results for Route Planning in Stochastic Transportation Networks

NASA Technical Reports Server (NTRS)

Boyan, Justin; Mitzenmacher, Michael

2000-01-01

In the bus network problem, the goal is to generate a plan for getting from point X to point Y within a city using buses in the smallest expected time. Because bus arrival times are not determined by a fixed schedule but instead may be random. the problem requires more than standard shortest path techniques. In recent work, Datar and Ranade provide algorithms in the case where bus arrivals are assumed to be independent and exponentially distributed. We offer solutions to two important generalizations of the problem, answering open questions posed by Datar and Ranade. First, we provide a polynomial time algorithm for a much wider class of arrival distributions, namely those with increasing failure rate. This class includes not only exponential distributions but also uniform, normal, and gamma distributions. Second, in the case where bus arrival times are independent and geometric discrete random variable,. we provide an algorithm for transportation networks of buses and trains, where trains run according to a fixed schedule.

Cloud Classification in Polar and Desert Regions and Smoke Classification from Biomass Burning Using a Hierarchical Neural Network

NASA Technical Reports Server (NTRS)

Alexander, June; Corwin, Edward; Lloyd, David; Logar, Antonette; Welch, Ronald

1996-01-01

This research focuses on a new neural network scene classification technique. The task is to identify scene elements in Advanced Very High Resolution Radiometry (AVHRR) data from three scene types: polar, desert and smoke from biomass burning in South America (smoke). The ultimate goal of this research is to design and implement a computer system which will identify the clouds present on a whole-Earth satellite view as a means of tracking global climate changes. Previous research has reported results for rule-based systems (Tovinkere et at 1992, 1993) for standard back propagation (Watters et at. 1993) and for a hierarchical approach (Corwin et al 1994) for polar data. This research uses a hierarchical neural network with don't care conditions and applies this technique to complex scenes. A hierarchical neural network consists of a switching network and a collection of leaf networks. The idea of the hierarchical neural network is that it is a simpler task to classify a certain pattern from a subset of patterns than it is to classify a pattern from the entire set. Therefore, the first task is to cluster the classes into groups. The switching, or decision network, performs an initial classification by selecting a leaf network. The leaf networks contain a reduced set of similar classes, and it is in the various leaf networks that the actual classification takes place. The grouping of classes in the various leaf networks is determined by applying an iterative clustering algorithm. Several clustering algorithms were investigated, but due to the size of the data sets, the exhaustive search algorithms were eliminated. A heuristic approach using a confusion matrix from a lightly trained neural network provided the basis for the clustering algorithm. Once the clusters have been identified, the hierarchical network can be trained. The approach of using don't care nodes results from the difficulty in generating extremely complex surfaces in order to separate one class from all of the others. This approach finds pairwise separating surfaces and forms the more complex separating surface from combinations of simpler surfaces. This technique both reduces training time and improves accuracy over the previously reported results. Accuracies of 97.47%, 95.70%, and 99.05% were achieved for the polar, desert and smoke data sets.

Convolutional Neural Network Based on Extreme Learning Machine for Maritime Ships Recognition in Infrared Images.

PubMed

Khellal, Atmane; Ma, Hongbin; Fei, Qing

2018-05-09

The success of Deep Learning models, notably convolutional neural networks (CNNs), makes them the favorable solution for object recognition systems in both visible and infrared domains. However, the lack of training data in the case of maritime ships research leads to poor performance due to the problem of overfitting. In addition, the back-propagation algorithm used to train CNN is very slow and requires tuning many hyperparameters. To overcome these weaknesses, we introduce a new approach fully based on Extreme Learning Machine (ELM) to learn useful CNN features and perform a fast and accurate classification, which is suitable for infrared-based recognition systems. The proposed approach combines an ELM based learning algorithm to train CNN for discriminative features extraction and an ELM based ensemble for classification. The experimental results on VAIS dataset, which is the largest dataset of maritime ships, confirm that the proposed approach outperforms the state-of-the-art models in term of generalization performance and training speed. For instance, the proposed model is up to 950 times faster than the traditional back-propagation based training of convolutional neural networks, primarily for low-level features extraction.

Human activity recognition based on feature selection in smart home using back-propagation algorithm.

PubMed

Fang, Hongqing; He, Lei; Si, Hao; Liu, Peng; Xie, Xiaolei

2014-09-01

In this paper, Back-propagation(BP) algorithm has been used to train the feed forward neural network for human activity recognition in smart home environments, and inter-class distance method for feature selection of observed motion sensor events is discussed and tested. And then, the human activity recognition performances of neural network using BP algorithm have been evaluated and compared with other probabilistic algorithms: Naïve Bayes(NB) classifier and Hidden Markov Model(HMM). The results show that different feature datasets yield different activity recognition accuracy. The selection of unsuitable feature datasets increases the computational complexity and degrades the activity recognition accuracy. Furthermore, neural network using BP algorithm has relatively better human activity recognition performances than NB classifier and HMM. Copyright © 2014 ISA. Published by Elsevier Ltd. All rights reserved.

The convergence analysis of SpikeProp algorithm with smoothing L1∕2 regularization.

PubMed

Zhao, Junhong; Zurada, Jacek M; Yang, Jie; Wu, Wei

2018-07-01

Unlike the first and the second generation artificial neural networks, spiking neural networks (SNNs) model the human brain by incorporating not only synaptic state but also a temporal component into their operating model. However, their intrinsic properties require expensive computation during training. This paper presents a novel algorithm to SpikeProp for SNN by introducing smoothing L 1∕2 regularization term into the error function. This algorithm makes the network structure sparse, with some smaller weights that can be eventually removed. Meanwhile, the convergence of this algorithm is proved under some reasonable conditions. The proposed algorithms have been tested for the convergence speed, the convergence rate and the generalization on the classical XOR-problem, Iris problem and Wisconsin Breast Cancer classification. Copyright © 2018 Elsevier Ltd. All rights reserved.

Decoding small surface codes with feedforward neural networks

NASA Astrophysics Data System (ADS)

Varsamopoulos, Savvas; Criger, Ben; Bertels, Koen

2018-01-01

Surface codes reach high error thresholds when decoded with known algorithms, but the decoding time will likely exceed the available time budget, especially for near-term implementations. To decrease the decoding time, we reduce the decoding problem to a classification problem that a feedforward neural network can solve. We investigate quantum error correction and fault tolerance at small code distances using neural network-based decoders, demonstrating that the neural network can generalize to inputs that were not provided during training and that they can reach similar or better decoding performance compared to previous algorithms. We conclude by discussing the time required by a feedforward neural network decoder in hardware.

Big Data: A Parallel Particle Swarm Optimization-Back-Propagation Neural Network Algorithm Based on MapReduce

PubMed Central

Cao, Jianfang; Cui, Hongyan; Shi, Hao; Jiao, Lijuan

2016-01-01

A back-propagation (BP) neural network can solve complicated random nonlinear mapping problems; therefore, it can be applied to a wide range of problems. However, as the sample size increases, the time required to train BP neural networks becomes lengthy. Moreover, the classification accuracy decreases as well. To improve the classification accuracy and runtime efficiency of the BP neural network algorithm, we proposed a parallel design and realization method for a particle swarm optimization (PSO)-optimized BP neural network based on MapReduce on the Hadoop platform using both the PSO algorithm and a parallel design. The PSO algorithm was used to optimize the BP neural network’s initial weights and thresholds and improve the accuracy of the classification algorithm. The MapReduce parallel programming model was utilized to achieve parallel processing of the BP algorithm, thereby solving the problems of hardware and communication overhead when the BP neural network addresses big data. Datasets on 5 different scales were constructed using the scene image library from the SUN Database. The classification accuracy of the parallel PSO-BP neural network algorithm is approximately 92%, and the system efficiency is approximately 0.85, which presents obvious advantages when processing big data. The algorithm proposed in this study demonstrated both higher classification accuracy and improved time efficiency, which represents a significant improvement obtained from applying parallel processing to an intelligent algorithm on big data. PMID:27304987

Low-cost autonomous perceptron neural network inspired by quantum computation

NASA Astrophysics Data System (ADS)

Zidan, Mohammed; Abdel-Aty, Abdel-Haleem; El-Sadek, Alaa; Zanaty, E. A.; Abdel-Aty, Mahmoud

2017-11-01

Achieving low cost learning with reliable accuracy is one of the important goals to achieve intelligent machines to save time, energy and perform learning process over limited computational resources machines. In this paper, we propose an efficient algorithm for a perceptron neural network inspired by quantum computing composite from a single neuron to classify inspirable linear applications after a single training iteration O(1). The algorithm is applied over a real world data set and the results are outer performs the other state-of-the art algorithms.

Smooth function approximation using neural networks.

PubMed

Ferrari, Silvia; Stengel, Robert F

2005-01-01

An algebraic approach for representing multidimensional nonlinear functions by feedforward neural networks is presented. In this paper, the approach is implemented for the approximation of smooth batch data containing the function's input, output, and possibly, gradient information. The training set is associated to the network adjustable parameters by nonlinear weight equations. The cascade structure of these equations reveals that they can be treated as sets of linear systems. Hence, the training process and the network approximation properties can be investigated via linear algebra. Four algorithms are developed to achieve exact or approximate matching of input-output and/or gradient-based training sets. Their application to the design of forward and feedback neurocontrollers shows that algebraic training is characterized by faster execution speeds and better generalization properties than contemporary optimization techniques.

Deep learning in the small sample size setting: cascaded feed forward neural networks for medical image segmentation

NASA Astrophysics Data System (ADS)

Gaonkar, Bilwaj; Hovda, David; Martin, Neil; Macyszyn, Luke

2016-03-01

Deep Learning, refers to large set of neural network based algorithms, have emerged as promising machine- learning tools in the general imaging and computer vision domains. Convolutional neural networks (CNNs), a specific class of deep learning algorithms, have been extremely effective in object recognition and localization in natural images. A characteristic feature of CNNs, is the use of a locally connected multi layer topology that is inspired by the animal visual cortex (the most powerful vision system in existence). While CNNs, perform admirably in object identification and localization tasks, typically require training on extremely large datasets. Unfortunately, in medical image analysis, large datasets are either unavailable or are extremely expensive to obtain. Further, the primary tasks in medical imaging are organ identification and segmentation from 3D scans, which are different from the standard computer vision tasks of object recognition. Thus, in order to translate the advantages of deep learning to medical image analysis, there is a need to develop deep network topologies and training methodologies, that are geared towards medical imaging related tasks and can work in a setting where dataset sizes are relatively small. In this paper, we present a technique for stacked supervised training of deep feed forward neural networks for segmenting organs from medical scans. Each `neural network layer' in the stack is trained to identify a sub region of the original image, that contains the organ of interest. By layering several such stacks together a very deep neural network is constructed. Such a network can be used to identify extremely small regions of interest in extremely large images, inspite of a lack of clear contrast in the signal or easily identifiable shape characteristics. What is even more intriguing is that the network stack achieves accurate segmentation even when it is trained on a single image with manually labelled ground truth. We validate this approach,using a publicly available head and neck CT dataset. We also show that a deep neural network of similar depth, if trained directly using backpropagation, cannot acheive the tasks achieved using our layer wise training paradigm.

Using Hybrid Algorithm to Improve Intrusion Detection in Multi Layer Feed Forward Neural Networks

ERIC Educational Resources Information Center

Ray, Loye Lynn

2014-01-01

The need for detecting malicious behavior on a computer networks continued to be important to maintaining a safe and secure environment. The purpose of this study was to determine the relationship of multilayer feed forward neural network architecture to the ability of detecting abnormal behavior in networks. This involved building, training, and…

Neural Network Classifier Architectures for Phoneme Recognition. CRC Technical Note No. CRC-TN-92-001.

ERIC Educational Resources Information Center

Treurniet, William

A study applied artificial neural networks, trained with the back-propagation learning algorithm, to modelling phonemes extracted from the DARPA TIMIT multi-speaker, continuous speech data base. A number of proposed network architectures were applied to the phoneme classification task, ranging from the simple feedforward multilayer network to more…

Deep learning for medical image segmentation - using the IBM TrueNorth neurosynaptic system

NASA Astrophysics Data System (ADS)

Moran, Steven; Gaonkar, Bilwaj; Whitehead, William; Wolk, Aidan; Macyszyn, Luke; Iyer, Subramanian S.

2018-03-01

Deep convolutional neural networks have found success in semantic image segmentation tasks in computer vision and medical imaging. These algorithms are executed on conventional von Neumann processor architectures or GPUs. This is suboptimal. Neuromorphic processors that replicate the structure of the brain are better-suited to train and execute deep learning models for image segmentation by relying on massively-parallel processing. However, given that they closely emulate the human brain, on-chip hardware and digital memory limitations also constrain them. Adapting deep learning models to execute image segmentation tasks on such chips, requires specialized training and validation. In this work, we demonstrate for the first-time, spinal image segmentation performed using a deep learning network implemented on neuromorphic hardware of the IBM TrueNorth Neurosynaptic System and validate the performance of our network by comparing it to human-generated segmentations of spinal vertebrae and disks. To achieve this on neuromorphic hardware, the training model constrains the coefficients of individual neurons to {-1,0,1} using the Energy Efficient Deep Neuromorphic (EEDN)1 networks training algorithm. Given the 1 million neurons and 256 million synapses, the scale and size of the neural network implemented by the IBM TrueNorth allows us to execute the requisite mapping between segmented images and non-uniform intensity MR images >20 times faster than on a GPU-accelerated network and using <0.1 W. This speed and efficiency implies that a trained neuromorphic chip can be deployed in intra-operative environments where real-time medical image segmentation is necessary.

Voltage control on a train system

DOEpatents

Gordon, Susanna P.; Evans, John A.

2004-01-20

The present invention provides methods for preventing low train voltages and managing interference, thereby improving the efficiency, reliability, and passenger comfort associated with commuter trains. An algorithm implementing neural network technology is used to predict low voltages before they occur. Once voltages are predicted, then multiple trains can be controlled to prevent low voltage events. Further, algorithms for managing inference are presented in the present invention. Different types of interference problems are addressed in the present invention such as "Interference During Acceleration", "Interference Near Station Stops", and "Interference During Delay Recovery." Managing such interference avoids unnecessary brake/acceleration cycles during acceleration, immediately before station stops, and after substantial delays. Algorithms are demonstrated to avoid oscillatory brake/acceleration cycles due to interference and to smooth the trajectories of closely following trains. This is achieved by maintaining sufficient following distances to avoid unnecessary braking/accelerating. These methods generate smooth train trajectories, making for a more comfortable ride, and improve train motor reliability by avoiding unnecessary mode-changes between propulsion and braking. These algorithms can also have a favorable impact on traction power system requirements and energy consumption.

Method of managing interference during delay recovery on a train system

DOEpatents

Gordon, Susanna P.; Evans, John A.

2005-12-27

The present invention provides methods for preventing low train voltages and managing interference, thereby improving the efficiency, reliability, and passenger comfort associated with commuter trains. An algorithm implementing neural network technology is used to predict low voltages before they occur. Once voltages are predicted, then multiple trains can be controlled to prevent low voltage events. Further, algorithms for managing inference are presented in the present invention. Different types of interference problems are addressed in the present invention such as "Interference During Acceleration", "Interference Near Station Stops", and "Interference During Delay Recovery." Managing such interference avoids unnecessary brake/acceleration cycles during acceleration, immediately before station stops, and after substantial delays. Algorithms are demonstrated to avoid oscillatory brake/acceleration cycles due to interference and to smooth the trajectories of closely following trains. This is achieved by maintaining sufficient following distances to avoid unnecessary braking/accelerating. These methods generate smooth train trajectories, making for a more comfortable ride, and improve train motor reliability by avoiding unnecessary mode-changes between propulsion and braking. These algorithms can also have a favorable impact on traction power system requirements and energy consumption.

Efficient high density train operations

DOEpatents

Gordon, Susanna P.; Evans, John A.

2001-01-01

The present invention provides methods for preventing low train voltages and managing interference, thereby improving the efficiency, reliability, and passenger comfort associated with commuter trains. An algorithm implementing neural network technology is used to predict low voltages before they occur. Once voltages are predicted, then multiple trains can be controlled to prevent low voltage events. Further, algorithms for managing inference are presented in the present invention. Different types of interference problems are addressed in the present invention such as "Interference. During Acceleration", "Interference Near Station Stops", and "Interference During Delay Recovery." Managing such interference avoids unnecessary brake/acceleration cycles during acceleration, immediately before station stops, and after substantial delays. Algorithms are demonstrated to avoid oscillatory brake/acceleration cycles due to interference and to smooth the trajectories of closely following trains. This is achieved by maintaining sufficient following distances to avoid unnecessary braking/accelerating. These methods generate smooth train trajectories, making for a more comfortable ride, and improve train motor reliability by avoiding unnecessary mode-changes between propulsion and braking. These algorithms can also have a favorable impact on traction power system requirements and energy consumption.

Hazardous gas detection for FTIR-based hyperspectral imaging system using DNN and CNN

NASA Astrophysics Data System (ADS)

Kim, Yong Chan; Yu, Hyeong-Geun; Lee, Jae-Hoon; Park, Dong-Jo; Nam, Hyun-Woo

2017-10-01

Recently, a hyperspectral imaging system (HIS) with a Fourier Transform InfraRed (FTIR) spectrometer has been widely used due to its strengths in detecting gaseous fumes. Even though numerous algorithms for detecting gaseous fumes have already been studied, it is still difficult to detect target gases properly because of atmospheric interference substances and unclear characteristics of low concentration gases. In this paper, we propose detection algorithms for classifying hazardous gases using a deep neural network (DNN) and a convolutional neural network (CNN). In both the DNN and CNN, spectral signal preprocessing, e.g., offset, noise, and baseline removal, are carried out. In the DNN algorithm, the preprocessed spectral signals are used as feature maps of the DNN with five layers, and it is trained by a stochastic gradient descent (SGD) algorithm (50 batch size) and dropout regularization (0.7 ratio). In the CNN algorithm, preprocessed spectral signals are trained with 1 × 3 convolution layers and 1 × 2 max-pooling layers. As a result, the proposed algorithms improve the classification accuracy rate by 1.5% over the existing support vector machine (SVM) algorithm for detecting and classifying hazardous gases.

A Large-Scale Multi-Hop Localization Algorithm Based on Regularized Extreme Learning for Wireless Networks.

PubMed

Zheng, Wei; Yan, Xiaoyong; Zhao, Wei; Qian, Chengshan

2017-12-20

A novel large-scale multi-hop localization algorithm based on regularized extreme learning is proposed in this paper. The large-scale multi-hop localization problem is formulated as a learning problem. Unlike other similar localization algorithms, the proposed algorithm overcomes the shortcoming of the traditional algorithms which are only applicable to an isotropic network, therefore has a strong adaptability to the complex deployment environment. The proposed algorithm is composed of three stages: data acquisition, modeling and location estimation. In data acquisition stage, the training information between nodes of the given network is collected. In modeling stage, the model among the hop-counts and the physical distances between nodes is constructed using regularized extreme learning. In location estimation stage, each node finds its specific location in a distributed manner. Theoretical analysis and several experiments show that the proposed algorithm can adapt to the different topological environments with low computational cost. Furthermore, high accuracy can be achieved by this method without setting complex parameters.

Design of thrust vectoring exhaust nozzles for real-time applications using neural networks

NASA Technical Reports Server (NTRS)

Prasanth, Ravi K.; Markin, Robert E.; Whitaker, Kevin W.

1991-01-01

Thrust vectoring continues to be an important issue in military aircraft system designs. A recently developed concept of vectoring aircraft thrust makes use of flexible exhaust nozzles. Subtle modifications in the nozzle wall contours produce a non-uniform flow field containing a complex pattern of shock and expansion waves. The end result, due to the asymmetric velocity and pressure distributions, is vectored thrust. Specification of the nozzle contours required for a desired thrust vector angle (an inverse design problem) has been achieved with genetic algorithms. This approach is computationally intensive and prevents the nozzles from being designed in real-time, which is necessary for an operational aircraft system. An investigation was conducted into using genetic algorithms to train a neural network in an attempt to obtain, in real-time, two-dimensional nozzle contours. Results show that genetic algorithm trained neural networks provide a viable, real-time alternative for designing thrust vectoring nozzles contours. Thrust vector angles up to 20 deg were obtained within an average error of 0.0914 deg. The error surfaces encountered were highly degenerate and thus the robustness of genetic algorithms was well suited for minimizing global errors.

Combined neural network/Phillips-Tikhonov approach to aerosol retrievals over land from the NASA Research Scanning Polarimeter

NASA Astrophysics Data System (ADS)

Di Noia, Antonio; Hasekamp, Otto P.; Wu, Lianghai; van Diedenhoven, Bastiaan; Cairns, Brian; Yorks, John E.

2017-11-01

In this paper, an algorithm for the retrieval of aerosol and land surface properties from airborne spectropolarimetric measurements - combining neural networks and an iterative scheme based on Phillips-Tikhonov regularization - is described. The algorithm - which is an extension of a scheme previously designed for ground-based retrievals - is applied to measurements from the Research Scanning Polarimeter (RSP) on board the NASA ER-2 aircraft. A neural network, trained on a large data set of synthetic measurements, is applied to perform aerosol retrievals from real RSP data, and the neural network retrievals are subsequently used as a first guess for the Phillips-Tikhonov retrieval. The resulting algorithm appears capable of accurately retrieving aerosol optical thickness, fine-mode effective radius and aerosol layer height from RSP data. Among the advantages of using a neural network as initial guess for an iterative algorithm are a decrease in processing time and an increase in the number of converging retrievals.

Mapping the Cortical Network Arising From Up-Regulated Amygdaloidal Activation Using -Louvain Algorithm.

PubMed

Liu, Ning; Yu, Xueli; Yao, Li; Zhao, Xiaojie

2018-06-01

The amygdala plays an important role in emotion processing. Several studies have proved that its activation can be regulated by real-time functional magnetic resonance imaging (rtfMRI)-based neurofeedback training. However, although studies have found brain regions that are functionally closely connected to the amygdala in the cortex, it is not clear whether these brain regions and the amygdala are structurally closely connected, and if they show the same training effect as the amygdala in the process of emotional regulation. In this paper, we instructed subjects to up-regulate the activation of the left amygdala (LA) through rtfMRI-based neurofeedback training. In order to fuse multimodal imaging data, we introduced a network analysis method called the -Louvain clustering algorithm. This method was used to integrate multimodal data from the training experiment and construct an LA-cortical network. Correlation analysis and main-effect analysis were conducted to determine the signal covariance associated with the activation of the target area; ultimately, we identified the left temporal pole superior as the amygdaloidal-cortical network region. As a deep nucleus in the brain, the treatment and stimulation of the amygdala remains challenging. Our results provide new insights for the regulation of activation in a deep nucleus using more neurofeedback techniques.

Prediction of properties of wheat dough using intelligent deep belief networks

NASA Astrophysics Data System (ADS)

Guha, Paramita; Bhatnagar, Taru; Pal, Ishan; Kamboj, Uma; Mishra, Sunita

2017-11-01

In this paper, the rheological and chemical properties of wheat dough are predicted using deep belief networks. Wheat grains are stored at controlled environmental conditions. The internal parameters of grains viz., protein, fat, carbohydrates, moisture, ash are determined using standard chemical analysis and viscosity of the dough is measured using Rheometer. Here, fat, carbohydrates, moisture, ash and temperature are considered as inputs whereas protein and viscosity are chosen as outputs. The prediction algorithm is developed using deep neural network where each layer is trained greedily using restricted Boltzmann machine (RBM) networks. The overall network is finally fine-tuned using standard neural network technique. In most literature, it has been found that fine-tuning is done using back-propagation technique. In this paper, a new algorithm is proposed in which each layer is tuned using RBM and the final network is fine-tuned using deep neural network (DNN). It has been observed that with the proposed algorithm, errors between the actual and predicted outputs are less compared to the conventional algorithm. Hence, the given network can be considered as beneficial as it predicts the outputs more accurately. Numerical results along with discussions are presented.

Neural network fusion capabilities for efficient implementation of tracking algorithms

NASA Astrophysics Data System (ADS)

Sundareshan, Malur K.; Amoozegar, Farid

1996-05-01

The ability to efficiently fuse information of different forms for facilitating intelligent decision-making is one of the major capabilities of trained multilayer neural networks that is being recognized int eh recent times. While development of innovative adaptive control algorithms for nonlinear dynamical plants which attempt to exploit these capabilities seems to be more popular, a corresponding development of nonlinear estimation algorithms using these approaches, particularly for application in target surveillance and guidance operations, has not received similar attention. In this paper we describe the capabilities and functionality of neural network algorithms for data fusion and implementation of nonlinear tracking filters. For a discussion of details and for serving as a vehicle for quantitative performance evaluations, the illustrative case of estimating the position and velocity of surveillance targets is considered. Efficient target tracking algorithms that can utilize data from a host of sensing modalities and are capable of reliably tracking even uncooperative targets executing fast and complex maneuvers are of interest in a number of applications. The primary motivation for employing neural networks in these applications comes form the efficiency with which more features extracted from different sensor measurements can be utilized as inputs for estimating target maneuvers. Such an approach results in an overall nonlinear tracking filter which has several advantages over the popular efforts at designing nonlinear estimation algorithms for tracking applications, the principle one being the reduction of mathematical and computational complexities. A system architecture that efficiently integrates the processing capabilities of a trained multilayer neural net with the tracking performance of a Kalman filter is described in this paper.

Neural network cloud top pressure and height for MODIS

NASA Astrophysics Data System (ADS)

Håkansson, Nina; Adok, Claudia; Thoss, Anke; Scheirer, Ronald; Hörnquist, Sara

2018-06-01

Cloud top height retrieval from imager instruments is important for nowcasting and for satellite climate data records. A neural network approach for cloud top height retrieval from the imager instrument MODIS (Moderate Resolution Imaging Spectroradiometer) is presented. The neural networks are trained using cloud top layer pressure data from the CALIOP (Cloud-Aerosol Lidar with Orthogonal Polarization) dataset. Results are compared with two operational reference algorithms for cloud top height: the MODIS Collection 6 Level 2 height product and the cloud top temperature and height algorithm in the 2014 version of the NWC SAF (EUMETSAT (European Organization for the Exploitation of Meteorological Satellites) Satellite Application Facility on Support to Nowcasting and Very Short Range Forecasting) PPS (Polar Platform System). All three techniques are evaluated using both CALIOP and CPR (Cloud Profiling Radar for CloudSat (CLOUD SATellite)) height. Instruments like AVHRR (Advanced Very High Resolution Radiometer) and VIIRS (Visible Infrared Imaging Radiometer Suite) contain fewer channels useful for cloud top height retrievals than MODIS, therefore several different neural networks are investigated to test how infrared channel selection influences retrieval performance. Also a network with only channels available for the AVHRR1 instrument is trained and evaluated. To examine the contribution of different variables, networks with fewer variables are trained. It is shown that variables containing imager information for neighboring pixels are very important. The error distributions of the involved cloud top height algorithms are found to be non-Gaussian. Different descriptive statistic measures are presented and it is exemplified that bias and SD (standard deviation) can be misleading for non-Gaussian distributions. The median and mode are found to better describe the tendency of the error distributions and IQR (interquartile range) and MAE (mean absolute error) are found to give the most useful information of the spread of the errors. For all descriptive statistics presented MAE, IQR, RMSE (root mean square error), SD, mode, median, bias and percentage of absolute errors above 0.25, 0.5, 1 and 2 km the neural network perform better than the reference algorithms both validated with CALIOP and CPR (CloudSat). The neural networks using the brightness temperatures at 11 and 12 µm show at least 32 % (or 623 m) lower MAE compared to the two operational reference algorithms when validating with CALIOP height. Validation with CPR (CloudSat) height gives at least 25 % (or 430 m) reduction of MAE.

Gradient calculations for dynamic recurrent neural networks: a survey.

PubMed

Pearlmutter, B A

1995-01-01

Surveys learning algorithms for recurrent neural networks with hidden units and puts the various techniques into a common framework. The authors discuss fixed point learning algorithms, namely recurrent backpropagation and deterministic Boltzmann machines, and nonfixed point algorithms, namely backpropagation through time, Elman's history cutoff, and Jordan's output feedback architecture. Forward propagation, an on-line technique that uses adjoint equations, and variations thereof, are also discussed. In many cases, the unified presentation leads to generalizations of various sorts. The author discusses advantages and disadvantages of temporally continuous neural networks in contrast to clocked ones continues with some "tricks of the trade" for training, using, and simulating continuous time and recurrent neural networks. The author presents some simulations, and at the end, addresses issues of computational complexity and learning speed.

A Computationally Inexpensive Optimal Guidance via Radial-Basis-Function Neural Network for Autonomous Soft Landing on Asteroids

PubMed Central

Zhang, Peng; Liu, Keping; Zhao, Bo; Li, Yuanchun

2015-01-01

Optimal guidance is essential for the soft landing task. However, due to its high computational complexities, it is hardly applied to the autonomous guidance. In this paper, a computationally inexpensive optimal guidance algorithm based on the radial basis function neural network (RBFNN) is proposed. The optimization problem of the trajectory for soft landing on asteroids is formulated and transformed into a two-point boundary value problem (TPBVP). Combining the database of initial states with the relative initial co-states, an RBFNN is trained offline. The optimal trajectory of the soft landing is determined rapidly by applying the trained network in the online guidance. The Monte Carlo simulations of soft landing on the Eros433 are performed to demonstrate the effectiveness of the proposed guidance algorithm. PMID:26367382

Nonlinear calibration for petroleum water content measurement using PSO

NASA Astrophysics Data System (ADS)

Li, Mingbao; Zhang, Jiawei

2008-10-01

A new algorithmic for strapdown inertial navigation system (SINS) state estimation based on neural networks is introduced. In training strategy, the error vector and its delay are introduced. This error vector is made of the position and velocity difference between the estimations of system and the outputs of GPS. After state prediction and state update, the states of the system are estimated. After off-line training, the network can approach the status switching of SINS and after on-line training, the state estimate precision can be improved further by reducing network output errors. Then the network convergence is discussed. In the end, several simulations with different noise are given. The results show that the neural network state estimator has lower noise sensitivity and better noise immunity than Kalman filter.

Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm.

PubMed

Lee, Jae-Hong; Kim, Do-Hyung; Jeong, Seong-Nyum; Choi, Seong-Ho

2018-04-01

The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%-91.2%) for premolars and 73.4% (95% CI, 59.9%-84.0%) for molars. We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

Neural network fusion capabilities for efficient implementation of tracking algorithms

NASA Astrophysics Data System (ADS)

Sundareshan, Malur K.; Amoozegar, Farid

1997-03-01

The ability to efficiently fuse information of different forms to facilitate intelligent decision making is one of the major capabilities of trained multilayer neural networks that is now being recognized. While development of innovative adaptive control algorithms for nonlinear dynamical plants that attempt to exploit these capabilities seems to be more popular, a corresponding development of nonlinear estimation algorithms using these approaches, particularly for application in target surveillance and guidance operations, has not received similar attention. We describe the capabilities and functionality of neural network algorithms for data fusion and implementation of tracking filters. To discuss details and to serve as a vehicle for quantitative performance evaluations, the illustrative case of estimating the position and velocity of surveillance targets is considered. Efficient target- tracking algorithms that can utilize data from a host of sensing modalities and are capable of reliably tracking even uncooperative targets executing fast and complex maneuvers are of interest in a number of applications. The primary motivation for employing neural networks in these applications comes from the efficiency with which more features extracted from different sensor measurements can be utilized as inputs for estimating target maneuvers. A system architecture that efficiently integrates the fusion capabilities of a trained multilayer neural net with the tracking performance of a Kalman filter is described. The innovation lies in the way the fusion of multisensor data is accomplished to facilitate improved estimation without increasing the computational complexity of the dynamical state estimator itself.

Cardiac Arrhythmia Classification by Multi-Layer Perceptron and Convolution Neural Networks.

PubMed

Savalia, Shalin; Emamian, Vahid

2018-05-04

The electrocardiogram (ECG) plays an imperative role in the medical field, as it records heart signal over time and is used to discover numerous cardiovascular diseases. If a documented ECG signal has a certain irregularity in its predefined features, this is called arrhythmia, the types of which include tachycardia, bradycardia, supraventricular arrhythmias, and ventricular, etc. This has encouraged us to do research that consists of distinguishing between several arrhythmias by using deep neural network algorithms such as multi-layer perceptron (MLP) and convolution neural network (CNN). The TensorFlow library that was established by Google for deep learning and machine learning is used in python to acquire the algorithms proposed here. The ECG databases accessible at PhysioBank.com and kaggle.com were used for training, testing, and validation of the MLP and CNN algorithms. The proposed algorithm consists of four hidden layers with weights, biases in MLP, and four-layer convolution neural networks which map ECG samples to the different classes of arrhythmia. The accuracy of the algorithm surpasses the performance of the current algorithms that have been developed by other cardiologists in both sensitivity and precision.

A Spatiotemporal Prediction Framework for Air Pollution Based on Deep RNN

NASA Astrophysics Data System (ADS)

Fan, J.; Li, Q.; Hou, J.; Feng, X.; Karimian, H.; Lin, S.

2017-10-01

Time series data in practical applications always contain missing values due to sensor malfunction, network failure, outliers etc. In order to handle missing values in time series, as well as the lack of considering temporal properties in machine learning models, we propose a spatiotemporal prediction framework based on missing value processing algorithms and deep recurrent neural network (DRNN). By using missing tag and missing interval to represent time series patterns, we implement three different missing value fixing algorithms, which are further incorporated into deep neural network that consists of LSTM (Long Short-term Memory) layers and fully connected layers. Real-world air quality and meteorological datasets (Jingjinji area, China) are used for model training and testing. Deep feed forward neural networks (DFNN) and gradient boosting decision trees (GBDT) are trained as baseline models against the proposed DRNN. Performances of three missing value fixing algorithms, as well as different machine learning models are evaluated and analysed. Experiments show that the proposed DRNN framework outperforms both DFNN and GBDT, therefore validating the capacity of the proposed framework. Our results also provides useful insights for better understanding of different strategies that handle missing values.

Deterministic convergence of chaos injection-based gradient method for training feedforward neural networks.

PubMed

Zhang, Huisheng; Zhang, Ying; Xu, Dongpo; Liu, Xiaodong

2015-06-01

It has been shown that, by adding a chaotic sequence to the weight update during the training of neural networks, the chaos injection-based gradient method (CIBGM) is superior to the standard backpropagation algorithm. This paper presents the theoretical convergence analysis of CIBGM for training feedforward neural networks. We consider both the case of batch learning as well as the case of online learning. Under mild conditions, we prove the weak convergence, i.e., the training error tends to a constant and the gradient of the error function tends to zero. Moreover, the strong convergence of CIBGM is also obtained with the help of an extra condition. The theoretical results are substantiated by a simulation example.

Thermoelastic steam turbine rotor control based on neural network

NASA Astrophysics Data System (ADS)

Rzadkowski, Romuald; Dominiczak, Krzysztof; Radulski, Wojciech; Szczepanik, R.

2015-12-01

Considered here are Nonlinear Auto-Regressive neural networks with eXogenous inputs (NARX) as a mathematical model of a steam turbine rotor for controlling steam turbine stress on-line. In order to obtain neural networks that locate critical stress and temperature points in the steam turbine during transient states, an FE rotor model was built. This model was used to train the neural networks on the basis of steam turbine transient operating data. The training included nonlinearity related to steam turbine expansion, heat exchange and rotor material properties during transients. Simultaneous neural networks are algorithms which can be implemented on PLC controllers. This allows for the application neural networks to control steam turbine stress in industrial power plants.

Genetic algorithm based adaptive neural network ensemble and its application in predicting carbon flux

USGS Publications Warehouse

Xue, Y.; Liu, S.; Hu, Y.; Yang, J.; Chen, Q.

2007-01-01

To improve the accuracy in prediction, Genetic Algorithm based Adaptive Neural Network Ensemble (GA-ANNE) is presented. Intersections are allowed between different training sets based on the fuzzy clustering analysis, which ensures the diversity as well as the accuracy of individual Neural Networks (NNs). Moreover, to improve the accuracy of the adaptive weights of individual NNs, GA is used to optimize the cluster centers. Empirical results in predicting carbon flux of Duke Forest reveal that GA-ANNE can predict the carbon flux more accurately than Radial Basis Function Neural Network (RBFNN), Bagging NN ensemble, and ANNE. ?? 2007 IEEE.

Scalable training of artificial neural networks with adaptive sparse connectivity inspired by network science.

PubMed

Mocanu, Decebal Constantin; Mocanu, Elena; Stone, Peter; Nguyen, Phuong H; Gibescu, Madeleine; Liotta, Antonio

2018-06-19

Through the success of deep learning in various domains, artificial neural networks are currently among the most used artificial intelligence methods. Taking inspiration from the network properties of biological neural networks (e.g. sparsity, scale-freeness), we argue that (contrary to general practice) artificial neural networks, too, should not have fully-connected layers. Here we propose sparse evolutionary training of artificial neural networks, an algorithm which evolves an initial sparse topology (Erdős-Rényi random graph) of two consecutive layers of neurons into a scale-free topology, during learning. Our method replaces artificial neural networks fully-connected layers with sparse ones before training, reducing quadratically the number of parameters, with no decrease in accuracy. We demonstrate our claims on restricted Boltzmann machines, multi-layer perceptrons, and convolutional neural networks for unsupervised and supervised learning on 15 datasets. Our approach has the potential to enable artificial neural networks to scale up beyond what is currently possible.

Segmentation of the hippocampus by transferring algorithmic knowledge for large cohort processing.

PubMed

Thyreau, Benjamin; Sato, Kazunori; Fukuda, Hiroshi; Taki, Yasuyuki

2018-01-01

The hippocampus is a particularly interesting target for neuroscience research studies due to its essential role within the human brain. In large human cohort studies, bilateral hippocampal structures are frequently identified and measured to gain insight into human behaviour or genomic variability in neuropsychiatric disorders of interest. Automatic segmentation is performed using various algorithms, with FreeSurfer being a popular option. In this manuscript, we present a method to segment the bilateral hippocampus using a deep-learned appearance model. Deep convolutional neural networks (ConvNets) have shown great success in recent years, due to their ability to learn meaningful features from a mass of training data. Our method relies on the following key novelties: (i) we use a wide and variable training set coming from multiple cohorts (ii) our training labels come in part from the output of the FreeSurfer algorithm, and (iii) we include synthetic data and use a powerful data augmentation scheme. Our method proves to be robust, and it has fast inference (<30s total per subject), with trained model available online (https://github.com/bthyreau/hippodeep). We depict illustrative results and show extensive qualitative and quantitative cohort-wide comparisons with FreeSurfer. Our work demonstrates that deep neural-network methods can easily encode, and even improve, existing anatomical knowledge, even when this knowledge exists in algorithmic form. Copyright © 2017 Elsevier B.V. All rights reserved.

Hybrid Neural-Network: Genetic Algorithm Technique for Aircraft Engine Performance Diagnostics Developed and Demonstrated

NASA Technical Reports Server (NTRS)

Kobayashi, Takahisa; Simon, Donald L.

2002-01-01

As part of the NASA Aviation Safety Program, a unique model-based diagnostics method that employs neural networks and genetic algorithms for aircraft engine performance diagnostics has been developed and demonstrated at the NASA Glenn Research Center against a nonlinear gas turbine engine model. Neural networks are applied to estimate the internal health condition of the engine, and genetic algorithms are used for sensor fault detection, isolation, and quantification. This hybrid architecture combines the excellent nonlinear estimation capabilities of neural networks with the capability to rank the likelihood of various faults given a specific sensor suite signature. The method requires a significantly smaller data training set than a neural network approach alone does, and it performs the combined engine health monitoring objectives of performance diagnostics and sensor fault detection and isolation in the presence of nominal and degraded engine health conditions.

Prediction of Aerodynamic Coefficients for Wind Tunnel Data using a Genetic Algorithm Optimized Neural Network

NASA Technical Reports Server (NTRS)

Rajkumar, T.; Aragon, Cecilia; Bardina, Jorge; Britten, Roy

2002-01-01

A fast, reliable way of predicting aerodynamic coefficients is produced using a neural network optimized by a genetic algorithm. Basic aerodynamic coefficients (e.g. lift, drag, pitching moment) are modelled as functions of angle of attack and Mach number. The neural network is first trained on a relatively rich set of data from wind tunnel tests of numerical simulations to learn an overall model. Most of the aerodynamic parameters can be well-fitted using polynomial functions. A new set of data, which can be relatively sparse, is then supplied to the network to produce a new model consistent with the previous model and the new data. Because the new model interpolates realistically between the sparse test data points, it is suitable for use in piloted simulations. The genetic algorithm is used to choose a neural network architecture to give best results, avoiding over-and under-fitting of the test data.

Artificial Intelligence in Prediction of Secondary Protein Structure Using CB513 Database

PubMed Central

Avdagic, Zikrija; Purisevic, Elvir; Omanovic, Samir; Coralic, Zlatan

2009-01-01

In this paper we describe CB513 a non-redundant dataset, suitable for development of algorithms for prediction of secondary protein structure. A program was made in Borland Delphi for transforming data from our dataset to make it suitable for learning of neural network for prediction of secondary protein structure implemented in MATLAB Neural-Network Toolbox. Learning (training and testing) of neural network is researched with different sizes of windows, different number of neurons in the hidden layer and different number of training epochs, while using dataset CB513. PMID:21347158

Tansig activation function (of MLP network) for cardiac abnormality detection

NASA Astrophysics Data System (ADS)

Adnan, Ja'afar; Daud, Nik Ghazali Nik; Ishak, Mohd Taufiq; Rizman, Zairi Ismael; Rahman, Muhammad Izzuddin Abd

2018-02-01

Heart abnormality often occurs regardless of gender, age and races. This problem sometimes does not show any symptoms and it can cause a sudden death to the patient. In general, heart abnormality is the irregular electrical activity of the heart. This paper attempts to develop a program that can detect heart abnormality activity through implementation of Multilayer Perceptron (MLP) network. A certain amount of data of the heartbeat signals from the electrocardiogram (ECG) will be used in this project to train the MLP network by using several training algorithms with Tansig activation function.

Particle analysis using laser ablation mass spectroscopy

DOEpatents

Parker, Eric P.; Rosenthal, Stephen E.; Trahan, Michael W.; Wagner, John S.

2003-09-09

The present invention provides a method of quickly identifying bioaerosols by class, even if the subject bioaerosol has not been previously encountered. The method begins by collecting laser ablation mass spectra from known particles. The spectra are correlated with the known particles, including the species of particle and the classification (e.g., bacteria). The spectra can then be used to train a neural network, for example using genetic algorithm-based training, to recognize each spectra and to recognize characteristics of the classifications. The spectra can also be used in a multivariate patch algorithm. Laser ablation mass specta from unknown particles can be presented as inputs to the trained neural net for identification as to classification. The description below first describes suitable intelligent algorithms and multivariate patch algorithms, then presents an example of the present invention including results.

Novel maximum-margin training algorithms for supervised neural networks.

PubMed

Ludwig, Oswaldo; Nunes, Urbano

2010-06-01

This paper proposes three novel training methods, two of them based on the backpropagation approach and a third one based on information theory for multilayer perceptron (MLP) binary classifiers. Both backpropagation methods are based on the maximal-margin (MM) principle. The first one, based on the gradient descent with adaptive learning rate algorithm (GDX) and named maximum-margin GDX (MMGDX), directly increases the margin of the MLP output-layer hyperplane. The proposed method jointly optimizes both MLP layers in a single process, backpropagating the gradient of an MM-based objective function, through the output and hidden layers, in order to create a hidden-layer space that enables a higher margin for the output-layer hyperplane, avoiding the testing of many arbitrary kernels, as occurs in case of support vector machine (SVM) training. The proposed MM-based objective function aims to stretch out the margin to its limit. An objective function based on Lp-norm is also proposed in order to take into account the idea of support vectors, however, overcoming the complexity involved in solving a constrained optimization problem, usually in SVM training. In fact, all the training methods proposed in this paper have time and space complexities O(N) while usual SVM training methods have time complexity O(N (3)) and space complexity O(N (2)) , where N is the training-data-set size. The second approach, named minimization of interclass interference (MICI), has an objective function inspired on the Fisher discriminant analysis. Such algorithm aims to create an MLP hidden output where the patterns have a desirable statistical distribution. In both training methods, the maximum area under ROC curve (AUC) is applied as stop criterion. The third approach offers a robust training framework able to take the best of each proposed training method. The main idea is to compose a neural model by using neurons extracted from three other neural networks, each one previously trained by MICI, MMGDX, and Levenberg-Marquard (LM), respectively. The resulting neural network was named assembled neural network (ASNN). Benchmark data sets of real-world problems have been used in experiments that enable a comparison with other state-of-the-art classifiers. The results provide evidence of the effectiveness of our methods regarding accuracy, AUC, and balanced error rate.

Resource constrained design of artificial neural networks using comparator neural network

NASA Technical Reports Server (NTRS)

Wah, Benjamin W.; Karnik, Tanay S.

1992-01-01

We present a systematic design method executed under resource constraints for automating the design of artificial neural networks using the back error propagation algorithm. Our system aims at finding the best possible configuration for solving the given application with proper tradeoff between the training time and the network complexity. The design of such a system is hampered by three related problems. First, there are infinitely many possible network configurations, each may take an exceedingly long time to train; hence, it is impossible to enumerate and train all of them to completion within fixed time, space, and resource constraints. Second, expert knowledge on predicting good network configurations is heuristic in nature and is application dependent, rendering it difficult to characterize fully in the design process. A learning procedure that refines this knowledge based on examples on training neural networks for various applications is, therefore, essential. Third, the objective of the network to be designed is ill-defined, as it is based on a subjective tradeoff between the training time and the network cost. A design process that proposes alternate configurations under different cost-performance tradeoff is important. We have developed a Design System which schedules the available time, divided into quanta, for testing alternative network configurations. Its goal is to select/generate and test alternative network configurations in each quantum, and find the best network when time is expended. Since time is limited, a dynamic schedule that determines the network configuration to be tested in each quantum is developed. The schedule is based on relative comparison of predicted training times of alternative network configurations using comparator network paradigm. The comparator network has been trained to compare training times for a large variety of traces of TSSE-versus-time collected during back-propagation learning of various applications.

A Hybrid Spectral Clustering and Deep Neural Network Ensemble Algorithm for Intrusion Detection in Sensor Networks

PubMed Central

Ma, Tao; Wang, Fen; Cheng, Jianjun; Yu, Yang; Chen, Xiaoyun

2016-01-01

The development of intrusion detection systems (IDS) that are adapted to allow routers and network defence systems to detect malicious network traffic disguised as network protocols or normal access is a critical challenge. This paper proposes a novel approach called SCDNN, which combines spectral clustering (SC) and deep neural network (DNN) algorithms. First, the dataset is divided into k subsets based on sample similarity using cluster centres, as in SC. Next, the distance between data points in a testing set and the training set is measured based on similarity features and is fed into the deep neural network algorithm for intrusion detection. Six KDD-Cup99 and NSL-KDD datasets and a sensor network dataset were employed to test the performance of the model. These experimental results indicate that the SCDNN classifier not only performs better than backpropagation neural network (BPNN), support vector machine (SVM), random forest (RF) and Bayes tree models in detection accuracy and the types of abnormal attacks found. It also provides an effective tool of study and analysis of intrusion detection in large networks. PMID:27754380

A Hybrid Spectral Clustering and Deep Neural Network Ensemble Algorithm for Intrusion Detection in Sensor Networks.

PubMed

Ma, Tao; Wang, Fen; Cheng, Jianjun; Yu, Yang; Chen, Xiaoyun

2016-10-13

The development of intrusion detection systems (IDS) that are adapted to allow routers and network defence systems to detect malicious network traffic disguised as network protocols or normal access is a critical challenge. This paper proposes a novel approach called SCDNN, which combines spectral clustering (SC) and deep neural network (DNN) algorithms. First, the dataset is divided into k subsets based on sample similarity using cluster centres, as in SC. Next, the distance between data points in a testing set and the training set is measured based on similarity features and is fed into the deep neural network algorithm for intrusion detection. Six KDD-Cup99 and NSL-KDD datasets and a sensor network dataset were employed to test the performance of the model. These experimental results indicate that the SCDNN classifier not only performs better than backpropagation neural network (BPNN), support vector machine (SVM), random forest (RF) and Bayes tree models in detection accuracy and the types of abnormal attacks found. It also provides an effective tool of study and analysis of intrusion detection in large networks.

Successful Reconstruction of a Physiological Circuit with Known Connectivity from Spiking Activity Alone

PubMed Central

Gerhard, Felipe; Kispersky, Tilman; Gutierrez, Gabrielle J.; Marder, Eve; Kramer, Mark; Eden, Uri

2013-01-01

Identifying the structure and dynamics of synaptic interactions between neurons is the first step to understanding neural network dynamics. The presence of synaptic connections is traditionally inferred through the use of targeted stimulation and paired recordings or by post-hoc histology. More recently, causal network inference algorithms have been proposed to deduce connectivity directly from electrophysiological signals, such as extracellularly recorded spiking activity. Usually, these algorithms have not been validated on a neurophysiological data set for which the actual circuitry is known. Recent work has shown that traditional network inference algorithms based on linear models typically fail to identify the correct coupling of a small central pattern generating circuit in the stomatogastric ganglion of the crab Cancer borealis. In this work, we show that point process models of observed spike trains can guide inference of relative connectivity estimates that match the known physiological connectivity of the central pattern generator up to a choice of threshold. We elucidate the necessary steps to derive faithful connectivity estimates from a model that incorporates the spike train nature of the data. We then apply the model to measure changes in the effective connectivity pattern in response to two pharmacological interventions, which affect both intrinsic neural dynamics and synaptic transmission. Our results provide the first successful application of a network inference algorithm to a circuit for which the actual physiological synapses between neurons are known. The point process methodology presented here generalizes well to larger networks and can describe the statistics of neural populations. In general we show that advanced statistical models allow for the characterization of effective network structure, deciphering underlying network dynamics and estimating information-processing capabilities. PMID:23874181

A Q-backpropagated time delay neural network for diagnosing severity of gait disturbances in Parkinson's disease.

PubMed

Nancy Jane, Y; Khanna Nehemiah, H; Arputharaj, Kannan

2016-04-01

Parkinson's disease (PD) is a movement disorder that affects the patient's nervous system and health-care applications mostly uses wearable sensors to collect these data. Since these sensors generate time stamped data, analyzing gait disturbances in PD becomes challenging task. The objective of this paper is to develop an effective clinical decision-making system (CDMS) that aids the physician in diagnosing the severity of gait disturbances in PD affected patients. This paper presents a Q-backpropagated time delay neural network (Q-BTDNN) classifier that builds a temporal classification model, which performs the task of classification and prediction in CDMS. The proposed Q-learning induced backpropagation (Q-BP) training algorithm trains the Q-BTDNN by generating a reinforced error signal. The network's weights are adjusted through backpropagating the generated error signal. For experimentation, the proposed work uses a PD gait database, which contains gait measures collected through wearable sensors from three different PD research studies. The experimental result proves the efficiency of Q-BP in terms of its improved classification accuracy of 91.49%, 92.19% and 90.91% with three datasets accordingly compared to other neural network training algorithms. Copyright © 2016 Elsevier Inc. All rights reserved.

High-Lift Optimization Design Using Neural Networks on a Multi-Element Airfoil

NASA Technical Reports Server (NTRS)

Greenman, Roxana M.; Roth, Karlin R.; Smith, Charles A. (Technical Monitor)

1998-01-01

The high-lift performance of a multi-element airfoil was optimized by using neural-net predictions that were trained using a computational data set. The numerical data was generated using a two-dimensional, incompressible, Navier-Stokes algorithm with the Spalart-Allmaras turbulence model. Because it is difficult to predict maximum lift for high-lift systems, an empirically-based maximum lift criteria was used in this study to determine both the maximum lift and the angle at which it occurs. Multiple input, single output networks were trained using the NASA Ames variation of the Levenberg-Marquardt algorithm for each of the aerodynamic coefficients (lift, drag, and moment). The artificial neural networks were integrated with a gradient-based optimizer. Using independent numerical simulations and experimental data for this high-lift configuration, it was shown that this design process successfully optimized flap deflection, gap, overlap, and angle of attack to maximize lift. Once the neural networks were trained and integrated with the optimizer, minimal additional computer resources were required to perform optimization runs with different initial conditions and parameters. Applying the neural networks within the high-lift rigging optimization process reduced the amount of computational time and resources by 83% compared with traditional gradient-based optimization procedures for multiple optimization runs.

Retrieval of ice thickness from polarimetric SAR data

NASA Technical Reports Server (NTRS)

Kwok, R.; Yueh, S. H.; Nghiem, S. V.; Huynh, D. D.

1993-01-01

We describe a potential procedure for retrieving ice thickness from multi-frequency polarimetric SAR data for thin ice. This procedure includes first masking out the thicker ice types with a simple classifier and then deriving the thickness of the remaining pixels using a model-inversion technique. The technique used to derive ice thickness from polarimetric observations is provided by a numerical estimator or neural network. A three-layer perceptron implemented with the backpropagation algorithm is used in this investigation with several improved aspects for a faster convergence rate and a better accuracy of the neural network. These improvements include weight initialization, normalization of the output range, the selection of offset constant, and a heuristic learning algorithm. The performance of the neural network is demonstrated by using training data generated by a theoretical scattering model for sea ice matched to the database of interest. The training data are comprised of the polarimetric backscattering coefficients of thin ice and the corresponding input ice parameters to the scattering model. The retrieved ice thickness from the theoretical backscattering coefficients is compare with the input ice thickness to the scattering model to illustrate the accuracy of the inversion method. Results indicate that the network convergence rate and accuracy are higher when multi-frequency training sets are presented. In addition, the dominant backscattering coefficients in retrieving ice thickness are found by comparing the behavior of the network trained backscattering data at various incidence angels. After the neural network is trained with the theoretical backscattering data at various incidence anges, the interconnection weights between nodes are saved and applied to the experimental data to be investigated. In this paper, we illustrate the effectiveness of this technique using polarimetric SAR data collected by the JPL DC-8 radar over a sea ice scene.

The Livermore Brain: Massive Deep Learning Networks Enabled by High Performance Computing

DOE Office of Scientific and Technical Information (OSTI.GOV)

Chen, Barry Y.

The proliferation of inexpensive sensor technologies like the ubiquitous digital image sensors has resulted in the collection and sharing of vast amounts of unsorted and unexploited raw data. Companies and governments who are able to collect and make sense of large datasets to help them make better decisions more rapidly will have a competitive advantage in the information era. Machine Learning technologies play a critical role for automating the data understanding process; however, to be maximally effective, useful intermediate representations of the data are required. These representations or “features” are transformations of the raw data into a form where patternsmore » are more easily recognized. Recent breakthroughs in Deep Learning have made it possible to learn these features from large amounts of labeled data. The focus of this project is to develop and extend Deep Learning algorithms for learning features from vast amounts of unlabeled data and to develop the HPC neural network training platform to support the training of massive network models. This LDRD project succeeded in developing new unsupervised feature learning algorithms for images and video and created a scalable neural network training toolkit for HPC. Additionally, this LDRD helped create the world’s largest freely-available image and video dataset supporting open multimedia research and used this dataset for training our deep neural networks. This research helped LLNL capture several work-for-others (WFO) projects, attract new talent, and establish collaborations with leading academic and commercial partners. Finally, this project demonstrated the successful training of the largest unsupervised image neural network using HPC resources and helped establish LLNL leadership at the intersection of Machine Learning and HPC research.« less

Automatic Classification of volcano-seismic events based on Deep Neural Networks.

NASA Astrophysics Data System (ADS)

Titos Luzón, M.; Bueno Rodriguez, A.; Garcia Martinez, L.; Benitez, C.; Ibáñez, J. M.

2017-12-01

Seismic monitoring of active volcanoes is a popular remote sensing technique to detect seismic activity, often associated to energy exchanges between the volcano and the environment. As a result, seismographs register a wide range of volcano-seismic signals that reflect the nature and underlying physics of volcanic processes. Machine learning and signal processing techniques provide an appropriate framework to analyze such data. In this research, we propose a new classification framework for seismic events based on deep neural networks. Deep neural networks are composed by multiple processing layers, and can discover intrinsic patterns from the data itself. Internal parameters can be initialized using a greedy unsupervised pre-training stage, leading to an efficient training of fully connected architectures. We aim to determine the robustness of these architectures as classifiers of seven different types of seismic events recorded at "Volcán de Fuego" (Colima, Mexico). Two deep neural networks with different pre-training strategies are studied: stacked denoising autoencoder and deep belief networks. Results are compared to existing machine learning algorithms (SVM, Random Forest, Multilayer Perceptron). We used 5 LPC coefficients over three non-overlapping segments as training features in order to characterize temporal evolution, avoid redundancy and encode the signal, regardless of its duration. Experimental results show that deep architectures can classify seismic events with higher accuracy than classical algorithms, attaining up to 92% recognition accuracy. Pre-training initialization helps these models to detect events that occur simultaneously in time (such explosions and rockfalls), increase robustness against noisy inputs, and provide better generalization. These results demonstrate deep neural networks are robust classifiers, and can be deployed in real-environments to monitor the seismicity of restless volcanoes.

Mastering the game of Go without human knowledge.

PubMed

Silver, David; Schrittwieser, Julian; Simonyan, Karen; Antonoglou, Ioannis; Huang, Aja; Guez, Arthur; Hubert, Thomas; Baker, Lucas; Lai, Matthew; Bolton, Adrian; Chen, Yutian; Lillicrap, Timothy; Hui, Fan; Sifre, Laurent; van den Driessche, George; Graepel, Thore; Hassabis, Demis

2017-10-18

A long-standing goal of artificial intelligence is an algorithm that learns, tabula rasa, superhuman proficiency in challenging domains. Recently, AlphaGo became the first program to defeat a world champion in the game of Go. The tree search in AlphaGo evaluated positions and selected moves using deep neural networks. These neural networks were trained by supervised learning from human expert moves, and by reinforcement learning from self-play. Here we introduce an algorithm based solely on reinforcement learning, without human data, guidance or domain knowledge beyond game rules. AlphaGo becomes its own teacher: a neural network is trained to predict AlphaGo's own move selections and also the winner of AlphaGo's games. This neural network improves the strength of the tree search, resulting in higher quality move selection and stronger self-play in the next iteration. Starting tabula rasa, our new program AlphaGo Zero achieved superhuman performance, winning 100-0 against the previously published, champion-defeating AlphaGo.

Mastering the game of Go without human knowledge

NASA Astrophysics Data System (ADS)

Silver, David; Schrittwieser, Julian; Simonyan, Karen; Antonoglou, Ioannis; Huang, Aja; Guez, Arthur; Hubert, Thomas; Baker, Lucas; Lai, Matthew; Bolton, Adrian; Chen, Yutian; Lillicrap, Timothy; Hui, Fan; Sifre, Laurent; van den Driessche, George; Graepel, Thore; Hassabis, Demis

2017-10-01

A long-standing goal of artificial intelligence is an algorithm that learns, tabula rasa, superhuman proficiency in challenging domains. Recently, AlphaGo became the first program to defeat a world champion in the game of Go. The tree search in AlphaGo evaluated positions and selected moves using deep neural networks. These neural networks were trained by supervised learning from human expert moves, and by reinforcement learning from self-play. Here we introduce an algorithm based solely on reinforcement learning, without human data, guidance or domain knowledge beyond game rules. AlphaGo becomes its own teacher: a neural network is trained to predict AlphaGo’s own move selections and also the winner of AlphaGo’s games. This neural network improves the strength of the tree search, resulting in higher quality move selection and stronger self-play in the next iteration. Starting tabula rasa, our new program AlphaGo Zero achieved superhuman performance, winning 100-0 against the previously published, champion-defeating AlphaGo.

Deep learning for computational chemistry.

PubMed

Goh, Garrett B; Hodas, Nathan O; Vishnu, Abhinav

2017-06-15

The rise and fall of artificial neural networks is well documented in the scientific literature of both computer science and computational chemistry. Yet almost two decades later, we are now seeing a resurgence of interest in deep learning, a machine learning algorithm based on multilayer neural networks. Within the last few years, we have seen the transformative impact of deep learning in many domains, particularly in speech recognition and computer vision, to the extent that the majority of expert practitioners in those field are now regularly eschewing prior established models in favor of deep learning models. In this review, we provide an introductory overview into the theory of deep neural networks and their unique properties that distinguish them from traditional machine learning algorithms used in cheminformatics. By providing an overview of the variety of emerging applications of deep neural networks, we highlight its ubiquity and broad applicability to a wide range of challenges in the field, including quantitative structure activity relationship, virtual screening, protein structure prediction, quantum chemistry, materials design, and property prediction. In reviewing the performance of deep neural networks, we observed a consistent outperformance against non-neural networks state-of-the-art models across disparate research topics, and deep neural network-based models often exceeded the "glass ceiling" expectations of their respective tasks. Coupled with the maturity of GPU-accelerated computing for training deep neural networks and the exponential growth of chemical data on which to train these networks on, we anticipate that deep learning algorithms will be a valuable tool for computational chemistry. © 2017 Wiley Periodicals, Inc. © 2017 Wiley Periodicals, Inc.

Deep learning for computational chemistry

DOE Office of Scientific and Technical Information (OSTI.GOV)

Goh, Garrett B.; Hodas, Nathan O.; Vishnu, Abhinav

The rise and fall of artificial neural networks is well documented in the scientific literature of both the fields of computer science and computational chemistry. Yet almost two decades later, we are now seeing a resurgence of interest in deep learning, a machine learning algorithm based on “deep” neural networks. Within the last few years, we have seen the transformative impact of deep learning the computer science domain, notably in speech recognition and computer vision, to the extent that the majority of practitioners in those field are now regularly eschewing prior established models in favor of deep learning models. Inmore » this review, we provide an introductory overview into the theory of deep neural networks and their unique properties as compared to traditional machine learning algorithms used in cheminformatics. By providing an overview of the variety of emerging applications of deep neural networks, we highlight its ubiquity and broad applicability to a wide range of challenges in the field, including QSAR, virtual screening, protein structure modeling, QM calculations, materials synthesis and property prediction. In reviewing the performance of deep neural networks, we observed a consistent outperformance against non neural networks state-of-the-art models across disparate research topics, and deep neural network based models often exceeded the “glass ceiling” expectations of their respective tasks. Coupled with the maturity of GPU-accelerated computing for training deep neural networks and the exponential growth of chemical data on which to train these networks on, we anticipate that deep learning algorithms will be a useful tool and may grow into a pivotal role for various challenges in the computational chemistry field.« less

Version 2 of the IASI NH3 neural network retrieval algorithm: near-real-time and reanalysed datasets

NASA Astrophysics Data System (ADS)

Van Damme, Martin; Whitburn, Simon; Clarisse, Lieven; Clerbaux, Cathy; Hurtmans, Daniel; Coheur, Pierre-François

2017-12-01

Recently, Whitburn et al.(2016) presented a neural-network-based algorithm for retrieving atmospheric ammonia (NH3) columns from Infrared Atmospheric Sounding Interferometer (IASI) satellite observations. In the past year, several improvements have been introduced, and the resulting new baseline version, Artificial Neural Network for IASI (ANNI)-NH3-v2.1, is documented here. One of the main changes to the algorithm is that separate neural networks were trained for land and sea observations, resulting in a better training performance for both groups. By reducing and transforming the input parameter space, performance is now also better for observations associated with favourable sounding conditions (i.e. enhanced thermal contrasts). Other changes relate to the introduction of a bias correction over land and sea and the treatment of the satellite zenith angle. In addition to these algorithmic changes, new recommendations for post-filtering the data and for averaging data in time or space are formulated. We also introduce a second dataset (ANNI-NH3-v2.1R-I) which relies on ERA-Interim ECMWF meteorological input data, along with surface temperature retrieved from a dedicated network, rather than the operationally provided Eumetsat IASI Level 2 (L2) data used for the standard near-real-time version. The need for such a dataset emerged after a series of sharp discontinuities were identified in the NH3 time series, which could be traced back to incremental changes in the IASI L2 algorithms for temperature and clouds. The reanalysed dataset is coherent in time and can therefore be used to study trends. Furthermore, both datasets agree reasonably well in the mean on recent data, after the date when the IASI meteorological L2 version 6 became operational (30 September 2014).

Automatic Classification Using Supervised Learning in a Medical Document Filtering Application.

ERIC Educational Resources Information Center

Mostafa, J.; Lam, W.

2000-01-01

Presents a multilevel model of the information filtering process that permits document classification. Evaluates a document classification approach based on a supervised learning algorithm, measures the accuracy of the algorithm in a neural network that was trained to classify medical documents on cell biology, and discusses filtering…

Quantum autoencoders for efficient compression of quantum data

NASA Astrophysics Data System (ADS)

Romero, Jonathan; Olson, Jonathan P.; Aspuru-Guzik, Alan

2017-12-01

Classical autoencoders are neural networks that can learn efficient low-dimensional representations of data in higher-dimensional space. The task of an autoencoder is, given an input x, to map x to a lower dimensional point y such that x can likely be recovered from y. The structure of the underlying autoencoder network can be chosen to represent the data on a smaller dimension, effectively compressing the input. Inspired by this idea, we introduce the model of a quantum autoencoder to perform similar tasks on quantum data. The quantum autoencoder is trained to compress a particular data set of quantum states, where a classical compression algorithm cannot be employed. The parameters of the quantum autoencoder are trained using classical optimization algorithms. We show an example of a simple programmable circuit that can be trained as an efficient autoencoder. We apply our model in the context of quantum simulation to compress ground states of the Hubbard model and molecular Hamiltonians.

Ads' click-through rates predicting based on gated recurrent unit neural networks

NASA Astrophysics Data System (ADS)

Chen, Qiaohong; Guo, Zixuan; Dong, Wen; Jin, Lingzi

2018-05-01

In order to improve the effect of online advertising and to increase the revenue of advertising, the gated recurrent unit neural networks(GRU) model is used as the ads' click through rates(CTR) predicting. Combined with the characteristics of gated unit structure and the unique of time sequence in data, using BPTT algorithm to train the model. Furthermore, by optimizing the step length algorithm of the gated unit recurrent neural networks, making the model reach optimal point better and faster in less iterative rounds. The experiment results show that the model based on the gated recurrent unit neural networks and its optimization of step length algorithm has the better effect on the ads' CTR predicting, which helps advertisers, media and audience achieve a win-win and mutually beneficial situation in Three-Side Game.

Atmospheric correction over coastal waters using multilayer neural networks

NASA Astrophysics Data System (ADS)

Fan, Y.; Li, W.; Charles, G.; Jamet, C.; Zibordi, G.; Schroeder, T.; Stamnes, K. H.

2017-12-01

Standard atmospheric correction (AC) algorithms work well in open ocean areas where the water inherent optical properties (IOPs) are correlated with pigmented particles. However, the IOPs of turbid coastal waters may independently vary with pigmented particles, suspended inorganic particles, and colored dissolved organic matter (CDOM). In turbid coastal waters standard AC algorithms often exhibit large inaccuracies that may lead to negative water-leaving radiances (Lw) or remote sensing reflectance (Rrs). We introduce a new atmospheric correction algorithm for coastal waters based on a multilayer neural network (MLNN) machine learning method. We use a coupled atmosphere-ocean radiative transfer model to simulate the Rayleigh-corrected radiance (Lrc) at the top of the atmosphere (TOA) and the Rrs just above the surface simultaneously, and train a MLNN to derive the aerosol optical depth (AOD) and Rrs directly from the TOA Lrc. The SeaDAS NIR algorithm, the SeaDAS NIR/SWIR algorithm, and the MODIS version of the Case 2 regional water - CoastColour (C2RCC) algorithm are included in the comparison with AERONET-OC measurements. The results show that the MLNN algorithm significantly improves retrieval of normalized Lw in blue bands (412 nm and 443 nm) and yields minor improvements in green and red bands. These results indicate that the MLNN algorithm is suitable for application in turbid coastal waters. Application of the MLNN algorithm to MODIS Aqua images in several coastal areas also shows that it is robust and resilient to contamination due to sunglint or adjacency effects of land and cloud edges. The MLNN algorithm is very fast once the neural network has been properly trained and is therefore suitable for operational use. A significant advantage of the MLNN algorithm is that it does not need SWIR bands, which implies significant cost reduction for dedicated OC missions. A recent effort has been made to extend the MLNN AC algorithm to extreme atmospheric conditions (i.e. heavy polluted continental aerosols) over coastal areas by including additional aerosol and ocean models to generate the training dataset. Preliminary tests show very good results. Results of applying the extended MLNN algorithm to VIIRS images over the Yellow Sea and East China Sea areas with extreme atmospheric and marine conditions will be provided.

Simulating and Synthesizing Substructures Using Neural Network and Genetic Algorithms

NASA Technical Reports Server (NTRS)

Liu, Youhua; Kapania, Rakesh K.; VanLandingham, Hugh F.

1997-01-01

The feasibility of simulating and synthesizing substructures by computational neural network models is illustrated by investigating a statically indeterminate beam, using both a 1-D and a 2-D plane stress modelling. The beam can be decomposed into two cantilevers with free-end loads. By training neural networks to simulate the cantilever responses to different loads, the original beam problem can be solved as a match-up between two subsystems under compatible interface conditions. The genetic algorithms are successfully used to solve the match-up problem. Simulated results are found in good agreement with the analytical or FEM solutions.

The algorithm study for using the back propagation neural network in CT image segmentation

NASA Astrophysics Data System (ADS)

Zhang, Peng; Liu, Jie; Chen, Chen; Li, Ying Qi

2017-01-01

Back propagation neural network(BP neural network) is a type of multi-layer feed forward network which spread positively, while the error spread backwardly. Since BP network has advantages in learning and storing the mapping between a large number of input and output layers without complex mathematical equations to describe the mapping relationship, it is most widely used. BP can iteratively compute the weight coefficients and thresholds of the network based on the training and back propagation of samples, which can minimize the error sum of squares of the network. Since the boundary of the computed tomography (CT) heart images is usually discontinuous, and it exist large changes in the volume and boundary of heart images, The conventional segmentation such as region growing and watershed algorithm can't achieve satisfactory results. Meanwhile, there are large differences between the diastolic and systolic images. The conventional methods can't accurately classify the two cases. In this paper, we introduced BP to handle the segmentation of heart images. We segmented a large amount of CT images artificially to obtain the samples, and the BP network was trained based on these samples. To acquire the appropriate BP network for the segmentation of heart images, we normalized the heart images, and extract the gray-level information of the heart. Then the boundary of the images was input into the network to compare the differences between the theoretical output and the actual output, and we reinput the errors into the BP network to modify the weight coefficients of layers. Through a large amount of training, the BP network tend to be stable, and the weight coefficients of layers can be determined, which means the relationship between the CT images and the boundary of heart.

Fast temporal neural learning using teacher forcing

NASA Technical Reports Server (NTRS)

Toomarian, Nikzad (Inventor); Bahren, Jacob (Inventor)

1992-01-01

A neural network is trained to output a time dependent target vector defined over a predetermined time interval in response to a time dependent input vector defined over the same time interval by applying corresponding elements of the error vector, or difference between the target vector and the actual neuron output vector, to the inputs of corresponding output neurons of the network as corrective feedback. This feedback decreases the error and quickens the learning process, so that a much smaller number of training cycles are required to complete the learning process. A conventional gradient descent algorithm is employed to update the neural network parameters at the end of the predetermined time interval. The foregoing process is repeated in repetitive cycles until the actual output vector corresponds to the target vector. In the preferred embodiment, as the overall error of the neural network output decreasing during successive training cycles, the portion of the error fed back to the output neurons is decreased accordingly, allowing the network to learn with greater freedom from teacher forcing as the network parameters converge to their optimum values. The invention may also be used to train a neural network with stationary training and target vectors.

Fast temporal neural learning using teacher forcing

NASA Technical Reports Server (NTRS)

Toomarian, Nikzad (Inventor); Bahren, Jacob (Inventor)

1995-01-01

A neural network is trained to output a time dependent target vector defined over a predetermined time interval in response to a time dependent input vector defined over the same time interval by applying corresponding elements of the error vector, or difference between the target vector and the actual neuron output vector, to the inputs of corresponding output neurons of the network as corrective feedback. This feedback decreases the error and quickens the learning process, so that a much smaller number of training cycles are required to complete the learning process. A conventional gradient descent algorithm is employed to update the neural network parameters at the end of the predetermined time interval. The foregoing process is repeated in repetitive cycles until the actual output vector corresponds to the target vector. In the preferred embodiment, as the overall error of the neural network output decreasing during successive training cycles, the portion of the error fed back to the output neurons is decreased accordingly, allowing the network to learn with greater freedom from teacher forcing as the network parameters converge to their optimum values. The invention may also be used to train a neural network with stationary training and target vectors.

Fuzzy logic and neural network technologies

NASA Technical Reports Server (NTRS)

Villarreal, James A.; Lea, Robert N.; Savely, Robert T.

1992-01-01

Applications of fuzzy logic technologies in NASA projects are reviewed to examine their advantages in the development of neural networks for aerospace and commercial expert systems and control. Examples of fuzzy-logic applications include a 6-DOF spacecraft controller, collision-avoidance systems, and reinforcement-learning techniques. The commercial applications examined include a fuzzy autofocusing system, an air conditioning system, and an automobile transmission application. The practical use of fuzzy logic is set in the theoretical context of artificial neural systems (ANSs) to give the background for an overview of ANS research programs at NASA. The research and application programs include the Network Execution and Training Simulator and faster training algorithms such as the Difference Optimized Training Scheme. The networks are well suited for pattern-recognition applications such as predicting sunspots, controlling posture maintenance, and conducting adaptive diagnoses.

Prediction of Flow Stress in Cadmium Using Constitutive Equation and Artificial Neural Network Approach

NASA Astrophysics Data System (ADS)

Sarkar, A.; Chakravartty, J. K.

2013-10-01

A model is developed to predict the constitutive flow behavior of cadmium during compression test using artificial neural network (ANN). The inputs of the neural network are strain, strain rate, and temperature, whereas flow stress is the output. Experimental data obtained from compression tests in the temperature range -30 to 70 °C, strain range 0.1 to 0.6, and strain rate range 10-3 to 1 s-1 are employed to develop the model. A three-layer feed-forward ANN is trained with Levenberg-Marquardt training algorithm. It has been shown that the developed ANN model can efficiently and accurately predict the deformation behavior of cadmium. This trained network could predict the flow stress better than a constitutive equation of the type.

Looking for underlying features in automatic and reviewed seismic bulletins through a neural network

NASA Astrophysics Data System (ADS)

Carluccio, R.; Console, R.; Chiappini, M.; Chiappini, S.

2009-12-01

SEL1 bulletins are, among all IDC products, a fundamental tool for NDCs in their task of national assessment of compliance with the CTBT. This is because SEL1s are expected to be disseminated within 2 hours from the occurrence of any detected waveform event, and the National Authorities are supposed to take a political decision in nearly real time, especially in the case when the event could triggers the request for an on site inspection. In this context not only the rapidity, but also the reliability of the SEL1 is a fundamental requirement. Our last years experience gained in the comparison between SEL1 and Italian Seismic Bulletin events has shown that SEL1s usually contain a big fraction of bogus events (sometimes close to 50%). This is due to many factors, all related to the availability of processing data and to the fast automatic algorithms involved. On the other hand, REBs are much more reliable as proved by our experience. Therefore, in spite of their relevant time delay by which they are distributed, which prevents their real-time use, REBs can be still useful in a retrospective way as reference information for comparison with SEL1s. This study tries to set up a sort of logical filter on the SEL1s that, while maintaining the rapidity requirements, improves their reliability. Our idea is based on the assumption that the SEL1s are produced by systematic algorithm of phase association and therefore some patterns among the input and output data could exist and be recognized. Our approach was initially based on a set of rules suggested by human experts on their personal experience, and its application on large datasets on a global scale. Other approaches not involving human interaction (data mining techniques) do exist. This study refers specifically to a semi-automatic approach: fitting of multi-parametric relationships hidden in the data set, through the application of neural networks by an algorithm of supervised learning. Full SEL1 and REB bulletins from Jan 2005 to Oct 2008 have been inserted in a database, together with IMS stations availability information. Part of these data have been used to create two sets of independent data (learning and verifying) used to train a "feed-forward" supervised neural network. A network supervised training algorithm using "confirmation flag" values has been used. In order to optimize network training input a significant, not redundant subset of input parameters has been looked for with the help of a genetic algorithm search tool. A suitable 12 input subset has been found and a network architecture of 12-20-1 has thus been chosen and trained on a 15094 records data set. Different runs of training sequences have been conducted, all showing CCR (Correct Classification Rate) values of the order of 75% - 80%. The trained network behavior is shown in term of ROC curve and input-out success-error matrices. The results of the analysis on our testing and validating data groups appear promising.

Equilibrium point control of a monkey arm simulator by a fast learning tree structured artificial neural network.

PubMed

Dornay, M; Sanger, T D

1993-01-01

A planar 17 muscle model of the monkey's arm based on realistic biomechanical measurements was simulated on a Symbolics Lisp Machine. The simulator implements the equilibrium point hypothesis for the control of arm movements. Given initial and final desired positions, it generates a minimum-jerk desired trajectory of the hand and uses the backdriving algorithm to determine an appropriate sequence of motor commands to the muscles (Flash 1987; Mussa-Ivaldi et al. 1991; Dornay 1991b). These motor commands specify a temporal sequence of stable (attractive) equilibrium positions which lead to the desired hand movement. A strong disadvantage of the simulator is that it has no memory of previous computations. Determining the desired trajectory using the minimum-jerk model is instantaneous, but the laborious backdriving algorithm is slow, and can take up to one hour for some trajectories. The complexity of the required computations makes it a poor model for biological motor control. We propose a computationally simpler and more biologically plausible method for control which achieves the benefits of the backdriving algorithm. A fast learning, tree-structured network (Sanger 1991c) was trained to remember the knowledge obtained by the backdriving algorithm. The neural network learned the nonlinear mapping from a 2-dimensional cartesian planar hand position (x,y) to a 17-dimensional motor command space (u1, . . ., u17). Learning 20 training trajectories, each composed of 26 sample points [[x,y], [u1, . . ., u17] took only 20 min on a Sun-4 Sparc workstation. After the learning stage, new, untrained test trajectories as well as the original trajectories of the hand were given to the neural network as input. The network calculated the required motor commands for these movements. The resulting movements were close to the desired ones for both the training and test cases.

Beyond Retinal Layers: A Deep Voting Model for Automated Geographic Atrophy Segmentation in SD-OCT Images

PubMed Central

Ji, Zexuan; Chen, Qiang; Niu, Sijie; Leng, Theodore; Rubin, Daniel L.

2018-01-01

Purpose To automatically and accurately segment geographic atrophy (GA) in spectral-domain optical coherence tomography (SD-OCT) images by constructing a voting system with deep neural networks without the use of retinal layer segmentation. Methods An automatic GA segmentation method for SD-OCT images based on the deep network was constructed. The structure of the deep network was composed of five layers, including one input layer, three hidden layers, and one output layer. During the training phase, the labeled A-scans with 1024 features were directly fed into the network as the input layer to obtain the deep representations. Then a soft-max classifier was trained to determine the label of each individual pixel. Finally, a voting decision strategy was used to refine the segmentation results among 10 trained models. Results Two image data sets with GA were used to evaluate the model. For the first dataset, our algorithm obtained a mean overlap ratio (OR) 86.94% ± 8.75%, absolute area difference (AAD) 11.49% ± 11.50%, and correlation coefficients (CC) 0.9857; for the second dataset, the mean OR, AAD, and CC of the proposed method were 81.66% ± 10.93%, 8.30% ± 9.09%, and 0.9952, respectively. The proposed algorithm was capable of improving over 5% and 10% segmentation accuracy, respectively, when compared with several state-of-the-art algorithms on two data sets. Conclusions Without retinal layer segmentation, the proposed algorithm could produce higher segmentation accuracy and was more stable when compared with state-of-the-art methods that relied on retinal layer segmentation results. Our model may provide reliable GA segmentations from SD-OCT images and be useful in the clinical diagnosis of advanced nonexudative AMD. Translational Relevance Based on the deep neural networks, this study presents an accurate GA segmentation method for SD-OCT images without using any retinal layer segmentation results, and may contribute to improved understanding of advanced nonexudative AMD. PMID:29302382

Beyond Retinal Layers: A Deep Voting Model for Automated Geographic Atrophy Segmentation in SD-OCT Images.

PubMed

Ji, Zexuan; Chen, Qiang; Niu, Sijie; Leng, Theodore; Rubin, Daniel L

2018-01-01

To automatically and accurately segment geographic atrophy (GA) in spectral-domain optical coherence tomography (SD-OCT) images by constructing a voting system with deep neural networks without the use of retinal layer segmentation. An automatic GA segmentation method for SD-OCT images based on the deep network was constructed. The structure of the deep network was composed of five layers, including one input layer, three hidden layers, and one output layer. During the training phase, the labeled A-scans with 1024 features were directly fed into the network as the input layer to obtain the deep representations. Then a soft-max classifier was trained to determine the label of each individual pixel. Finally, a voting decision strategy was used to refine the segmentation results among 10 trained models. Two image data sets with GA were used to evaluate the model. For the first dataset, our algorithm obtained a mean overlap ratio (OR) 86.94% ± 8.75%, absolute area difference (AAD) 11.49% ± 11.50%, and correlation coefficients (CC) 0.9857; for the second dataset, the mean OR, AAD, and CC of the proposed method were 81.66% ± 10.93%, 8.30% ± 9.09%, and 0.9952, respectively. The proposed algorithm was capable of improving over 5% and 10% segmentation accuracy, respectively, when compared with several state-of-the-art algorithms on two data sets. Without retinal layer segmentation, the proposed algorithm could produce higher segmentation accuracy and was more stable when compared with state-of-the-art methods that relied on retinal layer segmentation results. Our model may provide reliable GA segmentations from SD-OCT images and be useful in the clinical diagnosis of advanced nonexudative AMD. Based on the deep neural networks, this study presents an accurate GA segmentation method for SD-OCT images without using any retinal layer segmentation results, and may contribute to improved understanding of advanced nonexudative AMD.

Artificial neural network modeling and optimization of ultrahigh pressure extraction of green tea polyphenols.

PubMed

Xi, Jun; Xue, Yujing; Xu, Yinxiang; Shen, Yuhong

2013-11-01

In this study, the ultrahigh pressure extraction of green tea polyphenols was modeled and optimized by a three-layer artificial neural network. A feed-forward neural network trained with an error back-propagation algorithm was used to evaluate the effects of pressure, liquid/solid ratio and ethanol concentration on the total phenolic content of green tea extracts. The neural network coupled with genetic algorithms was also used to optimize the conditions needed to obtain the highest yield of tea polyphenols. The obtained optimal architecture of artificial neural network model involved a feed-forward neural network with three input neurons, one hidden layer with eight neurons and one output layer including single neuron. The trained network gave the minimum value in the MSE of 0.03 and the maximum value in the R(2) of 0.9571, which implied a good agreement between the predicted value and the actual value, and confirmed a good generalization of the network. Based on the combination of neural network and genetic algorithms, the optimum extraction conditions for the highest yield of green tea polyphenols were determined as follows: 498.8 MPa for pressure, 20.8 mL/g for liquid/solid ratio and 53.6% for ethanol concentration. The total phenolic content of the actual measurement under the optimum predicated extraction conditions was 582.4 ± 0.63 mg/g DW, which was well matched with the predicted value (597.2mg/g DW). This suggests that the artificial neural network model described in this work is an efficient quantitative tool to predict the extraction efficiency of green tea polyphenols. Crown Copyright © 2013. Published by Elsevier Ltd. All rights reserved.

Evolving spiking neural networks: a novel growth algorithm exhibits unintelligent design

NASA Astrophysics Data System (ADS)

Schaffer, J. David

2015-06-01

Spiking neural networks (SNNs) have drawn considerable excitement because of their computational properties, believed to be superior to conventional von Neumann machines, and sharing properties with living brains. Yet progress building these systems has been limited because we lack a design methodology. We present a gene-driven network growth algorithm that enables a genetic algorithm (evolutionary computation) to generate and test SNNs. The genome for this algorithm grows O(n) where n is the number of neurons; n is also evolved. The genome not only specifies the network topology, but all its parameters as well. Experiments show the algorithm producing SNNs that effectively produce a robust spike bursting behavior given tonic inputs, an application suitable for central pattern generators. Even though evolution did not include perturbations of the input spike trains, the evolved networks showed remarkable robustness to such perturbations. In addition, the output spike patterns retain evidence of the specific perturbation of the inputs, a feature that could be exploited by network additions that could use this information for refined decision making if required. On a second task, a sequence detector, a discriminating design was found that might be considered an example of "unintelligent design"; extra non-functional neurons were included that, while inefficient, did not hamper its proper functioning.

Two neural network algorithms for designing optimal terminal controllers with open final time

NASA Technical Reports Server (NTRS)

Plumer, Edward S.

1992-01-01

Multilayer neural networks, trained by the backpropagation through time algorithm (BPTT), have been used successfully as state-feedback controllers for nonlinear terminal control problems. Current BPTT techniques, however, are not able to deal systematically with open final-time situations such as minimum-time problems. Two approaches which extend BPTT to open final-time problems are presented. In the first, a neural network learns a mapping from initial-state to time-to-go. In the second, the optimal number of steps for each trial run is found using a line-search. Both methods are derived using Lagrange multiplier techniques. This theoretical framework is used to demonstrate that the derived algorithms are direct extensions of forward/backward sweep methods used in N-stage optimal control. The two algorithms are tested on a Zermelo problem and the resulting trajectories compare favorably to optimal control results.

A statistical framework for evaluating neural networks to predict recurrent events in breast cancer

NASA Astrophysics Data System (ADS)

Gorunescu, Florin; Gorunescu, Marina; El-Darzi, Elia; Gorunescu, Smaranda

2010-07-01

Breast cancer is the second leading cause of cancer deaths in women today. Sometimes, breast cancer can return after primary treatment. A medical diagnosis of recurrent cancer is often a more challenging task than the initial one. In this paper, we investigate the potential contribution of neural networks (NNs) to support health professionals in diagnosing such events. The NN algorithms are tested and applied to two different datasets. An extensive statistical analysis has been performed to verify our experiments. The results show that a simple network structure for both the multi-layer perceptron and radial basis function can produce equally good results, not all attributes are needed to train these algorithms and, finally, the classification performances of all algorithms are statistically robust. Moreover, we have shown that the best performing algorithm will strongly depend on the features of the datasets, and hence, there is not necessarily a single best classifier.

Comparison of Support Vector Machine, Neural Network, and CART Algorithms for the Land-Cover Classification Using Limited Training Data Points

EPA Science Inventory

Support vector machine (SVM) was applied for land-cover characterization using MODIS time-series data. Classification performance was examined with respect to training sample size, sample variability, and landscape homogeneity (purity). The results were compared to two convention...

Using Upper Layer Weights to Efficiently Construct and Train Feedforward Neural Networks Executing Backpropagation

DTIC Science & Technology

2011-03-01

algorithm is utilized by Belue, Steppe, & Bauer and Kocur , et al. (Belue, Steppe, & Bauer, April 1996) ( Kocur , et al., 1996). Bacauskiene and...Society. Cardiff, UK. Kocur , C., Roger, S., Myers, L., Burns, T., Hoffmeister, J., Bauer, K., et al. (1996). Using neural networks to select

Application of Artificial Neural Networks in the Design and Optimization of a Nanoparticulate Fingolimod Delivery System Based on Biodegradable Poly(3-Hydroxybutyrate-Co-3-Hydroxyvalerate).

PubMed

Shahsavari, Shadab; Rezaie Shirmard, Leila; Amini, Mohsen; Abedin Dokoosh, Farid

2017-01-01

Formulation of a nanoparticulate Fingolimod delivery system based on biodegradable poly(3-hydroxybutyrate-co-3-hydroxyvalerate) was optimized according to artificial neural networks (ANNs). Concentration of poly(3-hydroxybutyrate-co-3-hydroxyvalerate), PVA and amount of Fingolimod is considered as the input value, and the particle size, polydispersity index, loading capacity, and entrapment efficacy as output data in experimental design study. In vitro release study was carried out for best formulation according to statistical analysis. ANNs are employed to generate the best model to determine the relationships between various values. In order to specify the model with the best accuracy and proficiency for the in vitro release, a multilayer percepteron with different training algorithm has been examined. Three training model formulations including Levenberg-Marquardt (LM), gradient descent, and Bayesian regularization were employed for training the ANN models. It is demonstrated that the predictive ability of each training algorithm is in the order of LM > gradient descent > Bayesian regularization. Also, optimum formulation was achieved by LM training function with 15 hidden layers and 20 neurons. The transfer function of the hidden layer for this formulation and the output layer were tansig and purlin, respectively. Also, the optimization process was developed by minimizing the error among the predicted and observed values of training algorithm (about 0.0341). Copyright © 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Learning relevant features of data with multi-scale tensor networks

NASA Astrophysics Data System (ADS)

Miles Stoudenmire, E.

2018-07-01

Inspired by coarse-graining approaches used in physics, we show how similar algorithms can be adapted for data. The resulting algorithms are based on layered tree tensor networks and scale linearly with both the dimension of the input and the training set size. Computing most of the layers with an unsupervised algorithm, then optimizing just the top layer for supervised classification of the MNIST and fashion MNIST data sets gives very good results. We also discuss mixing a prior guess for supervised weights together with an unsupervised representation of the data, yielding a smaller number of features nevertheless able to give good performance.

Neural Networks and other Techniques for Fault Identification and Isolation of Aircraft Systems

NASA Technical Reports Server (NTRS)

Innocenti, M.; Napolitano, M.

2003-01-01

Fault identification, isolation, and accomodation have become critical issues in the overall performance of advanced aircraft systems. Neural Networks have shown to be a very attractive alternative to classic adaptation methods for identification and control of non-linear dynamic systems. The purpose of this paper is to show the improvements in neural network applications achievable through the use of learning algorithms more efficient than the classic Back-Propagation, and through the implementation of the neural schemes in parallel hardware. The results of the analysis of a scheme for Sensor Failure, Detection, Identification and Accommodation (SFDIA) using experimental flight data of a research aircraft model are presented. Conventional approaches to the problem are based on observers and Kalman Filters while more recent methods are based on neural approximators. The work described in this paper is based on the use of neural networks (NNs) as on-line learning non-linear approximators. The performances of two different neural architectures were compared. The first architecture is based on a Multi Layer Perceptron (MLP) NN trained with the Extended Back Propagation algorithm (EBPA). The second architecture is based on a Radial Basis Function (RBF) NN trained with the Extended-MRAN (EMRAN) algorithms. In addition, alternative methods for communications links fault detection and accomodation are presented, relative to multiple unmanned aircraft applications.

RBF neural network based PI pitch controller for a class of 5-MW wind turbines using particle swarm optimization algorithm.

PubMed

Poultangari, Iman; Shahnazi, Reza; Sheikhan, Mansour

2012-09-01

In order to control the pitch angle of blades in wind turbines, commonly the proportional and integral (PI) controller due to its simplicity and industrial usability is employed. The neural networks and evolutionary algorithms are tools that provide a suitable ground to determine the optimal PI gains. In this paper, a radial basis function (RBF) neural network based PI controller is proposed for collective pitch control (CPC) of a 5-MW wind turbine. In order to provide an optimal dataset to train the RBF neural network, particle swarm optimization (PSO) evolutionary algorithm is used. The proposed method does not need the complexities, nonlinearities and uncertainties of the system under control. The simulation results show that the proposed controller has satisfactory performance. Copyright © 2012 ISA. Published by Elsevier Ltd. All rights reserved.

Neural network modelling of planform geometry of headland-bay beaches

NASA Astrophysics Data System (ADS)

Iglesias, G.; López, I.; Castro, A.; Carballo, R.

2009-02-01

The shoreline of beaches in the lee of coastal salients or man-made structures, usually known as headland-bay beaches, has a distinctive curvature; wave fronts curve as a result of wave diffraction at the headland and in turn cause the shoreline to bend. The ensuing curved planform is of great interest both as a peculiar landform and in the context of engineering projects in which it is necessary to predict how a coastal structure will affect the sandy shoreline in its lee. A number of empirical models have been put forward, each based on a specific equation. A novel approach, based on the application of artificial neural networks, is presented in this work. Unlike the conventional method, no particular equation of the planform is embedded in the model. Instead, it is the model itself that learns about the problem from a series of examples of headland-bay beaches (the training set) and thereafter applies this self-acquired knowledge to other cases (the test set) for validation. Twenty-three headland-bay beaches from around the world were selected, of which sixteen and seven make up the training and test sets, respectively. As there is no well-developed theory for deciding upon the most convenient neural network architecture to deal with a particular data set, an experimental study was conducted in which ten different architectures with one and two hidden neuron layers and five training algorithms - 50 different options combining network architecture and training algorithm - were compared. Each of these options was implemented, trained and tested in order to find the best-performing approach for modelling the planform of headland-bay beaches. Finally, the selected neural network model was compared with a state-of-the-art planform model and was shown to outperform it.

Prediction of Aerodynamic Coefficient using Genetic Algorithm Optimized Neural Network for Sparse Data

NASA Technical Reports Server (NTRS)

Rajkumar, T.; Bardina, Jorge; Clancy, Daniel (Technical Monitor)

2002-01-01

Wind tunnels use scale models to characterize aerodynamic coefficients, Wind tunnel testing can be slow and costly due to high personnel overhead and intensive power utilization. Although manual curve fitting can be done, it is highly efficient to use a neural network to define the complex relationship between variables. Numerical simulation of complex vehicles on the wide range of conditions required for flight simulation requires static and dynamic data. Static data at low Mach numbers and angles of attack may be obtained with simpler Euler codes. Static data of stalled vehicles where zones of flow separation are usually present at higher angles of attack require Navier-Stokes simulations which are costly due to the large processing time required to attain convergence. Preliminary dynamic data may be obtained with simpler methods based on correlations and vortex methods; however, accurate prediction of the dynamic coefficients requires complex and costly numerical simulations. A reliable and fast method of predicting complex aerodynamic coefficients for flight simulation I'S presented using a neural network. The training data for the neural network are derived from numerical simulations and wind-tunnel experiments. The aerodynamic coefficients are modeled as functions of the flow characteristics and the control surfaces of the vehicle. The basic coefficients of lift, drag and pitching moment are expressed as functions of angles of attack and Mach number. The modeled and training aerodynamic coefficients show good agreement. This method shows excellent potential for rapid development of aerodynamic models for flight simulation. Genetic Algorithms (GA) are used to optimize a previously built Artificial Neural Network (ANN) that reliably predicts aerodynamic coefficients. Results indicate that the GA provided an efficient method of optimizing the ANN model to predict aerodynamic coefficients. The reliability of the ANN using the GA includes prediction of aerodynamic coefficients to an accuracy of 110% . In our problem, we would like to get an optimized neural network architecture and minimum data set. This has been accomplished within 500 training cycles of a neural network. After removing training pairs (outliers), the GA has produced much better results. The neural network constructed is a feed forward neural network with a back propagation learning mechanism. The main goal has been to free the network design process from constraints of human biases, and to discover better forms of neural network architectures. The automation of the network architecture search by genetic algorithms seems to have been the best way to achieve this goal.

Neural network pattern recognition of thermal-signature spectra for chemical defense

NASA Astrophysics Data System (ADS)

Carrieri, Arthur H.; Lim, Pascal I.

1995-05-01

We treat infrared patterns of absorption or emission by nerve and blister agent compounds (and simulants of this chemical group) as features for the training of neural networks to detect the compounds' liquid layers on the ground or their vapor plumes during evaporation by external heating. Training of a four-layer network architecture is composed of a backward-error-propagation algorithm and a gradient-descent paradigm. We conduct testing by feed-forwarding preprocessed spectra through the network in a scaled format consistent with the structure of the training-data-set representation. The best-performance weight matrix (spectral filter) evolved from final network training and testing with software simulation trials is electronically transferred to a set of eight artificial intelligence integrated circuits (ICs') in specific modular form (splitting of weight matrices). This form makes full use of all input-output IC nodes. This neural network computer serves an important real-time detection function when it is integrated into pre-and postprocessing data-handling units of a tactical prototype thermoluminescence sensor now under development at the Edgewood Research, Development, and Engineering Center.

Deep learning improves prediction of CRISPR-Cpf1 guide RNA activity.

PubMed

Kim, Hui Kwon; Min, Seonwoo; Song, Myungjae; Jung, Soobin; Choi, Jae Woo; Kim, Younggwang; Lee, Sangeun; Yoon, Sungroh; Kim, Hyongbum Henry

2018-03-01

We present two algorithms to predict the activity of AsCpf1 guide RNAs. Indel frequencies for 15,000 target sequences were used in a deep-learning framework based on a convolutional neural network to train Seq-deepCpf1. We then incorporated chromatin accessibility information to create the better-performing DeepCpf1 algorithm for cell lines for which such information is available and show that both algorithms outperform previous machine learning algorithms on our own and published data sets.

A rapid learning and dynamic stepwise updating algorithm for flat neural networks and the application to time-series prediction.

PubMed

Chen, C P; Wan, J Z

1999-01-01

A fast learning algorithm is proposed to find an optimal weights of the flat neural networks (especially, the functional-link network). Although the flat networks are used for nonlinear function approximation, they can be formulated as linear systems. Thus, the weights of the networks can be solved easily using a linear least-square method. This formulation makes it easier to update the weights instantly for both a new added pattern and a new added enhancement node. A dynamic stepwise updating algorithm is proposed to update the weights of the system on-the-fly. The model is tested on several time-series data including an infrared laser data set, a chaotic time-series, a monthly flour price data set, and a nonlinear system identification problem. The simulation results are compared to existing models in which more complex architectures and more costly training are needed. The results indicate that the proposed model is very attractive to real-time processes.

Modeling Self-Healing of Concrete Using Hybrid Genetic Algorithm-Artificial Neural Network.

PubMed

Ramadan Suleiman, Ahmed; Nehdi, Moncef L

2017-02-07

This paper presents an approach to predicting the intrinsic self-healing in concrete using a hybrid genetic algorithm-artificial neural network (GA-ANN). A genetic algorithm was implemented in the network as a stochastic optimizing tool for the initial optimal weights and biases. This approach can assist the network in achieving a global optimum and avoid the possibility of the network getting trapped at local optima. The proposed model was trained and validated using an especially built database using various experimental studies retrieved from the open literature. The model inputs include the cement content, water-to-cement ratio (w/c), type and dosage of supplementary cementitious materials, bio-healing materials, and both expansive and crystalline additives. Self-healing indicated by means of crack width is the model output. The results showed that the proposed GA-ANN model is capable of capturing the complex effects of various self-healing agents (e.g., biochemical material, silica-based additive, expansive and crystalline components) on the self-healing performance in cement-based materials.

Neural networks in data analysis and modeling for detecting littoral oil-spills by airborne laser fluorosensor remote sensing

NASA Astrophysics Data System (ADS)

Lin, Bin; An, Jubai; Brown, Carl E.; Chen, Weiwei

2003-05-01

In this paper an artificial neural network (ANN) approach, which is based on flexible nonlinear models for a very broad class of transfer functions, is applied for multi-spectral data analysis and modeling of airborne laser fluorosensor in order to differentiate between classes of oil on water surface. We use three types of algorithm: Perceptron Network, Back-Propagation (B-P) Network and Self-Organizing feature Maps (SOM) Network. Using the data in form of 64-channel spectra as inputs, the ANN presents the analysis and estimation results of the oil type on the basis of the type of background materials as outputs. The ANN is trained and tested using sample data set to the network. The results of the above 3 types of network are compared in this paper. It is proved that the training has developed a network that not only fits the training data, but also fits real-world data that the network will process operationally. The ANN model would play a significant role in the ocean oil-spill identification in the future.

Application of Particle Swarm Optimization Algorithm for Optimizing ANN Model in Recognizing Ripeness of Citrus

NASA Astrophysics Data System (ADS)

Diyana Rosli, Anis; Adenan, Nur Sabrina; Hashim, Hadzli; Ezan Abdullah, Noor; Sulaiman, Suhaimi; Baharudin, Rohaiza

2018-03-01

This paper shows findings of the application of Particle Swarm Optimization (PSO) algorithm in optimizing an Artificial Neural Network that could categorize between ripeness and unripeness stage of citrus suhuensis. The algorithm would adjust the network connections weights and adapt its values during training for best results at the output. Initially, citrus suhuensis fruit’s skin is measured using optically non-destructive method via spectrometer. The spectrometer would transmit VIS (visible spectrum) photonic light radiation to the surface (skin of citrus) of the sample. The reflected light from the sample’s surface would be received and measured by the same spectrometer in terms of reflectance percentage based on VIS range. These measured data are used to train and test the best optimized ANN model. The accuracy is based on receiver operating characteristic (ROC) performance. The result outcomes from this investigation have shown that the achieved accuracy for the optimized is 70.5% with a sensitivity and specificity of 60.1% and 80.0% respectively.

Application of complex discrete wavelet transform in classification of Doppler signals using complex-valued artificial neural network.

PubMed

Ceylan, Murat; Ceylan, Rahime; Ozbay, Yüksel; Kara, Sadik

2008-09-01

In biomedical signal classification, due to the huge amount of data, to compress the biomedical waveform data is vital. This paper presents two different structures formed using feature extraction algorithms to decrease size of feature set in training and test data. The proposed structures, named as wavelet transform-complex-valued artificial neural network (WT-CVANN) and complex wavelet transform-complex-valued artificial neural network (CWT-CVANN), use real and complex discrete wavelet transform for feature extraction. The aim of using wavelet transform is to compress data and to reduce training time of network without decreasing accuracy rate. In this study, the presented structures were applied to the problem of classification in carotid arterial Doppler ultrasound signals. Carotid arterial Doppler ultrasound signals were acquired from left carotid arteries of 38 patients and 40 healthy volunteers. The patient group included 22 males and 16 females with an established diagnosis of the early phase of atherosclerosis through coronary or aortofemoropopliteal (lower extremity) angiographies (mean age, 59 years; range, 48-72 years). Healthy volunteers were young non-smokers who seem to not bear any risk of atherosclerosis, including 28 males and 12 females (mean age, 23 years; range, 19-27 years). Sensitivity, specificity and average detection rate were calculated for comparison, after training and test phases of all structures finished. These parameters have demonstrated that training times of CVANN and real-valued artificial neural network (RVANN) were reduced using feature extraction algorithms without decreasing accuracy rate in accordance to our aim.

Design development of a neural network-based telemetry monitor

NASA Technical Reports Server (NTRS)

Lembeck, Michael F.

1992-01-01

This paper identifies the requirements and describes an architectural framework for an artificial neural network-based system that is capable of fulfilling monitoring and control requirements of future aerospace missions. Incorporated into this framework are a newly developed training algorithm and the concept of cooperative network architectures. The feasibility of such an approach is demonstrated for its ability to identify faults in low frequency waveforms.

Facial Emotions Recognition using Gabor Transform and Facial Animation Parameters with Neural Networks

NASA Astrophysics Data System (ADS)

Harit, Aditya; Joshi, J. C., Col; Gupta, K. K.

2018-03-01

The paper proposed an automatic facial emotion recognition algorithm which comprises of two main components: feature extraction and expression recognition. The algorithm uses a Gabor filter bank on fiducial points to find the facial expression features. The resulting magnitudes of Gabor transforms, along with 14 chosen FAPs (Facial Animation Parameters), compose the feature space. There are two stages: the training phase and the recognition phase. Firstly, for the present 6 different emotions, the system classifies all training expressions in 6 different classes (one for each emotion) in the training stage. In the recognition phase, it recognizes the emotion by applying the Gabor bank to a face image, then finds the fiducial points, and then feeds it to the trained neural architecture.

Computing single step operators of logic programming in radial basis function neural networks

NASA Astrophysics Data System (ADS)

Hamadneh, Nawaf; Sathasivam, Saratha; Choon, Ong Hong

2014-07-01

Logic programming is the process that leads from an original formulation of a computing problem to executable programs. A normal logic program consists of a finite set of clauses. A valuation I of logic programming is a mapping from ground atoms to false or true. The single step operator of any logic programming is defined as a function (Tp:I→I). Logic programming is well-suited to building the artificial intelligence systems. In this study, we established a new technique to compute the single step operators of logic programming in the radial basis function neural networks. To do that, we proposed a new technique to generate the training data sets of single step operators. The training data sets are used to build the neural networks. We used the recurrent radial basis function neural networks to get to the steady state (the fixed point of the operators). To improve the performance of the neural networks, we used the particle swarm optimization algorithm to train the networks.

An algorithm for generating modular hierarchical neural network classifiers: a step toward larger scale applications

NASA Astrophysics Data System (ADS)

Roverso, Davide

2003-08-01

Many-class learning is the problem of training a classifier to discriminate among a large number of target classes. Together with the problem of dealing with high-dimensional patterns (i.e. a high-dimensional input space), the many class problem (i.e. a high-dimensional output space) is a major obstacle to be faced when scaling-up classifier systems and algorithms from small pilot applications to large full-scale applications. The Autonomous Recursive Task Decomposition (ARTD) algorithm is here proposed as a solution to the problem of many-class learning. Example applications of ARTD to neural classifier training are also presented. In these examples, improvements in training time are shown to range from 4-fold to more than 30-fold in pattern classification tasks of both static and dynamic character.

Parameter estimates in binary black hole collisions using neural networks

NASA Astrophysics Data System (ADS)

Carrillo, M.; Gracia-Linares, M.; González, J. A.; Guzmán, F. S.

2016-10-01

We present an algorithm based on artificial neural networks (ANNs), that estimates the mass ratio in a binary black hole collision out of given gravitational wave (GW) strains. In this analysis, the ANN is trained with a sample of GW signals generated with numerical simulations. The effectiveness of the algorithm is evaluated with GWs generated also with simulations for given mass ratios unknown to the ANN. We measure the accuracy of the algorithm in the interpolation and extrapolation regimes. We present the results for noise free signals and signals contaminated with Gaussian noise, in order to foresee the dependence of the method accuracy in terms of the signal to noise ratio.

Research on optimization of combustion efficiency of thermal power unit based on genetic algorithm

NASA Astrophysics Data System (ADS)

Zhou, Qiongyang

2018-04-01

In order to improve the economic performance and reduce pollutant emissions of thermal power units, the characteristics of neural network in establishing boiler combustion model are analyzed based on the analysis of the main factors affecting boiler efficiency by using orthogonal method. In addition, on the basis of this model, the genetic algorithm is used to find the best control amount of the furnace combustion in a certain working condition. Through the genetic algorithm based on real number encoding and roulette selection is concluded: the best control quantity at a condition of furnace combustion can be combined with the boiler combustion system model for neural network training. The precision of the neural network model is further improved, and the basic work is laid for the research of the whole boiler combustion optimization system.

Performance Evaluation of Machine Learning Methods for Leaf Area Index Retrieval from Time-Series MODIS Reflectance Data

PubMed Central

Wang, Tongtong; Xiao, Zhiqiang; Liu, Zhigang

2017-01-01

Leaf area index (LAI) is an important biophysical parameter and the retrieval of LAI from remote sensing data is the only feasible method for generating LAI products at regional and global scales. However, most LAI retrieval methods use satellite observations at a specific time to retrieve LAI. Because of the impacts of clouds and aerosols, the LAI products generated by these methods are spatially incomplete and temporally discontinuous, and thus they cannot meet the needs of practical applications. To generate high-quality LAI products, four machine learning algorithms, including back-propagation neutral network (BPNN), radial basis function networks (RBFNs), general regression neutral networks (GRNNs), and multi-output support vector regression (MSVR) are proposed to retrieve LAI from time-series Moderate Resolution Imaging Spectroradiometer (MODIS) reflectance data in this study and performance of these machine learning algorithms is evaluated. The results demonstrated that GRNNs, RBFNs, and MSVR exhibited low sensitivity to training sample size, whereas BPNN had high sensitivity. The four algorithms performed slightly better with red, near infrared (NIR), and short wave infrared (SWIR) bands than red and NIR bands, and the results were significantly better than those obtained using single band reflectance data (red or NIR). Regardless of band composition, GRNNs performed better than the other three methods. Among the four algorithms, BPNN required the least training time, whereas MSVR needed the most for any sample size. PMID:28045443

Using deep convolutional neural networks to identify and classify tumor-associated stroma in diagnostic breast biopsies.

PubMed

Ehteshami Bejnordi, Babak; Mullooly, Maeve; Pfeiffer, Ruth M; Fan, Shaoqi; Vacek, Pamela M; Weaver, Donald L; Herschorn, Sally; Brinton, Louise A; van Ginneken, Bram; Karssemeijer, Nico; Beck, Andrew H; Gierach, Gretchen L; van der Laak, Jeroen A W M; Sherman, Mark E

2018-06-13

The breast stromal microenvironment is a pivotal factor in breast cancer development, growth and metastases. Although pathologists often detect morphologic changes in stroma by light microscopy, visual classification of such changes is subjective and non-quantitative, limiting its diagnostic utility. To gain insights into stromal changes associated with breast cancer, we applied automated machine learning techniques to digital images of 2387 hematoxylin and eosin stained tissue sections of benign and malignant image-guided breast biopsies performed to investigate mammographic abnormalities among 882 patients, ages 40-65 years, that were enrolled in the Breast Radiology Evaluation and Study of Tissues (BREAST) Stamp Project. Using deep convolutional neural networks, we trained an algorithm to discriminate between stroma surrounding invasive cancer and stroma from benign biopsies. In test sets (928 whole-slide images from 330 patients), this algorithm could distinguish biopsies diagnosed as invasive cancer from benign biopsies solely based on the stromal characteristics (area under the receiver operator characteristics curve = 0.962). Furthermore, without being trained specifically using ductal carcinoma in situ as an outcome, the algorithm detected tumor-associated stroma in greater amounts and at larger distances from grade 3 versus grade 1 ductal carcinoma in situ. Collectively, these results suggest that algorithms based on deep convolutional neural networks that evaluate only stroma may prove useful to classify breast biopsies and aid in understanding and evaluating the biology of breast lesions.

Performance Evaluation of Machine Learning Methods for Leaf Area Index Retrieval from Time-Series MODIS Reflectance Data.

PubMed

Wang, Tongtong; Xiao, Zhiqiang; Liu, Zhigang

2017-01-01

Leaf area index (LAI) is an important biophysical parameter and the retrieval of LAI from remote sensing data is the only feasible method for generating LAI products at regional and global scales. However, most LAI retrieval methods use satellite observations at a specific time to retrieve LAI. Because of the impacts of clouds and aerosols, the LAI products generated by these methods are spatially incomplete and temporally discontinuous, and thus they cannot meet the needs of practical applications. To generate high-quality LAI products, four machine learning algorithms, including back-propagation neutral network (BPNN), radial basis function networks (RBFNs), general regression neutral networks (GRNNs), and multi-output support vector regression (MSVR) are proposed to retrieve LAI from time-series Moderate Resolution Imaging Spectroradiometer (MODIS) reflectance data in this study and performance of these machine learning algorithms is evaluated. The results demonstrated that GRNNs, RBFNs, and MSVR exhibited low sensitivity to training sample size, whereas BPNN had high sensitivity. The four algorithms performed slightly better with red, near infrared (NIR), and short wave infrared (SWIR) bands than red and NIR bands, and the results were significantly better than those obtained using single band reflectance data (red or NIR). Regardless of band composition, GRNNs performed better than the other three methods. Among the four algorithms, BPNN required the least training time, whereas MSVR needed the most for any sample size.

Evolving optimised decision rules for intrusion detection using particle swarm paradigm

NASA Astrophysics Data System (ADS)

Sivatha Sindhu, Siva S.; Geetha, S.; Kannan, A.

2012-12-01

The aim of this article is to construct a practical intrusion detection system (IDS) that properly analyses the statistics of network traffic pattern and classify them as normal or anomalous class. The objective of this article is to prove that the choice of effective network traffic features and a proficient machine-learning paradigm enhances the detection accuracy of IDS. In this article, a rule-based approach with a family of six decision tree classifiers, namely Decision Stump, C4.5, Naive Baye's Tree, Random Forest, Random Tree and Representative Tree model to perform the detection of anomalous network pattern is introduced. In particular, the proposed swarm optimisation-based approach selects instances that compose training set and optimised decision tree operate over this trained set producing classification rules with improved coverage, classification capability and generalisation ability. Experiment with the Knowledge Discovery and Data mining (KDD) data set which have information on traffic pattern, during normal and intrusive behaviour shows that the proposed algorithm produces optimised decision rules and outperforms other machine-learning algorithm.

Character Recognition Using Genetically Trained Neural Networks

DOE Office of Scientific and Technical Information (OSTI.GOV)

Diniz, C.; Stantz, K.M.; Trahan, M.W.

1998-10-01

Computationally intelligent recognition of characters and symbols addresses a wide range of applications including foreign language translation and chemical formula identification. The combination of intelligent learning and optimization algorithms with layered neural structures offers powerful techniques for character recognition. These techniques were originally developed by Sandia National Laboratories for pattern and spectral analysis; however, their ability to optimize vast amounts of data make them ideal for character recognition. An adaptation of the Neural Network Designer soflsvare allows the user to create a neural network (NN_) trained by a genetic algorithm (GA) that correctly identifies multiple distinct characters. The initial successfidmore » recognition of standard capital letters can be expanded to include chemical and mathematical symbols and alphabets of foreign languages, especially Arabic and Chinese. The FIN model constructed for this project uses a three layer feed-forward architecture. To facilitate the input of characters and symbols, a graphic user interface (GUI) has been developed to convert the traditional representation of each character or symbol to a bitmap. The 8 x 8 bitmap representations used for these tests are mapped onto the input nodes of the feed-forward neural network (FFNN) in a one-to-one correspondence. The input nodes feed forward into a hidden layer, and the hidden layer feeds into five output nodes correlated to possible character outcomes. During the training period the GA optimizes the weights of the NN until it can successfully recognize distinct characters. Systematic deviations from the base design test the network's range of applicability. Increasing capacity, the number of letters to be recognized, requires a nonlinear increase in the number of hidden layer neurodes. Optimal character recognition performance necessitates a minimum threshold for the number of cases when genetically training the net. And, the amount of noise significantly degrades character recognition efficiency, some of which can be overcome by adding noise during training and optimizing the form of the network's activation fimction.« less

Quantum neural networks: Current status and prospects for development

NASA Astrophysics Data System (ADS)

Altaisky, M. V.; Kaputkina, N. E.; Krylov, V. A.

2014-11-01

The idea of quantum artificial neural networks, first formulated in [34], unites the artificial neural network concept with the quantum computation paradigm. Quantum artificial neural networks were first systematically considered in the PhD thesis by T. Menneer (1998). Based on the works of Menneer and Narayanan [42, 43], Kouda, Matsui, and Nishimura [35, 36], Altaisky [2, 68], Zhou [67], and others, quantum-inspired learning algorithms for neural networks were developed, and are now used in various training programs and computer games [29, 30]. The first practically realizable scaled hardware-implemented model of the quantum artificial neural network is obtained by D-Wave Systems, Inc. [33]. It is a quantum Hopfield network implemented on the basis of superconducting quantum interference devices (SQUIDs). In this work we analyze possibilities and underlying principles of an alternative way to implement quantum neural networks on the basis of quantum dots. A possibility of using quantum neural network algorithms in automated control systems, associative memory devices, and in modeling biological and social networks is examined.

Livermore Big Artificial Neural Network Toolkit

DOE Office of Scientific and Technical Information (OSTI.GOV)

Essen, Brian Van; Jacobs, Sam; Kim, Hyojin

2016-07-01

LBANN is a toolkit that is designed to train artificial neural networks efficiently on high performance computing architectures. It is optimized to take advantages of key High Performance Computing features to accelerate neural network training. Specifically it is optimized for low-latency, high bandwidth interconnects, node-local NVRAM, node-local GPU accelerators, and high bandwidth parallel file systems. It is built on top of the open source Elemental distributed-memory dense and spars-direct linear algebra and optimization library that is released under the BSD license. The algorithms contained within LBANN are drawn from the academic literature and implemented to work within a distributed-memory framework.

A Grey Wolf Optimizer for Modular Granular Neural Networks for Human Recognition

PubMed Central

Sánchez, Daniela; Melin, Patricia

2017-01-01

A grey wolf optimizer for modular neural network (MNN) with a granular approach is proposed. The proposed method performs optimal granulation of data and design of modular neural networks architectures to perform human recognition, and to prove its effectiveness benchmark databases of ear, iris, and face biometric measures are used to perform tests and comparisons against other works. The design of a modular granular neural network (MGNN) consists in finding optimal parameters of its architecture; these parameters are the number of subgranules, percentage of data for the training phase, learning algorithm, goal error, number of hidden layers, and their number of neurons. Nowadays, there is a great variety of approaches and new techniques within the evolutionary computing area, and these approaches and techniques have emerged to help find optimal solutions to problems or models and bioinspired algorithms are part of this area. In this work a grey wolf optimizer is proposed for the design of modular granular neural networks, and the results are compared against a genetic algorithm and a firefly algorithm in order to know which of these techniques provides better results when applied to human recognition. PMID:28894461

A Grey Wolf Optimizer for Modular Granular Neural Networks for Human Recognition.

PubMed

Sánchez, Daniela; Melin, Patricia; Castillo, Oscar

2017-01-01

A grey wolf optimizer for modular neural network (MNN) with a granular approach is proposed. The proposed method performs optimal granulation of data and design of modular neural networks architectures to perform human recognition, and to prove its effectiveness benchmark databases of ear, iris, and face biometric measures are used to perform tests and comparisons against other works. The design of a modular granular neural network (MGNN) consists in finding optimal parameters of its architecture; these parameters are the number of subgranules, percentage of data for the training phase, learning algorithm, goal error, number of hidden layers, and their number of neurons. Nowadays, there is a great variety of approaches and new techniques within the evolutionary computing area, and these approaches and techniques have emerged to help find optimal solutions to problems or models and bioinspired algorithms are part of this area. In this work a grey wolf optimizer is proposed for the design of modular granular neural networks, and the results are compared against a genetic algorithm and a firefly algorithm in order to know which of these techniques provides better results when applied to human recognition.

A study on the performance comparison of metaheuristic algorithms on the learning of neural networks

NASA Astrophysics Data System (ADS)

Lai, Kee Huong; Zainuddin, Zarita; Ong, Pauline

2017-08-01

The learning or training process of neural networks entails the task of finding the most optimal set of parameters, which includes translation vectors, dilation parameter, synaptic weights, and bias terms. Apart from the traditional gradient descent-based methods, metaheuristic methods can also be used for this learning purpose. Since the inception of genetic algorithm half a century ago, the last decade witnessed the explosion of a variety of novel metaheuristic algorithms, such as harmony search algorithm, bat algorithm, and whale optimization algorithm. Despite the proof of the no free lunch theorem in the discipline of optimization, a survey in the literature of machine learning gives contrasting results. Some researchers report that certain metaheuristic algorithms are superior to the others, whereas some others argue that different metaheuristic algorithms give comparable performance. As such, this paper aims to investigate if a certain metaheuristic algorithm will outperform the other algorithms. In this work, three metaheuristic algorithms, namely genetic algorithms, particle swarm optimization, and harmony search algorithm are considered. The algorithms are incorporated in the learning of neural networks and their classification results on the benchmark UCI machine learning data sets are compared. It is found that all three metaheuristic algorithms give similar and comparable performance, as captured in the average overall classification accuracy. The results corroborate the findings reported in the works done by previous researchers. Several recommendations are given, which include the need of statistical analysis to verify the results and further theoretical works to support the obtained empirical results.

Scaling of counter-current imbibition recovery curves using artificial neural networks

NASA Astrophysics Data System (ADS)

Jafari, Iman; Masihi, Mohsen; Nasiri Zarandi, Masoud

2018-06-01

Scaling imbibition curves are of great importance in the characterization and simulation of oil production from naturally fractured reservoirs. Different parameters such as matrix porosity and permeability, oil and water viscosities, matrix dimensions, and oil/water interfacial tensions have an effective on the imbibition process. Studies on the scaling imbibition curves along with the consideration of different assumptions have resulted in various scaling equations. In this work, using an artificial neural network (ANN) method, a novel technique is presented for scaling imbibition recovery curves, which can be used for scaling the experimental and field-scale imbibition cases. The imbibition recovery curves for training and testing the neural network were gathered through the simulation of different scenarios using a commercial reservoir simulator. In this ANN-based method, six parameters were assumed to have an effect on the imbibition process and were considered as the inputs for training the network. Using the ‘Bayesian regularization’ training algorithm, the network was trained and tested. Training and testing phases showed superior results in comparison with the other scaling methods. It is concluded that using the new technique is useful for scaling imbibition recovery curves, especially for complex cases, for which the common scaling methods are not designed.

Real time coarse orientation detection in MR scans using multi-planar deep convolutional neural networks

NASA Astrophysics Data System (ADS)

Bhatia, Parmeet S.; Reda, Fitsum; Harder, Martin; Zhan, Yiqiang; Zhou, Xiang Sean

2017-02-01

Automatically detecting anatomy orientation is an important task in medical image analysis. Specifically, the ability to automatically detect coarse orientation of structures is useful to minimize the effort of fine/accurate orientation detection algorithms, to initialize non-rigid deformable registration algorithms or to align models to target structures in model-based segmentation algorithms. In this work, we present a deep convolution neural network (DCNN)-based method for fast and robust detection of the coarse structure orientation, i.e., the hemi-sphere where the principal axis of a structure lies. That is, our algorithm predicts whether the principal orientation of a structure is in the northern hemisphere or southern hemisphere, which we will refer to as UP and DOWN, respectively, in the remainder of this manuscript. The only assumption of our method is that the entire structure is located within the scan's field-of-view (FOV). To efficiently solve the problem in 3D space, we formulated it as a multi-planar 2D deep learning problem. In the training stage, a large number coronal-sagittal slice pairs are constructed as 2-channel images to train a DCNN to classify whether a scan is UP or DOWN. During testing, we randomly sample a small number of coronal-sagittal 2-channel images and pass them through our trained network. Finally, coarse structure orientation is determined using majority voting. We tested our method on 114 Elbow MR Scans. Experimental results suggest that only five 2-channel images are sufficient to achieve a high success rate of 97.39%. Our method is also extremely fast and takes approximately 50 milliseconds per 3D MR scan. Our method is insensitive to the location of the structure in the FOV.

Neural network computer simulation of medical aerosols.

PubMed

Richardson, C J; Barlow, D J

1996-06-01

Preliminary investigations have been conducted to assess the potential for using artificial neural networks to simulate aerosol behaviour, with a view to employing this type of methodology in the evaluation and design of pulmonary drug-delivery systems. Details are presented of the general purpose software developed for these tasks; it implements a feed-forward back-propagation algorithm with weight decay and connection pruning, the user having complete run-time control of the network architecture and mode of training. A series of exploratory investigations is then reported in which different network structures and training strategies are assessed in terms of their ability to simulate known patterns of fluid flow in simple model systems. The first of these involves simulations of cellular automata-generated data for fluid flow through a partially obstructed two-dimensional pipe. The artificial neural networks are shown to be highly successful in simulating the behaviour of this simple linear system, but with important provisos relating to the information content of the training data and the criteria used to judge when the network is properly trained. A second set of investigations is then reported in which similar networks are used to simulate patterns of fluid flow through aerosol generation devices, using training data furnished through rigorous computational fluid dynamics modelling. These more complex three-dimensional systems are modelled with equal success. It is concluded that carefully tailored, well trained networks could provide valuable tools not just for predicting but also for analysing the spatial dynamics of pharmaceutical aerosols.

GXNOR-Net: Training deep neural networks with ternary weights and activations without full-precision memory under a unified discretization framework.

PubMed

Deng, Lei; Jiao, Peng; Pei, Jing; Wu, Zhenzhi; Li, Guoqi

2018-04-01

Although deep neural networks (DNNs) are being a revolutionary power to open up the AI era, the notoriously huge hardware overhead has challenged their applications. Recently, several binary and ternary networks, in which the costly multiply-accumulate operations can be replaced by accumulations or even binary logic operations, make the on-chip training of DNNs quite promising. Therefore there is a pressing need to build an architecture that could subsume these networks under a unified framework that achieves both higher performance and less overhead. To this end, two fundamental issues are yet to be addressed. The first one is how to implement the back propagation when neuronal activations are discrete. The second one is how to remove the full-precision hidden weights in the training phase to break the bottlenecks of memory/computation consumption. To address the first issue, we present a multi-step neuronal activation discretization method and a derivative approximation technique that enable the implementing the back propagation algorithm on discrete DNNs. While for the second issue, we propose a discrete state transition (DST) methodology to constrain the weights in a discrete space without saving the hidden weights. Through this way, we build a unified framework that subsumes the binary or ternary networks as its special cases, and under which a heuristic algorithm is provided at the website https://github.com/AcrossV/Gated-XNOR. More particularly, we find that when both the weights and activations become ternary values, the DNNs can be reduced to sparse binary networks, termed as gated XNOR networks (GXNOR-Nets) since only the event of non-zero weight and non-zero activation enables the control gate to start the XNOR logic operations in the original binary networks. This promises the event-driven hardware design for efficient mobile intelligence. We achieve advanced performance compared with state-of-the-art algorithms. Furthermore, the computational sparsity and the number of states in the discrete space can be flexibly modified to make it suitable for various hardware platforms. Copyright © 2018 Elsevier Ltd. All rights reserved.

MABAL: a Novel Deep-Learning Architecture for Machine-Assisted Bone Age Labeling.

PubMed

Mutasa, Simukayi; Chang, Peter D; Ruzal-Shapiro, Carrie; Ayyala, Rama

2018-02-05

Bone age assessment (BAA) is a commonly performed diagnostic study in pediatric radiology to assess skeletal maturity. The most commonly utilized method for assessment of BAA is the Greulich and Pyle method (Pediatr Radiol 46.9:1269-1274, 2016; Arch Dis Child 81.2:172-173, 1999) atlas. The evaluation of BAA can be a tedious and time-consuming process for the radiologist. As such, several computer-assisted detection/diagnosis (CAD) methods have been proposed for automation of BAA. Classical CAD tools have traditionally relied on hard-coded algorithmic features for BAA which suffer from a variety of drawbacks. Recently, the advent and proliferation of convolutional neural networks (CNNs) has shown promise in a variety of medical imaging applications. There have been at least two published applications of using deep learning for evaluation of bone age (Med Image Anal 36:41-51, 2017; JDI 1-5, 2017). However, current implementations are limited by a combination of both architecture design and relatively small datasets. The purpose of this study is to demonstrate the benefits of a customized neural network algorithm carefully calibrated to the evaluation of bone age utilizing a relatively large institutional dataset. In doing so, this study will aim to show that advanced architectures can be successfully trained from scratch in the medical imaging domain and can generate results that outperform any existing proposed algorithm. The training data consisted of 10,289 images of different skeletal age examinations, 8909 from the hospital Picture Archiving and Communication System at our institution and 1383 from the public Digital Hand Atlas Database. The data was separated into four cohorts, one each for male and female children above the age of 8, and one each for male and female children below the age of 10. The testing set consisted of 20 radiographs of each 1-year-age cohort from 0 to 1 years to 14-15+ years, half male and half female. The testing set included left-hand radiographs done for bone age assessment, trauma evaluation without significant findings, and skeletal surveys. A 14 hidden layer-customized neural network was designed for this study. The network included several state of the art techniques including residual-style connections, inception layers, and spatial transformer layers. Data augmentation was applied to the network inputs to prevent overfitting. A linear regression output was utilized. Mean square error was used as the network loss function and mean absolute error (MAE) was utilized as the primary performance metric. MAE accuracies on the validation and test sets for young females were 0.654 and 0.561 respectively. For older females, validation and test accuracies were 0.662 and 0.497 respectively. For young males, validation and test accuracies were 0.649 and 0.585 respectively. Finally, for older males, validation and test set accuracies were 0.581 and 0.501 respectively. The female cohorts were trained for 900 epochs each and the male cohorts were trained for 600 epochs. An eightfold cross-validation set was employed for hyperparameter tuning. Test error was obtained after training on a full data set with the selected hyperparameters. Using our proposed customized neural network architecture on our large available data, we achieved an aggregate validation and test set mean absolute errors of 0.637 and 0.536 respectively. To date, this is the best published performance on utilizing deep learning for bone age assessment. Our results support our initial hypothesis that customized, purpose-built neural networks provide improved performance over networks derived from pre-trained imaging data sets. We build on that initial work by showing that the addition of state-of-the-art techniques such as residual connections and inception architecture further improves prediction accuracy. This is important because the current assumption for use of residual and/or inception architectures is that a large pre-trained network is required for successful implementation given the relatively small datasets in medical imaging. Instead we show that a small, customized architecture incorporating advanced CNN strategies can indeed be trained from scratch, yielding significant improvements in algorithm accuracy. It should be noted that for all four cohorts, testing error outperformed validation error. One reason for this is that our ground truth for our test set was obtained by averaging two pediatric radiologist reads compared to our training data for which only a single read was used. This suggests that despite relatively noisy training data, the algorithm could successfully model the variation between observers and generate estimates that are close to the expected ground truth.

Development of a Deep Learning Algorithm for Automatic Diagnosis of Diabetic Retinopathy.

PubMed

Raju, Manoj; Pagidimarri, Venkatesh; Barreto, Ryan; Kadam, Amrit; Kasivajjala, Vamsichandra; Aswath, Arun

2017-01-01

This paper mainly focuses on the deep learning application in classifying the stage of diabetic retinopathy and detecting the laterality of the eye using funduscopic images. Diabetic retinopathy is a chronic, progressive, sight-threatening disease of the retinal blood vessels. Ophthalmologists diagnose diabetic retinopathy through early funduscopic screening. Normally, there is a time delay in reporting and intervention, apart from the financial cost and risk of blindness associated with it. Using a convolutional neural network based approach for automatic diagnosis of diabetic retinopathy, we trained the prediction network on the publicly available Kaggle dataset. Approximately 35,000 images were used to train the network, which observed a sensitivity of 80.28% and a specificity of 92.29% on the validation dataset of ~53,000 images. Using 8,810 images, the network was trained for detecting the laterality of the eye and observed an accuracy of 93.28% on the validation set of 8,816 images.

Classification of Parkinson's disease utilizing multi-edit nearest-neighbor and ensemble learning algorithms with speech samples.

PubMed

Zhang, He-Hua; Yang, Liuyang; Liu, Yuchuan; Wang, Pin; Yin, Jun; Li, Yongming; Qiu, Mingguo; Zhu, Xueru; Yan, Fang

2016-11-16

The use of speech based data in the classification of Parkinson disease (PD) has been shown to provide an effect, non-invasive mode of classification in recent years. Thus, there has been an increased interest in speech pattern analysis methods applicable to Parkinsonism for building predictive tele-diagnosis and tele-monitoring models. One of the obstacles in optimizing classifications is to reduce noise within the collected speech samples, thus ensuring better classification accuracy and stability. While the currently used methods are effect, the ability to invoke instance selection has been seldomly examined. In this study, a PD classification algorithm was proposed and examined that combines a multi-edit-nearest-neighbor (MENN) algorithm and an ensemble learning algorithm. First, the MENN algorithm is applied for selecting optimal training speech samples iteratively, thereby obtaining samples with high separability. Next, an ensemble learning algorithm, random forest (RF) or decorrelated neural network ensembles (DNNE), is used to generate trained samples from the collected training samples. Lastly, the trained ensemble learning algorithms are applied to the test samples for PD classification. This proposed method was examined using a more recently deposited public datasets and compared against other currently used algorithms for validation. Experimental results showed that the proposed algorithm obtained the highest degree of improved classification accuracy (29.44%) compared with the other algorithm that was examined. Furthermore, the MENN algorithm alone was found to improve classification accuracy by as much as 45.72%. Moreover, the proposed algorithm was found to exhibit a higher stability, particularly when combining the MENN and RF algorithms. This study showed that the proposed method could improve PD classification when using speech data and can be applied to future studies seeking to improve PD classification methods.

Inference of genetic network of Xenopus frog egg: improved genetic algorithm.

PubMed

Wu, Shinq-Jen; Chou, Chia-Hsien; Wu, Cheng-Tao; Lee, Tsu-Tian

2006-01-01

An improved genetic algorithm (IGA) is proposed to achieve S-system gene network modeling of Xenopus frog egg. Via the time-courses training datasets from Michaelis-Menten model, the optimal parameters are learned. The S-system can clearly describe activative and inhibitory interaction between genes as generating and consuming process. We concern the mitotic control in cell-cycle of Xenopus frog egg to realize cyclin-Cdc2 and Cdc25 for MPF activity. The proposed IGA can achieve global search with migration and keep the best chromosome with elitism operation. The generated gene regulatory networks can provide biological researchers for further experiments in Xenopus frog egg cell cycle control.

Pileup Mitigation with Machine Learning (PUMML)

NASA Astrophysics Data System (ADS)

Komiske, Patrick T.; Metodiev, Eric M.; Nachman, Benjamin; Schwartz, Matthew D.

2017-12-01

Pileup involves the contamination of the energy distribution arising from the primary collision of interest (leading vertex) by radiation from soft collisions (pileup). We develop a new technique for removing this contamination using machine learning and convolutional neural networks. The network takes as input the energy distribution of charged leading vertex particles, charged pileup particles, and all neutral particles and outputs the energy distribution of particles coming from leading vertex alone. The PUMML algorithm performs remarkably well at eliminating pileup distortion on a wide range of simple and complex jet observables. We test the robustness of the algorithm in a number of ways and discuss how the network can be trained directly on data.

Computational Models of Neuron-Astrocyte Interactions Lead to Improved Efficacy in the Performance of Neural Networks

PubMed Central

Alvarellos-González, Alberto; Pazos, Alejandro; Porto-Pazos, Ana B.

2012-01-01

The importance of astrocytes, one part of the glial system, for information processing in the brain has recently been demonstrated. Regarding information processing in multilayer connectionist systems, it has been shown that systems which include artificial neurons and astrocytes (Artificial Neuron-Glia Networks) have well-known advantages over identical systems including only artificial neurons. Since the actual impact of astrocytes in neural network function is unknown, we have investigated, using computational models, different astrocyte-neuron interactions for information processing; different neuron-glia algorithms have been implemented for training and validation of multilayer Artificial Neuron-Glia Networks oriented toward classification problem resolution. The results of the tests performed suggest that all the algorithms modelling astrocyte-induced synaptic potentiation improved artificial neural network performance, but their efficacy depended on the complexity of the problem. PMID:22649480

Computational models of neuron-astrocyte interactions lead to improved efficacy in the performance of neural networks.

PubMed

Alvarellos-González, Alberto; Pazos, Alejandro; Porto-Pazos, Ana B

2012-01-01

The importance of astrocytes, one part of the glial system, for information processing in the brain has recently been demonstrated. Regarding information processing in multilayer connectionist systems, it has been shown that systems which include artificial neurons and astrocytes (Artificial Neuron-Glia Networks) have well-known advantages over identical systems including only artificial neurons. Since the actual impact of astrocytes in neural network function is unknown, we have investigated, using computational models, different astrocyte-neuron interactions for information processing; different neuron-glia algorithms have been implemented for training and validation of multilayer Artificial Neuron-Glia Networks oriented toward classification problem resolution. The results of the tests performed suggest that all the algorithms modelling astrocyte-induced synaptic potentiation improved artificial neural network performance, but their efficacy depended on the complexity of the problem.

CNN universal machine as classificaton platform: an art-like clustering algorithm.

PubMed

Bálya, David

2003-12-01

Fast and robust classification of feature vectors is a crucial task in a number of real-time systems. A cellular neural/nonlinear network universal machine (CNN-UM) can be very efficient as a feature detector. The next step is to post-process the results for object recognition. This paper shows how a robust classification scheme based on adaptive resonance theory (ART) can be mapped to the CNN-UM. Moreover, this mapping is general enough to include different types of feed-forward neural networks. The designed analogic CNN algorithm is capable of classifying the extracted feature vectors keeping the advantages of the ART networks, such as robust, plastic and fault-tolerant behaviors. An analogic algorithm is presented for unsupervised classification with tunable sensitivity and automatic new class creation. The algorithm is extended for supervised classification. The presented binary feature vector classification is implemented on the existing standard CNN-UM chips for fast classification. The experimental evaluation shows promising performance after 100% accuracy on the training set.

Wavelet neural networks: a practical guide.

PubMed

Alexandridis, Antonios K; Zapranis, Achilleas D

2013-06-01

Wavelet networks (WNs) are a new class of networks which have been used with great success in a wide range of applications. However a general accepted framework for applying WNs is missing from the literature. In this study, we present a complete statistical model identification framework in order to apply WNs in various applications. The following subjects were thoroughly examined: the structure of a WN, training methods, initialization algorithms, variable significance and variable selection algorithms, model selection methods and finally methods to construct confidence and prediction intervals. In addition the complexity of each algorithm is discussed. Our proposed framework was tested in two simulated cases, in one chaotic time series described by the Mackey-Glass equation and in three real datasets described by daily temperatures in Berlin, daily wind speeds in New York and breast cancer classification. Our results have shown that the proposed algorithms produce stable and robust results indicating that our proposed framework can be applied in various applications. Copyright © 2013 Elsevier Ltd. All rights reserved.

Dreaming of Atmospheres

NASA Astrophysics Data System (ADS)

Waldmann, I. P.

2016-04-01

Here, we introduce the RobERt (Robotic Exoplanet Recognition) algorithm for the classification of exoplanetary emission spectra. Spectral retrieval of exoplanetary atmospheres frequently requires the preselection of molecular/atomic opacities to be defined by the user. In the era of open-source, automated, and self-sufficient retrieval algorithms, manual input should be avoided. User dependent input could, in worst-case scenarios, lead to incomplete models and biases in the retrieval. The RobERt algorithm is based on deep-belief neural (DBN) networks trained to accurately recognize molecular signatures for a wide range of planets, atmospheric thermal profiles, and compositions. Reconstructions of the learned features, also referred to as the “dreams” of the network, indicate good convergence and an accurate representation of molecular features in the DBN. Using these deep neural networks, we work toward retrieval algorithms that themselves understand the nature of the observed spectra, are able to learn from current and past data, and make sensible qualitative preselections of atmospheric opacities to be used for the quantitative stage of the retrieval process.

Single-hidden-layer feed-forward quantum neural network based on Grover learning.

PubMed

Liu, Cheng-Yi; Chen, Chein; Chang, Ching-Ter; Shih, Lun-Min

2013-09-01

In this paper, a novel single-hidden-layer feed-forward quantum neural network model is proposed based on some concepts and principles in the quantum theory. By combining the quantum mechanism with the feed-forward neural network, we defined quantum hidden neurons and connected quantum weights, and used them as the fundamental information processing unit in a single-hidden-layer feed-forward neural network. The quantum neurons make a wide range of nonlinear functions serve as the activation functions in the hidden layer of the network, and the Grover searching algorithm outstands the optimal parameter setting iteratively and thus makes very efficient neural network learning possible. The quantum neuron and weights, along with a Grover searching algorithm based learning, result in a novel and efficient neural network characteristic of reduced network, high efficient training and prospect application in future. Some simulations are taken to investigate the performance of the proposed quantum network and the result show that it can achieve accurate learning. Copyright © 2013 Elsevier Ltd. All rights reserved.

Artificial metaplasticity neural network applied to credit scoring.

PubMed

Marcano-Cedeño, Alexis; Marin-de-la-Barcena, A; Jimenez-Trillo, J; Piñuela, J A; Andina, D

2011-08-01

The assessment of the risk of default on credit is important for financial institutions. Different Artificial Neural Networks (ANN) have been suggested to tackle the credit scoring problem, however, the obtained error rates are often high. In the search for the best ANN algorithm for credit scoring, this paper contributes with the application of an ANN Training Algorithm inspired by the neurons' biological property of metaplasticity. This algorithm is especially efficient when few patterns of a class are available, or when information inherent to low probability events is crucial for a successful application, as weight updating is overemphasized in the less frequent activations than in the more frequent ones. Two well-known and readily available such as: Australia and German data sets has been used to test the algorithm. The results obtained by AMMLP shown have been superior to state-of-the-art classification algorithms in credit scoring.

LensFlow: A Convolutional Neural Network in Search of Strong Gravitational Lenses

NASA Astrophysics Data System (ADS)

Pourrahmani, Milad; Nayyeri, Hooshang; Cooray, Asantha

2018-03-01

In this work, we present our machine learning classification algorithm for identifying strong gravitational lenses from wide-area surveys using convolutional neural networks; LENSFLOW. We train and test the algorithm using a wide variety of strong gravitational lens configurations from simulations of lensing events. Images are processed through multiple convolutional layers that extract feature maps necessary to assign a lens probability to each image. LENSFLOW provides a ranking scheme for all sources that could be used to identify potential gravitational lens candidates by significantly reducing the number of images that have to be visually inspected. We apply our algorithm to the HST/ACS i-band observations of the COSMOS field and present our sample of identified lensing candidates. The developed machine learning algorithm is more computationally efficient and complimentary to classical lens identification algorithms and is ideal for discovering such events across wide areas from current and future surveys such as LSST and WFIRST.

Automatic speech recognition using a predictive echo state network classifier.

PubMed

Skowronski, Mark D; Harris, John G

2007-04-01

We have combined an echo state network (ESN) with a competitive state machine framework to create a classification engine called the predictive ESN classifier. We derive the expressions for training the predictive ESN classifier and show that the model was significantly more noise robust compared to a hidden Markov model in noisy speech classification experiments by 8+/-1 dB signal-to-noise ratio. The simple training algorithm and noise robustness of the predictive ESN classifier make it an attractive classification engine for automatic speech recognition.

Sequential Nonlinear Learning for Distributed Multiagent Systems via Extreme Learning Machines.

PubMed

Vanli, Nuri Denizcan; Sayin, Muhammed O; Delibalta, Ibrahim; Kozat, Suleyman Serdar

2017-03-01

We study online nonlinear learning over distributed multiagent systems, where each agent employs a single hidden layer feedforward neural network (SLFN) structure to sequentially minimize arbitrary loss functions. In particular, each agent trains its own SLFN using only the data that is revealed to itself. On the other hand, the aim of the multiagent system is to train the SLFN at each agent as well as the optimal centralized batch SLFN that has access to all the data, by exchanging information between neighboring agents. We address this problem by introducing a distributed subgradient-based extreme learning machine algorithm. The proposed algorithm provides guaranteed upper bounds on the performance of the SLFN at each agent and shows that each of these individual SLFNs asymptotically achieves the performance of the optimal centralized batch SLFN. Our performance guarantees explicitly distinguish the effects of data- and network-dependent parameters on the convergence rate of the proposed algorithm. The experimental results illustrate that the proposed algorithm achieves the oracle performance significantly faster than the state-of-the-art methods in the machine learning and signal processing literature. Hence, the proposed method is highly appealing for the applications involving big data.

Early driver fatigue detection from electroencephalography signals using artificial neural networks.

PubMed

King, L M; Nguyen, H T; Lal, S K L

2006-01-01

This paper describes a driver fatigue detection system using an artificial neural network (ANN). Using electroencephalogram (EEG) data sampled from 20 professional truck drivers and 35 non professional drivers, the time domain data are processed into alpha, beta, delta and theta bands and then presented to the neural network to detect the onset of driver fatigue. The neural network uses a training optimization technique called the magnified gradient function (MGF). This technique reduces the time required for training by modifying the standard back propagation (SBP) algorithm. The MGF is shown to classify professional driver fatigue with 81.49% accuracy (80.53% sensitivity, 82.44% specificity) and non-professional driver fatigue with 83.06% accuracy (84.04% sensitivity and 82.08% specificity).

Event-driven contrastive divergence for spiking neuromorphic systems.

PubMed

Neftci, Emre; Das, Srinjoy; Pedroni, Bruno; Kreutz-Delgado, Kenneth; Cauwenberghs, Gert

2013-01-01

Restricted Boltzmann Machines (RBMs) and Deep Belief Networks have been demonstrated to perform efficiently in a variety of applications, such as dimensionality reduction, feature learning, and classification. Their implementation on neuromorphic hardware platforms emulating large-scale networks of spiking neurons can have significant advantages from the perspectives of scalability, power dissipation and real-time interfacing with the environment. However, the traditional RBM architecture and the commonly used training algorithm known as Contrastive Divergence (CD) are based on discrete updates and exact arithmetics which do not directly map onto a dynamical neural substrate. Here, we present an event-driven variation of CD to train a RBM constructed with Integrate & Fire (I&F) neurons, that is constrained by the limitations of existing and near future neuromorphic hardware platforms. Our strategy is based on neural sampling, which allows us to synthesize a spiking neural network that samples from a target Boltzmann distribution. The recurrent activity of the network replaces the discrete steps of the CD algorithm, while Spike Time Dependent Plasticity (STDP) carries out the weight updates in an online, asynchronous fashion. We demonstrate our approach by training an RBM composed of leaky I&F neurons with STDP synapses to learn a generative model of the MNIST hand-written digit dataset, and by testing it in recognition, generation and cue integration tasks. Our results contribute to a machine learning-driven approach for synthesizing networks of spiking neurons capable of carrying out practical, high-level functionality.

Event-driven contrastive divergence for spiking neuromorphic systems

PubMed Central

Neftci, Emre; Das, Srinjoy; Pedroni, Bruno; Kreutz-Delgado, Kenneth; Cauwenberghs, Gert

2014-01-01

Restricted Boltzmann Machines (RBMs) and Deep Belief Networks have been demonstrated to perform efficiently in a variety of applications, such as dimensionality reduction, feature learning, and classification. Their implementation on neuromorphic hardware platforms emulating large-scale networks of spiking neurons can have significant advantages from the perspectives of scalability, power dissipation and real-time interfacing with the environment. However, the traditional RBM architecture and the commonly used training algorithm known as Contrastive Divergence (CD) are based on discrete updates and exact arithmetics which do not directly map onto a dynamical neural substrate. Here, we present an event-driven variation of CD to train a RBM constructed with Integrate & Fire (I&F) neurons, that is constrained by the limitations of existing and near future neuromorphic hardware platforms. Our strategy is based on neural sampling, which allows us to synthesize a spiking neural network that samples from a target Boltzmann distribution. The recurrent activity of the network replaces the discrete steps of the CD algorithm, while Spike Time Dependent Plasticity (STDP) carries out the weight updates in an online, asynchronous fashion. We demonstrate our approach by training an RBM composed of leaky I&F neurons with STDP synapses to learn a generative model of the MNIST hand-written digit dataset, and by testing it in recognition, generation and cue integration tasks. Our results contribute to a machine learning-driven approach for synthesizing networks of spiking neurons capable of carrying out practical, high-level functionality. PMID:24574952

Improved artificial neural networks in prediction of malignancy of lesions in contrast-enhanced MR-mammography.

PubMed

Vomweg, T W; Buscema, M; Kauczor, H U; Teifke, A; Intraligi, M; Terzi, S; Heussel, C P; Achenbach, T; Rieker, O; Mayer, D; Thelen, M

2003-09-01

The aim of this study was to evaluate the capability of improved artificial neural networks (ANN) and additional novel training methods in distinguishing between benign and malignant breast lesions in contrast-enhanced magnetic resonance-mammography (MRM). A total of 604 histologically proven cases of contrast-enhanced lesions of the female breast at MRI were analyzed. Morphological, dynamic and clinical parameters were collected and stored in a database. The data set was divided into several groups using random or experimental methods [Training & Testing (T&T) algorithm] to train and test different ANNs. An additional novel computer program for input variable selection was applied. Sensitivity and specificity were calculated and compared with a statistical method and an expert radiologist. After optimization of the distribution of cases among the training and testing sets by the T & T algorithm and the reduction of input variables by the Input Selection procedure a highly sophisticated ANN achieved a sensitivity of 93.6% and a specificity of 91.9% in predicting malignancy of lesions within an independent prediction sample set. The best statistical method reached a sensitivity of 90.5% and a specificity of 68.9%. An expert radiologist performed better than the statistical method but worse than the ANN (sensitivity 92.1%, specificity 85.6%). Features extracted out of dynamic contrast-enhanced MRM and additional clinical data can be successfully analyzed by advanced ANNs. The quality of the resulting network strongly depends on the training methods, which are improved by the use of novel training tools. The best results of an improved ANN outperform expert radiologists.

Interconnected growing self-organizing maps for auditory and semantic acquisition modeling.

PubMed

Cao, Mengxue; Li, Aijun; Fang, Qiang; Kaufmann, Emily; Kröger, Bernd J

2014-01-01

Based on the incremental nature of knowledge acquisition, in this study we propose a growing self-organizing neural network approach for modeling the acquisition of auditory and semantic categories. We introduce an Interconnected Growing Self-Organizing Maps (I-GSOM) algorithm, which takes associations between auditory information and semantic information into consideration, in this paper. Direct phonetic-semantic association is simulated in order to model the language acquisition in early phases, such as the babbling and imitation stages, in which no phonological representations exist. Based on the I-GSOM algorithm, we conducted experiments using paired acoustic and semantic training data. We use a cyclical reinforcing and reviewing training procedure to model the teaching and learning process between children and their communication partners. A reinforcing-by-link training procedure and a link-forgetting procedure are introduced to model the acquisition of associative relations between auditory and semantic information. Experimental results indicate that (1) I-GSOM has good ability to learn auditory and semantic categories presented within the training data; (2) clear auditory and semantic boundaries can be found in the network representation; (3) cyclical reinforcing and reviewing training leads to a detailed categorization as well as to a detailed clustering, while keeping the clusters that have already been learned and the network structure that has already been developed stable; and (4) reinforcing-by-link training leads to well-perceived auditory-semantic associations. Our I-GSOM model suggests that it is important to associate auditory information with semantic information during language acquisition. Despite its high level of abstraction, our I-GSOM approach can be interpreted as a biologically-inspired neurocomputational model.

Scaling Deep Learning on GPU and Knights Landing clusters

DOE PAGES

You, Yang; Buluc, Aydin; Demmel, James

2017-09-26

The speed of deep neural networks training has become a big bottleneck of deep learning research and development. For example, training GoogleNet by ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training process, the current deep learning systems heavily rely on the hardware accelerators. However, these accelerators have limited on-chip memory compared with CPUs. To handle large datasets, they need to fetch data from either CPU memory or remote processors. We use both self-hosted Intel Knights Landing (KNL) clusters and multi-GPU clusters as our target platforms. From an algorithm aspect, current distributed machine learningmore » systems are mainly designed for cloud systems. These methods are asynchronous because of the slow network and high fault-tolerance requirement on cloud systems. We focus on Elastic Averaging SGD (EASGD) to design algorithms for HPC clusters. Original EASGD used round-robin method for communication and updating. The communication is ordered by the machine rank ID, which is inefficient on HPC clusters. First, we redesign four efficient algorithms for HPC systems to improve EASGD's poor scaling on clusters. Async EASGD, Async MEASGD, and Hogwild EASGD are faster \\textcolor{black}{than} their existing counterparts (Async SGD, Async MSGD, and Hogwild SGD, resp.) in all the comparisons. Finally, we design Sync EASGD, which ties for the best performance among all the methods while being deterministic. In addition to the algorithmic improvements, we use some system-algorithm codesign techniques to scale up the algorithms. By reducing the percentage of communication from 87% to 14%, our Sync EASGD achieves 5.3x speedup over original EASGD on the same platform. We get 91.5% weak scaling efficiency on 4253 KNL cores, which is higher than the state-of-the-art implementation.« less

Scaling Deep Learning on GPU and Knights Landing clusters

DOE Office of Scientific and Technical Information (OSTI.GOV)

You, Yang; Buluc, Aydin; Demmel, James

The speed of deep neural networks training has become a big bottleneck of deep learning research and development. For example, training GoogleNet by ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training process, the current deep learning systems heavily rely on the hardware accelerators. However, these accelerators have limited on-chip memory compared with CPUs. To handle large datasets, they need to fetch data from either CPU memory or remote processors. We use both self-hosted Intel Knights Landing (KNL) clusters and multi-GPU clusters as our target platforms. From an algorithm aspect, current distributed machine learningmore » systems are mainly designed for cloud systems. These methods are asynchronous because of the slow network and high fault-tolerance requirement on cloud systems. We focus on Elastic Averaging SGD (EASGD) to design algorithms for HPC clusters. Original EASGD used round-robin method for communication and updating. The communication is ordered by the machine rank ID, which is inefficient on HPC clusters. First, we redesign four efficient algorithms for HPC systems to improve EASGD's poor scaling on clusters. Async EASGD, Async MEASGD, and Hogwild EASGD are faster \\textcolor{black}{than} their existing counterparts (Async SGD, Async MSGD, and Hogwild SGD, resp.) in all the comparisons. Finally, we design Sync EASGD, which ties for the best performance among all the methods while being deterministic. In addition to the algorithmic improvements, we use some system-algorithm codesign techniques to scale up the algorithms. By reducing the percentage of communication from 87% to 14%, our Sync EASGD achieves 5.3x speedup over original EASGD on the same platform. We get 91.5% weak scaling efficiency on 4253 KNL cores, which is higher than the state-of-the-art implementation.« less

DOE Office of Scientific and Technical Information (OSTI.GOV)

Hamadneh, Nawaf; Sathasivam, Saratha; Choon, Ong Hong

Logic programming is the process that leads from an original formulation of a computing problem to executable programs. A normal logic program consists of a finite set of clauses. A valuation I of logic programming is a mapping from ground atoms to false or true. The single step operator of any logic programming is defined as a function (T{sub p}:I→I). Logic programming is well-suited to building the artificial intelligence systems. In this study, we established a new technique to compute the single step operators of logic programming in the radial basis function neural networks. To do that, we proposed amore » new technique to generate the training data sets of single step operators. The training data sets are used to build the neural networks. We used the recurrent radial basis function neural networks to get to the steady state (the fixed point of the operators). To improve the performance of the neural networks, we used the particle swarm optimization algorithm to train the networks.« less

Neural model of gene regulatory network: a survey on supportive meta-heuristics.

PubMed

Biswas, Surama; Acharyya, Sriyankar

2016-06-01

Gene regulatory network (GRN) is produced as a result of regulatory interactions between different genes through their coded proteins in cellular context. Having immense importance in disease detection and drug finding, GRN has been modelled through various mathematical and computational schemes and reported in survey articles. Neural and neuro-fuzzy models have been the focus of attraction in bioinformatics. Predominant use of meta-heuristic algorithms in training neural models has proved its excellence. Considering these facts, this paper is organized to survey neural modelling schemes of GRN and the efficacy of meta-heuristic algorithms towards parameter learning (i.e. weighting connections) within the model. This survey paper renders two different structure-related approaches to infer GRN which are global structure approach and substructure approach. It also describes two neural modelling schemes, such as artificial neural network/recurrent neural network based modelling and neuro-fuzzy modelling. The meta-heuristic algorithms applied so far to learn the structure and parameters of neutrally modelled GRN have been reviewed here.

Real-time Adaptive Control Using Neural Generalized Predictive Control

NASA Technical Reports Server (NTRS)

Haley, Pam; Soloway, Don; Gold, Brian

1999-01-01

The objective of this paper is to demonstrate the feasibility of a Nonlinear Generalized Predictive Control algorithm by showing real-time adaptive control on a plant with relatively fast time-constants. Generalized Predictive Control has classically been used in process control where linear control laws were formulated for plants with relatively slow time-constants. The plant of interest for this paper is a magnetic levitation device that is nonlinear and open-loop unstable. In this application, the reference model of the plant is a neural network that has an embedded nominal linear model in the network weights. The control based on the linear model provides initial stability at the beginning of network training. In using a neural network the control laws are nonlinear and online adaptation of the model is possible to capture unmodeled or time-varying dynamics. Newton-Raphson is the minimization algorithm. Newton-Raphson requires the calculation of the Hessian, but even with this computational expense the low iteration rate make this a viable algorithm for real-time control.

Prediction of thermal conductivity of polyvinylpyrrolidone (PVP) electrospun nanocomposite fibers using artificial neural network and prey-predator algorithm.

PubMed

Khan, Waseem S; Hamadneh, Nawaf N; Khan, Waqar A

2017-01-01

In this study, multilayer perception neural network (MLPNN) was employed to predict thermal conductivity of PVP electrospun nanocomposite fibers with multiwalled carbon nanotubes (MWCNTs) and Nickel Zinc ferrites [(Ni0.6Zn0.4) Fe2O4]. This is the second attempt on the application of MLPNN with prey predator algorithm for the prediction of thermal conductivity of PVP electrospun nanocomposite fibers. The prey predator algorithm was used to train the neural networks to find the best models. The best models have the minimal of sum squared error between the experimental testing data and the corresponding models results. The minimal error was found to be 0.0028 for MWCNTs model and 0.00199 for Ni-Zn ferrites model. The predicted artificial neural networks (ANNs) responses were analyzed statistically using z-test, correlation coefficient, and the error functions for both inclusions. The predicted ANN responses for PVP electrospun nanocomposite fibers were compared with the experimental data and were found in good agreement.

Development of an Efficient Identifier for Nuclear Power Plant Transients Based on Latest Advances of Error Back-Propagation Learning Algorithm

NASA Astrophysics Data System (ADS)

Moshkbar-Bakhshayesh, Khalil; Ghofrani, Mohammad B.

2014-02-01

This study aims to improve the performance of nuclear power plants (NPPs) transients training and identification using the latest advances of error back-propagation (EBP) learning algorithm. To this end, elements of EBP, including input data, initial weights, learning rate, cost function, activation function, and weights updating procedure are investigated and an efficient neural network is developed. Usefulness of modular networks is also examined and appropriate identifiers, one for each transient, are employed. Furthermore, the effect of transient type on transient identifier performance is illustrated. Subsequently, the developed transient identifier is applied to Bushehr nuclear power plant (BNPP). Seven types of the plant events are probed to analyze the ability of the proposed identifier. The results reveal that identification occurs very early with only five plant variables, whilst in the previous studies a larger number of variables (typically 15 to 20) were required. Modular networks facilitated identification due to its sole dependency on the sign of each network output signal. Fast training of input patterns, extendibility for identification of more transients and reduction of false identification are other advantageous of the proposed identifier. Finally, the balance between the correct answer to the trained transients (memorization) and reasonable response to the test transients (generalization) is improved, meeting one of the primary design criteria of identifiers.

Next Day Building Load Predictions based on Limited Input Features Using an On-Line Laterally Primed Adaptive Resonance Theory Artificial Neural Network.

DOE Office of Scientific and Technical Information (OSTI.GOV)

Jones, Christian Birk; Robinson, Matt; Yasaei, Yasser

Optimal integration of thermal energy storage within commercial building applications requires accurate load predictions. Several methods exist that provide an estimate of a buildings future needs. Methods include component-based models and data-driven algorithms. This work implemented a previously untested algorithm for this application that is called a Laterally Primed Adaptive Resonance Theory (LAPART) artificial neural network (ANN). The LAPART algorithm provided accurate results over a two month period where minimal historical data and a small amount of input types were available. These results are significant, because common practice has often overlooked the implementation of an ANN. ANN have often beenmore » perceived to be too complex and require large amounts of data to provide accurate results. The LAPART neural network was implemented in an on-line learning manner. On-line learning refers to the continuous updating of training data as time occurs. For this experiment, training began with a singe day and grew to two months of data. This approach provides a platform for immediate implementation that requires minimal time and effort. The results from the LAPART algorithm were compared with statistical regression and a component-based model. The comparison was based on the predictions linear relationship with the measured data, mean squared error, mean bias error, and cost savings achieved by the respective prediction techniques. The results show that the LAPART algorithm provided a reliable and cost effective means to predict the building load for the next day.« less

Application of a neural network for reflectance spectrum classification

NASA Astrophysics Data System (ADS)

Yang, Gefei; Gartley, Michael

2017-05-01

Traditional reflectance spectrum classification algorithms are based on comparing spectrum across the electromagnetic spectrum anywhere from the ultra-violet to the thermal infrared regions. These methods analyze reflectance on a pixel by pixel basis. Inspired by high performance that Convolution Neural Networks (CNN) have demonstrated in image classification, we applied a neural network to analyze directional reflectance pattern images. By using the bidirectional reflectance distribution function (BRDF) data, we can reformulate the 4-dimensional into 2 dimensions, namely incident direction × reflected direction × channels. Meanwhile, RIT's micro-DIRSIG model is utilized to simulate additional training samples for improving the robustness of the neural networks training. Unlike traditional classification by using hand-designed feature extraction with a trainable classifier, neural networks create several layers to learn a feature hierarchy from pixels to classifier and all layers are trained jointly. Hence, the our approach of utilizing the angular features are different to traditional methods utilizing spatial features. Although training processing typically has a large computational cost, simple classifiers work well when subsequently using neural network generated features. Currently, most popular neural networks such as VGG, GoogLeNet and AlexNet are trained based on RGB spatial image data. Our approach aims to build a directional reflectance spectrum based neural network to help us to understand from another perspective. At the end of this paper, we compare the difference among several classifiers and analyze the trade-off among neural networks parameters.

Invariant-feature-based adaptive automatic target recognition in obscured 3D point clouds

NASA Astrophysics Data System (ADS)

Khuon, Timothy; Kershner, Charles; Mattei, Enrico; Alverio, Arnel; Rand, Robert

2014-06-01

Target recognition and classification in a 3D point cloud is a non-trivial process due to the nature of the data collected from a sensor system. The signal can be corrupted by noise from the environment, electronic system, A/D converter, etc. Therefore, an adaptive system with a desired tolerance is required to perform classification and recognition optimally. The feature-based pattern recognition algorithm architecture as described below is particularly devised for solving a single-sensor classification non-parametrically. Feature set is extracted from an input point cloud, normalized, and classifier a neural network classifier. For instance, automatic target recognition in an urban area would require different feature sets from one in a dense foliage area. The figure above (see manuscript) illustrates the architecture of the feature based adaptive signature extraction of 3D point cloud including LIDAR, RADAR, and electro-optical data. This network takes a 3D cluster and classifies it into a specific class. The algorithm is a supervised and adaptive classifier with two modes: the training mode and the performing mode. For the training mode, a number of novel patterns are selected from actual or artificial data. A particular 3D cluster is input to the network as shown above for the decision class output. The network consists of three sequential functional modules. The first module is for feature extraction that extracts the input cluster into a set of singular value features or feature vector. Then the feature vector is input into the feature normalization module to normalize and balance it before being fed to the neural net classifier for the classification. The neural net can be trained by actual or artificial novel data until each trained output reaches the declared output within the defined tolerance. In case new novel data is added after the neural net has been learned, the training is then resumed until the neural net has incrementally learned with the new novel data. The associative memory capability of the neural net enables the incremental learning. The back propagation algorithm or support vector machine can be utilized for the classification and recognition.

Causal Inference and Explaining Away in a Spiking Network

PubMed Central

Moreno-Bote, Rubén; Drugowitsch, Jan

2015-01-01

While the brain uses spiking neurons for communication, theoretical research on brain computations has mostly focused on non-spiking networks. The nature of spike-based algorithms that achieve complex computations, such as object probabilistic inference, is largely unknown. Here we demonstrate that a family of high-dimensional quadratic optimization problems with non-negativity constraints can be solved exactly and efficiently by a network of spiking neurons. The network naturally imposes the non-negativity of causal contributions that is fundamental to causal inference, and uses simple operations, such as linear synapses with realistic time constants, and neural spike generation and reset non-linearities. The network infers the set of most likely causes from an observation using explaining away, which is dynamically implemented by spike-based, tuned inhibition. The algorithm performs remarkably well even when the network intrinsically generates variable spike trains, the timing of spikes is scrambled by external sources of noise, or the network is mistuned. This type of network might underlie tasks such as odor identification and classification. PMID:26621426

Causal Inference and Explaining Away in a Spiking Network.

PubMed

Moreno-Bote, Rubén; Drugowitsch, Jan

2015-12-01

While the brain uses spiking neurons for communication, theoretical research on brain computations has mostly focused on non-spiking networks. The nature of spike-based algorithms that achieve complex computations, such as object probabilistic inference, is largely unknown. Here we demonstrate that a family of high-dimensional quadratic optimization problems with non-negativity constraints can be solved exactly and efficiently by a network of spiking neurons. The network naturally imposes the non-negativity of causal contributions that is fundamental to causal inference, and uses simple operations, such as linear synapses with realistic time constants, and neural spike generation and reset non-linearities. The network infers the set of most likely causes from an observation using explaining away, which is dynamically implemented by spike-based, tuned inhibition. The algorithm performs remarkably well even when the network intrinsically generates variable spike trains, the timing of spikes is scrambled by external sources of noise, or the network is mistuned. This type of network might underlie tasks such as odor identification and classification.

Network-based ranking methods for prediction of novel disease associated microRNAs.

PubMed

Le, Duc-Hau

2015-10-01

Many studies have shown roles of microRNAs on human disease and a number of computational methods have been proposed to predict such associations by ranking candidate microRNAs according to their relevance to a disease. Among them, machine learning-based methods usually have a limitation in specifying non-disease microRNAs as negative training samples. Meanwhile, network-based methods are becoming dominant since they well exploit a "disease module" principle in microRNA functional similarity networks. Of which, random walk with restart (RWR) algorithm-based method is currently state-of-the-art. The use of this algorithm was inspired from its success in predicting disease gene because the "disease module" principle also exists in protein interaction networks. Besides, many algorithms designed for webpage ranking have been successfully applied in ranking disease candidate genes because web networks share topological properties with protein interaction networks. However, these algorithms have not yet been utilized for disease microRNA prediction. We constructed microRNA functional similarity networks based on shared targets of microRNAs, and then we integrated them with a microRNA functional synergistic network, which was recently identified. After analyzing topological properties of these networks, in addition to RWR, we assessed the performance of (i) PRINCE (PRIoritizatioN and Complex Elucidation), which was proposed for disease gene prediction; (ii) PageRank with Priors (PRP) and K-Step Markov (KSM), which were used for studying web networks; and (iii) a neighborhood-based algorithm. Analyses on topological properties showed that all microRNA functional similarity networks are small-worldness and scale-free. The performance of each algorithm was assessed based on average AUC values on 35 disease phenotypes and average rankings of newly discovered disease microRNAs. As a result, the performance on the integrated network was better than that on individual ones. In addition, the performance of PRINCE, PRP and KSM was comparable with that of RWR, whereas it was worst for the neighborhood-based algorithm. Moreover, all the algorithms were stable with the change of parameters. Final, using the integrated network, we predicted six novel miRNAs (i.e., hsa-miR-101, hsa-miR-181d, hsa-miR-192, hsa-miR-423-3p, hsa-miR-484 and hsa-miR-98) associated with breast cancer. Network-based ranking algorithms, which were successfully applied for either disease gene prediction or for studying social/web networks, can be also used effectively for disease microRNA prediction. Copyright © 2015 Elsevier Ltd. All rights reserved.

Spiking neuron network Helmholtz machine.

PubMed

Sountsov, Pavel; Miller, Paul

2015-01-01

An increasing amount of behavioral and neurophysiological data suggests that the brain performs optimal (or near-optimal) probabilistic inference and learning during perception and other tasks. Although many machine learning algorithms exist that perform inference and learning in an optimal way, the complete description of how one of those algorithms (or a novel algorithm) can be implemented in the brain is currently incomplete. There have been many proposed solutions that address how neurons can perform optimal inference but the question of how synaptic plasticity can implement optimal learning is rarely addressed. This paper aims to unify the two fields of probabilistic inference and synaptic plasticity by using a neuronal network of realistic model spiking neurons to implement a well-studied computational model called the Helmholtz Machine. The Helmholtz Machine is amenable to neural implementation as the algorithm it uses to learn its parameters, called the wake-sleep algorithm, uses a local delta learning rule. Our spiking-neuron network implements both the delta rule and a small example of a Helmholtz machine. This neuronal network can learn an internal model of continuous-valued training data sets without supervision. The network can also perform inference on the learned internal models. We show how various biophysical features of the neural implementation constrain the parameters of the wake-sleep algorithm, such as the duration of the wake and sleep phases of learning and the minimal sample duration. We examine the deviations from optimal performance and tie them to the properties of the synaptic plasticity rule.

Spiking neuron network Helmholtz machine

PubMed Central

Sountsov, Pavel; Miller, Paul

2015-01-01

An increasing amount of behavioral and neurophysiological data suggests that the brain performs optimal (or near-optimal) probabilistic inference and learning during perception and other tasks. Although many machine learning algorithms exist that perform inference and learning in an optimal way, the complete description of how one of those algorithms (or a novel algorithm) can be implemented in the brain is currently incomplete. There have been many proposed solutions that address how neurons can perform optimal inference but the question of how synaptic plasticity can implement optimal learning is rarely addressed. This paper aims to unify the two fields of probabilistic inference and synaptic plasticity by using a neuronal network of realistic model spiking neurons to implement a well-studied computational model called the Helmholtz Machine. The Helmholtz Machine is amenable to neural implementation as the algorithm it uses to learn its parameters, called the wake-sleep algorithm, uses a local delta learning rule. Our spiking-neuron network implements both the delta rule and a small example of a Helmholtz machine. This neuronal network can learn an internal model of continuous-valued training data sets without supervision. The network can also perform inference on the learned internal models. We show how various biophysical features of the neural implementation constrain the parameters of the wake-sleep algorithm, such as the duration of the wake and sleep phases of learning and the minimal sample duration. We examine the deviations from optimal performance and tie them to the properties of the synaptic plasticity rule. PMID:25954191

Chaotic map clustering algorithm for EEG analysis

NASA Astrophysics Data System (ADS)

Bellotti, R.; De Carlo, F.; Stramaglia, S.

2004-03-01

The non-parametric chaotic map clustering algorithm has been applied to the analysis of electroencephalographic signals, in order to recognize the Huntington's disease, one of the most dangerous pathologies of the central nervous system. The performance of the method has been compared with those obtained through parametric algorithms, as K-means and deterministic annealing, and supervised multi-layer perceptron. While supervised neural networks need a training phase, performed by means of data tagged by the genetic test, and the parametric methods require a prior choice of the number of classes to find, the chaotic map clustering gives a natural evidence of the pathological class, without any training or supervision, thus providing a new efficient methodology for the recognition of patterns affected by the Huntington's disease.

Bidirectional extreme learning machine for regression problem and its learning effectiveness.

PubMed

Yang, Yimin; Wang, Yaonan; Yuan, Xiaofang

2012-09-01

It is clear that the learning effectiveness and learning speed of neural networks are in general far slower than required, which has been a major bottleneck for many applications. Recently, a simple and efficient learning method, referred to as extreme learning machine (ELM), was proposed by Huang , which has shown that, compared to some conventional methods, the training time of neural networks can be reduced by a thousand times. However, one of the open problems in ELM research is whether the number of hidden nodes can be further reduced without affecting learning effectiveness. This brief proposes a new learning algorithm, called bidirectional extreme learning machine (B-ELM), in which some hidden nodes are not randomly selected. In theory, this algorithm tends to reduce network output error to 0 at an extremely early learning stage. Furthermore, we find a relationship between the network output error and the network output weights in the proposed B-ELM. Simulation results demonstrate that the proposed method can be tens to hundreds of times faster than other incremental ELM algorithms.

Classification of multispectral image data by the Binary Diamond neural network and by nonparametric, pixel-by-pixel methods

NASA Technical Reports Server (NTRS)

Salu, Yehuda; Tilton, James

1993-01-01

The classification of multispectral image data obtained from satellites has become an important tool for generating ground cover maps. This study deals with the application of nonparametric pixel-by-pixel classification methods in the classification of pixels, based on their multispectral data. A new neural network, the Binary Diamond, is introduced, and its performance is compared with a nearest neighbor algorithm and a back-propagation network. The Binary Diamond is a multilayer, feed-forward neural network, which learns from examples in unsupervised, 'one-shot' mode. It recruits its neurons according to the actual training set, as it learns. The comparisons of the algorithms were done by using a realistic data base, consisting of approximately 90,000 Landsat 4 Thematic Mapper pixels. The Binary Diamond and the nearest neighbor performances were close, with some advantages to the Binary Diamond. The performance of the back-propagation network lagged behind. An efficient nearest neighbor algorithm, the binned nearest neighbor, is described. Ways for improving the performances, such as merging categories, and analyzing nonboundary pixels, are addressed and evaluated.

A False Alarm Reduction Method for a Gas Sensor Based Electronic Nose

PubMed Central

Rahman, Mohammad Mizanur; Suksompong, Prapun; Toochinda, Pisanu; Taparugssanagorn, Attaphongse

2017-01-01

Electronic noses (E-Noses) are becoming popular for food and fruit quality assessment due to their robustness and repeated usability without fatigue, unlike human experts. An E-Nose equipped with classification algorithms and having open ended classification boundaries such as the k-nearest neighbor (k-NN), support vector machine (SVM), and multilayer perceptron neural network (MLPNN), are found to suffer from false classification errors of irrelevant odor data. To reduce false classification and misclassification errors, and to improve correct rejection performance; algorithms with a hyperspheric boundary, such as a radial basis function neural network (RBFNN) and generalized regression neural network (GRNN) with a Gaussian activation function in the hidden layer should be used. The simulation results presented in this paper show that GRNN has more correct classification efficiency and false alarm reduction capability compared to RBFNN. As the design of a GRNN and RBFNN is complex and expensive due to large numbers of neuron requirements, a simple hyperspheric classification method based on minimum, maximum, and mean (MMM) values of each class of the training dataset was presented. The MMM algorithm was simple and found to be fast and efficient in correctly classifying data of training classes, and correctly rejecting data of extraneous odors, and thereby reduced false alarms. PMID:28895910

A False Alarm Reduction Method for a Gas Sensor Based Electronic Nose.

PubMed

Rahman, Mohammad Mizanur; Charoenlarpnopparut, Chalie; Suksompong, Prapun; Toochinda, Pisanu; Taparugssanagorn, Attaphongse

2017-09-12

Electronic noses (E-Noses) are becoming popular for food and fruit quality assessment due to their robustness and repeated usability without fatigue, unlike human experts. An E-Nose equipped with classification algorithms and having open ended classification boundaries such as the k -nearest neighbor ( k -NN), support vector machine (SVM), and multilayer perceptron neural network (MLPNN), are found to suffer from false classification errors of irrelevant odor data. To reduce false classification and misclassification errors, and to improve correct rejection performance; algorithms with a hyperspheric boundary, such as a radial basis function neural network (RBFNN) and generalized regression neural network (GRNN) with a Gaussian activation function in the hidden layer should be used. The simulation results presented in this paper show that GRNN has more correct classification efficiency and false alarm reduction capability compared to RBFNN. As the design of a GRNN and RBFNN is complex and expensive due to large numbers of neuron requirements, a simple hyperspheric classification method based on minimum, maximum, and mean (MMM) values of each class of the training dataset was presented. The MMM algorithm was simple and found to be fast and efficient in correctly classifying data of training classes, and correctly rejecting data of extraneous odors, and thereby reduced false alarms.

Neural network error correction for solving coupled ordinary differential equations

NASA Technical Reports Server (NTRS)

Shelton, R. O.; Darsey, J. A.; Sumpter, B. G.; Noid, D. W.

1992-01-01

A neural network is presented to learn errors generated by a numerical algorithm for solving coupled nonlinear differential equations. The method is based on using a neural network to correctly learn the error generated by, for example, Runge-Kutta on a model molecular dynamics (MD) problem. The neural network programs used in this study were developed by NASA. Comparisons are made for training the neural network using backpropagation and a new method which was found to converge with fewer iterations. The neural net programs, the MD model and the calculations are discussed.

Application of the SP algorithm to the INTERMAGNET magnetograms of the disturbed geomagnetic field

NASA Astrophysics Data System (ADS)

Sidorov, R. V.; Soloviev, A. A.; Bogoutdinov, Sh. R.

2012-05-01

The algorithmic system developed in the Laboratory of Geoinformatics at the Geophysical Center, Russian Academy of Sciences, which is intended for recognizing spikes on the magnetograms from the global network INTERMAGNET provides the possibility to carry out retrospective analysis of the magnetograms from the World Data Centers. Application of this system to the analysis of the magnetograms allows automating the job of the experts-interpreters on identifying the artificial spikes in the INTERMAGNET data. The present paper is focused on the SP algorithm (abbreviated from SPIKE) which recognizes artificial spikes on the records of the geomagnetic field. Initially, this algorithm was trained on the magnetograms of 2007 and 2008, which recorded the quiet geomagnetic field. The results of training and testing showed that the algorithm is quite efficient. Applying this method to the problem of recognizing spikes on the data for periods of enhanced geomagnetic activity is a separate task. In this short communication, we present the results of applying the SP algorithm trained on the data of 2007 to the INTERMAGNET magnetograms for 2003 and 2005 sampled every minute. This analysis shows that the SP algorithm does not exhibit a worse performance if applied to the records of a disturbed geomagnetic field.

Loop Mirror Laser Neural Network with a Fast Liquid-Crystal Display

NASA Astrophysics Data System (ADS)

Mos, Evert C.; Schleipen, Jean J. H. B.; de Waardt, Huug; Khoe, Djan G. D.

1999-07-01

In our laser neural network (LNN) all-optical threshold action is obtained by application of controlled optical feedback to a laser diode. Here an extended experimental LNN is presented with as many as 32 neurons and 12 inputs. In the setup we use a fast liquid-crystal display to implement an optical matrix vector multiplier. This display, based on ferroelectric liquid-crystal material, enables us to present 125 training examples s to the LNN. To maximize the optical feedback efficiency of the setup, a loop mirror is introduced. We use a -rule learning algorithm to train the network to perform a number of functions toward the application area of telecommunication data switching.

Classification of breast MRI lesions using small-size training sets: comparison of deep learning approaches

NASA Astrophysics Data System (ADS)

Amit, Guy; Ben-Ari, Rami; Hadad, Omer; Monovich, Einat; Granot, Noa; Hashoul, Sharbell

2017-03-01

Diagnostic interpretation of breast MRI studies requires meticulous work and a high level of expertise. Computerized algorithms can assist radiologists by automatically characterizing the detected lesions. Deep learning approaches have shown promising results in natural image classification, but their applicability to medical imaging is limited by the shortage of large annotated training sets. In this work, we address automatic classification of breast MRI lesions using two different deep learning approaches. We propose a novel image representation for dynamic contrast enhanced (DCE) breast MRI lesions, which combines the morphological and kinetics information in a single multi-channel image. We compare two classification approaches for discriminating between benign and malignant lesions: training a designated convolutional neural network and using a pre-trained deep network to extract features for a shallow classifier. The domain-specific trained network provided higher classification accuracy, compared to the pre-trained model, with an area under the ROC curve of 0.91 versus 0.81, and an accuracy of 0.83 versus 0.71. Similar accuracy was achieved in classifying benign lesions, malignant lesions, and normal tissue images. The trained network was able to improve accuracy by using the multi-channel image representation, and was more robust to reductions in the size of the training set. A small-size convolutional neural network can learn to accurately classify findings in medical images using only a few hundred images from a few dozen patients. With sufficient data augmentation, such a network can be trained to outperform a pre-trained out-of-domain classifier. Developing domain-specific deep-learning models for medical imaging can facilitate technological advancements in computer-aided diagnosis.

An Emotional ANN (EANN) approach to modeling rainfall-runoff process

NASA Astrophysics Data System (ADS)

Nourani, Vahid

2017-01-01

This paper presents the first hydrological implementation of Emotional Artificial Neural Network (EANN), as a new generation of Artificial Intelligence-based models for daily rainfall-runoff (r-r) modeling of the watersheds. Inspired by neurophysiological form of brain, in addition to conventional weights and bias, an EANN includes simulated emotional parameters aimed at improving the network learning process. EANN trained by a modified version of back-propagation (BP) algorithm was applied to single and multi-step-ahead runoff forecasting of two watersheds with two distinct climatic conditions. Also to evaluate the ability of EANN trained by smaller training data set, three data division strategies with different number of training samples were considered for the training purpose. The overall comparison of the obtained results of the r-r modeling indicates that the EANN could outperform the conventional feed forward neural network (FFNN) model up to 13% and 34% in terms of training and verification efficiency criteria, respectively. The superiority of EANN over classic ANN is due to its ability to recognize and distinguish dry (rainless days) and wet (rainy days) situations using hormonal parameters of the artificial emotional system.

Semi-supervised anomaly detection - towards model-independent searches of new physics

NASA Astrophysics Data System (ADS)

Kuusela, Mikael; Vatanen, Tommi; Malmi, Eric; Raiko, Tapani; Aaltonen, Timo; Nagai, Yoshikazu

2012-06-01

Most classification algorithms used in high energy physics fall under the category of supervised machine learning. Such methods require a training set containing both signal and background events and are prone to classification errors should this training data be systematically inaccurate for example due to the assumed MC model. To complement such model-dependent searches, we propose an algorithm based on semi-supervised anomaly detection techniques, which does not require a MC training sample for the signal data. We first model the background using a multivariate Gaussian mixture model. We then search for deviations from this model by fitting to the observations a mixture of the background model and a number of additional Gaussians. This allows us to perform pattern recognition of any anomalous excess over the background. We show by a comparison to neural network classifiers that such an approach is a lot more robust against misspecification of the signal MC than supervised classification. In cases where there is an unexpected signal, a neural network might fail to correctly identify it, while anomaly detection does not suffer from such a limitation. On the other hand, when there are no systematic errors in the training data, both methods perform comparably.

Function approximation and documentation of sampling data using artificial neural networks.

PubMed

Zhang, Wenjun; Barrion, Albert

2006-11-01

Biodiversity studies in ecology often begin with the fitting and documentation of sampling data. This study is conducted to make function approximation on sampling data and to document the sampling information using artificial neural network algorithms, based on the invertebrate data sampled in the irrigated rice field. Three types of sampling data, i.e., the curve species richness vs. the sample size, the curve rarefaction, and the curve mean abundance of newly sampled species vs.the sample size, are fitted and documented using BP (Backpropagation) network and RBF (Radial Basis Function) network. As the comparisons, The Arrhenius model, and rarefaction model, and power function are tested for their ability to fit these data. The results show that the BP network and RBF network fit the data better than these models with smaller errors. BP network and RBF network can fit non-linear functions (sampling data) with specified accuracy and don't require mathematical assumptions. In addition to the interpolation, BP network is used to extrapolate the functions and the asymptote of the sampling data can be drawn. BP network cost a longer time to train the network and the results are always less stable compared to the RBF network. RBF network require more neurons to fit functions and generally it may not be used to extrapolate the functions. The mathematical function for sampling data can be exactly fitted using artificial neural network algorithms by adjusting the desired accuracy and maximum iterations. The total numbers of functional species of invertebrates in the tropical irrigated rice field are extrapolated as 140 to 149 using trained BP network, which are similar to the observed richness.

Application of artificial neural networks to the design optimization of aerospace structural components

NASA Technical Reports Server (NTRS)

Berke, Laszlo; Patnaik, Surya N.; Murthy, Pappu L. N.

1993-01-01

The application of artificial neural networks to capture structural design expertise is demonstrated. The principal advantage of a trained neural network is that it requires trivial computational effort to produce an acceptable new design. For the class of problems addressed, the development of a conventional expert system would be extremely difficult. In the present effort, a structural optimization code with multiple nonlinear programming algorithms and an artificial neural network code NETS were used. A set of optimum designs for a ring and two aircraft wings for static and dynamic constraints were generated by using the optimization codes. The optimum design data were processed to obtain input and output pairs, which were used to develop a trained artificial neural network with the code NETS. Optimum designs for new design conditions were predicted by using the trained network. Neural net prediction of optimum designs was found to be satisfactory for most of the output design parameters. However, results from the present study indicate that caution must be exercised to ensure that all design variables are within selected error bounds.

Optimum Design of Aerospace Structural Components Using Neural Networks

NASA Technical Reports Server (NTRS)

Berke, L.; Patnaik, S. N.; Murthy, P. L. N.

1993-01-01

The application of artificial neural networks to capture structural design expertise is demonstrated. The principal advantage of a trained neural network is that it requires a trivial computational effort to produce an acceptable new design. For the class of problems addressed, the development of a conventional expert system would be extremely difficult. In the present effort, a structural optimization code with multiple nonlinear programming algorithms and an artificial neural network code NETS were used. A set of optimum designs for a ring and two aircraft wings for static and dynamic constraints were generated using the optimization codes. The optimum design data were processed to obtain input and output pairs, which were used to develop a trained artificial neural network using the code NETS. Optimum designs for new design conditions were predicted using the trained network. Neural net prediction of optimum designs was found to be satisfactory for the majority of the output design parameters. However, results from the present study indicate that caution must be exercised to ensure that all design variables are within selected error bounds.

Identifying presence of correlated errors in GRACE monthly harmonic coefficients using machine learning algorithms

NASA Astrophysics Data System (ADS)

Piretzidis, Dimitrios; Sra, Gurveer; Karantaidis, George; Sideris, Michael G.

2017-04-01

A new method for identifying correlated errors in Gravity Recovery and Climate Experiment (GRACE) monthly harmonic coefficients has been developed and tested. Correlated errors are present in the differences between monthly GRACE solutions, and can be suppressed using a de-correlation filter. In principle, the de-correlation filter should be implemented only on coefficient series with correlated errors to avoid losing useful geophysical information. In previous studies, two main methods of implementing the de-correlation filter have been utilized. In the first one, the de-correlation filter is implemented starting from a specific minimum order until the maximum order of the monthly solution examined. In the second one, the de-correlation filter is implemented only on specific coefficient series, the selection of which is based on statistical testing. The method proposed in the present study exploits the capabilities of supervised machine learning algorithms such as neural networks and support vector machines (SVMs). The pattern of correlated errors can be described by several numerical and geometric features of the harmonic coefficient series. The features of extreme cases of both correlated and uncorrelated coefficients are extracted and used for the training of the machine learning algorithms. The trained machine learning algorithms are later used to identify correlated errors and provide the probability of a coefficient series to be correlated. Regarding SVMs algorithms, an extensive study is performed with various kernel functions in order to find the optimal training model for prediction. The selection of the optimal training model is based on the classification accuracy of the trained SVM algorithm on the same samples used for training. Results show excellent performance of all algorithms with a classification accuracy of 97% - 100% on a pre-selected set of training samples, both in the validation stage of the training procedure and in the subsequent use of the trained algorithms to classify independent coefficients. This accuracy is also confirmed by the external validation of the trained algorithms using the hydrology model GLDAS NOAH. The proposed method meet the requirement of identifying and de-correlating only coefficients with correlated errors. Also, there is no need of applying statistical testing or other techniques that require prior de-correlation of the harmonic coefficients.

Learning Traffic as Images: A Deep Convolutional Neural Network for Large-Scale Transportation Network Speed Prediction.

PubMed

Ma, Xiaolei; Dai, Zhuang; He, Zhengbing; Ma, Jihui; Wang, Yong; Wang, Yunpeng

2017-04-10

This paper proposes a convolutional neural network (CNN)-based method that learns traffic as images and predicts large-scale, network-wide traffic speed with a high accuracy. Spatiotemporal traffic dynamics are converted to images describing the time and space relations of traffic flow via a two-dimensional time-space matrix. A CNN is applied to the image following two consecutive steps: abstract traffic feature extraction and network-wide traffic speed prediction. The effectiveness of the proposed method is evaluated by taking two real-world transportation networks, the second ring road and north-east transportation network in Beijing, as examples, and comparing the method with four prevailing algorithms, namely, ordinary least squares, k-nearest neighbors, artificial neural network, and random forest, and three deep learning architectures, namely, stacked autoencoder, recurrent neural network, and long-short-term memory network. The results show that the proposed method outperforms other algorithms by an average accuracy improvement of 42.91% within an acceptable execution time. The CNN can train the model in a reasonable time and, thus, is suitable for large-scale transportation networks.

Learning Traffic as Images: A Deep Convolutional Neural Network for Large-Scale Transportation Network Speed Prediction

PubMed Central

Ma, Xiaolei; Dai, Zhuang; He, Zhengbing; Ma, Jihui; Wang, Yong; Wang, Yunpeng

2017-01-01

This paper proposes a convolutional neural network (CNN)-based method that learns traffic as images and predicts large-scale, network-wide traffic speed with a high accuracy. Spatiotemporal traffic dynamics are converted to images describing the time and space relations of traffic flow via a two-dimensional time-space matrix. A CNN is applied to the image following two consecutive steps: abstract traffic feature extraction and network-wide traffic speed prediction. The effectiveness of the proposed method is evaluated by taking two real-world transportation networks, the second ring road and north-east transportation network in Beijing, as examples, and comparing the method with four prevailing algorithms, namely, ordinary least squares, k-nearest neighbors, artificial neural network, and random forest, and three deep learning architectures, namely, stacked autoencoder, recurrent neural network, and long-short-term memory network. The results show that the proposed method outperforms other algorithms by an average accuracy improvement of 42.91% within an acceptable execution time. The CNN can train the model in a reasonable time and, thus, is suitable for large-scale transportation networks. PMID:28394270

Neural network modeling of drying of rice in BAU-STR dryer

NASA Astrophysics Data System (ADS)

Alam, Md. Ashraful; Saha, Chayan Kumer; Alam, Md. Monjurul; Ashraf, Md. Ali; Bala, Bilash Kanti; Harvey, Jagger

2018-05-01

The experimental performance and artificial neural network modeling of rice drying in BAU-STR dryer is presented in this paper. The dryer consists of a biomass stove as a heat source, a perforated inner bin and a perforated outer bin with annular space for grains, and a blower (1 hp) to supply heated air. The dryer capacity was 500 kg of freshly harvested rice. Twenty experimental runs were conducted to investigate the experimental performance of the dryer for drying of rice. An independent multilayer neural network approach was used to predict the performance of the BAU-STR dryer for drying of rice. Ten sets of experimental data were used for training using back propagation algorithm and another ten sets of data were used for testing the artificial neural network model. The prediction of the performance of the dryer was found to be excellent after it was adequately trained. The statistical analysis showed that the errors (MSE and RMSE) were within and acceptable range of ±5% with a coefficient of determination (R2) of 99%. The model can be used to predict the potential of the dryer for different locations, and can also be used in a predictive optimal control algorithm.

Real time selective harmonic minimization for multilevel inverters using genetic algorithm and artifical neural network angle generation

DOE Office of Scientific and Technical Information (OSTI.GOV)

Filho, Faete J; Tolbert, Leon M; Ozpineci, Burak

2012-01-01

The work developed here proposes a methodology for calculating switching angles for varying DC sources in a multilevel cascaded H-bridges converter. In this approach the required fundamental is achieved, the lower harmonics are minimized, and the system can be implemented in real time with low memory requirements. Genetic algorithm (GA) is the stochastic search method to find the solution for the set of equations where the input voltages are the known variables and the switching angles are the unknown variables. With the dataset generated by GA, an artificial neural network (ANN) is trained to store the solutions without excessive memorymore » storage requirements. This trained ANN then senses the voltage of each cell and produces the switching angles in order to regulate the fundamental at 120 V and eliminate or minimize the low order harmonics while operating in real time.« less

Machine learning of network metrics in ATLAS Distributed Data Management

NASA Astrophysics Data System (ADS)

Lassnig, Mario; Toler, Wesley; Vamosi, Ralf; Bogado, Joaquin; ATLAS Collaboration

2017-10-01

The increasing volume of physics data poses a critical challenge to the ATLAS experiment. In anticipation of high luminosity physics, automation of everyday data management tasks has become necessary. Previously many of these tasks required human decision-making and operation. Recent advances in hardware and software have made it possible to entrust more complicated duties to automated systems using models trained by machine learning algorithms. In this contribution we show results from one of our ongoing automation efforts that focuses on network metrics. First, we describe our machine learning framework built atop the ATLAS Analytics Platform. This framework can automatically extract and aggregate data, train models with various machine learning algorithms, and eventually score the resulting models and parameters. Second, we use these models to forecast metrics relevant for networkaware job scheduling and data brokering. We show the characteristics of the data and evaluate the forecasting accuracy of our models.

Wind power prediction based on genetic neural network

NASA Astrophysics Data System (ADS)

Zhang, Suhan

2017-04-01

The scale of grid connected wind farms keeps increasing. To ensure the stability of power system operation, make a reasonable scheduling scheme and improve the competitiveness of wind farm in the electricity generation market, it's important to accurately forecast the short-term wind power. To reduce the influence of the nonlinear relationship between the disturbance factor and the wind power, the improved prediction model based on genetic algorithm and neural network method is established. To overcome the shortcomings of long training time of BP neural network and easy to fall into local minimum and improve the accuracy of the neural network, genetic algorithm is adopted to optimize the parameters and topology of neural network. The historical data is used as input to predict short-term wind power. The effectiveness and feasibility of the method is verified by the actual data of a certain wind farm as an example.

Effectiveness of feature and classifier algorithms in character recognition systems

NASA Astrophysics Data System (ADS)

Wilson, Charles L.

1993-04-01

At the first Census Optical Character Recognition Systems Conference, NIST generated accuracy data for more than character recognition systems. Most systems were tested on the recognition of isolated digits and upper and lower case alphabetic characters. The recognition experiments were performed on sample sizes of 58,000 digits, and 12,000 upper and lower case alphabetic characters. The algorithms used by the 26 conference participants included rule-based methods, image-based methods, statistical methods, and neural networks. The neural network methods included Multi-Layer Perceptron's, Learned Vector Quantitization, Neocognitrons, and cascaded neural networks. In this paper 11 different systems are compared using correlations between the answers of different systems, comparing the decrease in error rate as a function of confidence of recognition, and comparing the writer dependence of recognition. This comparison shows that methods that used different algorithms for feature extraction and recognition performed with very high levels of correlation. This is true for neural network systems, hybrid systems, and statistically based systems, and leads to the conclusion that neural networks have not yet demonstrated a clear superiority to more conventional statistical methods. Comparison of these results with the models of Vapnick (for estimation problems), MacKay (for Bayesian statistical models), Moody (for effective parameterization), and Boltzmann models (for information content) demonstrate that as the limits of training data variance are approached, all classifier systems have similar statistical properties. The limiting condition can only be approached for sufficiently rich feature sets because the accuracy limit is controlled by the available information content of the training set, which must pass through the feature extraction process prior to classification.

Mirrored STDP Implements Autoencoder Learning in a Network of Spiking Neurons.

PubMed

Burbank, Kendra S

2015-12-01

The autoencoder algorithm is a simple but powerful unsupervised method for training neural networks. Autoencoder networks can learn sparse distributed codes similar to those seen in cortical sensory areas such as visual area V1, but they can also be stacked to learn increasingly abstract representations. Several computational neuroscience models of sensory areas, including Olshausen & Field's Sparse Coding algorithm, can be seen as autoencoder variants, and autoencoders have seen extensive use in the machine learning community. Despite their power and versatility, autoencoders have been difficult to implement in a biologically realistic fashion. The challenges include their need to calculate differences between two neuronal activities and their requirement for learning rules which lead to identical changes at feedforward and feedback connections. Here, we study a biologically realistic network of integrate-and-fire neurons with anatomical connectivity and synaptic plasticity that closely matches that observed in cortical sensory areas. Our choice of synaptic plasticity rules is inspired by recent experimental and theoretical results suggesting that learning at feedback connections may have a different form from learning at feedforward connections, and our results depend critically on this novel choice of plasticity rules. Specifically, we propose that plasticity rules at feedforward versus feedback connections are temporally opposed versions of spike-timing dependent plasticity (STDP), leading to a symmetric combined rule we call Mirrored STDP (mSTDP). We show that with mSTDP, our network follows a learning rule that approximately minimizes an autoencoder loss function. When trained with whitened natural image patches, the learned synaptic weights resemble the receptive fields seen in V1. Our results use realistic synaptic plasticity rules to show that the powerful autoencoder learning algorithm could be within the reach of real biological networks.

Mirrored STDP Implements Autoencoder Learning in a Network of Spiking Neurons

PubMed Central

Burbank, Kendra S.

2015-01-01

The autoencoder algorithm is a simple but powerful unsupervised method for training neural networks. Autoencoder networks can learn sparse distributed codes similar to those seen in cortical sensory areas such as visual area V1, but they can also be stacked to learn increasingly abstract representations. Several computational neuroscience models of sensory areas, including Olshausen & Field’s Sparse Coding algorithm, can be seen as autoencoder variants, and autoencoders have seen extensive use in the machine learning community. Despite their power and versatility, autoencoders have been difficult to implement in a biologically realistic fashion. The challenges include their need to calculate differences between two neuronal activities and their requirement for learning rules which lead to identical changes at feedforward and feedback connections. Here, we study a biologically realistic network of integrate-and-fire neurons with anatomical connectivity and synaptic plasticity that closely matches that observed in cortical sensory areas. Our choice of synaptic plasticity rules is inspired by recent experimental and theoretical results suggesting that learning at feedback connections may have a different form from learning at feedforward connections, and our results depend critically on this novel choice of plasticity rules. Specifically, we propose that plasticity rules at feedforward versus feedback connections are temporally opposed versions of spike-timing dependent plasticity (STDP), leading to a symmetric combined rule we call Mirrored STDP (mSTDP). We show that with mSTDP, our network follows a learning rule that approximately minimizes an autoencoder loss function. When trained with whitened natural image patches, the learned synaptic weights resemble the receptive fields seen in V1. Our results use realistic synaptic plasticity rules to show that the powerful autoencoder learning algorithm could be within the reach of real biological networks. PMID:26633645

Maximum entropy methods for extracting the learned features of deep neural networks.

PubMed

Finnegan, Alex; Song, Jun S

2017-10-01

New architectures of multilayer artificial neural networks and new methods for training them are rapidly revolutionizing the application of machine learning in diverse fields, including business, social science, physical sciences, and biology. Interpreting deep neural networks, however, currently remains elusive, and a critical challenge lies in understanding which meaningful features a network is actually learning. We present a general method for interpreting deep neural networks and extracting network-learned features from input data. We describe our algorithm in the context of biological sequence analysis. Our approach, based on ideas from statistical physics, samples from the maximum entropy distribution over possible sequences, anchored at an input sequence and subject to constraints implied by the empirical function learned by a network. Using our framework, we demonstrate that local transcription factor binding motifs can be identified from a network trained on ChIP-seq data and that nucleosome positioning signals are indeed learned by a network trained on chemical cleavage nucleosome maps. Imposing a further constraint on the maximum entropy distribution also allows us to probe whether a network is learning global sequence features, such as the high GC content in nucleosome-rich regions. This work thus provides valuable mathematical tools for interpreting and extracting learned features from feed-forward neural networks.

Neural Network Based Intrusion Detection System for Critical Infrastructures

DOE Office of Scientific and Technical Information (OSTI.GOV)

Todd Vollmer; Ondrej Linda; Milos Manic

2009-07-01

Resiliency and security in control systems such as SCADA and Nuclear plant’s in today’s world of hackers and malware are a relevant concern. Computer systems used within critical infrastructures to control physical functions are not immune to the threat of cyber attacks and may be potentially vulnerable. Tailoring an intrusion detection system to the specifics of critical infrastructures can significantly improve the security of such systems. The IDS-NNM – Intrusion Detection System using Neural Network based Modeling, is presented in this paper. The main contributions of this work are: 1) the use and analyses of real network data (data recordedmore » from an existing critical infrastructure); 2) the development of a specific window based feature extraction technique; 3) the construction of training dataset using randomly generated intrusion vectors; 4) the use of a combination of two neural network learning algorithms – the Error-Back Propagation and Levenberg-Marquardt, for normal behavior modeling. The presented algorithm was evaluated on previously unseen network data. The IDS-NNM algorithm proved to be capable of capturing all intrusion attempts presented in the network communication while not generating any false alerts.« less

Automatic classification of schizophrenia using resting-state functional language network via an adaptive learning algorithm

NASA Astrophysics Data System (ADS)

Zhu, Maohu; Jie, Nanfeng; Jiang, Tianzi

2014-03-01

A reliable and precise classification of schizophrenia is significant for its diagnosis and treatment of schizophrenia. Functional magnetic resonance imaging (fMRI) is a novel tool increasingly used in schizophrenia research. Recent advances in statistical learning theory have led to applying pattern classification algorithms to access the diagnostic value of functional brain networks, discovered from resting state fMRI data. The aim of this study was to propose an adaptive learning algorithm to distinguish schizophrenia patients from normal controls using resting-state functional language network. Furthermore, here the classification of schizophrenia was regarded as a sample selection problem where a sparse subset of samples was chosen from the labeled training set. Using these selected samples, which we call informative vectors, a classifier for the clinic diagnosis of schizophrenia was established. We experimentally demonstrated that the proposed algorithm incorporating resting-state functional language network achieved 83.6% leaveone- out accuracy on resting-state fMRI data of 27 schizophrenia patients and 28 normal controls. In contrast with KNearest- Neighbor (KNN), Support Vector Machine (SVM) and l1-norm, our method yielded better classification performance. Moreover, our results suggested that a dysfunction of resting-state functional language network plays an important role in the clinic diagnosis of schizophrenia.

Traffic sign recognition by color segmentation and neural network

NASA Astrophysics Data System (ADS)

Surinwarangkoon, Thongchai; Nitsuwat, Supot; Moore, Elvin J.

2011-12-01

An algorithm is proposed for traffic sign detection and identification based on color filtering, color segmentation and neural networks. Traffic signs in Thailand are classified by color into four types: namely, prohibitory signs (red or blue), general warning signs (yellow) and construction area warning signs (amber). A color filtering method is first used to detect traffic signs and classify them by type. Then color segmentation methods adapted for each color type are used to extract inner features, e.g., arrows, bars etc. Finally, neural networks trained to recognize signs in each color type are used to identify any given traffic sign. Experiments show that the algorithm can improve the accuracy of traffic sign detection and recognition for the traffic signs used in Thailand.

Finger language recognition based on ensemble artificial neural network learning using armband EMG sensors.

PubMed

Kim, Seongjung; Kim, Jongman; Ahn, Soonjae; Kim, Youngho

2018-04-18

Deaf people use sign or finger languages for communication, but these methods of communication are very specialized. For this reason, the deaf can suffer from social inequalities and financial losses due to their communication restrictions. In this study, we developed a finger language recognition algorithm based on an ensemble artificial neural network (E-ANN) using an armband system with 8-channel electromyography (EMG) sensors. The developed algorithm was composed of signal acquisition, filtering, segmentation, feature extraction and an E-ANN based classifier that was evaluated with the Korean finger language (14 consonants, 17 vowels and 7 numbers) in 17 subjects. E-ANN was categorized according to the number of classifiers (1 to 10) and size of training data (50 to 1500). The accuracy of the E-ANN-based classifier was obtained by 5-fold cross validation and compared with an artificial neural network (ANN)-based classifier. As the number of classifiers (1 to 8) and size of training data (50 to 300) increased, the average accuracy of the E-ANN-based classifier increased and the standard deviation decreased. The optimal E-ANN was composed with eight classifiers and 300 size of training data, and the accuracy of the E-ANN was significantly higher than that of the general ANN.

False alarm reduction by the And-ing of multiple multivariate Gaussian classifiers

NASA Astrophysics Data System (ADS)

Dobeck, Gerald J.; Cobb, J. Tory

2003-09-01

The high-resolution sonar is one of the principal sensors used by the Navy to detect and classify sea mines in minehunting operations. For such sonar systems, substantial effort has been devoted to the development of automated detection and classification (D/C) algorithms. These have been spurred by several factors including (1) aids for operators to reduce work overload, (2) more optimal use of all available data, and (3) the introduction of unmanned minehunting systems. The environments where sea mines are typically laid (harbor areas, shipping lanes, and the littorals) give rise to many false alarms caused by natural, biologic, and man-made clutter. The objective of the automated D/C algorithms is to eliminate most of these false alarms while still maintaining a very high probability of mine detection and classification (PdPc). In recent years, the benefits of fusing the outputs of multiple D/C algorithms have been studied. We refer to this as Algorithm Fusion. The results have been remarkable, including reliable robustness to new environments. This paper describes a method for training several multivariate Gaussian classifiers such that their And-ing dramatically reduces false alarms while maintaining a high probability of classification. This training approach is referred to as the Focused- Training method. This work extends our 2001-2002 work where the Focused-Training method was used with three other types of classifiers: the Attractor-based K-Nearest Neighbor Neural Network (a type of radial-basis, probabilistic neural network), the Optimal Discrimination Filter Classifier (based linear discrimination theory), and the Quadratic Penalty Function Support Vector Machine (QPFSVM). Although our experience has been gained in the area of sea mine detection and classification, the principles described herein are general and can be applied to a wide range of pattern recognition and automatic target recognition (ATR) problems.

Application of artificial neural networks to identify equilibration in computer simulations

NASA Astrophysics Data System (ADS)

Leibowitz, Mitchell H.; Miller, Evan D.; Henry, Michael M.; Jankowski, Eric

2017-11-01

Determining which microstates generated by a thermodynamic simulation are representative of the ensemble for which sampling is desired is a ubiquitous, underspecified problem. Artificial neural networks are one type of machine learning algorithm that can provide a reproducible way to apply pattern recognition heuristics to underspecified problems. Here we use the open-source TensorFlow machine learning library and apply it to the problem of identifying which hypothetical observation sequences from a computer simulation are “equilibrated” and which are not. We generate training populations and test populations of observation sequences with embedded linear and exponential correlations. We train a two-neuron artificial network to distinguish the correlated and uncorrelated sequences. We find that this simple network is good enough for > 98% accuracy in identifying exponentially-decaying energy trajectories from molecular simulations.

A continually online-trained neural network controller for brushless DC motor drives

DOE Office of Scientific and Technical Information (OSTI.GOV)

Rubaai, A.; Kotaru, R.; Kankam, M.D.

2000-04-01

In this paper, a high-performance controller with simultaneous online identification and control is designed for brushless dc motor drives. The dynamics of the motor/load are modeled online, and controlled using two different neural network based identification and control schemes, as the system is in operation. In the first scheme, an attempt is made to control the rotor angular speed, utilizing a single three-hidden-layer network. The second scheme attempts to control the stator currents, using a predetermined control law as a function of the estimated states. This schemes incorporates three multilayered feedforward neural networks that are online trained, using the Levenburg-Marquadtmore » training algorithm. The control of the direct and quadrature components of the stator current successfully tracked a wide variety of trajectories after relatively short online training periods. The control strategy adapts to the uncertainties of the motor/load dynamics and, in addition, learns their inherent nonlinearities. Simulation results illustrated that a neurocontroller used in conjunction with adaptive control schemes can result in a flexible control device which may be utilized in a wide range of environments.« less

Analysis of Artificial Neural Network Backpropagation Using Conjugate Gradient Fletcher Reeves In The Predicting Process

NASA Astrophysics Data System (ADS)

Wanto, Anjar; Zarlis, Muhammad; Sawaluddin; Hartama, Dedy

2017-12-01

Backpropagation is a good artificial neural network algorithm used to predict, one of which is to predict the rate of Consumer Price Index (CPI) based on the foodstuff sector. While conjugate gradient fletcher reeves is a suitable optimization method when juxtaposed with backpropagation method, because this method can shorten iteration without reducing the quality of training and testing result. Consumer Price Index (CPI) data that will be predicted to come from the Central Statistics Agency (BPS) Pematangsiantar. The results of this study will be expected to contribute to the government in making policies to improve economic growth. In this study, the data obtained will be processed by conducting training and testing with artificial neural network backpropagation by using parameter learning rate 0,01 and target error minimum that is 0.001-0,09. The training network is built with binary and bipolar sigmoid activation functions. After the results with backpropagation are obtained, it will then be optimized using the conjugate gradient fletcher reeves method by conducting the same training and testing based on 5 predefined network architectures. The result, the method used can increase the speed and accuracy result.

A Software Package for Neural Network Applications Development

NASA Technical Reports Server (NTRS)

Baran, Robert H.

1993-01-01

Original Backprop (Version 1.2) is an MS-DOS package of four stand-alone C-language programs that enable users to develop neural network solutions to a variety of practical problems. Original Backprop generates three-layer, feed-forward (series-coupled) networks which map fixed-length input vectors into fixed length output vectors through an intermediate (hidden) layer of binary threshold units. Version 1.2 can handle up to 200 input vectors at a time, each having up to 128 real-valued components. The first subprogram, TSET, appends a number (up to 16) of classification bits to each input, thus creating a training set of input output pairs. The second subprogram, BACKPROP, creates a trilayer network to do the prescribed mapping and modifies the weights of its connections incrementally until the training set is leaned. The learning algorithm is the 'back-propagating error correction procedures first described by F. Rosenblatt in 1961. The third subprogram, VIEWNET, lets the trained network be examined, tested, and 'pruned' (by the deletion of unnecessary hidden units). The fourth subprogram, DONET, makes a TSR routine by which the finished product of the neural net design-and-training exercise can be consulted under other MS-DOS applications.

Classification of conductance traces with recurrent neural networks

NASA Astrophysics Data System (ADS)

Lauritzen, Kasper P.; Magyarkuti, András; Balogh, Zoltán; Halbritter, András; Solomon, Gemma C.

2018-02-01

We present a new automated method for structural classification of the traces obtained in break junction experiments. Using recurrent neural networks trained on the traces of minimal cross-sectional area in molecular dynamics simulations, we successfully separate the traces into two classes: point contact or nanowire. This is done without any assumptions about the expected features of each class. The trained neural network is applied to experimental break junction conductance traces, and it separates the classes as well as the previously used experimental methods. The effect of using partial conductance traces is explored, and we show that the method performs equally well using full or partial traces (as long as the trace just prior to breaking is included). When only the initial part of the trace is included, the results are still better than random chance. Finally, we show that the neural network classification method can be used to classify experimental conductance traces without using simulated results for training, but instead training the network on a few representative experimental traces. This offers a tool to recognize some characteristic motifs of the traces, which can be hard to find by simple data selection algorithms.

Short term load forecasting using a self-supervised adaptive neural network

DOE Office of Scientific and Technical Information (OSTI.GOV)

Yoo, H.; Pimmel, R.L.

The authors developed a self-supervised adaptive neural network to perform short term load forecasts (STLF) for a large power system covering a wide service area with several heavy load centers. They used the self-supervised network to extract correlational features from temperature and load data. In using data from the calendar year 1993 as a test case, they found a 0.90 percent error for hour-ahead forecasting and 1.92 percent error for day-ahead forecasting. These levels of error compare favorably with those obtained by other techniques. The algorithm ran in a couple of minutes on a PC containing an Intel Pentium --more » 120 MHz CPU. Since the algorithm included searching the historical database, training the network, and actually performing the forecasts, this approach provides a real-time, portable, and adaptable STLF.« less

Using input feature information to improve ultraviolet retrieval in neural networks

NASA Astrophysics Data System (ADS)

Sun, Zhibin; Chang, Ni-Bin; Gao, Wei; Chen, Maosi; Zempila, Melina

2017-09-01

In neural networks, the training/predicting accuracy and algorithm efficiency can be improved significantly via accurate input feature extraction. In this study, some spatial features of several important factors in retrieving surface ultraviolet (UV) are extracted. An extreme learning machine (ELM) is used to retrieve the surface UV of 2014 in the continental United States, using the extracted features. The results conclude that more input weights can improve the learning capacities of neural networks.

Semi-supervised prediction of gene regulatory networks using machine learning algorithms.

PubMed

Patel, Nihir; Wang, Jason T L

2015-10-01

Use of computational methods to predict gene regulatory networks (GRNs) from gene expression data is a challenging task. Many studies have been conducted using unsupervised methods to fulfill the task; however, such methods usually yield low prediction accuracies due to the lack of training data. In this article, we propose semi-supervised methods for GRN prediction by utilizing two machine learning algorithms, namely, support vector machines (SVM) and random forests (RF). The semi-supervised methods make use of unlabelled data for training. We investigated inductive and transductive learning approaches, both of which adopt an iterative procedure to obtain reliable negative training data from the unlabelled data. We then applied our semi-supervised methods to gene expression data of Escherichia coli and Saccharomyces cerevisiae, and evaluated the performance of our methods using the expression data. Our analysis indicated that the transductive learning approach outperformed the inductive learning approach for both organisms. However, there was no conclusive difference identified in the performance of SVM and RF. Experimental results also showed that the proposed semi-supervised methods performed better than existing supervised methods for both organisms.

An experimental study of interstitial lung tissue classification in HRCT images using ANN and role of cost functions

NASA Astrophysics Data System (ADS)

Dash, Jatindra K.; Kale, Mandar; Mukhopadhyay, Sudipta; Khandelwal, Niranjan; Prabhakar, Nidhi; Garg, Mandeep; Kalra, Naveen

2017-03-01

In this paper, we investigate the effect of the error criteria used during a training phase of the artificial neural network (ANN) on the accuracy of the classifier for classification of lung tissues affected with Interstitial Lung Diseases (ILD). Mean square error (MSE) and the cross-entropy (CE) criteria are chosen being most popular choice in state-of-the-art implementations. The classification experiment performed on the six interstitial lung disease (ILD) patterns viz. Consolidation, Emphysema, Ground Glass Opacity, Micronodules, Fibrosis and Healthy from MedGIFT database. The texture features from an arbitrary region of interest (AROI) are extracted using Gabor filter. Two different neural networks are trained with the scaled conjugate gradient back propagation algorithm with MSE and CE error criteria function respectively for weight updation. Performance is evaluated in terms of average accuracy of these classifiers using 4 fold cross-validation. Each network is trained for five times for each fold with randomly initialized weight vectors and accuracies are computed. Significant improvement in classification accuracy is observed when ANN is trained by using CE (67.27%) as error function compared to MSE (63.60%). Moreover, standard deviation of the classification accuracy for the network trained with CE (6.69) error criteria is found less as compared to network trained with MSE (10.32) criteria.

On adaptive learning rate that guarantees convergence in feedforward networks.

PubMed

Behera, Laxmidhar; Kumar, Swagat; Patnaik, Awhan

2006-09-01

This paper investigates new learning algorithms (LF I and LF II) based on Lyapunov function for the training of feedforward neural networks. It is observed that such algorithms have interesting parallel with the popular backpropagation (BP) algorithm where the fixed learning rate is replaced by an adaptive learning rate computed using convergence theorem based on Lyapunov stability theory. LF II, a modified version of LF I, has been introduced with an aim to avoid local minima. This modification also helps in improving the convergence speed in some cases. Conditions for achieving global minimum for these kind of algorithms have been studied in detail. The performances of the proposed algorithms are compared with BP algorithm and extended Kalman filtering (EKF) on three bench-mark function approximation problems: XOR, 3-bit parity, and 8-3 encoder. The comparisons are made in terms of number of learning iterations and computational time required for convergence. It is found that the proposed algorithms (LF I and II) are much faster in convergence than other two algorithms to attain same accuracy. Finally, the comparison is made on a complex two-dimensional (2-D) Gabor function and effect of adaptive learning rate for faster convergence is verified. In a nutshell, the investigations made in this paper help us better understand the learning procedure of feedforward neural networks in terms of adaptive learning rate, convergence speed, and local minima.

Efficient generation of image chips for training deep learning algorithms

NASA Astrophysics Data System (ADS)

Han, Sanghui; Fafard, Alex; Kerekes, John; Gartley, Michael; Ientilucci, Emmett; Savakis, Andreas; Law, Charles; Parhan, Jason; Turek, Matt; Fieldhouse, Keith; Rovito, Todd

2017-05-01

Training deep convolutional networks for satellite or aerial image analysis often requires a large amount of training data. For a more robust algorithm, training data need to have variations not only in the background and target, but also radiometric variations in the image such as shadowing, illumination changes, atmospheric conditions, and imaging platforms with different collection geometry. Data augmentation is a commonly used approach to generating additional training data. However, this approach is often insufficient in accounting for real world changes in lighting, location or viewpoint outside of the collection geometry. Alternatively, image simulation can be an efficient way to augment training data that incorporates all these variations, such as changing backgrounds, that may be encountered in real data. The Digital Imaging and Remote Sensing Image Image Generation (DIRSIG) model is a tool that produces synthetic imagery using a suite of physics-based radiation propagation modules. DIRSIG can simulate images taken from different sensors with variation in collection geometry, spectral response, solar elevation and angle, atmospheric models, target, and background. Simulation of Urban Mobility (SUMO) is a multi-modal traffic simulation tool that explicitly models vehicles that move through a given road network. The output of the SUMO model was incorporated into DIRSIG to generate scenes with moving vehicles. The same approach was used when using helicopters as targets, but with slight modifications. Using the combination of DIRSIG and SUMO, we quickly generated many small images, with the target at the center with different backgrounds. The simulations generated images with vehicles and helicopters as targets, and corresponding images without targets. Using parallel computing, 120,000 training images were generated in about an hour. Some preliminary results show an improvement in the deep learning algorithm when real image training data are augmented with the simulated images, especially when obtaining sufficient real data was particularly challenging.

Interconnected growing self-organizing maps for auditory and semantic acquisition modeling

PubMed Central

Cao, Mengxue; Li, Aijun; Fang, Qiang; Kaufmann, Emily; Kröger, Bernd J.

2014-01-01

Based on the incremental nature of knowledge acquisition, in this study we propose a growing self-organizing neural network approach for modeling the acquisition of auditory and semantic categories. We introduce an Interconnected Growing Self-Organizing Maps (I-GSOM) algorithm, which takes associations between auditory information and semantic information into consideration, in this paper. Direct phonetic–semantic association is simulated in order to model the language acquisition in early phases, such as the babbling and imitation stages, in which no phonological representations exist. Based on the I-GSOM algorithm, we conducted experiments using paired acoustic and semantic training data. We use a cyclical reinforcing and reviewing training procedure to model the teaching and learning process between children and their communication partners. A reinforcing-by-link training procedure and a link-forgetting procedure are introduced to model the acquisition of associative relations between auditory and semantic information. Experimental results indicate that (1) I-GSOM has good ability to learn auditory and semantic categories presented within the training data; (2) clear auditory and semantic boundaries can be found in the network representation; (3) cyclical reinforcing and reviewing training leads to a detailed categorization as well as to a detailed clustering, while keeping the clusters that have already been learned and the network structure that has already been developed stable; and (4) reinforcing-by-link training leads to well-perceived auditory–semantic associations. Our I-GSOM model suggests that it is important to associate auditory information with semantic information during language acquisition. Despite its high level of abstraction, our I-GSOM approach can be interpreted as a biologically-inspired neurocomputational model. PMID:24688478

High pressure air compressor valve fault diagnosis using feedforward neural networks

NASA Astrophysics Data System (ADS)

James Li, C.; Yu, Xueli

1995-09-01

Feedforward neural networks (FNNs) are developed and implemented to classify a four-stage high pressure air compressor into one of the following conditions: baseline, suction or exhaust valve faults. These FNNs are used for the compressor's automatic condition monitoring and fault diagnosis. Measurements of 39 variables are obtained under different baseline conditions and third-stage suction and exhaust valve faults. These variables include pressures and temperatures at all stages, voltage between phase aand phase b, voltage between phase band phase c, total three-phase real power, cooling water flow rate, etc. To reduce the number of variables, the amount of their discriminatory information is quantified by scattering matrices to identify statistical significant ones. Measurements of the selected variables are then used by a fully automatic structural and weight learning algorithm to construct three-layer FNNs to classify the compressor's condition. This learning algorithm requires neither guesses of initial weight values nor number of neurons in the hidden layer of an FNN. It takes an incremental approach in which a hidden neuron is trained by exemplars and then augmented to the existing network. These exemplars are then made orthogonal to the newly identified hidden neuron. They are subsequently used for the training of the next hidden neuron. The betterment continues until a desired accuracy is reached. After the neural networks are established, novel measurements from various conditions that haven't been previously seen by the FNNs are then used to evaluate their ability in fault diagnosis. The trained neural networks provide very accurate diagnosis for suction and discharge valve defects.

Using the ADAP Learning Algorithm to Forecast the Onset of Diabetes Mellitus

PubMed Central

Smith, Jack W.; Everhart, J.E.; Dickson, W.C.; Knowler, W.C.; Johannes, R.S.

1988-01-01

Neural networks or connectionist models for parallel processing are not new. However, a resurgence of interest in the past half decade has occurred. In part, this is related to a better understanding of what are now referred to as hidden nodes. These algorithms are considered to be of marked value in pattern recognition problems. Because of that, we tested the ability of an early neural network model, ADAP, to forecast the onset of diabetes mellitus in a high risk population of Pima Indians. The algorithm's performance was analyzed using standard measures for clinical tests: sensitivity, specificity, and a receiver operating characteristic curve. The crossover point for sensitivity and specificity is 0.76. We are currently further examining these methods by comparing the ADAP results with those obtained from logistic regression and linear perceptron models using precisely the same training and forecasting sets. A description of the algorithm is included.

Determination of the mechanical and physical properties of cartilage by coupling poroelastic-based finite element models of indentation with artificial neural networks.

PubMed

Arbabi, Vahid; Pouran, Behdad; Campoli, Gianni; Weinans, Harrie; Zadpoor, Amir A

2016-03-21

One of the most widely used techniques to determine the mechanical properties of cartilage is based on indentation tests and interpretation of the obtained force-time or displacement-time data. In the current computational approaches, one needs to simulate the indentation test with finite element models and use an optimization algorithm to estimate the mechanical properties of cartilage. The modeling procedure is cumbersome, and the simulations need to be repeated for every new experiment. For the first time, we propose a method for fast and accurate estimation of the mechanical and physical properties of cartilage as a poroelastic material with the aid of artificial neural networks. In our study, we used finite element models to simulate the indentation for poroelastic materials with wide combinations of mechanical and physical properties. The obtained force-time curves are then divided into three parts: the first two parts of the data is used for training and validation of an artificial neural network, while the third part is used for testing the trained network. The trained neural network receives the force-time curves as the input and provides the properties of cartilage as the output. We observed that the trained network could accurately predict the properties of cartilage within the range of properties for which it was trained. The mechanical and physical properties of cartilage could therefore be estimated very fast, since no additional finite element modeling is required once the neural network is trained. The robustness of the trained artificial neural network in determining the properties of cartilage based on noisy force-time data was assessed by introducing noise to the simulated force-time data. We found that the training procedure could be optimized so as to maximize the robustness of the neural network against noisy force-time data. Copyright © 2016 Elsevier Ltd. All rights reserved.

Development and Validation of a Deep Neural Network Model for Prediction of Postoperative In-hospital Mortality.

PubMed

Lee, Christine K; Hofer, Ira; Gabel, Eilon; Baldi, Pierre; Cannesson, Maxime

2018-04-17

The authors tested the hypothesis that deep neural networks trained on intraoperative features can predict postoperative in-hospital mortality. The data used to train and validate the algorithm consists of 59,985 patients with 87 features extracted at the end of surgery. Feed-forward networks with a logistic output were trained using stochastic gradient descent with momentum. The deep neural networks were trained on 80% of the data, with 20% reserved for testing. The authors assessed improvement of the deep neural network by adding American Society of Anesthesiologists (ASA) Physical Status Classification and robustness of the deep neural network to a reduced feature set. The networks were then compared to ASA Physical Status, logistic regression, and other published clinical scores including the Surgical Apgar, Preoperative Score to Predict Postoperative Mortality, Risk Quantification Index, and the Risk Stratification Index. In-hospital mortality in the training and test sets were 0.81% and 0.73%. The deep neural network with a reduced feature set and ASA Physical Status classification had the highest area under the receiver operating characteristics curve, 0.91 (95% CI, 0.88 to 0.93). The highest logistic regression area under the curve was found with a reduced feature set and ASA Physical Status (0.90, 95% CI, 0.87 to 0.93). The Risk Stratification Index had the highest area under the receiver operating characteristics curve, at 0.97 (95% CI, 0.94 to 0.99). Deep neural networks can predict in-hospital mortality based on automatically extractable intraoperative data, but are not (yet) superior to existing methods.

On-line, adaptive state estimator for active noise control

NASA Technical Reports Server (NTRS)

Lim, Tae W.

1994-01-01

Dynamic characteristics of airframe structures are expected to vary as aircraft flight conditions change. Accurate knowledge of the changing dynamic characteristics is crucial to enhancing the performance of the active noise control system using feedback control. This research investigates the development of an adaptive, on-line state estimator using a neural network concept to conduct active noise control. In this research, an algorithm has been developed that can be used to estimate displacement and velocity responses at any locations on the structure from a limited number of acceleration measurements and input force information. The algorithm employs band-pass filters to extract from the measurement signal the frequency contents corresponding to a desired mode. The filtered signal is then used to train a neural network which consists of a linear neuron with three weights. The structure of the neural network is designed as simple as possible to increase the sampling frequency as much as possible. The weights obtained through neural network training are then used to construct the transfer function of a mode in z-domain and to identify modal properties of each mode. By using the identified transfer function and interpolating the mode shape obtained at sensor locations, the displacement and velocity responses are estimated with reasonable accuracy at any locations on the structure. The accuracy of the response estimates depends on the number of modes incorporated in the estimates and the number of sensors employed to conduct mode shape interpolation. Computer simulation demonstrates that the algorithm is capable of adapting to the varying dynamic characteristics of structural properties. Experimental implementation of the algorithm on a DSP (digital signal processing) board for a plate structure is underway. The algorithm is expected to reach the sampling frequency range of about 10 kHz to 20 kHz which needs to be maintained for a typical active noise control application.

Training feed-forward neural networks with gain constraints

PubMed

Hartman

2000-04-01

Inaccurate input-output gains (partial derivatives of outputs with respect to inputs) are common in neural network models when input variables are correlated or when data are incomplete or inaccurate. Accurate gains are essential for optimization, control, and other purposes. We develop and explore a method for training feedforward neural networks subject to inequality or equality-bound constraints on the gains of the learned mapping. Gain constraints are implemented as penalty terms added to the objective function, and training is done using gradient descent. Adaptive and robust procedures are devised for balancing the relative strengths of the various terms in the objective function, which is essential when the constraints are inconsistent with the data. The approach has the virtue that the model domain of validity can be extended via extrapolation training, which can dramatically improve generalization. The algorithm is demonstrated here on artificial and real-world problems with very good results and has been advantageously applied to dozens of models currently in commercial use.

Artificial immune system algorithm in VLSI circuit configuration

NASA Astrophysics Data System (ADS)

Mansor, Mohd. Asyraf; Sathasivam, Saratha; Kasihmuddin, Mohd Shareduwan Mohd

2017-08-01

In artificial intelligence, the artificial immune system is a robust bio-inspired heuristic method, extensively used in solving many constraint optimization problems, anomaly detection, and pattern recognition. This paper discusses the implementation and performance of artificial immune system (AIS) algorithm integrated with Hopfield neural networks for VLSI circuit configuration based on 3-Satisfiability problems. Specifically, we emphasized on the clonal selection technique in our binary artificial immune system algorithm. We restrict our logic construction to 3-Satisfiability (3-SAT) clauses in order to outfit with the transistor configuration in VLSI circuit. The core impetus of this research is to find an ideal hybrid model to assist in the VLSI circuit configuration. In this paper, we compared the artificial immune system (AIS) algorithm (HNN-3SATAIS) with the brute force algorithm incorporated with Hopfield neural network (HNN-3SATBF). Microsoft Visual C++ 2013 was used as a platform for training, simulating and validating the performances of the proposed network. The results depict that the HNN-3SATAIS outperformed HNN-3SATBF in terms of circuit accuracy and CPU time. Thus, HNN-3SATAIS can be used to detect an early error in the VLSI circuit design.

A network-based training environment: a medical image processing paradigm.

PubMed

Costaridou, L; Panayiotakis, G; Sakellaropoulos, P; Cavouras, D; Dimopoulos, J

1998-01-01

The capability of interactive multimedia and Internet technologies is investigated with respect to the implementation of a distance learning environment. The system is built according to a client-server architecture, based on the Internet infrastructure, composed of server nodes conceptually modelled as WWW sites. Sites are implemented by customization of available components. The environment integrates network-delivered interactive multimedia courses, network-based tutoring, SIG support, information databases of professional interest, as well as course and tutoring management. This capability has been demonstrated by means of an implemented system, validated with digital image processing content, specifically image enhancement. Image enhancement methods are theoretically described and applied to mammograms. Emphasis is given to the interactive presentation of the effects of algorithm parameters on images. The system end-user access depends on available bandwidth, so high-speed access can be achieved via LAN or local ISDN connections. Network based training offers new means of improved access and sharing of learning resources and expertise, as promising supplements in training.

Fully automatic time-window selection using machine learning for global adjoint tomography

NASA Astrophysics Data System (ADS)

Chen, Y.; Hill, J.; Lei, W.; Lefebvre, M. P.; Bozdag, E.; Komatitsch, D.; Tromp, J.

2017-12-01

Selecting time windows from seismograms such that the synthetic measurements (from simulations) and measured observations are sufficiently close is indispensable in a global adjoint tomography framework. The increasing amount of seismic data collected everyday around the world demands "intelligent" algorithms for seismic window selection. While the traditional FLEXWIN algorithm can be "automatic" to some extent, it still requires both human input and human knowledge or experience, and thus is not deemed to be fully automatic. The goal of intelligent window selection is to automatically select windows based on a learnt engine that is built upon a huge number of existing windows generated through the adjoint tomography project. We have formulated the automatic window selection problem as a classification problem. All possible misfit calculation windows are classified as either usable or unusable. Given a large number of windows with a known selection mode (select or not select), we train a neural network to predict the selection mode of an arbitrary input window. Currently, the five features we extract from the windows are its cross-correlation value, cross-correlation time lag, amplitude ratio between observed and synthetic data, window length, and minimum STA/LTA value. More features can be included in the future. We use these features to characterize each window for training a multilayer perceptron neural network (MPNN). Training the MPNN is equivalent to solve a non-linear optimization problem. We use backward propagation to derive the gradient of the loss function with respect to the weighting matrices and bias vectors and use the mini-batch stochastic gradient method to iteratively optimize the MPNN. Numerical tests show that with a careful selection of the training data and a sufficient amount of training data, we are able to train a robust neural network that is capable of detecting the waveforms in an arbitrary earthquake data with negligible detection error compared to existing selection methods (e.g. FLEXWIN). We will introduce in detail the mathematical formulation of the window-selection-oriented MPNN and show very encouraging results when applying the new algorithm to real earthquake data.

New model for prediction binary mixture of antihistamine decongestant using artificial neural networks and least squares support vector machine by spectrophotometry method

NASA Astrophysics Data System (ADS)

Mofavvaz, Shirin; Sohrabi, Mahmoud Reza; Nezamzadeh-Ejhieh, Alireza

2017-07-01

In the present study, artificial neural networks (ANNs) and least squares support vector machines (LS-SVM) as intelligent methods based on absorption spectra in the range of 230-300 nm have been used for determination of antihistamine decongestant contents. In the first step, one type of network (feed-forward back-propagation) from the artificial neural network with two different training algorithms, Levenberg-Marquardt (LM) and gradient descent with momentum and adaptive learning rate back-propagation (GDX) algorithm, were employed and their performance was evaluated. The performance of the LM algorithm was better than the GDX algorithm. In the second one, the radial basis network was utilized and results compared with the previous network. In the last one, the other intelligent method named least squares support vector machine was proposed to construct the antihistamine decongestant prediction model and the results were compared with two of the aforementioned networks. The values of the statistical parameters mean square error (MSE), Regression coefficient (R2), correlation coefficient (r) and also mean recovery (%), relative standard deviation (RSD) used for selecting the best model between these methods. Moreover, the proposed methods were compared to the high- performance liquid chromatography (HPLC) as a reference method. One way analysis of variance (ANOVA) test at the 95% confidence level applied to the comparison results of suggested and reference methods that there were no significant differences between them.

Predicting Electrocardiogram and Arterial Blood Pressure Waveforms with Different Echo State Network Architectures

DTIC Science & Technology

2014-11-01

networks were trained to predict an individual’s electrocardiogram (ECG) and arterial blood pressure ( ABP ) waveform data, which can potentially help...various ESN architectures for prediction tasks, and establishes the benefits of using ESN architecture designs for predicting ECG and ABP waveforms...arterial blood pressure ( ABP ) waveforms immediately prior to the machine generated alarms. When tested, the algorithm suppressed approximately 59.7

DREAMING OF ATMOSPHERES

DOE Office of Scientific and Technical Information (OSTI.GOV)

Waldmann, I. P., E-mail: ingo@star.ucl.ac.uk

Here, we introduce the RobERt (Robotic Exoplanet Recognition) algorithm for the classification of exoplanetary emission spectra. Spectral retrieval of exoplanetary atmospheres frequently requires the preselection of molecular/atomic opacities to be defined by the user. In the era of open-source, automated, and self-sufficient retrieval algorithms, manual input should be avoided. User dependent input could, in worst-case scenarios, lead to incomplete models and biases in the retrieval. The RobERt algorithm is based on deep-belief neural (DBN) networks trained to accurately recognize molecular signatures for a wide range of planets, atmospheric thermal profiles, and compositions. Reconstructions of the learned features, also referred to as themore » “dreams” of the network, indicate good convergence and an accurate representation of molecular features in the DBN. Using these deep neural networks, we work toward retrieval algorithms that themselves understand the nature of the observed spectra, are able to learn from current and past data, and make sensible qualitative preselections of atmospheric opacities to be used for the quantitative stage of the retrieval process.« less

A comparison of neural network and fuzzy clustering techniques in segmenting magnetic resonance images of the brain

NASA Technical Reports Server (NTRS)

Hall, Lawrence O.; Bensaid, Amine M.; Clarke, Laurence P.; Velthuizen, Robert P.; Silbiger, Martin S.; Bezdek, James C.

1992-01-01

Magnetic resonance (MR) brain section images are segmented and then synthetically colored to give visual representations of the original data with three approaches: the literal and approximate fuzzy c-means unsupervised clustering algorithms and a supervised computational neural network, a dynamic multilayered perception trained with the cascade correlation learning algorithm. Initial clinical results are presented on both normal volunteers and selected patients with brain tumors surrounded by edema. Supervised and unsupervised segmentation techniques provide broadly similar results. Unsupervised fuzzy algorithms were visually observed to show better segmentation when compared with raw image data for volunteer studies. However, for a more complex segmentation problem with tumor/edema or cerebrospinal fluid boundary, where the tissues have similar MR relaxation behavior, inconsistency in rating among experts was observed.

Decoding of finger trajectory from ECoG using deep learning.

PubMed

Xie, Ziqian; Schwartz, Odelia; Prasad, Abhishek

2018-06-01

Conventional decoding pipeline for brain-machine interfaces (BMIs) consists of chained different stages of feature extraction, time-frequency analysis and statistical learning models. Each of these stages uses a different algorithm trained in a sequential manner, which makes it difficult to make the whole system adaptive. The goal was to create an adaptive online system with a single objective function and a single learning algorithm so that the whole system can be trained in parallel to increase the decoding performance. Here, we used deep neural networks consisting of convolutional neural networks (CNN) and a special kind of recurrent neural network (RNN) called long short term memory (LSTM) to address these needs. We used electrocorticography (ECoG) data collected by Kubanek et al. The task consisted of individual finger flexions upon a visual cue. Our model combined a hierarchical feature extractor CNN and a RNN that was able to process sequential data and recognize temporal dynamics in the neural data. CNN was used as the feature extractor and LSTM was used as the regression algorithm to capture the temporal dynamics of the signal. We predicted the finger trajectory using ECoG signals and compared results for the least angle regression (LARS), CNN-LSTM, random forest, LSTM model (LSTM_HC, for using hard-coded features) and a decoding pipeline consisting of band-pass filtering, energy extraction, feature selection and linear regression. The results showed that the deep learning models performed better than the commonly used linear model. The deep learning models not only gave smoother and more realistic trajectories but also learned the transition between movement and rest state. This study demonstrated a decoding network for BMI that involved a convolutional and recurrent neural network model. It integrated the feature extraction pipeline into the convolution and pooling layer and used LSTM layer to capture the state transitions. The discussed network eliminated the need to separately train the model at each step in the decoding pipeline. The whole system can be jointly optimized using stochastic gradient descent and is capable of online learning.

Decoding of finger trajectory from ECoG using deep learning

NASA Astrophysics Data System (ADS)

Xie, Ziqian; Schwartz, Odelia; Prasad, Abhishek

2018-06-01

Objective. Conventional decoding pipeline for brain-machine interfaces (BMIs) consists of chained different stages of feature extraction, time-frequency analysis and statistical learning models. Each of these stages uses a different algorithm trained in a sequential manner, which makes it difficult to make the whole system adaptive. The goal was to create an adaptive online system with a single objective function and a single learning algorithm so that the whole system can be trained in parallel to increase the decoding performance. Here, we used deep neural networks consisting of convolutional neural networks (CNN) and a special kind of recurrent neural network (RNN) called long short term memory (LSTM) to address these needs. Approach. We used electrocorticography (ECoG) data collected by Kubanek et al. The task consisted of individual finger flexions upon a visual cue. Our model combined a hierarchical feature extractor CNN and a RNN that was able to process sequential data and recognize temporal dynamics in the neural data. CNN was used as the feature extractor and LSTM was used as the regression algorithm to capture the temporal dynamics of the signal. Main results. We predicted the finger trajectory using ECoG signals and compared results for the least angle regression (LARS), CNN-LSTM, random forest, LSTM model (LSTM_HC, for using hard-coded features) and a decoding pipeline consisting of band-pass filtering, energy extraction, feature selection and linear regression. The results showed that the deep learning models performed better than the commonly used linear model. The deep learning models not only gave smoother and more realistic trajectories but also learned the transition between movement and rest state. Significance. This study demonstrated a decoding network for BMI that involved a convolutional and recurrent neural network model. It integrated the feature extraction pipeline into the convolution and pooling layer and used LSTM layer to capture the state transitions. The discussed network eliminated the need to separately train the model at each step in the decoding pipeline. The whole system can be jointly optimized using stochastic gradient descent and is capable of online learning.

An NN-Based SRD Decomposition Algorithm and Its Application in Nonlinear Compensation

PubMed Central

Yan, Honghang; Deng, Fang; Sun, Jian; Chen, Jie

2014-01-01

In this study, a neural network-based square root of descending (SRD) order decomposition algorithm for compensating for nonlinear data generated by sensors is presented. The study aims at exploring the optimized decomposition of data 1.00,0.00,0.00 and minimizing the computational complexity and memory space of the training process. A linear decomposition algorithm, which automatically finds the optimal decomposition of N subparts and reduces the training time to 1N and memory cost to 1N, has been implemented on nonlinear data obtained from an encoder. Particular focus is given to the theoretical access of estimating the numbers of hidden nodes and the precision of varying the decomposition method. Numerical experiments are designed to evaluate the effect of this algorithm. Moreover, a designed device for angular sensor calibration is presented. We conduct an experiment that samples the data of an encoder and compensates for the nonlinearity of the encoder to testify this novel algorithm. PMID:25232912

Rule Extraction Based on Extreme Learning Machine and an Improved Ant-Miner Algorithm for Transient Stability Assessment.

PubMed

Li, Yang; Li, Guoqing; Wang, Zhenhao

2015-01-01

In order to overcome the problems of poor understandability of the pattern recognition-based transient stability assessment (PRTSA) methods, a new rule extraction method based on extreme learning machine (ELM) and an improved Ant-miner (IAM) algorithm is presented in this paper. First, the basic principles of ELM and Ant-miner algorithm are respectively introduced. Then, based on the selected optimal feature subset, an example sample set is generated by the trained ELM-based PRTSA model. And finally, a set of classification rules are obtained by IAM algorithm to replace the original ELM network. The novelty of this proposal is that transient stability rules are extracted from an example sample set generated by the trained ELM-based transient stability assessment model by using IAM algorithm. The effectiveness of the proposed method is shown by the application results on the New England 39-bus power system and a practical power system--the southern power system of Hebei province.

Unsupervised algorithms for intrusion detection and identification in wireless ad hoc sensor networks

NASA Astrophysics Data System (ADS)

Hortos, William S.

2009-05-01

In previous work by the author, parameters across network protocol layers were selected as features in supervised algorithms that detect and identify certain intrusion attacks on wireless ad hoc sensor networks (WSNs) carrying multisensor data. The algorithms improved the residual performance of the intrusion prevention measures provided by any dynamic key-management schemes and trust models implemented among network nodes. The approach of this paper does not train algorithms on the signature of known attack traffic, but, instead, the approach is based on unsupervised anomaly detection techniques that learn the signature of normal network traffic. Unsupervised learning does not require the data to be labeled or to be purely of one type, i.e., normal or attack traffic. The approach can be augmented to add any security attributes and quantified trust levels, established during data exchanges among nodes, to the set of cross-layer features from the WSN protocols. A two-stage framework is introduced for the security algorithms to overcome the problems of input size and resource constraints. The first stage is an unsupervised clustering algorithm which reduces the payload of network data packets to a tractable size. The second stage is a traditional anomaly detection algorithm based on a variation of support vector machines (SVMs), whose efficiency is improved by the availability of data in the packet payload. In the first stage, selected algorithms are adapted to WSN platforms to meet system requirements for simple parallel distributed computation, distributed storage and data robustness. A set of mobile software agents, acting like an ant colony in securing the WSN, are distributed at the nodes to implement the algorithms. The agents move among the layers involved in the network response to the intrusions at each active node and trustworthy neighborhood, collecting parametric values and executing assigned decision tasks. This minimizes the need to move large amounts of audit-log data through resource-limited nodes and locates routines closer to that data. Performance of the unsupervised algorithms is evaluated against the network intrusions of black hole, flooding, Sybil and other denial-of-service attacks in simulations of published scenarios. Results for scenarios with intentionally malfunctioning sensors show the robustness of the two-stage approach to intrusion anomalies.

Prediction of composite fatigue life under variable amplitude loading using artificial neural network trained by genetic algorithm

NASA Astrophysics Data System (ADS)

Rohman, Muhamad Nur; Hidayat, Mas Irfan P.; Purniawan, Agung

2018-04-01

Neural networks (NN) have been widely used in application of fatigue life prediction. In the use of fatigue life prediction for polymeric-base composite, development of NN model is necessary with respect to the limited fatigue data and applicable to be used to predict the fatigue life under varying stress amplitudes in the different stress ratios. In the present paper, Multilayer-Perceptrons (MLP) model of neural network is developed, and Genetic Algorithm was employed to optimize the respective weights of NN for prediction of polymeric-base composite materials under variable amplitude loading. From the simulation result obtained with two different composite systems, named E-glass fabrics/epoxy (layups [(±45)/(0)2]S), and E-glass/polyester (layups [90/0/±45/0]S), NN model were trained with fatigue data from two different stress ratios, which represent limited fatigue data, can be used to predict another four and seven stress ratios respectively, with high accuracy of fatigue life prediction. The accuracy of NN prediction were quantified with the small value of mean square error (MSE). When using 33% from the total fatigue data for training, the NN model able to produce high accuracy for all stress ratios. When using less fatigue data during training (22% from the total fatigue data), the NN model still able to produce high coefficient of determination between the prediction result compared with obtained by experiment.

Two-Dimensional High-Lift Aerodynamic Optimization Using Neural Networks

NASA Technical Reports Server (NTRS)

Greenman, Roxana M.

1998-01-01

The high-lift performance of a multi-element airfoil was optimized by using neural-net predictions that were trained using a computational data set. The numerical data was generated using a two-dimensional, incompressible, Navier-Stokes algorithm with the Spalart-Allmaras turbulence model. Because it is difficult to predict maximum lift for high-lift systems, an empirically-based maximum lift criteria was used in this study to determine both the maximum lift and the angle at which it occurs. The 'pressure difference rule,' which states that the maximum lift condition corresponds to a certain pressure difference between the peak suction pressure and the pressure at the trailing edge of the element, was applied and verified with experimental observations for this configuration. Multiple input, single output networks were trained using the NASA Ames variation of the Levenberg-Marquardt algorithm for each of the aerodynamic coefficients (lift, drag and moment). The artificial neural networks were integrated with a gradient-based optimizer. Using independent numerical simulations and experimental data for this high-lift configuration, it was shown that this design process successfully optimized flap deflection, gap, overlap, and angle of attack to maximize lift. Once the neural nets were trained and integrated with the optimizer, minimal additional computer resources were required to perform optimization runs with different initial conditions and parameters. Applying the neural networks within the high-lift rigging optimization process reduced the amount of computational time and resources by 44% compared with traditional gradient-based optimization procedures for multiple optimization runs.

Java-Based Diabetes Type 2 Prediction Tool for Better Diagnosis

PubMed Central

Odedra, Devang; Mallick, Medhavi; Shukla, Prateek; Samanta, Subir; Vidyarthi, Ambarish S.

2012-01-01

Abstract Background The concept of classification of clinical data can be utilized in the development of an effective diagnosis system by taking the advantage of computational intelligence. Diabetes disease diagnosis via proper interpretation of the diabetes data is an important problem in neural networks. Unfortunately, although several classification studies have been carried out with significant performance, many of the current methods often fail to reach out to patients. Graphical user interface-enabled tools need to be developed through which medical practitioners can simply enter the health profiles of their patients and receive an instant diabetes prediction with an acceptable degree of confidence. Methods In this study, the neural network approach was used for a dataset of 768 persons from a Pima Indian population living near Phoenix, AZ. A neural network mixture of experts model was trained with these data using the expectation-minimization algorithm. Results The mixture of experts method was used to train the algorithm with 97% accuracy. A graphical user interface was developed that would work in conjunction with the trained network to provide the output in a presentable format. Conclusions This study provides a machine-implementable approach that can be used by physicians and patients to minimize the extent of error in diagnosis. The authors are hopeful that replication of results of this study in other populations may lead to improved diagnosis. Physicians can simply enter the health profile of patients and get the diagnosis for diabetes type 2. PMID:22059431

PONS2train: tool for testing the MLP architecture and local traning methods for runoff forecast

NASA Astrophysics Data System (ADS)

Maca, P.; Pavlasek, J.; Pech, P.

2012-04-01

The purpose of presented poster is to introduce the PONS2train developed for runoff prediction via multilayer perceptron - MLP. The software application enables the implementation of 12 different MLP's transfer functions, comparison of 9 local training algorithms and finally the evaluation the MLP performance via 17 selected model evaluation metrics. The PONS2train software is written in C++ programing language. Its implementation consists of 4 classes. The NEURAL_NET and NEURON classes implement the MLP, the CRITERIA class estimates model evaluation metrics and for model performance evaluation via testing and validation datasets. The DATA_PATTERN class prepares the validation, testing and calibration datasets. The software application uses the LAPACK, BLAS and ARMADILLO C++ linear algebra libraries. The PONS2train implements the first order local optimization algorithms: standard on-line and batch back-propagation with learning rate combined with momentum and its variants with the regularization term, Rprop and standard batch back-propagation with variable momentum and learning rate. The second order local training algorithms represents: the Levenberg-Marquardt algorithm with and without regularization and four variants of scaled conjugate gradients. The other important PONS2train features are: the multi-run, the weight saturation control, early stopping of trainings, and the MLP weights analysis. The weights initialization is done via two different methods: random sampling from uniform distribution on open interval or Nguyen Widrow method. The data patterns can be transformed via linear and nonlinear transformation. The runoff forecast case study focuses on PONS2train implementation and shows the different aspects of the MLP training, the MLP architecture estimation, the neural network weights analysis and model uncertainty estimation.

Application of neural networks to group technology

NASA Astrophysics Data System (ADS)

Caudell, Thomas P.; Smith, Scott D. G.; Johnson, G. C.; Wunsch, Donald C., II

1991-08-01

Adaptive resonance theory (ART) neural networks are being developed for application to the industrial engineering problem of group technology--the reuse of engineering designs. Two- and three-dimensional representations of engineering designs are input to ART-1 neural networks to produce groups or families of similar parts. These representations, in their basic form, amount to bit maps of the part, and can become very large when the part is represented in high resolution. This paper describes an enhancement to an algorithmic form of ART-1 that allows it to operate directly on compressed input representations and to generate compressed memory templates. The performance of this compressed algorithm is compared to that of the regular algorithm on real engineering designs and a significant savings in memory storage as well as a speed up in execution is observed. In additions, a `neural database'' system under development is described. This system demonstrates the feasibility of training an ART-1 network to first cluster designs into families, and then to recall the family when presented a similar design. This application is of large practical value to industry, making it possible to avoid duplication of design efforts.

Prediction of heart abnormality using MLP network

NASA Astrophysics Data System (ADS)

Hashim, Fakroul Ridzuan; Januar, Yulni; Mat, Muhammad Hadzren; Rizman, Zairi Ismael; Awang, Mat Kamil

2018-02-01

Heart abnormality does not choose gender, age and races when it strikes. With no warning signs or symptoms, it can result to a sudden death of the patient. Generally, heart's irregular electrical activity is defined as heart abnormality. Via implementation of Multilayer Perceptron (MLP) network, this paper tries to develop a program that allows the detection of heart abnormality activity. Utilizing several training algorithms with Purelin activation function, an amount of heartbeat signals received through the electrocardiogram (ECG) will be employed to condition the MLP network.

Scaling deep learning on GPU and knights landing clusters

DOE Office of Scientific and Technical Information (OSTI.GOV)

You, Yang; Buluc, Aydin; Demmel, James

Training neural networks has become a big bottleneck. For example, training ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training process, the current deep learning systems heavily rely on the hardware accelerators. However, these accelerators have limited on-chip memory compared with CPUs. We use both self-host Intel Knights Landing (KNL) clusters and multi-GPU clusters as our target platforms. From the algorithm aspect, we focus on Elastic Averaging SGD (EASGD) to design algorithms for HPC clusters. We redesign four efficient algorithms for HPC systems to improve EASGD's poor scaling on clusters. Async EASGD, Async MEASGD,more » and Hogwild EASGD are faster than existing counter-part methods (Async SGD, Async MSGD, and Hogwild SGD) in all comparisons. Sync EASGD achieves 5.3X speedup over original EASGD on the same platform. We achieve 91.5% weak scaling efficiency on 4253 KNL cores, which is higher than the state-of-the-art implementation.« less

Multilayer Perceptron for Robust Nonlinear Interval Regression Analysis Using Genetic Algorithms

PubMed Central

2014-01-01

On the basis of fuzzy regression, computational models in intelligence such as neural networks have the capability to be applied to nonlinear interval regression analysis for dealing with uncertain and imprecise data. When training data are not contaminated by outliers, computational models perform well by including almost all given training data in the data interval. Nevertheless, since training data are often corrupted by outliers, robust learning algorithms employed to resist outliers for interval regression analysis have been an interesting area of research. Several approaches involving computational intelligence are effective for resisting outliers, but the required parameters for these approaches are related to whether the collected data contain outliers or not. Since it seems difficult to prespecify the degree of contamination beforehand, this paper uses multilayer perceptron to construct the robust nonlinear interval regression model using the genetic algorithm. Outliers beyond or beneath the data interval will impose slight effect on the determination of data interval. Simulation results demonstrate that the proposed method performs well for contaminated datasets. PMID:25110755

Multilayer perceptron for robust nonlinear interval regression analysis using genetic algorithms.

PubMed

Hu, Yi-Chung

2014-01-01

On the basis of fuzzy regression, computational models in intelligence such as neural networks have the capability to be applied to nonlinear interval regression analysis for dealing with uncertain and imprecise data. When training data are not contaminated by outliers, computational models perform well by including almost all given training data in the data interval. Nevertheless, since training data are often corrupted by outliers, robust learning algorithms employed to resist outliers for interval regression analysis have been an interesting area of research. Several approaches involving computational intelligence are effective for resisting outliers, but the required parameters for these approaches are related to whether the collected data contain outliers or not. Since it seems difficult to prespecify the degree of contamination beforehand, this paper uses multilayer perceptron to construct the robust nonlinear interval regression model using the genetic algorithm. Outliers beyond or beneath the data interval will impose slight effect on the determination of data interval. Simulation results demonstrate that the proposed method performs well for contaminated datasets.

Application of artificial neural network to predict clay sensitivity in a high landslide prone area using CPTu data- A case study in Southwest of Sweden

NASA Astrophysics Data System (ADS)

Shahri, Abbas; Mousavinaseri, Mahsasadat; Naderi, Shima; Espersson, Maria

2015-04-01

Application of Artificial Neural Networks (ANNs) in many areas of engineering, in particular to geotechnical engineering problems such as site characterization has demonstrated some degree of success. The present paper aims to evaluate the feasibility of several various types of ANN models to predict the clay sensitivity of soft clays form piezocone penetration test data (CPTu). To get the aim, a research database of CPTu data of 70 test points around the Göta River near the Lilli Edet in the southwest of Sweden which is a high prone land slide area were collected and considered as input for ANNs. For training algorithms the quick propagation, conjugate gradient descent, quasi-Newton, limited memory quasi-Newton and Levenberg-Marquardt were developed tested and trained using the CPTu data to provide a comparison between the results of field investigation and ANN models to estimate the clay sensitivity. The reason of using the clay sensitivity parameter in this study is due to its relation to landslides in Sweden.A special high sensitive clay namely quick clay is considered as the main responsible for experienced landslides in Sweden which has high sensitivity and prone to slide. The training and testing program was started with 3-2-1 ANN architecture structure. By testing and trying several various architecture structures and changing the hidden layer in order to have a higher output resolution the 3-4-4-3-1 architecture structure for ANN in this study was confirmed. The tested algorithm showed that increasing the hidden layers up to 4 layers in ANN can improve the results and the 3-4-4-3-1 architecture structure ANNs for prediction of clay sensitivity represent reliable and reasonable response. The obtained results showed that the conjugate gradient descent algorithm with R2=0.897 has the best performance among the tested algorithms. Keywords: clay sensitivity, landslide, Artificial Neural Network

Towards representation of a perceptual color manifold using associative memory for color constancy.

PubMed

Seow, Ming-Jung; Asari, Vijayan K

2009-01-01

In this paper, we propose the concept of a manifold of color perception through empirical observation that the center-surround properties of images in a perceptually similar environment define a manifold in the high dimensional space. Such a manifold representation can be learned using a novel recurrent neural network based learning algorithm. Unlike the conventional recurrent neural network model in which the memory is stored in an attractive fixed point at discrete locations in the state space, the dynamics of the proposed learning algorithm represent memory as a nonlinear line of attraction. The region of convergence around the nonlinear line is defined by the statistical characteristics of the training data. This learned manifold can then be used as a basis for color correction of the images having different color perception to the learned color perception. Experimental results show that the proposed recurrent neural network learning algorithm is capable of color balance the lighting variations in images captured in different environments successfully.

Online Learning Flight Control for Intelligent Flight Control Systems (IFCS)

NASA Technical Reports Server (NTRS)

Niewoehner, Kevin R.; Carter, John (Technical Monitor)

2001-01-01

The research accomplishments for the cooperative agreement 'Online Learning Flight Control for Intelligent Flight Control Systems (IFCS)' include the following: (1) previous IFC program data collection and analysis; (2) IFC program support site (configured IFC systems support network, configured Tornado/VxWorks OS development system, made Configuration and Documentation Management Systems Internet accessible); (3) Airborne Research Test Systems (ARTS) II Hardware (developed hardware requirements specification, developing environmental testing requirements, hardware design, and hardware design development); (4) ARTS II software development laboratory unit (procurement of lab style hardware, configured lab style hardware, and designed interface module equivalent to ARTS II faceplate); (5) program support documentation (developed software development plan, configuration management plan, and software verification and validation plan); (6) LWR algorithm analysis (performed timing and profiling on algorithm); (7) pre-trained neural network analysis; (8) Dynamic Cell Structures (DCS) Neural Network Analysis (performing timing and profiling on algorithm); and (9) conducted technical interchange and quarterly meetings to define IFC research goals.

Face recognition via Gabor and convolutional neural network

NASA Astrophysics Data System (ADS)

Lu, Tongwei; Wu, Menglu; Lu, Tao

2018-04-01

In recent years, the powerful feature learning and classification ability of convolutional neural network have attracted widely attention. Compared with the deep learning, the traditional machine learning algorithm has a good explanatory which deep learning does not have. Thus, In this paper, we propose a method to extract the feature of the traditional algorithm as the input of convolution neural network. In order to reduce the complexity of the network, the kernel function of Gabor wavelet is used to extract the feature from different position, frequency and direction of target image. It is sensitive to edge of image which can provide good direction and scale selection. The extraction of the image from eight directions on a scale are as the input of network that we proposed. The network have the advantage of weight sharing and local connection and texture feature of the input image can reduce the influence of facial expression, gesture and illumination. At the same time, we introduced a layer which combined the results of the pooling and convolution can extract deeper features. The training network used the open source caffe framework which is beneficial to feature extraction. The experiment results of the proposed method proved that the network structure effectively overcame the barrier of illumination and had a good robustness as well as more accurate and rapid than the traditional algorithm.

DCS-Neural-Network Program for Aircraft Control and Testing

NASA Technical Reports Server (NTRS)

Jorgensen, Charles C.

2006-01-01

A computer program implements a dynamic-cell-structure (DCS) artificial neural network that can perform such tasks as learning selected aerodynamic characteristics of an airplane from wind-tunnel test data and computing real-time stability and control derivatives of the airplane for use in feedback linearized control. A DCS neural network is one of several types of neural networks that can incorporate additional nodes in order to rapidly learn increasingly complex relationships between inputs and outputs. In the DCS neural network implemented by the present program, the insertion of nodes is based on accumulated error. A competitive Hebbian learning rule (a supervised-learning rule in which connection weights are adjusted to minimize differences between actual and desired outputs for training examples) is used. A Kohonen-style learning rule (derived from a relatively simple training algorithm, implements a Delaunay triangulation layout of neurons) is used to adjust node positions during training. Neighborhood topology determines which nodes are used to estimate new values. The network learns, starting with two nodes, and adds new nodes sequentially in locations chosen to maximize reductions in global error. At any given time during learning, the error becomes homogeneously distributed over all nodes.

Algorithm for predicting the evolution of series of dynamics of complex systems in solving information problems

NASA Astrophysics Data System (ADS)

Kasatkina, T. I.; Dushkin, A. V.; Pavlov, V. A.; Shatovkin, R. R.

2018-03-01

In the development of information, systems and programming to predict the series of dynamics, neural network methods have recently been applied. They are more flexible, in comparison with existing analogues and are capable of taking into account the nonlinearities of the series. In this paper, we propose a modified algorithm for predicting the series of dynamics, which includes a method for training neural networks, an approach to describing and presenting input data, based on the prediction by the multilayer perceptron method. To construct a neural network, the values of a series of dynamics at the extremum points and time values corresponding to them, formed based on the sliding window method, are used as input data. The proposed algorithm can act as an independent approach to predicting the series of dynamics, and be one of the parts of the forecasting system. The efficiency of predicting the evolution of the dynamics series for a short-term one-step and long-term multi-step forecast by the classical multilayer perceptron method and a modified algorithm using synthetic and real data is compared. The result of this modification was the minimization of the magnitude of the iterative error that arises from the previously predicted inputs to the inputs to the neural network, as well as the increase in the accuracy of the iterative prediction of the neural network.

Receptive field optimisation and supervision of a fuzzy spiking neural network.

PubMed

Glackin, Cornelius; Maguire, Liam; McDaid, Liam; Sayers, Heather

2011-04-01

This paper presents a supervised training algorithm that implements fuzzy reasoning on a spiking neural network. Neuron selectivity is facilitated using receptive fields that enable individual neurons to be responsive to certain spike train firing rates and behave in a similar manner as fuzzy membership functions. The connectivity of the hidden and output layers in the fuzzy spiking neural network (FSNN) is representative of a fuzzy rule base. Fuzzy C-Means clustering is utilised to produce clusters that represent the antecedent part of the fuzzy rule base that aid classification of the feature data. Suitable cluster widths are determined using two strategies; subjective thresholding and evolutionary thresholding respectively. The former technique typically results in compact solutions in terms of the number of neurons, and is shown to be particularly suited to small data sets. In the latter technique a pool of cluster candidates is generated using Fuzzy C-Means clustering and then a genetic algorithm is employed to select the most suitable clusters and to specify cluster widths. In both scenarios, the network is supervised but learning only occurs locally as in the biological case. The advantages and disadvantages of the network topology for the Fisher Iris and Wisconsin Breast Cancer benchmark classification tasks are demonstrated and directions of current and future work are discussed. Copyright © 2010 Elsevier Ltd. All rights reserved.

Reverse engineering highlights potential principles of large gene regulatory network design and learning.

PubMed

Carré, Clément; Mas, André; Krouk, Gabriel

2017-01-01

Inferring transcriptional gene regulatory networks from transcriptomic datasets is a key challenge of systems biology, with potential impacts ranging from medicine to agronomy. There are several techniques used presently to experimentally assay transcription factors to target relationships, defining important information about real gene regulatory networks connections. These techniques include classical ChIP-seq, yeast one-hybrid, or more recently, DAP-seq or target technologies. These techniques are usually used to validate algorithm predictions. Here, we developed a reverse engineering approach based on mathematical and computer simulation to evaluate the impact that this prior knowledge on gene regulatory networks may have on training machine learning algorithms. First, we developed a gene regulatory networks-simulating engine called FRANK (Fast Randomizing Algorithm for Network Knowledge) that is able to simulate large gene regulatory networks (containing 10 4 genes) with characteristics of gene regulatory networks observed in vivo. FRANK also generates stable or oscillatory gene expression directly produced by the simulated gene regulatory networks. The development of FRANK leads to important general conclusions concerning the design of large and stable gene regulatory networks harboring scale free properties (built ex nihilo). In combination with supervised (accepting prior knowledge) support vector machine algorithm we (i) address biologically oriented questions concerning our capacity to accurately reconstruct gene regulatory networks and in particular we demonstrate that prior-knowledge structure is crucial for accurate learning, and (ii) draw conclusions to inform experimental design to performed learning able to solve gene regulatory networks in the future. By demonstrating that our predictions concerning the influence of the prior-knowledge structure on support vector machine learning capacity holds true on real data ( Escherichia coli K14 network reconstruction using network and transcriptomic data), we show that the formalism used to build FRANK can to some extent be a reasonable model for gene regulatory networks in real cells.

A Supervised Approach to Windowing Detection on Dynamic Networks

DTIC Science & Technology

2017-07-01

A supervised approach to windowing detection on dynamic networks Benjamin Fish University of Illinois at Chicago 1200 W. Harrison St. Chicago...Using this framework, we introduce windowing algorithms that take a supervised approach : they leverage ground truth on training data to find a good...windowing of the test data. We compare the supervised approach to previous approaches and several baselines on real data. ACM Reference format: Benjamin

Learning spatially coherent properties of the visual world in connectionist networks

NASA Astrophysics Data System (ADS)

Becker, Suzanna; Hinton, Geoffrey E.

1991-10-01

In the unsupervised learning paradigm, a network of neuron-like units is presented with an ensemble of input patterns from a structured environment, such as the visual world, and learns to represent the regularities in that input. The major goal in developing unsupervised learning algorithms is to find objective functions that characterize the quality of the network's representation without explicitly specifying the desired outputs of any of the units. The sort of objective functions considered cause a unit to become tuned to spatially coherent features of visual images (such as texture, depth, shading, and surface orientation), by learning to predict the outputs of other units which have spatially adjacent receptive fields. Simulations show that using an information-theoretic algorithm called IMAX, a network can be trained to represent depth by observing random dot stereograms of surfaces with continuously varying disparities. Once a layer of depth-tuned units has developed, subsequent layers are trained to perform surface interpolation of curved surfaces, by learning to predict the depth of one image region based on depth measurements in surrounding regions. An extension of the basic model allows a population of competing neurons to learn a distributed code for disparity, which naturally gives rise to a representation of discontinuities.

Parallel protein secondary structure prediction based on neural networks.

PubMed

Zhong, Wei; Altun, Gulsah; Tian, Xinmin; Harrison, Robert; Tai, Phang C; Pan, Yi

2004-01-01

Protein secondary structure prediction has a fundamental influence on today's bioinformatics research. In this work, binary and tertiary classifiers of protein secondary structure prediction are implemented on Denoeux belief neural network (DBNN) architecture. Hydrophobicity matrix, orthogonal matrix, BLOSUM62 and PSSM (position specific scoring matrix) are experimented separately as the encoding schemes for DBNN. The experimental results contribute to the design of new encoding schemes. New binary classifier for Helix versus not Helix ( approximately H) for DBNN produces prediction accuracy of 87% when PSSM is used for the input profile. The performance of DBNN binary classifier is comparable to other best prediction methods. The good test results for binary classifiers open a new approach for protein structure prediction with neural networks. Due to the time consuming task of training the neural networks, Pthread and OpenMP are employed to parallelize DBNN in the hyperthreading enabled Intel architecture. Speedup for 16 Pthreads is 4.9 and speedup for 16 OpenMP threads is 4 in the 4 processors shared memory architecture. Both speedup performance of OpenMP and Pthread is superior to that of other research. With the new parallel training algorithm, thousands of amino acids can be processed in reasonable amount of time. Our research also shows that hyperthreading technology for Intel architecture is efficient for parallel biological algorithms.

NeuroCa: integrated framework for systematic analysis of spatiotemporal neuronal activity patterns from large-scale optical recording data

PubMed Central

Jang, Min Jee; Nam, Yoonkey

2015-01-01

Abstract. Optical recording facilitates monitoring the activity of a large neural network at the cellular scale, but the analysis and interpretation of the collected data remain challenging. Here, we present a MATLAB-based toolbox, named NeuroCa, for the automated processing and quantitative analysis of large-scale calcium imaging data. Our tool includes several computational algorithms to extract the calcium spike trains of individual neurons from the calcium imaging data in an automatic fashion. Two algorithms were developed to decompose the imaging data into the activity of individual cells and subsequently detect calcium spikes from each neuronal signal. Applying our method to dense networks in dissociated cultures, we were able to obtain the calcium spike trains of ∼1000 neurons in a few minutes. Further analyses using these data permitted the quantification of neuronal responses to chemical stimuli as well as functional mapping of spatiotemporal patterns in neuronal firing within the spontaneous, synchronous activity of a large network. These results demonstrate that our method not only automates time-consuming, labor-intensive tasks in the analysis of neural data obtained using optical recording techniques but also provides a systematic way to visualize and quantify the collective dynamics of a network in terms of its cellular elements. PMID:26229973

A Direct Position-Determination Approach for Multiple Sources Based on Neural Network Computation.

PubMed

Chen, Xin; Wang, Ding; Yin, Jiexin; Wu, Ying

2018-06-13

The most widely used localization technology is the two-step method that localizes transmitters by measuring one or more specified positioning parameters. Direct position determination (DPD) is a promising technique that directly localizes transmitters from sensor outputs and can offer superior localization performance. However, existing DPD algorithms such as maximum likelihood (ML)-based and multiple signal classification (MUSIC)-based estimations are computationally expensive, making it difficult to satisfy real-time demands. To solve this problem, we propose the use of a modular neural network for multiple-source DPD. In this method, the area of interest is divided into multiple sub-areas. Multilayer perceptron (MLP) neural networks are employed to detect the presence of a source in a sub-area and filter sources in other sub-areas, and radial basis function (RBF) neural networks are utilized for position estimation. Simulation results show that a number of appropriately trained neural networks can be successfully used for DPD. The performance of the proposed MLP-MLP-RBF method is comparable to the performance of the conventional MUSIC-based DPD algorithm for various signal-to-noise ratios and signal power ratios. Furthermore, the MLP-MLP-RBF network is less computationally intensive than the classical DPD algorithm and is therefore an attractive choice for real-time applications.

Convolutional Neural Network-Based Embarrassing Situation Detection under Camera for Social Robot in Smart Homes

PubMed Central

Sheng, Weihua; Junior, Francisco Erivaldo Fernandes; Li, Shaobo

2018-01-01

Recent research has shown that the ubiquitous use of cameras and voice monitoring equipment in a home environment can raise privacy concerns and affect human mental health. This can be a major obstacle to the deployment of smart home systems for elderly or disabled care. This study uses a social robot to detect embarrassing situations. Firstly, we designed an improved neural network structure based on the You Only Look Once (YOLO) model to obtain feature information. By focusing on reducing area redundancy and computation time, we proposed a bounding-box merging algorithm based on region proposal networks (B-RPN), to merge the areas that have similar features and determine the borders of the bounding box. Thereafter, we designed a feature extraction algorithm based on our improved YOLO and B-RPN, called F-YOLO, for our training datasets, and then proposed a real-time object detection algorithm based on F-YOLO (RODA-FY). We implemented RODA-FY and compared models on our MAT social robot. Secondly, we considered six types of situations in smart homes, and developed training and validation datasets, containing 2580 and 360 images, respectively. Meanwhile, we designed three types of experiments with four types of test datasets composed of 960 sample images. Thirdly, we analyzed how a different number of training iterations affects our prediction estimation, and then we explored the relationship between recognition accuracy and learning rates. Our results show that our proposed privacy detection system can recognize designed situations in the smart home with an acceptable recognition accuracy of 94.48%. Finally, we compared the results among RODA-FY, Inception V3, and YOLO, which indicate that our proposed RODA-FY outperforms the other comparison models in recognition accuracy. PMID:29757211

Boosted ARTMAP: modifications to fuzzy ARTMAP motivated by boosting theory.

PubMed

Verzi, Stephen J; Heileman, Gregory L; Georgiopoulos, Michael

2006-05-01

In this paper, several modifications to the Fuzzy ARTMAP neural network architecture are proposed for conducting classification in complex, possibly noisy, environments. The goal of these modifications is to improve upon the generalization performance of Fuzzy ART-based neural networks, such as Fuzzy ARTMAP, in these situations. One of the major difficulties of employing Fuzzy ARTMAP on such learning problems involves over-fitting of the training data. Structural risk minimization is a machine-learning framework that addresses the issue of over-fitting by providing a backbone for analysis as well as an impetus for the design of better learning algorithms. The theory of structural risk minimization reveals a trade-off between training error and classifier complexity in reducing generalization error, which will be exploited in the learning algorithms proposed in this paper. Boosted ART extends Fuzzy ART by allowing the spatial extent of each cluster formed to be adjusted independently. Boosted ARTMAP generalizes upon Fuzzy ARTMAP by allowing non-zero training error in an effort to reduce the hypothesis complexity and hence improve overall generalization performance. Although Boosted ARTMAP is strictly speaking not a boosting algorithm, the changes it encompasses were motivated by the goals that one strives to achieve when employing boosting. Boosted ARTMAP is an on-line learner, it does not require excessive parameter tuning to operate, and it reduces precisely to Fuzzy ARTMAP for particular parameter values. Another architecture described in this paper is Structural Boosted ARTMAP, which uses both Boosted ART and Boosted ARTMAP to perform structural risk minimization learning. Structural Boosted ARTMAP will allow comparison of the capabilities of off-line versus on-line learning as well as empirical risk minimization versus structural risk minimization using Fuzzy ARTMAP-based neural network architectures. Both empirical and theoretical results are presented to enhance the understanding of these architectures.

Convolutional Neural Network-Based Embarrassing Situation Detection under Camera for Social Robot in Smart Homes.

PubMed

Yang, Guanci; Yang, Jing; Sheng, Weihua; Junior, Francisco Erivaldo Fernandes; Li, Shaobo

2018-05-12

Recent research has shown that the ubiquitous use of cameras and voice monitoring equipment in a home environment can raise privacy concerns and affect human mental health. This can be a major obstacle to the deployment of smart home systems for elderly or disabled care. This study uses a social robot to detect embarrassing situations. Firstly, we designed an improved neural network structure based on the You Only Look Once (YOLO) model to obtain feature information. By focusing on reducing area redundancy and computation time, we proposed a bounding-box merging algorithm based on region proposal networks (B-RPN), to merge the areas that have similar features and determine the borders of the bounding box. Thereafter, we designed a feature extraction algorithm based on our improved YOLO and B-RPN, called F-YOLO, for our training datasets, and then proposed a real-time object detection algorithm based on F-YOLO (RODA-FY). We implemented RODA-FY and compared models on our MAT social robot. Secondly, we considered six types of situations in smart homes, and developed training and validation datasets, containing 2580 and 360 images, respectively. Meanwhile, we designed three types of experiments with four types of test datasets composed of 960 sample images. Thirdly, we analyzed how a different number of training iterations affects our prediction estimation, and then we explored the relationship between recognition accuracy and learning rates. Our results show that our proposed privacy detection system can recognize designed situations in the smart home with an acceptable recognition accuracy of 94.48%. Finally, we compared the results among RODA-FY, Inception V3, and YOLO, which indicate that our proposed RODA-FY outperforms the other comparison models in recognition accuracy.

Engine cylinder pressure reconstruction using crank kinematics and recurrently-trained neural networks

NASA Astrophysics Data System (ADS)

Bennett, C.; Dunne, J. F.; Trimby, S.; Richardson, D.

2017-02-01

A recurrent non-linear autoregressive with exogenous input (NARX) neural network is proposed, and a suitable fully-recurrent training methodology is adapted and tuned, for reconstructing cylinder pressure in multi-cylinder IC engines using measured crank kinematics. This type of indirect sensing is important for cost effective closed-loop combustion control and for On-Board Diagnostics. The challenge addressed is to accurately predict cylinder pressure traces within the cycle under generalisation conditions: i.e. using data not previously seen by the network during training. This involves direct construction and calibration of a suitable inverse crank dynamic model, which owing to singular behaviour at top-dead-centre (TDC), has proved difficult via physical model construction, calibration, and inversion. The NARX architecture is specialised and adapted to cylinder pressure reconstruction, using a fully-recurrent training methodology which is needed because the alternatives are too slow and unreliable for practical network training on production engines. The fully-recurrent Robust Adaptive Gradient Descent (RAGD) algorithm, is tuned initially using synthesised crank kinematics, and then tested on real engine data to assess the reconstruction capability. Real data is obtained from a 1.125 l, 3-cylinder, in-line, direct injection spark ignition (DISI) engine involving synchronised measurements of crank kinematics and cylinder pressure across a range of steady-state speed and load conditions. The paper shows that a RAGD-trained NARX network using both crank velocity and crank acceleration as input information, provides fast and robust training. By using the optimum epoch identified during RAGD training, acceptably accurate cylinder pressures, and especially accurate location-of-peak-pressure, can be reconstructed robustly under generalisation conditions, making it the most practical NARX configuration and recurrent training methodology for use on production engines.

A novel topology control approach to maintain the node degree in dynamic wireless sensor networks.

PubMed

Huang, Yuanjiang; Martínez, José-Fernán; Díaz, Vicente Hernández; Sendra, Juana

2014-03-07

Topology control is an important technique to improve the connectivity and the reliability of Wireless Sensor Networks (WSNs) by means of adjusting the communication range of wireless sensor nodes. In this paper, a novel Fuzzy-logic Topology Control (FTC) is proposed to achieve any desired average node degree by adaptively changing communication range, thus improving the network connectivity, which is the main target of FTC. FTC is a fully localized control algorithm, and does not rely on location information of neighbors. Instead of designing membership functions and if-then rules for fuzzy-logic controller, FTC is constructed from the training data set to facilitate the design process. FTC is proved to be accurate, stable and has short settling time. In order to compare it with other representative localized algorithms (NONE, FLSS, k-Neighbor and LTRT), FTC is evaluated through extensive simulations. The simulation results show that: firstly, similar to k-Neighbor algorithm, FTC is the best to achieve the desired average node degree as node density varies; secondly, FTC is comparable to FLSS and k-Neighbor in terms of energy-efficiency, but is better than LTRT and NONE; thirdly, FTC has the lowest average maximum communication range than other algorithms, which indicates that the most energy-consuming node in the network consumes the lowest power.

Relationship between isoseismal area and magnitude of historical earthquakes in Greece by a hybrid fuzzy neural network method

NASA Astrophysics Data System (ADS)

Tselentis, G.-A.; Sokos, E.

2012-01-01

In this paper we suggest the use of diffusion-neural-networks, (neural networks with intrinsic fuzzy logic abilities) to assess the relationship between isoseismal area and earthquake magnitude for the region of Greece. It is of particular importance to study historical earthquakes for which we often have macroseismic information in the form of isoseisms but it is statistically incomplete to assess magnitudes from an isoseismal area or to train conventional artificial neural networks for magnitude estimation. Fuzzy relationships are developed and used to train a feed forward neural network with a back propagation algorithm to obtain the final relationships. Seismic intensity data from 24 earthquakes in Greece have been used. Special attention is being paid to the incompleteness and contradictory patterns in scanty historical earthquake records. The results show that the proposed processing model is very effective, better than applying classical artificial neural networks since the magnitude macroseismic intensity target function has a strong nonlinearity and in most cases the macroseismic datasets are very small.

Incremental Structured Dictionary Learning for Video Sensor-Based Object Tracking

PubMed Central

Xue, Ming; Yang, Hua; Zheng, Shibao; Zhou, Yi; Yu, Zhenghua

2014-01-01

To tackle robust object tracking for video sensor-based applications, an online discriminative algorithm based on incremental discriminative structured dictionary learning (IDSDL-VT) is presented. In our framework, a discriminative dictionary combining both positive, negative and trivial patches is designed to sparsely represent the overlapped target patches. Then, a local update (LU) strategy is proposed for sparse coefficient learning. To formulate the training and classification process, a multiple linear classifier group based on a K-combined voting (KCV) function is proposed. As the dictionary evolves, the models are also trained to timely adapt the target appearance variation. Qualitative and quantitative evaluations on challenging image sequences compared with state-of-the-art algorithms demonstrate that the proposed tracking algorithm achieves a more favorable performance. We also illustrate its relay application in visual sensor networks. PMID:24549252

Piecewise convexity of artificial neural networks.

PubMed

Rister, Blaine; Rubin, Daniel L

2017-10-01

Although artificial neural networks have shown great promise in applications including computer vision and speech recognition, there remains considerable practical and theoretical difficulty in optimizing their parameters. The seemingly unreasonable success of gradient descent methods in minimizing these non-convex functions remains poorly understood. In this work we offer some theoretical guarantees for networks with piecewise affine activation functions, which have in recent years become the norm. We prove three main results. First, that the network is piecewise convex as a function of the input data. Second, that the network, considered as a function of the parameters in a single layer, all others held constant, is again piecewise convex. Third, that the network as a function of all its parameters is piecewise multi-convex, a generalization of biconvexity. From here we characterize the local minima and stationary points of the training objective, showing that they minimize the objective on certain subsets of the parameter space. We then analyze the performance of two optimization algorithms on multi-convex problems: gradient descent, and a method which repeatedly solves a number of convex sub-problems. We prove necessary convergence conditions for the first algorithm and both necessary and sufficient conditions for the second, after introducing regularization to the objective. Finally, we remark on the remaining difficulty of the global optimization problem. Under the squared error objective, we show that by varying the training data, a single rectifier neuron admits local minima arbitrarily far apart, both in objective value and parameter space. Copyright © 2017 Elsevier Ltd. All rights reserved.

Reinforced two-step-ahead weight adjustment technique for online training of recurrent neural networks.

PubMed

Chang, Li-Chiu; Chen, Pin-An; Chang, Fi-John

2012-08-01

A reliable forecast of future events possesses great value. The main purpose of this paper is to propose an innovative learning technique for reinforcing the accuracy of two-step-ahead (2SA) forecasts. The real-time recurrent learning (RTRL) algorithm for recurrent neural networks (RNNs) can effectively model the dynamics of complex processes and has been used successfully in one-step-ahead forecasts for various time series. A reinforced RTRL algorithm for 2SA forecasts using RNNs is proposed in this paper, and its performance is investigated by two famous benchmark time series and a streamflow during flood events in Taiwan. Results demonstrate that the proposed reinforced 2SA RTRL algorithm for RNNs can adequately forecast the benchmark (theoretical) time series, significantly improve the accuracy of flood forecasts, and effectively reduce time-lag effects.

Radial basis function neural networks in non-destructive determination of compound aspirin tablets on NIR spectroscopy.

PubMed

Dou, Ying; Mi, Hong; Zhao, Lingzhi; Ren, Yuqiu; Ren, Yulin

2006-09-01

The application of the second most popular artificial neural networks (ANNs), namely, the radial basis function (RBF) networks, has been developed for quantitative analysis of drugs during the last decade. In this paper, the two components (aspirin and phenacetin) were simultaneously determined in compound aspirin tablets by using near-infrared (NIR) spectroscopy and RBF networks. The total database was randomly divided into a training set (50) and a testing set (17). Different preprocessing methods (standard normal variate (SNV), multiplicative scatter correction (MSC), first-derivative and second-derivative) were applied to two sets of NIR spectra of compound aspirin tablets with different concentrations of two active components and compared each other. After that, the performance of RBF learning algorithm adopted the nearest neighbor clustering algorithm (NNCA) and the criterion for selection used a cross-validation technique. Results show that using RBF networks to quantificationally analyze tablets is reliable, and the best RBF model was obtained by first-derivative spectra.

Evolving RBF neural networks for adaptive soft-sensor design.

PubMed

Alexandridis, Alex

2013-12-01

This work presents an adaptive framework for building soft-sensors based on radial basis function (RBF) neural network models. The adaptive fuzzy means algorithm is utilized in order to evolve an RBF network, which approximates the unknown system based on input-output data from it. The methodology gradually builds the RBF network model, based on two separate levels of adaptation: On the first level, the structure of the hidden layer is modified by adding or deleting RBF centers, while on the second level, the synaptic weights are adjusted with the recursive least squares with exponential forgetting algorithm. The proposed approach is tested on two different systems, namely a simulated nonlinear DC Motor and a real industrial reactor. The results show that the produced soft-sensors can be successfully applied to model the two nonlinear systems. A comparison with two different adaptive modeling techniques, namely a dynamic evolving neural-fuzzy inference system (DENFIS) and neural networks trained with online backpropagation, highlights the advantages of the proposed methodology.

Sample selection via angular distance in the space of the arguments of an artificial neural network

NASA Astrophysics Data System (ADS)

Fernández Jaramillo, J. M.; Mayerle, R.

2018-05-01

In the construction of an artificial neural network (ANN) a proper data splitting of the available samples plays a major role in the training process. This selection of subsets for training, testing and validation affects the generalization ability of the neural network. Also the number of samples has an impact in the time required for the design of the ANN and the training. This paper introduces an efficient and simple method for reducing the set of samples used for training a neural network. The method reduces the required time to calculate the network coefficients, while keeping the diversity and avoiding overtraining the ANN due the presence of similar samples. The proposed method is based on the calculation of the angle between two vectors, each one representing one input of the neural network. When the angle formed among samples is smaller than a defined threshold only one input is accepted for the training. The accepted inputs are scattered throughout the sample space. Tidal records are used to demonstrate the proposed method. The results of a cross-validation show that with few inputs the quality of the outputs is not accurate and depends on the selection of the first sample, but as the number of inputs increases the accuracy is improved and differences among the scenarios with a different starting sample have and important reduction. A comparison with the K-means clustering algorithm shows that for this application the proposed method with a smaller number of samples is producing a more accurate network.

Metaheuristic and Machine Learning Models for TFE-731-2, PW4056, and JT8D-9 Cruise Thrust

NASA Astrophysics Data System (ADS)

Baklacioglu, Tolga

2017-08-01

The requirement for an accurate engine thrust model has a major antecedence in airline fuel saving programs, assessment of environmental effects of fuel consumption, emissions reduction studies, and air traffic management applications. In this study, utilizing engine manufacturers' real data, a metaheuristic model based on genetic algorithms (GAs) and a machine learning model based on neural networks (NNs) trained with Levenberg-Marquardt (LM), delta-bar-delta (DBD), and conjugate gradient (CG) algorithms were accomplished to incorporate the effect of both flight altitude and Mach number in the estimation of thrust. For the GA model, the analysis of population size impact on the model's accuracy and effect of number of data on model coefficients were also performed. For the NN model, design of optimum topology was searched for one- and two-hidden-layer networks. Predicted thrust values presented a close agreement with real thrust data for both models, among which LM trained NNs gave the best accuracies.

Propagation based phase retrieval of simulated intensity measurements using artificial neural networks

NASA Astrophysics Data System (ADS)

Kemp, Z. D. C.

2018-04-01

Determining the phase of a wave from intensity measurements has many applications in fields such as electron microscopy, visible light optics, and medical imaging. Propagation based phase retrieval, where the phase is obtained from defocused images, has shown significant promise. There are, however, limitations in the accuracy of the retrieved phase arising from such methods. Sources of error include shot noise, image misalignment, and diffraction artifacts. We explore the use of artificial neural networks (ANNs) to improve the accuracy of propagation based phase retrieval algorithms applied to simulated intensity measurements. We employ a phase retrieval algorithm based on the transport-of-intensity equation to obtain the phase from simulated micrographs of procedurally generated specimens. We then train an ANN with pairs of retrieved and exact phases, and use the trained ANN to process a test set of retrieved phase maps. The total error in the phase is significantly reduced using this method. We also discuss a variety of potential extensions to this work.

Optimal design approach for heating irregular-shaped objects in three-dimensional radiant furnaces using a hybrid genetic algorithm-artificial neural network method

NASA Astrophysics Data System (ADS)

Darvishvand, Leila; Kamkari, Babak; Kowsary, Farshad

2018-03-01

In this article, a new hybrid method based on the combination of the genetic algorithm (GA) and artificial neural network (ANN) is developed to optimize the design of three-dimensional (3-D) radiant furnaces. A 3-D irregular shape design body (DB) heated inside a 3-D radiant furnace is considered as a case study. The uniform thermal conditions on the DB surfaces are obtained by minimizing an objective function. An ANN is developed to predict the objective function value which is trained through the data produced by applying the Monte Carlo method. The trained ANN is used in conjunction with the GA to find the optimal design variables. The results show that the computational time using the GA-ANN approach is significantly less than that of the conventional method. It is concluded that the integration of the ANN with GA is an efficient technique for optimization of the radiant furnaces.

Detection of single and multilayer clouds in an artificial neural network approach

NASA Astrophysics Data System (ADS)

Sun-Mack, Sunny; Minnis, Patrick; Smith, William L.; Hong, Gang; Chen, Yan

2017-10-01

Determining whether a scene observed with a satellite imager is composed of a thin cirrus over a water cloud or thick cirrus contiguous with underlying layers of ice and water clouds is often difficult because of similarities in the observed radiance values. In this paper an artificial neural network (ANN) algorithm, employing several Aqua MODIS infrared channels and the retrieved total cloud visible optical depth, is trained to detect multilayer ice-over-water cloud systems as identified by matched April 2009 CloudSat and CALIPSO (CC) data. The CC lidar and radar profiles provide the vertical structure that serves as output truth for a multilayer ANN, or MLANN, algorithm. Applying the trained MLANN to independent July 2008 MODIS data resulted in a combined ML and single layer hit rate of 75% (72%) for nonpolar regions during the day (night). The results are comparable to or more accurate than currently available methods. Areas of improvement are identified and will be addressed in future versions of the MLANN.

An Indoor Pedestrian Positioning Method Using HMM with a Fuzzy Pattern Recognition Algorithm in a WLAN Fingerprint System

PubMed Central

Ni, Yepeng; Liu, Jianbo; Liu, Shan; Bai, Yaxin

2016-01-01

With the rapid development of smartphones and wireless networks, indoor location-based services have become more and more prevalent. Due to the sophisticated propagation of radio signals, the Received Signal Strength Indicator (RSSI) shows a significant variation during pedestrian walking, which introduces critical errors in deterministic indoor positioning. To solve this problem, we present a novel method to improve the indoor pedestrian positioning accuracy by embedding a fuzzy pattern recognition algorithm into a Hidden Markov Model. The fuzzy pattern recognition algorithm follows the rule that the RSSI fading has a positive correlation to the distance between the measuring point and the AP location even during a dynamic positioning measurement. Through this algorithm, we use the RSSI variation trend to replace the specific RSSI value to achieve a fuzzy positioning. The transition probability of the Hidden Markov Model is trained by the fuzzy pattern recognition algorithm with pedestrian trajectories. Using the Viterbi algorithm with the trained model, we can obtain a set of hidden location states. In our experiments, we demonstrate that, compared with the deterministic pattern matching algorithm, our method can greatly improve the positioning accuracy and shows robust environmental adaptability. PMID:27618053

An assessment of support vector machines for land cover classification

USGS Publications Warehouse

Huang, C.; Davis, L.S.; Townshend, J.R.G.

2002-01-01

The support vector machine (SVM) is a group of theoretically superior machine learning algorithms. It was found competitive with the best available machine learning algorithms in classifying high-dimensional data sets. This paper gives an introduction to the theoretical development of the SVM and an experimental evaluation of its accuracy, stability and training speed in deriving land cover classifications from satellite images. The SVM was compared to three other popular classifiers, including the maximum likelihood classifier (MLC), neural network classifiers (NNC) and decision tree classifiers (DTC). The impacts of kernel configuration on the performance of the SVM and of the selection of training data and input variables on the four classifiers were also evaluated in this experiment.

A neural network construction method for surrogate modeling of physics-based analysis

NASA Astrophysics Data System (ADS)

Sung, Woong Je

In this thesis existing methodologies related to the developmental methods of neural networks have been surveyed and their approaches to network sizing and structuring are carefully observed. This literature review covers the constructive methods, the pruning methods, and the evolutionary methods and questions about the basic assumption intrinsic to the conventional neural network learning paradigm, which is primarily devoted to optimization of connection weights (or synaptic strengths) for the pre-determined connection structure of the network. The main research hypothesis governing this thesis is that, without breaking a prevailing dichotomy between weights and connectivity of the network during learning phase, the efficient design of a task-specific neural network is hard to achieve because, as long as connectivity and weights are searched by separate means, a structural optimization of the neural network requires either repetitive re-training procedures or computationally expensive topological meta-search cycles. The main contribution of this thesis is designing and testing a novel learning mechanism which efficiently learns not only weight parameters but also connection structure from a given training data set, and positioning this learning mechanism within the surrogate modeling practice. In this work, a simple and straightforward extension to the conventional error Back-Propagation (BP) algorithm has been formulated to enable a simultaneous learning for both connectivity and weights of the Generalized Multilayer Perceptron (GMLP) in supervised learning tasks. A particular objective is to achieve a task-specific network having reasonable generalization performance with a minimal training time. The dichotomy between architectural design and weight optimization is reconciled by a mechanism establishing a new connection for a neuron pair which has potentially higher error-gradient than one of the existing connections. Interpreting an instance of the absence of connection as a zero-weight connection, the potential contribution to training error reduction of any present or absent connection can readily be evaluated using the BP algorithm. Instead of being broken, the connections that contribute less remain frozen with constant weight values optimized to that point but they are excluded from further weight optimization until reselected. In this way, a selective weight optimization is executed only for the dynamically maintained pool of high gradient connections. By searching the rapidly changing weights and concentrating optimization resources on them, the learning process is accelerated without either a significant increase in computational cost or a need for re-training. This results in a more task-adapted network connection structure. Combined with another important criterion for the division of a neuron which adds a new computational unit to a network, a highly fitted network can be grown out of the minimal random structure. This particular learning strategy can belong to a more broad class of the variable connectivity learning scheme and the devised algorithm has been named Optimal Brain Growth (OBG). The OBG algorithm has been tested on two canonical problems; a regression analysis using the Complicated Interaction Regression Function and a classification of the Two-Spiral Problem. A comparative study with conventional Multilayer Perceptrons (MLPs) consisting of single- and double-hidden layers shows that OBG is less sensitive to random initial conditions and generalizes better with only a minimal increase in computational time. This partially proves that a variable connectivity learning scheme has great potential to enhance computational efficiency and reduce efforts to select proper network architecture. To investigate the applicability of the OBG to more practical surrogate modeling tasks, the geometry-to-pressure mapping of a particular class of airfoils in the transonic flow regime has been sought using both the conventional MLP networks with pre-defined architecture and the OBG-developed networks started from the same initial MLP networks. Considering wide variety in airfoil geometry and diversity of flow conditions distributed over a range of flow Mach numbers and angles of attack, the new method shows a great potential to capture fundamentally nonlinear flow phenomena especially related to the occurrence of shock waves on airfoil surfaces in transonic flow regime. (Abstract shortened by UMI.).

An "intelligent" approach based on side-by-side cascade-correlation neural networks for estimating thermophysical properties from photothermal responses

NASA Astrophysics Data System (ADS)

Grieu, Stéphane; Faugeroux, Olivier; Traoré, Adama; Claudet, Bernard; Bodnar, Jean-Luc

2015-01-01

In the present paper, an artificial-intelligence-based approach dealing with the estimation of thermophysical properties is designed and evaluated. This new and "intelligent" approach makes use of photothermal responses obtained when subjecting materials to a light flux. So, the main objective of the present work was to estimate simultaneously both the thermal diffusivity and conductivity of materials, from front-face or rear-face photothermal responses to pseudo random binary signals. To this end, we used side-by-side feedforward neural networks trained with the cascade-correlation algorithm. In addition, computation time was a key point to consider. That is why the developed algorithms are computationally tractable.

Efficient Online Learning Algorithms Based on LSTM Neural Networks.

PubMed

Ergen, Tolga; Kozat, Suleyman Serdar

2017-09-13

We investigate online nonlinear regression and introduce novel regression structures based on the long short term memory (LSTM) networks. For the introduced structures, we also provide highly efficient and effective online training methods. To train these novel LSTM-based structures, we put the underlying architecture in a state space form and introduce highly efficient and effective particle filtering (PF)-based updates. We also provide stochastic gradient descent and extended Kalman filter-based updates. Our PF-based training method guarantees convergence to the optimal parameter estimation in the mean square error sense provided that we have a sufficient number of particles and satisfy certain technical conditions. More importantly, we achieve this performance with a computational complexity in the order of the first-order gradient-based methods by controlling the number of particles. Since our approach is generic, we also introduce a gated recurrent unit (GRU)-based approach by directly replacing the LSTM architecture with the GRU architecture, where we demonstrate the superiority of our LSTM-based approach in the sequential prediction task via different real life data sets. In addition, the experimental results illustrate significant performance improvements achieved by the introduced algorithms with respect to the conventional methods over several different benchmark real life data sets.

Application of artificial neural networks to composite ply micromechanics

NASA Technical Reports Server (NTRS)

Brown, D. A.; Murthy, P. L. N.; Berke, L.

1991-01-01

Artificial neural networks can provide improved computational efficiency relative to existing methods when an algorithmic description of functional relationships is either totally unavailable or is complex in nature. For complex calculations, significant reductions in elapsed computation time are possible. The primary goal is to demonstrate the applicability of artificial neural networks to composite material characterization. As a test case, a neural network was trained to accurately predict composite hygral, thermal, and mechanical properties when provided with basic information concerning the environment, constituent materials, and component ratios used in the creation of the composite. A brief introduction on neural networks is provided along with a description of the project itself.

Genetic training of network using chaos concept: application to QSAR studies of vibration modes of tetrahedral halides.

PubMed

Lu, Qingzhang; Shen, Guoli; Yu, Ruqin

2002-11-15

The chaotic dynamical system is introduced in genetic algorithm to train ANN to formulate the CGANN algorithm. Logistic mapping as one of the most important chaotic dynamic mappings provides each new generation a high chance to hold GA's population diversity. This enhances the ability to overcome overfitting in training an ANN. The proposed CGANN has been used for QSAR studies to predict the tetrahedral modes (nu(1)(A1) and nu(2)(E)) of halides [MX(4)](epsilon). The frequencies predicted by QSAR were compared with those calculated by quantum chemistry methods including PM3, AM1, and MNDO/d. The possibility of improving the predictive ability of QSAR by including quantum chemistry parameters as feature variables has been investigated using tetrahedral tetrahalide examples. Copyright 2002 Wiley Periodicals, Inc.

Mapping soil landscape as spatial continua: The Neural Network Approach

NASA Astrophysics Data System (ADS)

Zhu, A.-Xing

2000-03-01

A neural network approach was developed to populate a soil similarity model that was designed to represent soil landscape as spatial continua for hydroecological modeling at watersheds of mesoscale size. The approach employs multilayer feed forward neural networks. The input to the network was data on a set of soil formative environmental factors; the output from the network was a set of similarity values to a set of prescribed soil classes. The network was trained using a conjugate gradient algorithm in combination with a simulated annealing technique to learn the relationships between a set of prescribed soils and their environmental factors. Once trained, the network was used to compute for every location in an area the similarity values of the soil to the set of prescribed soil classes. The similarity values were then used to produce detailed soil spatial information. The approach also included a Geographic Information System procedure for selecting representative training and testing samples and a process of determining the network internal structure. The approach was applied to soil mapping in a watershed, the Lubrecht Experimental Forest, in western Montana. The case study showed that the soil spatial information derived using the neural network approach reveals much greater spatial detail and has a higher quality than that derived from the conventional soil map. Implications of this detailed soil spatial information for hydroecological modeling at the watershed scale are also discussed.

Automated red blood cells extraction from holographic images using fully convolutional neural networks.

PubMed

Yi, Faliu; Moon, Inkyu; Javidi, Bahram

2017-10-01

In this paper, we present two models for automatically extracting red blood cells (RBCs) from RBCs holographic images based on a deep learning fully convolutional neural network (FCN) algorithm. The first model, called FCN-1, only uses the FCN algorithm to carry out RBCs prediction, whereas the second model, called FCN-2, combines the FCN approach with the marker-controlled watershed transform segmentation scheme to achieve RBCs extraction. Both models achieve good segmentation accuracy. In addition, the second model has much better performance in terms of cell separation than traditional segmentation methods. In the proposed methods, the RBCs phase images are first numerically reconstructed from RBCs holograms recorded with off-axis digital holographic microscopy. Then, some RBCs phase images are manually segmented and used as training data to fine-tune the FCN. Finally, each pixel in new input RBCs phase images is predicted into either foreground or background using the trained FCN models. The RBCs prediction result from the first model is the final segmentation result, whereas the result from the second model is used as the internal markers of the marker-controlled transform algorithm for further segmentation. Experimental results show that the given schemes can automatically extract RBCs from RBCs phase images and much better RBCs separation results are obtained when the FCN technique is combined with the marker-controlled watershed segmentation algorithm.

Automated red blood cells extraction from holographic images using fully convolutional neural networks

PubMed Central

Yi, Faliu; Moon, Inkyu; Javidi, Bahram

2017-01-01

In this paper, we present two models for automatically extracting red blood cells (RBCs) from RBCs holographic images based on a deep learning fully convolutional neural network (FCN) algorithm. The first model, called FCN-1, only uses the FCN algorithm to carry out RBCs prediction, whereas the second model, called FCN-2, combines the FCN approach with the marker-controlled watershed transform segmentation scheme to achieve RBCs extraction. Both models achieve good segmentation accuracy. In addition, the second model has much better performance in terms of cell separation than traditional segmentation methods. In the proposed methods, the RBCs phase images are first numerically reconstructed from RBCs holograms recorded with off-axis digital holographic microscopy. Then, some RBCs phase images are manually segmented and used as training data to fine-tune the FCN. Finally, each pixel in new input RBCs phase images is predicted into either foreground or background using the trained FCN models. The RBCs prediction result from the first model is the final segmentation result, whereas the result from the second model is used as the internal markers of the marker-controlled transform algorithm for further segmentation. Experimental results show that the given schemes can automatically extract RBCs from RBCs phase images and much better RBCs separation results are obtained when the FCN technique is combined with the marker-controlled watershed segmentation algorithm. PMID:29082078

SortNet: learning to rank by a neural preference function.

PubMed

Rigutini, Leonardo; Papini, Tiziano; Maggini, Marco; Scarselli, Franco

2011-09-01

Relevance ranking consists in sorting a set of objects with respect to a given criterion. However, in personalized retrieval systems, the relevance criteria may usually vary among different users and may not be predefined. In this case, ranking algorithms that adapt their behavior from users' feedbacks must be devised. Two main approaches are proposed in the literature for learning to rank: the use of a scoring function, learned by examples, that evaluates a feature-based representation of each object yielding an absolute relevance score, a pairwise approach, where a preference function is learned to determine the object that has to be ranked first in a given pair. In this paper, we present a preference learning method for learning to rank. A neural network, the comparative neural network (CmpNN), is trained from examples to approximate the comparison function for a pair of objects. The CmpNN adopts a particular architecture designed to implement the symmetries naturally present in a preference function. The learned preference function can be embedded as the comparator into a classical sorting algorithm to provide a global ranking of a set of objects. To improve the ranking performances, an active-learning procedure is devised, that aims at selecting the most informative patterns in the training set. The proposed algorithm is evaluated on the LETOR dataset showing promising performances in comparison with other state-of-the-art algorithms.

Using electronic medical records to increase the efficiency of catheter-associated urinary tract infection surveillance for National Health and Safety Network reporting.

PubMed

Shepard, John; Hadhazy, Eric; Frederick, John; Nicol, Spencer; Gade, Padmaja; Cardon, Andrew; Wilson, Jorge; Vetteth, Yohan; Madison, Sasha

2014-03-01

Streamlining health care-associated infection surveillance is essential for health care facilities owing to the continuing increases in reporting requirements. Stanford Hospital, a 583-bed adult tertiary care center, used their electronic medical record (EMR) to develop an electronic algorithm to reduce the time required to conduct catheter-associated urinary tract infection (CAUTI) surveillance in adults. The algorithm provides inclusion and exclusion criteria, using the National Healthcare Safety Network definitions, for patients with a CAUTI. The algorithm was validated by trained infection preventionists through complete chart review for a random sample of cultures collected during the study period, September 1, 2012, to February 28, 2013. During the study period, a total of 6,379 positive urine cultures were identified. The Stanford Hospital electronic CAUTI algorithm identified 6,101 of these positive cultures (95.64%) as not a CAUTI, 191 (2.99%) as a possible CAUTI requiring further validation, and 87 (1.36%) as a definite CAUTI. Overall, use of the algorithm reduced CAUTI surveillance requirements at Stanford Hospital by 97.01%. The electronic algorithm proved effective in increasing the efficiency of CAUTI surveillance. The data suggest that CAUTI surveillance using the National Healthcare Safety Network definitions can be fully automated. Copyright © 2014 Association for Professionals in Infection Control and Epidemiology, Inc. All rights reserved.

Thermalnet: a Deep Convolutional Network for Synthetic Thermal Image Generation

NASA Astrophysics Data System (ADS)

Kniaz, V. V.; Gorbatsevich, V. S.; Mizginov, V. A.

2017-05-01

Deep convolutional neural networks have dramatically changed the landscape of the modern computer vision. Nowadays methods based on deep neural networks show the best performance among image recognition and object detection algorithms. While polishing of network architectures received a lot of scholar attention, from the practical point of view the preparation of a large image dataset for a successful training of a neural network became one of major challenges. This challenge is particularly profound for image recognition in wavelengths lying outside the visible spectrum. For example no infrared or radar image datasets large enough for successful training of a deep neural network are available to date in public domain. Recent advances of deep neural networks prove that they are also capable to do arbitrary image transformations such as super-resolution image generation, grayscale image colorisation and imitation of style of a given artist. Thus a natural question arise: how could be deep neural networks used for augmentation of existing large image datasets? This paper is focused on the development of the Thermalnet deep convolutional neural network for augmentation of existing large visible image datasets with synthetic thermal images. The Thermalnet network architecture is inspired by colorisation deep neural networks.

Adaptive Optimization of Aircraft Engine Performance Using Neural Networks

NASA Technical Reports Server (NTRS)

Simon, Donald L.; Long, Theresa W.

1995-01-01

Preliminary results are presented on the development of an adaptive neural network based control algorithm to enhance aircraft engine performance. This work builds upon a previous National Aeronautics and Space Administration (NASA) effort known as Performance Seeking Control (PSC). PSC is an adaptive control algorithm which contains a model of the aircraft's propulsion system which is updated on-line to match the operation of the aircraft's actual propulsion system. Information from the on-line model is used to adapt the control system during flight to allow optimal operation of the aircraft's propulsion system (inlet, engine, and nozzle) to improve aircraft engine performance without compromising reliability or operability. Performance Seeking Control has been shown to yield reductions in fuel flow, increases in thrust, and reductions in engine fan turbine inlet temperature. The neural network based adaptive control, like PSC, will contain a model of the propulsion system which will be used to calculate optimal control commands on-line. Hopes are that it will be able to provide some additional benefits above and beyond those of PSC. The PSC algorithm is computationally intensive, it is valid only at near steady-state flight conditions, and it has no way to adapt or learn on-line. These issues are being addressed in the development of the optimal neural controller. Specialized neural network processing hardware is being developed to run the software, the algorithm will be valid at steady-state and transient conditions, and will take advantage of the on-line learning capability of neural networks. Future plans include testing the neural network software and hardware prototype against an aircraft engine simulation. In this paper, the proposed neural network software and hardware is described and preliminary neural network training results are presented.

The fatigue life prediction of aluminium alloy using genetic algorithm and neural network

NASA Astrophysics Data System (ADS)

Susmikanti, Mike

2013-09-01

The behavior of the fatigue life of the industrial materials is very important. In many cases, the material with experiencing fatigue life cannot be avoided, however, there are many ways to control their behavior. Many investigations of the fatigue life phenomena of alloys have been done, but it is high cost and times consuming computation. This paper report the modeling and simulation approaches to predict the fatigue life behavior of Aluminum Alloys and resolves some problems of computation. First, the simulation using genetic algorithm was utilized to optimize the load to obtain the stress values. These results can be used to provide N-cycle fatigue life of the material. Furthermore, the experimental data was applied as input data in the neural network learning, while the samples data were applied for testing of the training data. Finally, the multilayer perceptron algorithm is applied to predict whether the given data sets in accordance with the fatigue life of the alloy. To achieve rapid convergence, the Levenberg-Marquardt algorithm was also employed. The simulations results shows that the fatigue behaviors of aluminum under pressure can be predicted. In addition, implementation of neural networks successfully identified a model for material fatigue life.

The Passive Microwave Neural Network Precipitation Retrieval (PNPR) for AMSU/MHS and ATMS cross-track scanning radiometers

NASA Astrophysics Data System (ADS)

Sano', Paolo; Casella, Daniele; Panegrossi, Giulia; Cinzia Marra, Anna; Dietrich, Stefano

2016-04-01

Spaceborne microwave cross-track scanning radiometers, originally developed for temperature and humidity sounding, have shown great capabilities to provide a significant contribution in precipitation monitoring both in terms of measurement quality and spatial/temporal coverage. The Passive microwave Neural network Precipitation Retrieval (PNPR) algorithm for cross-track scanning radiometers, originally developed for the Advanced Microwave Sounding Unit/Microwave Humidity Sounder (AMSU-A/MHS) radiometers (on board the European MetOp and U.S. NOAA satellites), was recently newly designed to exploit the Advanced Technology Microwave Sounder (ATMS) on board the Suomi-NPP satellite and the future JPSS satellites. The PNPR algorithm is based on the Artificial Neural Network (ANN) approach. The main PNPR-ATMS algorithm changes with respect to PNPR-AMSU/MHS are the design and implementation of a new ANN able to manage the information derived from the additional ATMS channels (respect to the AMSU-A/MHS radiometer) and a new screening procedure for not-precipitating pixels. In order to achieve maximum consistency of the retrieved surface precipitation, both PNPR algorithms are based on the same physical foundation. The PNPR is optimized for the European and the African area. The neural network was trained using a cloud-radiation database built upon 94 cloud-resolving simulations over Europe and the Mediterranean and over the African area and radiative transfer model simulations of TB vectors consistent with the AMSU-A/MHS and ATMS channel frequencies, viewing angles, and view-angle dependent IFOV sizes along the scan projections. As opposed to other ANN precipitation retrieval algorithms, PNPR uses a unique ANN that retrieves the surface precipitation rate for all types of surface backgrounds represented in the training database, i.e., land (vegetated or arid), ocean, snow/ice or coast. This approach prevents different precipitation estimates from being inconsistent with one another when an observed precipitation system extends over two or more types of surfaces. As input data, the PNPR algorithm incorporates the TBs from selected channels, and various additional TBs-derived variables. Ancillary geographical/geophysical inputs (i.e., latitude, terrain height, surface type, season) are also considered during the training phase. The PNPR algorithm outputs consist of both the surface precipitation rate (along with the information on precipitation phase: liquid, mixed, solid) and a pixel-based quality index. We will illustrate the main features of the PNPR algorithm and will show results of a verification study over Europe and Africa. The study is based on the available ground-based radar and/or rain gauge network observations over the European area. In addition, results of the comparison with rainfall products available from the NASA/JAXA Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) (over the African area) and Global Precipitation Measurement (GPM) Dual frequency Precipitation Radar (DPR) will be shown. The analysis is built upon a two-years coincidence dataset of AMSU/MHS and ATMS observations with PR (2013-2014) and DPR (2014-2015). The PNPR is developed within the EUMETSAT H/SAF program (Satellite Application Facility for Operational Hydrology and Water Management), where it is used operationally towards the full exploitation of all microwave radiometers available in the GPM era. The algorithm will be tailored to the future European Microwave Sounder (MWS) onboard the MetOp-Second Generation (MetOp-SG) satellites.

A stable second order method for training back propagation networks

NASA Technical Reports Server (NTRS)

Nachtsheim, Philip R.

1993-01-01

A simple method for improving the learning rate of the back-propagation algorithm is described. The basis of the method is that approximate second order corrections can be incorporated in the output units. The extended method leads to significant improvements in the convergence rate.

A convolutional neural network-based screening tool for X-ray serial crystallography

PubMed Central

Ke, Tsung-Wei; Brewster, Aaron S.; Yu, Stella X.; Ushizima, Daniela; Yang, Chao; Sauter, Nicholas K.

2018-01-01

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization. PMID:29714177

A convolutional neural network-based screening tool for X-ray serial crystallography.

PubMed

Ke, Tsung Wei; Brewster, Aaron S; Yu, Stella X; Ushizima, Daniela; Yang, Chao; Sauter, Nicholas K

2018-05-01

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization. open access.

A convolutional neural network-based screening tool for X-ray serial crystallography

DOE PAGES

Ke, Tsung-Wei; Brewster, Aaron S.; Yu, Stella X.; ...

2018-04-24

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization.

A convolutional neural network-based screening tool for X-ray serial crystallography

DOE Office of Scientific and Technical Information (OSTI.GOV)

Ke, Tsung-Wei; Brewster, Aaron S.; Yu, Stella X.

A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization.

Hybrid ANN optimized artificial fish swarm algorithm based classifier for classification of suspicious lesions in breast DCE-MRI

NASA Astrophysics Data System (ADS)

Janaki Sathya, D.; Geetha, K.

2017-12-01

Automatic mass or lesion classification systems are developed to aid in distinguishing between malignant and benign lesions present in the breast DCE-MR images, the systems need to improve both the sensitivity and specificity of DCE-MR image interpretation in order to be successful for clinical use. A new classifier (a set of features together with a classification method) based on artificial neural networks trained using artificial fish swarm optimization (AFSO) algorithm is proposed in this paper. The basic idea behind the proposed classifier is to use AFSO algorithm for searching the best combination of synaptic weights for the neural network. An optimal set of features based on the statistical textural features is presented. The investigational outcomes of the proposed suspicious lesion classifier algorithm therefore confirm that the resulting classifier performs better than other such classifiers reported in the literature. Therefore this classifier demonstrates that the improvement in both the sensitivity and specificity are possible through automated image analysis.

Network intrusion detection based on a general regression neural network optimized by an improved artificial immune algorithm.

PubMed

Wu, Jianfa; Peng, Dahao; Li, Zhuping; Zhao, Li; Ling, Huanzhang

2015-01-01

To effectively and accurately detect and classify network intrusion data, this paper introduces a general regression neural network (GRNN) based on the artificial immune algorithm with elitist strategies (AIAE). The elitist archive and elitist crossover were combined with the artificial immune algorithm (AIA) to produce the AIAE-GRNN algorithm, with the aim of improving its adaptivity and accuracy. In this paper, the mean square errors (MSEs) were considered the affinity function. The AIAE was used to optimize the smooth factors of the GRNN; then, the optimal smooth factor was solved and substituted into the trained GRNN. Thus, the intrusive data were classified. The paper selected a GRNN that was separately optimized using a genetic algorithm (GA), particle swarm optimization (PSO), and fuzzy C-mean clustering (FCM) to enable a comparison of these approaches. As shown in the results, the AIAE-GRNN achieves a higher classification accuracy than PSO-GRNN, but the running time of AIAE-GRNN is long, which was proved first. FCM and GA-GRNN were eliminated because of their deficiencies in terms of accuracy and convergence. To improve the running speed, the paper adopted principal component analysis (PCA) to reduce the dimensions of the intrusive data. With the reduction in dimensionality, the PCA-AIAE-GRNN decreases in accuracy less and has better convergence than the PCA-PSO-GRNN, and the running speed of the PCA-AIAE-GRNN was relatively improved. The experimental results show that the AIAE-GRNN has a higher robustness and accuracy than the other algorithms considered and can thus be used to classify the intrusive data.

Segmentation of organs-at-risks in head and neck CT images using convolutional neural networks

PubMed Central

Ibragimov, Bulat; Xing, Lei

2017-01-01

Purpose Accurate segmentation of organs-at-risks (OARs) is the key step for efficient planning of radiation therapy for head and neck (HaN) cancer treatment. In the work, we proposed the first deep learning-based algorithm, for segmentation of OARs in HaN CT images, and compared its performance against state-of-the-art automated segmentation algorithms, commercial software and inter-observer variability. Methods Convolutional neural networks (CNNs) – a concept from the field of deep learning – were used to study consistent intensity patterns of OARs from training CT images and to segment the OAR in a previously unseen test CT image. For CNN training, we extracted a representative number of positive intensity patches around voxels that belong to the OAR of interest in training CT images, and negative intensity patches around voxels that belong to the surrounding structures. These patches then passed through a sequence of CNN layers that captured local image features such as corners, end-points and edges, and combined them into more complex high-order features that can efficiently describe the OAR. The trained network was applied to classify voxels in a region of interest in the test image where the corresponding OAR is expected to be located. We then smoothed the obtained classification results by using Markov random fields algorithm. We finally extracted the largest connected component of the smoothed voxels classified as the OAR by CNN, performed dilate-erode operations to remov cavities of the component, which resulted in segmentation of the OAR in the test image. Results The performance of CNNs was validated on segmentation of spinal cord, mandible, parotid glands, submandibular glands, larynx, pharynx, eye globes, optic nerves and optic chiasm using 50 CT images. The obtained segmentation results varied from 37.4% Dice coefficient (DSC) for chiasm to 89.5% DSC for mandible. We also analyzed the performance of state-of-the-art algorithms and commercial software reported in the literature, and observed that CNNs demonstrate similar or superior performance on segmentation of spinal cord, mandible, parotid glands, larynx, pharynx, eye globes and optic nerves, but inferior performance on segmentation of submandibular glands and optic chiasm. Conclusion We concluded that convolution neural networks can accurately segment most of OARs using a representative database of 50 HaN CT images. At the same time, inclusion of additional information, e.g. MR images, may be beneficial for some OARs with poorly-visible boundaries. PMID:28205307

An incremental approach to genetic-algorithms-based classification.

PubMed

Guan, Sheng-Uei; Zhu, Fangming

2005-04-01

Incremental learning has been widely addressed in the machine learning literature to cope with learning tasks where the learning environment is ever changing or training samples become available over time. However, most research work explores incremental learning with statistical algorithms or neural networks, rather than evolutionary algorithms. The work in this paper employs genetic algorithms (GAs) as basic learning algorithms for incremental learning within one or more classifier agents in a multiagent environment. Four new approaches with different initialization schemes are proposed. They keep the old solutions and use an "integration" operation to integrate them with new elements to accommodate new attributes, while biased mutation and crossover operations are adopted to further evolve a reinforced solution. The simulation results on benchmark classification data sets show that the proposed approaches can deal with the arrival of new input attributes and integrate them with the original input space. It is also shown that the proposed approaches can be successfully used for incremental learning and improve classification rates as compared to the retraining GA. Possible applications for continuous incremental training and feature selection are also discussed.

Real-Time Adaptive Color Segmentation by Neural Networks

NASA Technical Reports Server (NTRS)

Duong, Tuan A.

2004-01-01

Artificial neural networks that would utilize the cascade error projection (CEP) algorithm have been proposed as means of autonomous, real-time, adaptive color segmentation of images that change with time. In the original intended application, such a neural network would be used to analyze digitized color video images of terrain on a remote planet as viewed from an uninhabited spacecraft approaching the planet. During descent toward the surface of the planet, information on the segmentation of the images into differently colored areas would be updated adaptively in real time to capture changes in contrast, brightness, and resolution, all in an effort to identify a safe and scientifically productive landing site and provide control feedback to steer the spacecraft toward that site. Potential terrestrial applications include monitoring images of crops to detect insect invasions and monitoring of buildings and other facilities to detect intruders. The CEP algorithm is reliable and is well suited to implementation in very-large-scale integrated (VLSI) circuitry. It was chosen over other neural-network learning algorithms because it is better suited to realtime learning: It provides a self-evolving neural-network structure, requires fewer iterations to converge and is more tolerant to low resolution (that is, fewer bits) in the quantization of neural-network synaptic weights. Consequently, a CEP neural network learns relatively quickly, and the circuitry needed to implement it is relatively simple. Like other neural networks, a CEP neural network includes an input layer, hidden units, and output units (see figure). As in other neural networks, a CEP network is presented with a succession of input training patterns, giving rise to a set of outputs that are compared with the desired outputs. Also as in other neural networks, the synaptic weights are updated iteratively in an effort to bring the outputs closer to target values. A distinctive feature of the CEP neural network and algorithm is that each update of synaptic weights takes place in conjunction with the addition of another hidden unit, which then remains in place as still other hidden units are added on subsequent iterations. For a given training pattern, the synaptic weight between (1) the inputs and the previously added hidden units and (2) the newly added hidden unit is updated by an amount proportional to the partial derivative of a quadratic error function with respect to the synaptic weight. The synaptic weight between the newly added hidden unit and each output unit is given by a more complex function that involves the errors between the outputs and their target values, the transfer functions (hyperbolic tangents) of the neural units, and the derivatives of the transfer functions.

Segmentation of dermoscopy images using wavelet networks.

PubMed

Sadri, Amir Reza; Zekri, Maryam; Sadri, Saeed; Gheissari, Niloofar; Mokhtari, Mojgan; Kolahdouzan, Farzaneh

2013-04-01

This paper introduces a new approach for the segmentation of skin lesions in dermoscopic images based on wavelet network (WN). The WN presented here is a member of fixed-grid WNs that is formed with no need of training. In this WN, after formation of wavelet lattice, determining shift and scale parameters of wavelets with two screening stage and selecting effective wavelets, orthogonal least squares algorithm is used to calculate the network weights and to optimize the network structure. The existence of two stages of screening increases globality of the wavelet lattice and provides a better estimation of the function especially for larger scales. R, G, and B values of a dermoscopy image are considered as the network inputs and the network structure formation. Then, the image is segmented and the skin lesions exact boundary is determined accordingly. The segmentation algorithm were applied to 30 dermoscopic images and evaluated with 11 different metrics, using the segmentation result obtained by a skilled pathologist as the ground truth. Experimental results show that our method acts more effectively in comparison with some modern techniques that have been successfully used in many medical imaging problems.

Online learning algorithm for time series forecasting suitable for low cost wireless sensor networks nodes.

PubMed

Pardo, Juan; Zamora-Martínez, Francisco; Botella-Rocamora, Paloma

2015-04-21

Time series forecasting is an important predictive methodology which can be applied to a wide range of problems. Particularly, forecasting the indoor temperature permits an improved utilization of the HVAC (Heating, Ventilating and Air Conditioning) systems in a home and thus a better energy efficiency. With such purpose the paper describes how to implement an Artificial Neural Network (ANN) algorithm in a low cost system-on-chip to develop an autonomous intelligent wireless sensor network. The present paper uses a Wireless Sensor Networks (WSN) to monitor and forecast the indoor temperature in a smart home, based on low resources and cost microcontroller technology as the 8051MCU. An on-line learning approach, based on Back-Propagation (BP) algorithm for ANNs, has been developed for real-time time series learning. It performs the model training with every new data that arrive to the system, without saving enormous quantities of data to create a historical database as usual, i.e., without previous knowledge. Consequently to validate the approach a simulation study through a Bayesian baseline model have been tested in order to compare with a database of a real application aiming to see the performance and accuracy. The core of the paper is a new algorithm, based on the BP one, which has been described in detail, and the challenge was how to implement a computational demanding algorithm in a simple architecture with very few hardware resources.

Online Learning Algorithm for Time Series Forecasting Suitable for Low Cost Wireless Sensor Networks Nodes

PubMed Central

Pardo, Juan; Zamora-Martínez, Francisco; Botella-Rocamora, Paloma

2015-01-01

Time series forecasting is an important predictive methodology which can be applied to a wide range of problems. Particularly, forecasting the indoor temperature permits an improved utilization of the HVAC (Heating, Ventilating and Air Conditioning) systems in a home and thus a better energy efficiency. With such purpose the paper describes how to implement an Artificial Neural Network (ANN) algorithm in a low cost system-on-chip to develop an autonomous intelligent wireless sensor network. The present paper uses a Wireless Sensor Networks (WSN) to monitor and forecast the indoor temperature in a smart home, based on low resources and cost microcontroller technology as the 8051MCU. An on-line learning approach, based on Back-Propagation (BP) algorithm for ANNs, has been developed for real-time time series learning. It performs the model training with every new data that arrive to the system, without saving enormous quantities of data to create a historical database as usual, i.e., without previous knowledge. Consequently to validate the approach a simulation study through a Bayesian baseline model have been tested in order to compare with a database of a real application aiming to see the performance and accuracy. The core of the paper is a new algorithm, based on the BP one, which has been described in detail, and the challenge was how to implement a computational demanding algorithm in a simple architecture with very few hardware resources. PMID:25905698

Learning a single-hidden layer feedforward neural network using a rank correlation-based strategy with application to high dimensional gene expression and proteomic spectra datasets in cancer detection.

PubMed

Belciug, Smaranda; Gorunescu, Florin

2018-06-08

Methods based on microarrays (MA), mass spectrometry (MS), and machine learning (ML) algorithms have evolved rapidly in recent years, allowing for early detection of several types of cancer. A pitfall of these approaches, however, is the overfitting of data due to large number of attributes and small number of instances -- a phenomenon known as the 'curse of dimensionality'. A potentially fruitful idea to avoid this drawback is to develop algorithms that combine fast computation with a filtering module for the attributes. The goal of this paper is to propose a statistical strategy to initiate the hidden nodes of a single-hidden layer feedforward neural network (SLFN) by using both the knowledge embedded in data and a filtering mechanism for attribute relevance. In order to attest its feasibility, the proposed model has been tested on five publicly available high-dimensional datasets: breast, lung, colon, and ovarian cancer regarding gene expression and proteomic spectra provided by cDNA arrays, DNA microarray, and MS. The novel algorithm, called adaptive SLFN (aSLFN), has been compared with four major classification algorithms: traditional ELM, radial basis function network (RBF), single-hidden layer feedforward neural network trained by backpropagation algorithm (BP-SLFN), and support vector-machine (SVM). Experimental results showed that the classification performance of aSLFN is competitive with the comparison models. Copyright © 2018. Published by Elsevier Inc.

Learning Perfectly Secure Cryptography to Protect Communications with Adversarial Neural Cryptography

PubMed Central

2018-01-01

Researches in Artificial Intelligence (AI) have achieved many important breakthroughs, especially in recent years. In some cases, AI learns alone from scratch and performs human tasks faster and better than humans. With the recent advances in AI, it is natural to wonder whether Artificial Neural Networks will be used to successfully create or break cryptographic algorithms. Bibliographic review shows the main approach to this problem have been addressed throughout complex Neural Networks, but without understanding or proving the security of the generated model. This paper presents an analysis of the security of cryptographic algorithms generated by a new technique called Adversarial Neural Cryptography (ANC). Using the proposed network, we show limitations and directions to improve the current approach of ANC. Training the proposed Artificial Neural Network with the improved model of ANC, we show that artificially intelligent agents can learn the unbreakable One-Time Pad (OTP) algorithm, without human knowledge, to communicate securely through an insecure communication channel. This paper shows in which conditions an AI agent can learn a secure encryption scheme. However, it also shows that, without a stronger adversary, it is more likely to obtain an insecure one. PMID:29695066

Learning Perfectly Secure Cryptography to Protect Communications with Adversarial Neural Cryptography.

PubMed

Coutinho, Murilo; de Oliveira Albuquerque, Robson; Borges, Fábio; García Villalba, Luis Javier; Kim, Tai-Hoon

2018-04-24

Researches in Artificial Intelligence (AI) have achieved many important breakthroughs, especially in recent years. In some cases, AI learns alone from scratch and performs human tasks faster and better than humans. With the recent advances in AI, it is natural to wonder whether Artificial Neural Networks will be used to successfully create or break cryptographic algorithms. Bibliographic review shows the main approach to this problem have been addressed throughout complex Neural Networks, but without understanding or proving the security of the generated model. This paper presents an analysis of the security of cryptographic algorithms generated by a new technique called Adversarial Neural Cryptography (ANC). Using the proposed network, we show limitations and directions to improve the current approach of ANC. Training the proposed Artificial Neural Network with the improved model of ANC, we show that artificially intelligent agents can learn the unbreakable One-Time Pad (OTP) algorithm, without human knowledge, to communicate securely through an insecure communication channel. This paper shows in which conditions an AI agent can learn a secure encryption scheme. However, it also shows that, without a stronger adversary, it is more likely to obtain an insecure one.

Fault-tolerant nonlinear adaptive flight control using sliding mode online learning.

PubMed

Krüger, Thomas; Schnetter, Philipp; Placzek, Robin; Vörsmann, Peter

2012-08-01

An expanded nonlinear model inversion flight control strategy using sliding mode online learning for neural networks is presented. The proposed control strategy is implemented for a small unmanned aircraft system (UAS). This class of aircraft is very susceptible towards nonlinearities like atmospheric turbulence, model uncertainties and of course system failures. Therefore, these systems mark a sensible testbed to evaluate fault-tolerant, adaptive flight control strategies. Within this work the concept of feedback linearization is combined with feed forward neural networks to compensate for inversion errors and other nonlinear effects. Backpropagation-based adaption laws of the network weights are used for online training. Within these adaption laws the standard gradient descent backpropagation algorithm is augmented with the concept of sliding mode control (SMC). Implemented as a learning algorithm, this nonlinear control strategy treats the neural network as a controlled system and allows a stable, dynamic calculation of the learning rates. While considering the system's stability, this robust online learning method therefore offers a higher speed of convergence, especially in the presence of external disturbances. The SMC-based flight controller is tested and compared with the standard gradient descent backpropagation algorithm in the presence of system failures. Copyright © 2012 Elsevier Ltd. All rights reserved.

Regional shape-based feature space for segmenting biomedical images using neural networks

NASA Astrophysics Data System (ADS)

Sundaramoorthy, Gopal; Hoford, John D.; Hoffman, Eric A.

1993-07-01

In biomedical images, structure of interest, particularly the soft tissue structures, such as the heart, airways, bronchial and arterial trees often have grey-scale and textural characteristics similar to other structures in the image, making it difficult to segment them using only gray- scale and texture information. However, these objects can be visually recognized by their unique shapes and sizes. In this paper we discuss, what we believe to be, a novel, simple scheme for extracting features based on regional shapes. To test the effectiveness of these features for image segmentation (classification), we use an artificial neural network and a statistical cluster analysis technique. The proposed shape-based feature extraction algorithm computes regional shape vectors (RSVs) for all pixels that meet a certain threshold criteria. The distance from each such pixel to a boundary is computed in 8 directions (or in 26 directions for a 3-D image). Together, these 8 (or 26) values represent the pixel's (or voxel's) RSV. All RSVs from an image are used to train a multi-layered perceptron neural network which uses these features to 'learn' a suitable classification strategy. To clearly distinguish the desired object from other objects within an image, several examples from inside and outside the desired object are used for training. Several examples are presented to illustrate the strengths and weaknesses of our algorithm. Both synthetic and actual biomedical images are considered. Future extensions to this algorithm are also discussed.

Applying Gradient Descent in Convolutional Neural Networks

NASA Astrophysics Data System (ADS)

Cui, Nan

2018-04-01

With the development of the integrated circuit and computer science, people become caring more about solving practical issues via information technologies. Along with that, a new subject called Artificial Intelligent (AI) comes up. One popular research interest of AI is about recognition algorithm. In this paper, one of the most common algorithms, Convolutional Neural Networks (CNNs) will be introduced, for image recognition. Understanding its theory and structure is of great significance for every scholar who is interested in this field. Convolution Neural Network is an artificial neural network which combines the mathematical method of convolution and neural network. The hieratical structure of CNN provides it reliable computer speed and reasonable error rate. The most significant characteristics of CNNs are feature extraction, weight sharing and dimension reduction. Meanwhile, combining with the Back Propagation (BP) mechanism and the Gradient Descent (GD) method, CNNs has the ability to self-study and in-depth learning. Basically, BP provides an opportunity for backwardfeedback for enhancing reliability and GD is used for self-training process. This paper mainly discusses the CNN and the related BP and GD algorithms, including the basic structure and function of CNN, details of each layer, the principles and features of BP and GD, and some examples in practice with a summary in the end.

A Parallel Adaboost-Backpropagation Neural Network for Massive Image Dataset Classification

NASA Astrophysics Data System (ADS)

Cao, Jianfang; Chen, Lichao; Wang, Min; Shi, Hao; Tian, Yun

2016-12-01

Image classification uses computers to simulate human understanding and cognition of images by automatically categorizing images. This study proposes a faster image classification approach that parallelizes the traditional Adaboost-Backpropagation (BP) neural network using the MapReduce parallel programming model. First, we construct a strong classifier by assembling the outputs of 15 BP neural networks (which are individually regarded as weak classifiers) based on the Adaboost algorithm. Second, we design Map and Reduce tasks for both the parallel Adaboost-BP neural network and the feature extraction algorithm. Finally, we establish an automated classification model by building a Hadoop cluster. We use the Pascal VOC2007 and Caltech256 datasets to train and test the classification model. The results are superior to those obtained using traditional Adaboost-BP neural network or parallel BP neural network approaches. Our approach increased the average classification accuracy rate by approximately 14.5% and 26.0% compared to the traditional Adaboost-BP neural network and parallel BP neural network, respectively. Furthermore, the proposed approach requires less computation time and scales very well as evaluated by speedup, sizeup and scaleup. The proposed approach may provide a foundation for automated large-scale image classification and demonstrates practical value.

A Parallel Adaboost-Backpropagation Neural Network for Massive Image Dataset Classification.

PubMed

Cao, Jianfang; Chen, Lichao; Wang, Min; Shi, Hao; Tian, Yun

2016-12-01

Image classification uses computers to simulate human understanding and cognition of images by automatically categorizing images. This study proposes a faster image classification approach that parallelizes the traditional Adaboost-Backpropagation (BP) neural network using the MapReduce parallel programming model. First, we construct a strong classifier by assembling the outputs of 15 BP neural networks (which are individually regarded as weak classifiers) based on the Adaboost algorithm. Second, we design Map and Reduce tasks for both the parallel Adaboost-BP neural network and the feature extraction algorithm. Finally, we establish an automated classification model by building a Hadoop cluster. We use the Pascal VOC2007 and Caltech256 datasets to train and test the classification model. The results are superior to those obtained using traditional Adaboost-BP neural network or parallel BP neural network approaches. Our approach increased the average classification accuracy rate by approximately 14.5% and 26.0% compared to the traditional Adaboost-BP neural network and parallel BP neural network, respectively. Furthermore, the proposed approach requires less computation time and scales very well as evaluated by speedup, sizeup and scaleup. The proposed approach may provide a foundation for automated large-scale image classification and demonstrates practical value.

A Parallel Adaboost-Backpropagation Neural Network for Massive Image Dataset Classification

PubMed Central

Cao, Jianfang; Chen, Lichao; Wang, Min; Shi, Hao; Tian, Yun

2016-01-01

Image classification uses computers to simulate human understanding and cognition of images by automatically categorizing images. This study proposes a faster image classification approach that parallelizes the traditional Adaboost-Backpropagation (BP) neural network using the MapReduce parallel programming model. First, we construct a strong classifier by assembling the outputs of 15 BP neural networks (which are individually regarded as weak classifiers) based on the Adaboost algorithm. Second, we design Map and Reduce tasks for both the parallel Adaboost-BP neural network and the feature extraction algorithm. Finally, we establish an automated classification model by building a Hadoop cluster. We use the Pascal VOC2007 and Caltech256 datasets to train and test the classification model. The results are superior to those obtained using traditional Adaboost-BP neural network or parallel BP neural network approaches. Our approach increased the average classification accuracy rate by approximately 14.5% and 26.0% compared to the traditional Adaboost-BP neural network and parallel BP neural network, respectively. Furthermore, the proposed approach requires less computation time and scales very well as evaluated by speedup, sizeup and scaleup. The proposed approach may provide a foundation for automated large-scale image classification and demonstrates practical value. PMID:27905520

An Indoor Positioning Technique Based on a Feed-Forward Artificial Neural Network Using Levenberg-Marquardt Learning Method

NASA Astrophysics Data System (ADS)

Pahlavani, P.; Gholami, A.; Azimi, S.

2017-09-01

This paper presents an indoor positioning technique based on a multi-layer feed-forward (MLFF) artificial neural networks (ANN). Most of the indoor received signal strength (RSS)-based WLAN positioning systems use the fingerprinting technique that can be divided into two phases: the offline (calibration) phase and the online (estimation) phase. In this paper, RSSs were collected for all references points in four directions and two periods of time (Morning and Evening). Hence, RSS readings were sampled at a regular time interval and specific orientation at each reference point. The proposed ANN based model used Levenberg-Marquardt algorithm for learning and fitting the network to the training data. This RSS readings in all references points and the known position of these references points was prepared for training phase of the proposed MLFF neural network. Eventually, the average positioning error for this network using 30% check and validation data was computed approximately 2.20 meter.

A regularization approach to continuous learning with an application to financial derivatives pricing.

PubMed

Ormoneit, D

1999-12-01

We consider the training of neural networks in cases where the nonlinear relationship of interest gradually changes over time. One possibility to deal with this problem is by regularization where a variation penalty is added to the usual mean squared error criterion. To learn the regularized network weights we suggest the Iterative Extended Kalman Filter (IEKF) as a learning rule, which may be derived from a Bayesian perspective on the regularization problem. A primary application of our algorithm is in financial derivatives pricing, where neural networks may be used to model the dependency of the derivatives' price on one or several underlying assets. After giving a brief introduction to the problem of derivatives pricing we present experiments with German stock index options data showing that a regularized neural network trained with the IEKF outperforms several benchmark models and alternative learning procedures. In particular, the performance may be greatly improved using a newly designed neural network architecture that accounts for no-arbitrage pricing restrictions.

Path optimisation of a mobile robot using an artificial neural network controller

NASA Astrophysics Data System (ADS)

Singh, M. K.; Parhi, D. R.

2011-01-01

This article proposed a novel approach for design of an intelligent controller for an autonomous mobile robot using a multilayer feed forward neural network, which enables the robot to navigate in a real world dynamic environment. The inputs to the proposed neural controller consist of left, right and front obstacle distance with respect to its position and target angle. The output of the neural network is steering angle. A four layer neural network has been designed to solve the path and time optimisation problem of mobile robots, which deals with the cognitive tasks such as learning, adaptation, generalisation and optimisation. A back propagation algorithm is used to train the network. This article also analyses the kinematic design of mobile robots for dynamic movements. The simulation results are compared with experimental results, which are satisfactory and show very good agreement. The training of the neural nets and the control performance analysis has been done in a real experimental setup.

Acoustic emission localization based on FBG sensing network and SVR algorithm

NASA Astrophysics Data System (ADS)

Sai, Yaozhang; Zhao, Xiuxia; Hou, Dianli; Jiang, Mingshun

2017-03-01

In practical application, carbon fiber reinforced plastics (CFRP) structures are easy to appear all sorts of invisible damages. So the damages should be timely located and detected for the safety of CFPR structures. In this paper, an acoustic emission (AE) localization system based on fiber Bragg grating (FBG) sensing network and support vector regression (SVR) is proposed for damage localization. AE signals, which are caused by damage, are acquired by high speed FBG interrogation. According to the Shannon wavelet transform, time differences between AE signals are extracted for localization algorithm based on SVR. According to the SVR model, the coordinate of AE source can be accurately predicted without wave velocity. The FBG system and localization algorithm are verified on a 500 mm×500 mm×2 mm CFRP plate. The experimental results show that the average error of localization system is 2.8 mm and the training time is 0.07 s.

Fast learning method for convolutional neural networks using extreme learning machine and its application to lane detection.

PubMed

Kim, Jihun; Kim, Jonghong; Jang, Gil-Jin; Lee, Minho

2017-03-01

Deep learning has received significant attention recently as a promising solution to many problems in the area of artificial intelligence. Among several deep learning architectures, convolutional neural networks (CNNs) demonstrate superior performance when compared to other machine learning methods in the applications of object detection and recognition. We use a CNN for image enhancement and the detection of driving lanes on motorways. In general, the process of lane detection consists of edge extraction and line detection. A CNN can be used to enhance the input images before lane detection by excluding noise and obstacles that are irrelevant to the edge detection result. However, training conventional CNNs requires considerable computation and a big dataset. Therefore, we suggest a new learning algorithm for CNNs using an extreme learning machine (ELM). The ELM is a fast learning method used to calculate network weights between output and hidden layers in a single iteration and thus, can dramatically reduce learning time while producing accurate results with minimal training data. A conventional ELM can be applied to networks with a single hidden layer; as such, we propose a stacked ELM architecture in the CNN framework. Further, we modify the backpropagation algorithm to find the targets of hidden layers and effectively learn network weights while maintaining performance. Experimental results confirm that the proposed method is effective in reducing learning time and improving performance. Copyright © 2016 Elsevier Ltd. All rights reserved.

A CNN based neurobiology inspired approach for retinal image quality assessment.

PubMed

Mahapatra, Dwarikanath; Roy, Pallab K; Sedai, Suman; Garnavi, Rahil

2016-08-01

Retinal image quality assessment (IQA) algorithms use different hand crafted features for training classifiers without considering the working of the human visual system (HVS) which plays an important role in IQA. We propose a convolutional neural network (CNN) based approach that determines image quality using the underlying principles behind the working of the HVS. CNNs provide a principled approach to feature learning and hence higher accuracy in decision making. Experimental results demonstrate the superior performance of our proposed algorithm over competing methods.

A Novel Topology Control Approach to Maintain the Node Degree in Dynamic Wireless Sensor Networks

PubMed Central

Huang, Yuanjiang; Martínez, José-Fernán; Díaz, Vicente Hernández; Sendra, Juana

2014-01-01

Topology control is an important technique to improve the connectivity and the reliability of Wireless Sensor Networks (WSNs) by means of adjusting the communication range of wireless sensor nodes. In this paper, a novel Fuzzy-logic Topology Control (FTC) is proposed to achieve any desired average node degree by adaptively changing communication range, thus improving the network connectivity, which is the main target of FTC. FTC is a fully localized control algorithm, and does not rely on location information of neighbors. Instead of designing membership functions and if-then rules for fuzzy-logic controller, FTC is constructed from the training data set to facilitate the design process. FTC is proved to be accurate, stable and has short settling time. In order to compare it with other representative localized algorithms (NONE, FLSS, k-Neighbor and LTRT), FTC is evaluated through extensive simulations. The simulation results show that: firstly, similar to k-Neighbor algorithm, FTC is the best to achieve the desired average node degree as node density varies; secondly, FTC is comparable to FLSS and k-Neighbor in terms of energy-efficiency, but is better than LTRT and NONE; thirdly, FTC has the lowest average maximum communication range than other algorithms, which indicates that the most energy-consuming node in the network consumes the lowest power. PMID:24608008

Broad Absorption Line Quasar catalogues with Supervised Neural Networks

DOE Office of Scientific and Technical Information (OSTI.GOV)

Scaringi, Simone; Knigge, Christian; Cottis, Christopher E.

2008-12-05

We have applied a Learning Vector Quantization (LVQ) algorithm to SDSS DR5 quasar spectra in order to create a large catalogue of broad absorption line quasars (BALQSOs). We first discuss the problems with BALQSO catalogues constructed using the conventional balnicity and/or absorption indices (BI and AI), and then describe the supervised LVQ network we have trained to recognise BALQSOs. The resulting BALQSO catalogue should be substantially more robust and complete than BI-or AI-based ones.

Using Fuzzy Logic for Performance Evaluation in Reinforcement Learning

NASA Technical Reports Server (NTRS)

Berenji, Hamid R.; Khedkar, Pratap S.

1992-01-01

Current reinforcement learning algorithms require long training periods which generally limit their applicability to small size problems. A new architecture is described which uses fuzzy rules to initialize its two neural networks: a neural network for performance evaluation and another for action selection. This architecture is applied to control of dynamic systems and it is demonstrated that it is possible to start with an approximate prior knowledge and learn to refine it through experiments using reinforcement learning.

Siamese convolutional networks for tracking the spine motion

NASA Astrophysics Data System (ADS)

Liu, Yuan; Sui, Xiubao; Sun, Yicheng; Liu, Chengwei; Hu, Yong

2017-09-01

Deep learning models have demonstrated great success in various computer vision tasks such as image classification and object tracking. However, tracking the lumbar spine by digitalized video fluoroscopic imaging (DVFI), which can quantitatively analyze the motion mode of spine to diagnose lumbar instability, has not yet been well developed due to the lack of steady and robust tracking method. In this paper, we propose a novel visual tracking algorithm of the lumbar vertebra motion based on a Siamese convolutional neural network (CNN) model. We train a full-convolutional neural network offline to learn generic image features. The network is trained to learn a similarity function that compares the labeled target in the first frame with the candidate patches in the current frame. The similarity function returns a high score if the two images depict the same object. Once learned, the similarity function is used to track a previously unseen object without any adapting online. In the current frame, our tracker is performed by evaluating the candidate rotated patches sampled around the previous frame target position and presents a rotated bounding box to locate the predicted target precisely. Results indicate that the proposed tracking method can detect the lumbar vertebra steadily and robustly. Especially for images with low contrast and cluttered background, the presented tracker can still achieve good tracking performance. Further, the proposed algorithm operates at high speed for real time tracking.

Using deep recurrent neural network for direct beam solar irradiance cloud screening

NASA Astrophysics Data System (ADS)

Chen, Maosi; Davis, John M.; Liu, Chaoshun; Sun, Zhibin; Zempila, Melina Maria; Gao, Wei

2017-09-01

Cloud screening is an essential procedure for in-situ calibration and atmospheric properties retrieval on (UV-)MultiFilter Rotating Shadowband Radiometer [(UV-)MFRSR]. Previous study has explored a cloud screening algorithm for direct-beam (UV-)MFRSR voltage measurements based on the stability assumption on a long time period (typically a half day or a whole day). To design such an algorithm requires in-depth understanding of radiative transfer and delicate data manipulation. Recent rapid developments on deep neural network and computation hardware have opened a window for modeling complicated End-to-End systems with a standardized strategy. In this study, a multi-layer dynamic bidirectional recurrent neural network is built for determining the cloudiness on each time point with a 17-year training dataset and tested with another 1-year dataset. The dataset is the daily 3-minute cosine corrected voltages, airmasses, and the corresponding cloud/clear-sky labels at two stations of the USDA UV-B Monitoring and Research Program. The results show that the optimized neural network model (3-layer, 250 hidden units, and 80 epochs of training) has an overall test accuracy of 97.87% (97.56% for the Oklahoma site and 98.16% for the Hawaii site). Generally, the neural network model grasps the key concept of the original model to use data in the entire day rather than short nearby measurements to perform cloud screening. A scrutiny of the logits layer suggests that the neural network model automatically learns a way to calculate a quantity similar to total optical depth and finds an appropriate threshold for cloud screening.

Artificial Neural Network applied to lightning flashes

NASA Astrophysics Data System (ADS)

Gin, R. B.; Guedes, D.; Bianchi, R.

2013-05-01

The development of video cameras enabled cientists to study lightning discharges comportment with more precision. The main goal of this project is to create a system able to detect images of lightning discharges stored in videos and classify them using an Artificial Neural Network (ANN)using C Language and OpenCV libraries. The developed system, can be split in two different modules: detection module and classification module. The detection module uses OpenCV`s computer vision libraries and image processing techniques to detect if there are significant differences between frames in a sequence, indicating that something, still not classified, occurred. Whenever there is a significant difference between two consecutive frames, two main algorithms are used to analyze the frame image: brightness and shape algorithms. These algorithms detect both shape and brightness of the event, removing irrelevant events like birds, as well as detecting the relevant events exact position, allowing the system to track it over time. The classification module uses a neural network to classify the relevant events as horizontal or vertical lightning, save the event`s images and calculates his number of discharges. The Neural Network was implemented using the backpropagation algorithm, and was trained with 42 training images , containing 57 lightning events (one image can have more than one lightning). TheANN was tested with one to five hidden layers, with up to 50 neurons each. The best configuration achieved a success rate of 95%, with one layer containing 20 neurons (33 test images with 42 events were used in this phase). This configuration was implemented in the developed system to analyze 20 video files, containing 63 lightning discharges previously manually detected. Results showed that all the lightning discharges were detected, many irrelevant events were unconsidered, and the event's number of discharges was correctly computed. The neural network used in this project achieved a success rate of 90%. The videos used in this experiment were acquired by seven video cameras installed in São Bernardo do Campo, Brazil, that continuously recorded lightning events during the summer. The cameras were disposed in a 360 loop, recording all data at a time resolution of 33ms. During this period, several convective storms were recorded.

Visualizing deep neural network by alternately image blurring and deblurring.

PubMed

Wang, Feng; Liu, Haijun; Cheng, Jian

2018-01-01

Visualization from trained deep neural networks has drawn massive public attention in recent. One of the visualization approaches is to train images maximizing the activation of specific neurons. However, directly maximizing the activation would lead to unrecognizable images, which cannot provide any meaningful information. In this paper, we introduce a simple but effective technique to constrain the optimization route of the visualization. By adding two totally inverse transformations, image blurring and deblurring, to the optimization procedure, recognizable images can be created. Our algorithm is good at extracting the details in the images, which are usually filtered by previous methods in the visualizations. Extensive experiments on AlexNet, VGGNet and GoogLeNet illustrate that we can better understand the neural networks utilizing the knowledge obtained by the visualization. Copyright © 2017 Elsevier Ltd. All rights reserved.

Network model of chemical-sensing system inspired by mouse taste buds.

PubMed

Tateno, Katsumi; Igarashi, Jun; Ohtubo, Yoshitaka; Nakada, Kazuki; Miki, Tsutomu; Yoshii, Kiyonori

2011-07-01

Taste buds endure extreme changes in temperature, pH, osmolarity, so on. Even though taste bud cells are replaced in a short span, they contribute to consistent taste reception. Each taste bud consists of about 50 cells whose networks are assumed to process taste information, at least preliminarily. In this article, we describe a neural network model inspired by the taste bud cells of mice. It consists of two layers. In the first layer, the chemical stimulus is transduced into an irregular spike train. The synchronization of the output impulses is induced by the irregular spike train at the second layer. These results show that the intensity of the chemical stimulus is encoded as the degree of the synchronization of output impulses. The present algorithms for signal processing result in a robust chemical-sensing system.

Deep greedy learning under thermal variability in full diurnal cycles

NASA Astrophysics Data System (ADS)

Rauss, Patrick; Rosario, Dalton

2017-08-01

We study the generalization and scalability behavior of a deep belief network (DBN) applied to a challenging long-wave infrared hyperspectral dataset, consisting of radiance from several manmade and natural materials within a fixed site located 500 m from an observation tower. The collections cover multiple full diurnal cycles and include different atmospheric conditions. Using complementary priors, a DBN uses a greedy algorithm that can learn deep, directed belief networks one layer at a time and has two layers form to provide undirected associative memory. The greedy algorithm initializes a slower learning procedure, which fine-tunes the weights, using a contrastive version of the wake-sleep algorithm. After fine-tuning, a network with three hidden layers forms a very good generative model of the joint distribution of spectral data and their labels, despite significant data variability between and within classes due to environmental and temperature variation occurring within and between full diurnal cycles. We argue, however, that more questions than answers are raised regarding the generalization capacity of these deep nets through experiments aimed at investigating their training and augmented learning behavior.

Automatic welding detection by an intelligent tool pipe inspection

NASA Astrophysics Data System (ADS)

Arizmendi, C. J.; Garcia, W. L.; Quintero, M. A.

2015-07-01

This work provide a model based on machine learning techniques in welds recognition, based on signals obtained through in-line inspection tool called “smart pig” in Oil and Gas pipelines. The model uses a signal noise reduction phase by means of pre-processing algorithms and attribute-selection techniques. The noise reduction techniques were selected after a literature review and testing with survey data. Subsequently, the model was trained using recognition and classification algorithms, specifically artificial neural networks and support vector machines. Finally, the trained model was validated with different data sets and the performance was measured with cross validation and ROC analysis. The results show that is possible to identify welding automatically with an efficiency between 90 and 98 percent.

Adaptively combined FIR and functional link artificial neural network equalizer for nonlinear communication channel.

PubMed

Zhao, Haiquan; Zhang, Jiashu

2009-04-01

This paper proposes a novel computational efficient adaptive nonlinear equalizer based on combination of finite impulse response (FIR) filter and functional link artificial neural network (CFFLANN) to compensate linear and nonlinear distortions in nonlinear communication channel. This convex nonlinear combination results in improving the speed while retaining the lower steady-state error. In addition, since the CFFLANN needs not the hidden layers, which exist in conventional neural-network-based equalizers, it exhibits a simpler structure than the traditional neural networks (NNs) and can require less computational burden during the training mode. Moreover, appropriate adaptation algorithm for the proposed equalizer is derived by the modified least mean square (MLMS). Results obtained from the simulations clearly show that the proposed equalizer using the MLMS algorithm can availably eliminate various intensity linear and nonlinear distortions, and be provided with better anti-jamming performance. Furthermore, comparisons of the mean squared error (MSE), the bit error rate (BER), and the effect of eigenvalue ratio (EVR) of input correlation matrix are presented.

Calibration of neural networks using genetic algorithms, with application to optimal path planning

NASA Technical Reports Server (NTRS)

Smith, Terence R.; Pitney, Gilbert A.; Greenwood, Daniel

1987-01-01

Genetic algorithms (GA) are used to search the synaptic weight space of artificial neural systems (ANS) for weight vectors that optimize some network performance function. GAs do not suffer from some of the architectural constraints involved with other techniques and it is straightforward to incorporate terms into the performance function concerning the metastructure of the ANS. Hence GAs offer a remarkably general approach to calibrating ANS. GAs are applied to the problem of calibrating an ANS that finds optimal paths over a given surface. This problem involves training an ANS on a relatively small set of paths and then examining whether the calibrated ANS is able to find good paths between arbitrary start and end points on the surface.

Multilayer perceptron, fuzzy sets, and classification

NASA Technical Reports Server (NTRS)

Pal, Sankar K.; Mitra, Sushmita

1992-01-01

A fuzzy neural network model based on the multilayer perceptron, using the back-propagation algorithm, and capable of fuzzy classification of patterns is described. The input vector consists of membership values to linguistic properties while the output vector is defined in terms of fuzzy class membership values. This allows efficient modeling of fuzzy or uncertain patterns with appropriate weights being assigned to the backpropagated errors depending upon the membership values at the corresponding outputs. During training, the learning rate is gradually decreased in discrete steps until the network converges to a minimum error solution. The effectiveness of the algorithm is demonstrated on a speech recognition problem. The results are compared with those of the conventional MLP, the Bayes classifier, and the other related models.

Quicklook Constituent Abundance and Stretch Parameter Retrieval for the Juno Microwave Radiometer using Neural Networks

NASA Astrophysics Data System (ADS)

Bellotti, A.; Steffes, P. G.

2016-12-01

The Juno Microwave Radiometer (MWR) has six channels ranging from 1.36-50 cm and the ability to peer deep into the Jovian atmosphere. An Artifical Neural Network algorithm has been developed to rapidly perform inversion for the deep abundance of ammonia, the deep abundance of water vapor, and atmospheric "stretch" (a parameter that reflects the deviation from a wet adiabate in the higher atmosphere). This algorithm is "trained" by using simulated emissions at the six wavelengths computed using the Juno atmospheric microwave radiative transfer (JAMRT) model presented by Oyafuso et al. (This meeting). By exploiting the emission measurements conducted at six wavelengths and at various incident angles, the neural network can provide preliminary results to a useful precison in a computational method hundreds of times faster than conventional methods. This can quickly provide important insights into the variability and structure of the Jovian atmosphere.

Study on loading path optimization of internal high pressure forming process

NASA Astrophysics Data System (ADS)

Jiang, Shufeng; Zhu, Hengda; Gao, Fusheng

2017-09-01

In the process of internal high pressure forming, there is no formula to describe the process parameters and forming results. The article use numerical simulation to obtain several input parameters and corresponding output result, use the BP neural network to found their mapping relationship, and with weighted summing method make each evaluating parameters to set up a formula which can evaluate quality. Then put the training BP neural network into the particle swarm optimization, and take the evaluating formula of the quality as adapting formula of particle swarm optimization, finally do the optimization and research at the range of each parameters. The results show that the parameters obtained by the BP neural network algorithm and the particle swarm optimization algorithm can meet the practical requirements. The method can solve the optimization of the process parameters in the internal high pressure forming process.

Enhanced Higgs boson to τ(+)τ(-) search with deep learning.

PubMed

Baldi, P; Sadowski, P; Whiteson, D

2015-03-20

The Higgs boson is thought to provide the interaction that imparts mass to the fundamental fermions, but while measurements at the Large Hadron Collider (LHC) are consistent with this hypothesis, current analysis techniques lack the statistical power to cross the traditional 5σ significance barrier without more data. Deep learning techniques have the potential to increase the statistical power of this analysis by automatically learning complex, high-level data representations. In this work, deep neural networks are used to detect the decay of the Higgs boson to a pair of tau leptons. A Bayesian optimization algorithm is used to tune the network architecture and training algorithm hyperparameters, resulting in a deep network of eight nonlinear processing layers that improves upon the performance of shallow classifiers even without the use of features specifically engineered by physicists for this application. The improvement in discovery significance is equivalent to an increase in the accumulated data set of 25%.

The relevance of network micro-structure for neural dynamics.

PubMed

Pernice, Volker; Deger, Moritz; Cardanobile, Stefano; Rotter, Stefan

2013-01-01

The activity of cortical neurons is determined by the input they receive from presynaptic neurons. Many previous studies have investigated how specific aspects of the statistics of the input affect the spike trains of single neurons and neurons in recurrent networks. However, typically very simple random network models are considered in such studies. Here we use a recently developed algorithm to construct networks based on a quasi-fractal probability measure which are much more variable than commonly used network models, and which therefore promise to sample the space of recurrent networks in a more exhaustive fashion than previously possible. We use the generated graphs as the underlying network topology in simulations of networks of integrate-and-fire neurons in an asynchronous and irregular state. Based on an extensive dataset of networks and neuronal simulations we assess statistical relations between features of the network structure and the spiking activity. Our results highlight the strong influence that some details of the network structure have on the activity dynamics of both single neurons and populations, even if some global network parameters are kept fixed. We observe specific and consistent relations between activity characteristics like spike-train irregularity or correlations and network properties, for example the distributions of the numbers of in- and outgoing connections or clustering. Exploiting these relations, we demonstrate that it is possible to estimate structural characteristics of the network from activity data. We also assess higher order correlations of spiking activity in the various networks considered here, and find that their occurrence strongly depends on the network structure. These results provide directions for further theoretical studies on recurrent networks, as well as new ways to interpret spike train recordings from neural circuits.

Construction of Gene Regulatory Networks Using Recurrent Neural Networks and Swarm Intelligence.

PubMed

Khan, Abhinandan; Mandal, Sudip; Pal, Rajat Kumar; Saha, Goutam

2016-01-01

We have proposed a methodology for the reverse engineering of biologically plausible gene regulatory networks from temporal genetic expression data. We have used established information and the fundamental mathematical theory for this purpose. We have employed the Recurrent Neural Network formalism to extract the underlying dynamics present in the time series expression data accurately. We have introduced a new hybrid swarm intelligence framework for the accurate training of the model parameters. The proposed methodology has been first applied to a small artificial network, and the results obtained suggest that it can produce the best results available in the contemporary literature, to the best of our knowledge. Subsequently, we have implemented our proposed framework on experimental (in vivo) datasets. Finally, we have investigated two medium sized genetic networks (in silico) extracted from GeneNetWeaver, to understand how the proposed algorithm scales up with network size. Additionally, we have implemented our proposed algorithm with half the number of time points. The results indicate that a reduction of 50% in the number of time points does not have an effect on the accuracy of the proposed methodology significantly, with a maximum of just over 15% deterioration in the worst case.

Developing algorithm for the critical care physician scheduling

NASA Astrophysics Data System (ADS)

Lee, Hyojun; Pah, Adam; Amaral, Luis; Northwestern Memorial Hospital Collaboration

Understanding the social network has enabled us to quantitatively study social phenomena such as behaviors in adoption and propagation of information. However, most work has been focusing on networks of large heterogeneous communities, and little attention has been paid to how work-relevant information spreads within networks of small and homogeneous groups of highly trained individuals, such as physicians. Within the professionals, the behavior patterns and the transmission of information relevant to the job are dependent not only on the social network between the employees but also on the schedules and teams that work together. In order to systematically investigate the dependence of the spread of ideas and adoption of innovations on a work-environment network, we sought to construct a model for the interaction network of critical care physicians at Northwestern Memorial Hospital (NMH) based on their work schedules. We inferred patterns and hidden rules from past work schedules such as turnover rates. Using the characteristics of the work schedules of the physicians and their turnover rates, we were able to create multi-year synthetic work schedules for a generic intensive care unit. The algorithm for creating shift schedules can be applied to other schedule dependent networks ARO1.

Iris double recognition based on modified evolutionary neural network

NASA Astrophysics Data System (ADS)

Liu, Shuai; Liu, Yuan-Ning; Zhu, Xiao-Dong; Huo, Guang; Liu, Wen-Tao; Feng, Jia-Kai

2017-11-01

Aiming at multicategory iris recognition under illumination and noise interference, this paper proposes a method of iris double recognition based on a modified evolutionary neural network. An equalization histogram and Laplace of Gaussian operator are used to process the iris to suppress illumination and noise interference and Haar wavelet to convert the iris feature to binary feature encoding. Calculate the Hamming distance for the test iris and template iris , and compare with classification threshold, determine the type of iris. If the iris cannot be identified as a different type, there needs to be a secondary recognition. The connection weights in back-propagation (BP) neural network use modified evolutionary neural network to adaptively train. The modified neural network is composed of particle swarm optimization with mutation operator and BP neural network. According to different iris libraries in different circumstances of experimental results, under illumination and noise interference, the correct recognition rate of this algorithm is higher, the ROC curve is closer to the coordinate axis, the training and recognition time is shorter, and the stability and the robustness are better.

Use of qualitative and quantitative information in neural networks for assessing agricultural chemical contamination of domestic wells

USGS Publications Warehouse

Mishra, A.; Ray, C.; Kolpin, D.W.

2004-01-01

A neural network analysis of agrichemical occurrence in groundwater was conducted using data from a pilot study of 192 small-diameter drilled and driven wells and 115 dug and bored wells in Illinois, a regional reconnaissance network of 303 wells across 12 Midwestern states, and a study of 687 domestic wells across Iowa. Potential factors contributing to well contamination (e.g., depth to aquifer material, well depth, and distance to cropland) were investigated. These contributing factors were available in either numeric (actual or categorical) or descriptive (yes or no) format. A method was devised to use the numeric and descriptive values simultaneously. Training of the network was conducted using a standard backpropagation algorithm. Approximately 15% of the data was used for testing. Analysis indicated that training error was quite low for most data. Testing results indicated that it was possible to predict the contamination potential of a well with pesticides. However, predicting the actual level of contamination was more difficult. For pesticide occurrence in drilled and driven wells, the network predictions were good. The performance of the network was poorer for predicting nitrate occurrence in dug and bored wells. Although the data set for Iowa was large, the prediction ability of the trained network was poor, due to descriptive or categorical input parameters, compared with smaller data sets such as that for Illinois, which contained more numeric information.

A biologically inspired neural network model to transformation invariant object recognition

NASA Astrophysics Data System (ADS)

Iftekharuddin, Khan M.; Li, Yaqin; Siddiqui, Faraz

2007-09-01

Transformation invariant image recognition has been an active research area due to its widespread applications in a variety of fields such as military operations, robotics, medical practices, geographic scene analysis, and many others. The primary goal for this research is detection of objects in the presence of image transformations such as changes in resolution, rotation, translation, scale and occlusion. We investigate a biologically-inspired neural network (NN) model for such transformation-invariant object recognition. In a classical training-testing setup for NN, the performance is largely dependent on the range of transformation or orientation involved in training. However, an even more serious dilemma is that there may not be enough training data available for successful learning or even no training data at all. To alleviate this problem, a biologically inspired reinforcement learning (RL) approach is proposed. In this paper, the RL approach is explored for object recognition with different types of transformations such as changes in scale, size, resolution and rotation. The RL is implemented in an adaptive critic design (ACD) framework, which approximates the neuro-dynamic programming of an action network and a critic network, respectively. Two ACD algorithms such as Heuristic Dynamic Programming (HDP) and Dual Heuristic dynamic Programming (DHP) are investigated to obtain transformation invariant object recognition. The two learning algorithms are evaluated statistically using simulated transformations in images as well as with a large-scale UMIST face database with pose variations. In the face database authentication case, the 90° out-of-plane rotation of faces from 20 different subjects in the UMIST database is used. Our simulations show promising results for both designs for transformation-invariant object recognition and authentication of faces. Comparing the two algorithms, DHP outperforms HDP in learning capability, as DHP takes fewer steps to perform a successful recognition task in general. Further, the residual critic error in DHP is generally smaller than that of HDP, and DHP achieves a 100% success rate more frequently than HDP for individual objects/subjects. On the other hand, HDP is more robust than the DHP as far as success rate across the database is concerned when applied in a stochastic and uncertain environment, and the computational time involved in DHP is more.

Application of a feedforward neural network in the search for kuroko deposits in the Hokuroku District, Japan

USGS Publications Warehouse

Singer, Donald A.; Kouda, Ryoichi

1996-01-01

A feedforward neural network with one hidden layer and five neurons was trained to recognize the distance to kuroko mineral deposits. Average amounts per hole of pyrite, sericite, and gypsum plus anhydrite as measured by X-rays in 69 drillholes were used to train the net. Drillholes near and between the Fukazawa, Furutobe, and Shakanai mines were used. The training data were selected carefully to represent well-explored areas where some confidence of the distance to ore was assured. A logarithmic transform was applied to remove the skewness of distance and each variable was scaled and centered by subtracting the median and dividing by the interquartile range. The learning algorithm of annealing plus conjugate gradients was used to minimize the mean squared error of the scaled distance to ore. The trained network then was applied to all of the 152 drillholes that had measured gypsum, sericite, and pyrite. A contour plot of the neural net predicted distance to ore shows fairly wide areas of 1 km or less to ore; each of the known deposit groups is within the 1 km contour. The high and low distances on the margins of the contoured distance plot are in part the result of boundary effects of the contouring algorithm. For example, the short distances to ore predicted west of the Shakanai (Hanaoka) deposits are in basement. However, the short distances to ore predicted northeast of Furotobe, just off the figure, coincide with the location of the Nurukawa kuroko deposit and the Omaki deposit, south of the Shakanai-Hanaoka deposits, seems to be on an extension of short distance to ore contour, but is beyond the 3 km limit from drillholes. Also of interest are some areas only a few kilometers from the Fukazawa and Shakanai groups of deposits that are estimated to be many kilometers from ore, apparently reflecting the network's recognition of the extreme local variability of the geology near some deposits.

Evaluation of Porosity and Permeability for an Oil Prospect, Offshore Vietnam by using Artificial Neural Networks

NASA Astrophysics Data System (ADS)

Bui, H. T.; Ho, L. T.; Ushijima, K.; Nur, A.

2006-12-01

Determination of porosity and permeability plays a key role either in characterization of a reservoir or in development of an oil field. Their distribution helps to predict the major faults or fractured zones that are related to high porosity area in order to reduce drilling hazards. Porosity and permeability of the rock can be determined directly from the core sample or obtained from well log data such as: sonic, density, neutron or resistivity. These input parameters depend not only on porosity (?) but also on the rock matrix, fluids contained in the rocks, clay mineral component, or geometry of pore structures. Therefore, it is not easy to estimate exactly porosity and permeability since having corrected those values by conventional well log interpretation method. In this study, the Artificial Neural Networks (ANNs) have been used to derive porosity and permeability directly from well log data for Vung Dong oil prospect, southern offshore Vietnam. Firstly, we designed a training patterns for ANNs from neutron porosity, bulk density, P-sonic, deep resistivity, shallow resistivity and MSFL log curves. Then, ANNs were trained by core samples data for porosity and permeability. Several ANNs paradigms have been tried on a basis of trial and error. The batch back- propagation algorithm was found more proficient in training porosity network meanwhile the quick propagation algorithm is more effective in the permeability network. Secondly, trained ANNs was tested and applied for real data set of some wells to calculate and reveal the distribution maps of porosity or permeability. Distributions of porosity and permeability have been correlated with seismic data interpretation to map the faults and fractured zones in the study. The ANNs showed good results of porosity and permeability distribution with high reliability, fast, accurate and low cost features. Therefore, the ANNs should be widely applied in oil and gas industry.

Hardware friendly probabilistic spiking neural network with long-term and short-term plasticity.

PubMed

Hsieh, Hung-Yi; Tang, Kea-Tiong

2013-12-01

This paper proposes a probabilistic spiking neural network (PSNN) with unimodal weight distribution, possessing long- and short-term plasticity. The proposed algorithm is derived by both the arithmetic gradient decent calculation and bioinspired algorithms. The algorithm is benchmarked by the Iris and Wisconsin breast cancer (WBC) data sets. The network features fast convergence speed and high accuracy. In the experiment, the PSNN took not more than 40 epochs for convergence. The average testing accuracy for Iris and WBC data is 96.7% and 97.2%, respectively. To test the usefulness of the PSNN to real world application, the PSNN was also tested with the odor data, which was collected by our self-developed electronic nose (e-nose). Compared with the algorithm (K-nearest neighbor) that has the highest classification accuracy in the e-nose for the same odor data, the classification accuracy of the PSNN is only 1.3% less but the memory requirement can be reduced at least 40%. All the experiments suggest that the PSNN is hardware friendly. First, it requires only nine-bits weight resolution for training and testing. Second, the PSNN can learn complex data sets with a little number of neurons that in turn reduce the cost of VLSI implementation. In addition, the algorithm is insensitive to synaptic noise and the parameter variation induced by the VLSI fabrication. Therefore, the algorithm can be implemented by either software or hardware, making it suitable for wider application.

Comparison of RF spectrum prediction methods for dynamic spectrum access

NASA Astrophysics Data System (ADS)

Kovarskiy, Jacob A.; Martone, Anthony F.; Gallagher, Kyle A.; Sherbondy, Kelly D.; Narayanan, Ram M.

2017-05-01

Dynamic spectrum access (DSA) refers to the adaptive utilization of today's busy electromagnetic spectrum. Cognitive radio/radar technologies require DSA to intelligently transmit and receive information in changing environments. Predicting radio frequency (RF) activity reduces sensing time and energy consumption for identifying usable spectrum. Typical spectrum prediction methods involve modeling spectral statistics with Hidden Markov Models (HMM) or various neural network structures. HMMs describe the time-varying state probabilities of Markov processes as a dynamic Bayesian network. Neural Networks model biological brain neuron connections to perform a wide range of complex and often non-linear computations. This work compares HMM, Multilayer Perceptron (MLP), and Recurrent Neural Network (RNN) algorithms and their ability to perform RF channel state prediction. Monte Carlo simulations on both measured and simulated spectrum data evaluate the performance of these algorithms. Generalizing spectrum occupancy as an alternating renewal process allows Poisson random variables to generate simulated data while energy detection determines the occupancy state of measured RF spectrum data for testing. The results suggest that neural networks achieve better prediction accuracy and prove more adaptable to changing spectral statistics than HMMs given sufficient training data.

Lesion Detection in CT Images Using Deep Learning Semantic Segmentation Technique

NASA Astrophysics Data System (ADS)

Kalinovsky, A.; Liauchuk, V.; Tarasau, A.

2017-05-01

In this paper, the problem of automatic detection of tuberculosis lesion on 3D lung CT images is considered as a benchmark for testing out algorithms based on a modern concept of Deep Learning. For training and testing of the algorithms a domestic dataset of 338 3D CT scans of tuberculosis patients with manually labelled lesions was used. The algorithms which are based on using Deep Convolutional Networks were implemented and applied in three different ways including slice-wise lesion detection in 2D images using semantic segmentation, slice-wise lesion detection in 2D images using sliding window technique as well as straightforward detection of lesions via semantic segmentation in whole 3D CT scans. The algorithms demonstrate superior performance compared to algorithms based on conventional image analysis methods.

Network Intrusion Detection Based on a General Regression Neural Network Optimized by an Improved Artificial Immune Algorithm

PubMed Central

Wu, Jianfa; Peng, Dahao; Li, Zhuping; Zhao, Li; Ling, Huanzhang

2015-01-01

To effectively and accurately detect and classify network intrusion data, this paper introduces a general regression neural network (GRNN) based on the artificial immune algorithm with elitist strategies (AIAE). The elitist archive and elitist crossover were combined with the artificial immune algorithm (AIA) to produce the AIAE-GRNN algorithm, with the aim of improving its adaptivity and accuracy. In this paper, the mean square errors (MSEs) were considered the affinity function. The AIAE was used to optimize the smooth factors of the GRNN; then, the optimal smooth factor was solved and substituted into the trained GRNN. Thus, the intrusive data were classified. The paper selected a GRNN that was separately optimized using a genetic algorithm (GA), particle swarm optimization (PSO), and fuzzy C-mean clustering (FCM) to enable a comparison of these approaches. As shown in the results, the AIAE-GRNN achieves a higher classification accuracy than PSO-GRNN, but the running time of AIAE-GRNN is long, which was proved first. FCM and GA-GRNN were eliminated because of their deficiencies in terms of accuracy and convergence. To improve the running speed, the paper adopted principal component analysis (PCA) to reduce the dimensions of the intrusive data. With the reduction in dimensionality, the PCA-AIAE-GRNN decreases in accuracy less and has better convergence than the PCA-PSO-GRNN, and the running speed of the PCA-AIAE-GRNN was relatively improved. The experimental results show that the AIAE-GRNN has a higher robustness and accuracy than the other algorithms considered and can thus be used to classify the intrusive data. PMID:25807466

Simple techniques for improving deep neural network outcomes on commodity hardware

NASA Astrophysics Data System (ADS)

Colina, Nicholas Christopher A.; Perez, Carlos E.; Paraan, Francis N. C.

2017-08-01

We benchmark improvements in the performance of deep neural networks (DNN) on the MNIST data test upon imple-menting two simple modifications to the algorithm that have little overhead computational cost. First is GPU parallelization on a commodity graphics card, and second is initializing the DNN with random orthogonal weight matrices prior to optimization. Eigenspectra analysis of the weight matrices reveal that the initially orthogonal matrices remain nearly orthogonal after training. The probability distributions from which these orthogonal matrices are drawn are also shown to significantly affect the performance of these deep neural networks.

Enhanced online convolutional neural networks for object tracking

NASA Astrophysics Data System (ADS)

Zhang, Dengzhuo; Gao, Yun; Zhou, Hao; Li, Tianwen

2018-04-01

In recent several years, object tracking based on convolution neural network has gained more and more attention. The initialization and update of convolution filters can directly affect the precision of object tracking effective. In this paper, a novel object tracking via an enhanced online convolution neural network without offline training is proposed, which initializes the convolution filters by a k-means++ algorithm and updates the filters by an error back-propagation. The comparative experiments of 7 trackers on 15 challenging sequences showed that our tracker can perform better than other trackers in terms of AUC and precision.

Target recognition based on convolutional neural network

NASA Astrophysics Data System (ADS)

Wang, Liqiang; Wang, Xin; Xi, Fubiao; Dong, Jian

2017-11-01

One of the important part of object target recognition is the feature extraction, which can be classified into feature extraction and automatic feature extraction. The traditional neural network is one of the automatic feature extraction methods, while it causes high possibility of over-fitting due to the global connection. The deep learning algorithm used in this paper is a hierarchical automatic feature extraction method, trained with the layer-by-layer convolutional neural network (CNN), which can extract the features from lower layers to higher layers. The features are more discriminative and it is beneficial to the object target recognition.

Application of Artificial Neural Networks to the Development of Improved Multi-Sensor Retrievals of Near-Surface Air Temperature and Humidity Over Ocean

NASA Technical Reports Server (NTRS)

Roberts, J. Brent; Robertson, Franklin R.; Clayson, Carol Anne

2012-01-01

Improved estimates of near-surface air temperature and air humidity are critical to the development of more accurate turbulent surface heat fluxes over the ocean. Recent progress in retrieving these parameters has been made through the application of artificial neural networks (ANN) and the use of multi-sensor passive microwave observations. Details are provided on the development of an improved retrieval algorithm that applies the nonlinear statistical ANN methodology to a set of observations from the Advanced Microwave Scanning Radiometer (AMSR-E) and the Advanced Microwave Sounding Unit (AMSU-A) that are currently available from the NASA AQUA satellite platform. Statistical inversion techniques require an adequate training dataset to properly capture embedded physical relationships. The development of multiple training datasets containing only in-situ observations, only synthetic observations produced using the Community Radiative Transfer Model (CRTM), or a mixture of each is discussed. An intercomparison of results using each training dataset is provided to highlight the relative advantages and disadvantages of each methodology. Particular emphasis will be placed on the development of retrievals in cloudy versus clear-sky conditions. Near-surface air temperature and humidity retrievals using the multi-sensor ANN algorithms are compared to previous linear and non-linear retrieval schemes.

Query construction, entropy, and generalization in neural-network models

NASA Astrophysics Data System (ADS)

Sollich, Peter

1994-05-01

We study query construction algorithms, which aim at improving the generalization ability of systems that learn from examples by choosing optimal, nonredundant training sets. We set up a general probabilistic framework for deriving such algorithms from the requirement of optimizing a suitable objective function; specifically, we consider the objective functions entropy (or information gain) and generalization error. For two learning scenarios, the high-low game and the linear perceptron, we evaluate the generalization performance obtained by applying the corresponding query construction algorithms and compare it to training on random examples. We find qualitative differences between the two scenarios due to the different structure of the underlying rules (nonlinear and ``noninvertible'' versus linear); in particular, for the linear perceptron, random examples lead to the same generalization ability as a sequence of queries in the limit of an infinite number of examples. We also investigate learning algorithms which are ill matched to the learning environment and find that, in this case, minimum entropy queries can in fact yield a lower generalization ability than random examples. Finally, we study the efficiency of single queries and its dependence on the learning history, i.e., on whether the previous training examples were generated randomly or by querying, and the difference between globally and locally optimal query construction.

Application of adaptive boosting to EP-derived multilayer feed-forward neural networks (MLFN) to improve benign/malignant breast cancer classification

NASA Astrophysics Data System (ADS)

Land, Walker H., Jr.; Masters, Timothy D.; Lo, Joseph Y.; McKee, Dan

2001-07-01

A new neural network technology was developed for improving the benign/malignant diagnosis of breast cancer using mammogram findings. A new paradigm, Adaptive Boosting (AB), uses a markedly different theory in solutioning Computational Intelligence (CI) problems. AB, a new machine learning paradigm, focuses on finding weak learning algorithm(s) that initially need to provide slightly better than random performance (i.e., approximately 55%) when processing a mammogram training set. Then, by successive development of additional architectures (using the mammogram training set), the adaptive boosting process improves the performance of the basic Evolutionary Programming derived neural network architectures. The results of these several EP-derived hybrid architectures are then intelligently combined and tested using a similar validation mammogram data set. Optimization focused on improving specificity and positive predictive value at very high sensitivities, where an analysis of the performance of the hybrid would be most meaningful. Using the DUKE mammogram database of 500 biopsy proven samples, on average this hybrid was able to achieve (under statistical 5-fold cross-validation) a specificity of 48.3% and a positive predictive value (PPV) of 51.8% while maintaining 100% sensitivity. At 97% sensitivity, a specificity of 56.6% and a PPV of 55.8% were obtained.

The effect of lossy image compression on image classification

NASA Technical Reports Server (NTRS)

Paola, Justin D.; Schowengerdt, Robert A.

1995-01-01

We have classified four different images, under various levels of JPEG compression, using the following classification algorithms: minimum-distance, maximum-likelihood, and neural network. The training site accuracy and percent difference from the original classification were tabulated for each image compression level, with maximum-likelihood showing the poorest results. In general, as compression ratio increased, the classification retained its overall appearance, but much of the pixel-to-pixel detail was eliminated. We also examined the effect of compression on spatial pattern detection using a neural network.

Neural Decoder for Topological Codes

NASA Astrophysics Data System (ADS)

Torlai, Giacomo; Melko, Roger G.

2017-07-01

We present an algorithm for error correction in topological codes that exploits modern machine learning techniques. Our decoder is constructed from a stochastic neural network called a Boltzmann machine, of the type extensively used in deep learning. We provide a general prescription for the training of the network and a decoding strategy that is applicable to a wide variety of stabilizer codes with very little specialization. We demonstrate the neural decoder numerically on the well-known two-dimensional toric code with phase-flip errors.

Efficient and self-adaptive in-situ learning in multilayer memristor neural networks.

PubMed

Li, Can; Belkin, Daniel; Li, Yunning; Yan, Peng; Hu, Miao; Ge, Ning; Jiang, Hao; Montgomery, Eric; Lin, Peng; Wang, Zhongrui; Song, Wenhao; Strachan, John Paul; Barnell, Mark; Wu, Qing; Williams, R Stanley; Yang, J Joshua; Xia, Qiangfei

2018-06-19

Memristors with tunable resistance states are emerging building blocks of artificial neural networks. However, in situ learning on a large-scale multiple-layer memristor network has yet to be demonstrated because of challenges in device property engineering and circuit integration. Here we monolithically integrate hafnium oxide-based memristors with a foundry-made transistor array into a multiple-layer neural network. We experimentally demonstrate in situ learning capability and achieve competitive classification accuracy on a standard machine learning dataset, which further confirms that the training algorithm allows the network to adapt to hardware imperfections. Our simulation using the experimental parameters suggests that a larger network would further increase the classification accuracy. The memristor neural network is a promising hardware platform for artificial intelligence with high speed-energy efficiency.

Advanced obstacle avoidance for a laser based wheelchair using optimised Bayesian neural networks.

PubMed

Trieu, Hoang T; Nguyen, Hung T; Willey, Keith

2008-01-01

In this paper we present an advanced method of obstacle avoidance for a laser based intelligent wheelchair using optimized Bayesian neural networks. Three neural networks are designed for three separate sub-tasks: passing through a door way, corridor and wall following and general obstacle avoidance. The accurate usable accessible space is determined by including the actual wheelchair dimensions in a real-time map used as inputs to each networks. Data acquisitions are performed separately to collect the patterns required for specified sub-tasks. Bayesian frame work is used to determine the optimal neural network structure in each case. Then these networks are trained under the supervision of Bayesian rule. Experiment results showed that compare to the VFH algorithm our neural networks navigated a smoother path following a near optimum trajectory.

Inversion of 2-D DC resistivity data using rapid optimization and minimal complexity neural network

NASA Astrophysics Data System (ADS)

Singh, U. K.; Tiwari, R. K.; Singh, S. B.

2010-02-01

The backpropagation (BP) artificial neural network (ANN) technique of optimization based on steepest descent algorithm is known to be inept for its poor performance and does not ensure global convergence. Nonlinear and complex DC resistivity data require efficient ANN model and more intensive optimization procedures for better results and interpretations. Improvements in the computational ANN modeling process are described with the goals of enhancing the optimization process and reducing ANN model complexity. Well-established optimization methods, such as Radial basis algorithm (RBA) and Levenberg-Marquardt algorithms (LMA) have frequently been used to deal with complexity and nonlinearity in such complex geophysical records. We examined here the efficiency of trained LMA and RB networks by using 2-D synthetic resistivity data and then finally applied to the actual field vertical electrical resistivity sounding (VES) data collected from the Puga Valley, Jammu and Kashmir, India. The resulting ANN reconstruction resistivity results are compared with the result of existing inversion approaches, which are in good agreement. The depths and resistivity structures obtained by the ANN methods also correlate well with the known drilling results and geologic boundaries. The application of the above ANN algorithms proves to be robust and could be used for fast estimation of resistive structures for other complex earth model also.

3D Cloud Field Prediction using A-Train Data and Machine Learning Techniques

NASA Astrophysics Data System (ADS)

Johnson, C. L.

2017-12-01

Validation of cloud process parameterizations used in global climate models (GCMs) would greatly benefit from observed 3D cloud fields at the size comparable to that of a GCM grid cell. For the highest resolution simulations, surface grid cells are on the order of 100 km by 100 km. CloudSat/CALIPSO data provides 1 km width of detailed vertical cloud fraction profile (CFP) and liquid and ice water content (LWC/IWC). This work utilizes four machine learning algorithms to create nonlinear regressions of CFP, LWC, and IWC data using radiances, surface type and location of measurement as predictors and applies the regression equations to off-track locations generating 3D cloud fields for 100 km by 100 km domains. The CERES-CloudSat-CALIPSO-MODIS (C3M) merged data set for February 2007 is used. Support Vector Machines, Artificial Neural Networks, Gaussian Processes and Decision Trees are trained on 1000 km of continuous C3M data. Accuracy is computed using existing vertical profiles that are excluded from the training data and occur within 100 km of the training data. Accuracy of the four algorithms is compared. Average accuracy for one day of predicted data is 86% for the most successful algorithm. The methodology for training the algorithms, determining valid prediction regions and applying the equations off-track is discussed. Predicted 3D cloud fields are provided as inputs to the Ed4 NASA LaRC Fu-Liou radiative transfer code and resulting TOA radiances compared to observed CERES/MODIS radiances. Differences in computed radiances using predicted profiles and observed radiances are compared.

Autonomous reinforcement learning with experience replay.

PubMed

Wawrzyński, Paweł; Tanwani, Ajay Kumar

2013-05-01

This paper considers the issues of efficiency and autonomy that are required to make reinforcement learning suitable for real-life control tasks. A real-time reinforcement learning algorithm is presented that repeatedly adjusts the control policy with the use of previously collected samples, and autonomously estimates the appropriate step-sizes for the learning updates. The algorithm is based on the actor-critic with experience replay whose step-sizes are determined on-line by an enhanced fixed point algorithm for on-line neural network training. An experimental study with simulated octopus arm and half-cheetah demonstrates the feasibility of the proposed algorithm to solve difficult learning control problems in an autonomous way within reasonably short time. Copyright © 2012 Elsevier Ltd. All rights reserved.

Optimization of internet content filtering-Combined with KNN and OCAT algorithms

NASA Astrophysics Data System (ADS)

Guo, Tianze; Wu, Lingjing; Liu, Jiaming

2018-04-01

The face of the status quo that rampant illegal content in the Internet, the result of traditional way to filter information, keyword recognition and manual screening, is getting worse. Based on this, this paper uses OCAT algorithm nested by KNN classification algorithm to construct a corpus training library that can dynamically learn and update, which can be improved on the filter corpus for constantly updated illegal content of the network, including text and pictures, and thus can better filter and investigate illegal content and its source. After that, the research direction will focus on the simplified updating of recognition and comparison algorithms and the optimization of the corpus learning ability in order to improve the efficiency of filtering, save time and resources.

Detection of Wildfires with Artificial Neural Networks

NASA Astrophysics Data System (ADS)

Umphlett, B.; Leeman, J.; Morrissey, M. L.

2011-12-01

Currently fire detection for the National Oceanic and Atmospheric Administration (NOAA) using satellite data is accomplished with algorithms and error checking human analysts. Artificial neural networks (ANNs) have been shown to be more accurate than algorithms or statistical methods for applications dealing with multiple datasets of complex observed data in the natural sciences. ANNs also deal well with multiple data sources that are not all equally reliable or equally informative to the problem. An ANN was tested to evaluate its accuracy in detecting wildfires utilizing polar orbiter numerical data from the Advanced Very High Resolution Radiometer (AVHRR). Datasets containing locations of known fires were gathered from the NOAA's polar orbiting satellites via the Comprehensive Large Array-data Stewardship System (CLASS). The data was then calibrated and navigation corrected using the Environment for Visualizing Images (ENVI). Fires were located with the aid of shapefiles generated via ArcGIS. Afterwards, several smaller ten pixel by ten pixel datasets were created for each fire (using the ENVI corrected data). Several datasets were created for each fire in order to vary fire position and avoid training the ANN to look only at fires in the center of an image. Datasets containing no fires were also created. A basic pattern recognition neural network was established with the MATLAB neural network toolbox. The datasets were then randomly separated into categories used to train, validate, and test the ANN. To prevent over fitting of the data, the mean squared error (MSE) of the network was monitored and training was stopped when the MSE began to rise. Networks were tested using each channel of the AVHRR data independently, channels 3a and 3b combined, and all six channels. The number of hidden neurons for each input set was also varied between 5-350 in steps of 5 neurons. Each configuration was run 10 times, totaling about 4,200 individual network evaluations. Thirty network parameters were recorded to characterize performance. These parameters were plotted with various data display techniques to determine which network configuration was not only most accurate in fire classification, but also the most computationally efficient. The most accurate fire classification network used all six channels of AVHRR data to achieve an accuracy ranging from 73-90%.

Artificial intelligence: Neural network model as the multidisciplinary team member in clinical decision support to avoid medical mistakes.

PubMed

Buzaev, Igor Vyacheslavovich; Plechev, Vladimir Vyacheslavovich; Nikolaeva, Irina Evgenievna; Galimova, Rezida Maratovna

2016-09-01

The continuous uninterrupted feedback system is the essential part of any well-organized system. We propose aLYNX concept that is a possibility to use an artificial intelligence algorithm or a neural network model in decision-making system so as to avoid possible mistakes and to remind the doctors to review tactics once more in selected cases. aLYNX system includes: registry with significant factors, decisions and results; machine learning process based on this registry data; the use of the machine learning results as the adviser. We show a possibility to build a computer adviser with a neural network model for making a choice between coronary aortic bypass surgery (CABG) and percutaneous coronary intervention (PCI) in order to achieve a higher 5-year survival rate in patients with angina based on the experience of 5107 patients. The neural network was trained by 4679 patients who achieved 5-year survival. Among them, 2390 patients underwent PCI and 2289 CABG. After training, the correlation coefficient ( r ) of the network was 0.74 for training, 0.67 for validation, 0.71 for test and 0.73 for total. Simulation of the neural network function has been performed after training in the two groups of patients with known 5-year outcome. The disagreement rate was significantly higher in the dead patient group than that in the survivor group between neural network model and heart team [16.8% (787/4679) vs. 20.3% (87/428), P  = 0.065)]. The study shows the possibility to build a computer adviser with a neural network model for making a choice between CABG and PCI in order to achieve a higher 5-year survival rate in patients with angina.

Optimization of neural network architecture for classification of radar jamming FM signals

NASA Astrophysics Data System (ADS)

Soto, Alberto; Mendoza, Ariadna; Flores, Benjamin C.

2017-05-01

The purpose of this study is to investigate several artificial Neural Network (NN) architectures in order to design a cognitive radar system capable of optimally distinguishing linear Frequency-Modulated (FM) signals from bandlimited Additive White Gaussian Noise (AWGN). The goal is to create a theoretical framework to determine an optimal NN architecture to achieve a Probability of Detection (PD) of 95% or higher and a Probability of False Alarm (PFA) of 1.5% or lower at 5 dB Signal to Noise Ratio (SNR). Literature research reveals that the frequency-domain power spectral densities characterize a signal more efficiently than its time-domain counterparts. Therefore, the input data is preprocessed by calculating the magnitude square of the Discrete Fourier Transform of the digitally sampled bandlimited AWGN and linear FM signals to populate a matrix containing N number of samples and M number of spectra. This matrix is used as input for the NN, and the spectra are divided as follows: 70% for training, 15% for validation, and 15% for testing. The study begins by experimentally deducing the optimal number of hidden neurons (1-40 neurons), then the optimal number of hidden layers (1-5 layers), and lastly, the most efficient learning algorithm. The training algorithms examined are: Resilient Backpropagation, Scaled Conjugate Gradient, Conjugate Gradient with Powell/Beale Restarts, Polak-Ribiére Conjugate Gradient, and Variable Learning Rate Backpropagation. We determine that an architecture with ten hidden neurons (or higher), one hidden layer, and a Scaled Conjugate Gradient for training algorithm encapsulates an optimal architecture for our application.

A novel stock forecasting model based on High-order-fuzzy-fluctuation Trends and Back Propagation Neural Network

PubMed Central

Dai, Zongli; Zhao, Aiwu; He, Jie

2018-01-01

In this paper, we propose a hybrid method to forecast the stock prices called High-order-fuzzy-fluctuation-Trends-based Back Propagation(HTBP)Neural Network model. First, we compare each value of the historical training data with the previous day's value to obtain a fluctuation trend time series (FTTS). On this basis, the FTTS blur into fuzzy time series (FFTS) based on the fluctuation of the increasing, equality, decreasing amplitude and direction. Since the relationship between FFTS and future wave trends is nonlinear, the HTBP neural network algorithm is used to find the mapping rules in the form of self-learning. Finally, the results of the algorithm output are used to predict future fluctuations. The proposed model provides some innovative features:(1)It combines fuzzy set theory and neural network algorithm to avoid overfitting problems existed in traditional models. (2)BP neural network algorithm can intelligently explore the internal rules of the actual existence of sequential data, without the need to analyze the influence factors of specific rules and the path of action. (3)The hybrid modal can reasonably remove noises from the internal rules by proper fuzzy treatment. This paper takes the TAIEX data set of Taiwan stock exchange as an example, and compares and analyzes the prediction performance of the model. The experimental results show that this method can predict the stock market in a very simple way. At the same time, we use this method to predict the Shanghai stock exchange composite index, and further verify the effectiveness and universality of the method. PMID:29420584

A novel stock forecasting model based on High-order-fuzzy-fluctuation Trends and Back Propagation Neural Network.

PubMed

Guan, Hongjun; Dai, Zongli; Zhao, Aiwu; He, Jie

2018-01-01

In this paper, we propose a hybrid method to forecast the stock prices called High-order-fuzzy-fluctuation-Trends-based Back Propagation(HTBP)Neural Network model. First, we compare each value of the historical training data with the previous day's value to obtain a fluctuation trend time series (FTTS). On this basis, the FTTS blur into fuzzy time series (FFTS) based on the fluctuation of the increasing, equality, decreasing amplitude and direction. Since the relationship between FFTS and future wave trends is nonlinear, the HTBP neural network algorithm is used to find the mapping rules in the form of self-learning. Finally, the results of the algorithm output are used to predict future fluctuations. The proposed model provides some innovative features:(1)It combines fuzzy set theory and neural network algorithm to avoid overfitting problems existed in traditional models. (2)BP neural network algorithm can intelligently explore the internal rules of the actual existence of sequential data, without the need to analyze the influence factors of specific rules and the path of action. (3)The hybrid modal can reasonably remove noises from the internal rules by proper fuzzy treatment. This paper takes the TAIEX data set of Taiwan stock exchange as an example, and compares and analyzes the prediction performance of the model. The experimental results show that this method can predict the stock market in a very simple way. At the same time, we use this method to predict the Shanghai stock exchange composite index, and further verify the effectiveness and universality of the method.

Automated mammographic breast density estimation using a fully convolutional network.

PubMed

Lee, Juhun; Nishikawa, Robert M

2018-03-01

The purpose of this study was to develop a fully automated algorithm for mammographic breast density estimation using deep learning. Our algorithm used a fully convolutional network, which is a deep learning framework for image segmentation, to segment both the breast and the dense fibroglandular areas on mammographic images. Using the segmented breast and dense areas, our algorithm computed the breast percent density (PD), which is the faction of dense area in a breast. Our dataset included full-field digital screening mammograms of 604 women, which included 1208 mediolateral oblique (MLO) and 1208 craniocaudal (CC) views. We allocated 455, 58, and 91 of 604 women and their exams into training, testing, and validation datasets, respectively. We established ground truth for the breast and the dense fibroglandular areas via manual segmentation and segmentation using a simple thresholding based on BI-RADS density assessments by radiologists, respectively. Using the mammograms and ground truth, we fine-tuned a pretrained deep learning network to train the network to segment both the breast and the fibroglandular areas. Using the validation dataset, we evaluated the performance of the proposed algorithm against radiologists' BI-RADS density assessments. Specifically, we conducted a correlation analysis between a BI-RADS density assessment of a given breast and its corresponding PD estimate by the proposed algorithm. In addition, we evaluated our algorithm in terms of its ability to classify the BI-RADS density using PD estimates, and its ability to provide consistent PD estimates for the left and the right breast and the MLO and CC views of the same women. To show the effectiveness of our algorithm, we compared the performance of our algorithm against a state of the art algorithm, laboratory for individualized breast radiodensity assessment (LIBRA). The PD estimated by our algorithm correlated well with BI-RADS density ratings by radiologists. Pearson's rho values of our algorithm for CC view, MLO view, and CC-MLO-averaged were 0.81, 0.79, and 0.85, respectively, while those of LIBRA were 0.58, 0.71, and 0.69, respectively. For CC view and CC-MLO averaged cases, the difference in rho values between the proposed algorithm and LIBRA showed statistical significance (P < 0.006). In addition, our algorithm provided reliable PD estimates for the left and the right breast (Pearson's ρ > 0.87) and for the MLO and CC views (Pearson's ρ = 0.76). However, LIBRA showed a lower Pearson's rho value (0.66) for both the left and right breasts for the CC view. In addition, our algorithm showed an excellent ability to separate each sub BI-RADS breast density class (statistically significant, p-values = 0.0001 or less); only one comparison pair, density 1 and density 2 in the CC view, was not statistically significant (P = 0.54). However, LIBRA failed to separate breasts in density 1 and 2 for both the CC and MLO views (P > 0.64). We have developed a new deep learning based algorithm for breast density segmentation and estimation. We showed that the proposed algorithm correlated well with BI-RADS density assessments by radiologists and outperformed an existing state of the art algorithm. © 2018 American Association of Physicists in Medicine.

Using human brain activity to guide machine learning.

PubMed

Fong, Ruth C; Scheirer, Walter J; Cox, David D

2018-03-29

Machine learning is a field of computer science that builds algorithms that learn. In many cases, machine learning algorithms are used to recreate a human ability like adding a caption to a photo, driving a car, or playing a game. While the human brain has long served as a source of inspiration for machine learning, little effort has been made to directly use data collected from working brains as a guide for machine learning algorithms. Here we demonstrate a new paradigm of "neurally-weighted" machine learning, which takes fMRI measurements of human brain activity from subjects viewing images, and infuses these data into the training process of an object recognition learning algorithm to make it more consistent with the human brain. After training, these neurally-weighted classifiers are able to classify images without requiring any additional neural data. We show that our neural-weighting approach can lead to large performance gains when used with traditional machine vision features, as well as to significant improvements with already high-performing convolutional neural network features. The effectiveness of this approach points to a path forward for a new class of hybrid machine learning algorithms which take both inspiration and direct constraints from neuronal data.

Deep Learning in Medical Imaging: General Overview

PubMed Central

Lee, June-Goo; Jun, Sanghoon; Cho, Young-Won; Lee, Hyunna; Kim, Guk Bae

2017-01-01

The artificial neural network (ANN)–a machine learning technique inspired by the human neuronal synapse system–was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and healthcare, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging. PMID:28670152

Deep Learning in Medical Imaging: General Overview.

PubMed

Lee, June-Goo; Jun, Sanghoon; Cho, Young-Won; Lee, Hyunna; Kim, Guk Bae; Seo, Joon Beom; Kim, Namkug

2017-01-01

The artificial neural network (ANN)-a machine learning technique inspired by the human neuronal synapse system-was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and healthcare, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging.

Application of a hybrid model of neural networks and genetic algorithms to evaluate landslide susceptibility

NASA Astrophysics Data System (ADS)

Wang, H. B.; Li, J. W.; Zhou, B.; Yuan, Z. Q.; Chen, Y. P.

2013-03-01

In the last few decades, the development of Geographical Information Systems (GIS) technology has provided a method for the evaluation of landslide susceptibility and hazard. Slope units were found to be appropriate for the fundamental morphological elements in landslide susceptibility evaluation. Following the DEM construction in a loess area susceptible to landslides, the direct-reverse DEM technology was employed to generate 216 slope units in the studied area. After a detailed investigation, the landslide inventory was mapped in which 39 landslides, including paleo-landslides, old landslides and recent landslides, were present. Of the 216 slope units, 123 involved landslides. To analyze the mechanism of these landslides, six environmental factors were selected to evaluate landslide occurrence: slope angle, aspect, the height and shape of the slope, distance to river and human activities. These factors were extracted in terms of the slope unit within the ArcGIS software. The spatial analysis demonstrates that most of the landslides are located on convex slopes at an elevation of 100-150 m with slope angles from 135°-225° and 40°-60°. Landslide occurrence was then checked according to these environmental factors using an artificial neural network with back propagation, optimized by genetic algorithms. A dataset of 120 slope units was chosen for training the neural network model, i.e., 80 units with landslide presence and 40 units without landslide presence. The parameters of genetic algorithms and neural networks were then set: population size of 100, crossover probability of 0.65, mutation probability of 0.01, momentum factor of 0.60, learning rate of 0.7, max learning number of 10 000, and target error of 0.000001. After training on the datasets, the susceptibility of landslides was mapped for the land-use plan and hazard mitigation. Comparing the susceptibility map with landslide inventory, it was noted that the prediction accuracy of landslide occurrence is 93.02%, whereas units without landslide occurrence are predicted with an accuracy of 81.13%. To sum up, the verification shows satisfactory agreement with an accuracy of 86.46% between the susceptibility map and the landslide locations. In the landslide susceptibility assessment, ten new slopes were predicted to show potential for failure, which can be confirmed by the engineering geological conditions of these slopes. It was also observed that some disadvantages could be overcome in the application of the neural networks with back propagation, for example, the low convergence rate and local minimum, after the network was optimized using genetic algorithms. To conclude, neural networks with back propagation that are optimized by genetic algorithms are an effective method to predict landslide susceptibility with high accuracy.

Mastering the game of Go with deep neural networks and tree search.

PubMed

Silver, David; Huang, Aja; Maddison, Chris J; Guez, Arthur; Sifre, Laurent; van den Driessche, George; Schrittwieser, Julian; Antonoglou, Ioannis; Panneershelvam, Veda; Lanctot, Marc; Dieleman, Sander; Grewe, Dominik; Nham, John; Kalchbrenner, Nal; Sutskever, Ilya; Lillicrap, Timothy; Leach, Madeleine; Kavukcuoglu, Koray; Graepel, Thore; Hassabis, Demis

2016-01-28

The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses 'value networks' to evaluate board positions and 'policy networks' to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.

A new algorithm to detect earthquakes outside the seismic network: preliminary results

NASA Astrophysics Data System (ADS)

Giudicepietro, Flora; Esposito, Antonietta Maria; Ricciolino, Patrizia

2017-04-01

In this text we are going to present a new technique for detecting earthquakes outside the seismic network, which are often the cause of fault of automatic analysis system. Our goal is to develop a robust method that provides the discrimination result as quickly as possible. We discriminate local earthquakes from regional earthquakes, both recorded at SGG station, equipped with short period sensors, operated by Osservatorio Vesuviano (INGV) in the Southern Apennines (Italy). The technique uses a Multi Layer Perceptron (MLP) neural network with an architecture composed by an input layer, a hidden layer and a single node output layer. We pre-processed the data using the Linear Predictive Coding (LPC) technique to extract the spectral features of the signals in a compact form. We performed several experiments by shortening the signal window length. In particular, we used windows of 4, 2 and 1 seconds containing the onset of the local and the regional earthquakes. We used a dataset of 103 local earthquakes and 79 regional earthquakes, most of which occurred in Greece, Albania and Crete. We split the dataset into a training set, for the network training, and a testing set to evaluate the network's capacity of discrimination. In order to assess the network stability, we repeated this procedure six times, randomly changing the data composition of the training and testing set and the initial weights of the net. We estimated the performance of this method by calculating the average of correct detection percentages obtained for each of the six permutations. The average performances are 99.02%, 98.04% and 98.53%, which concern respectively the experiments carried out on 4, 2 and 1 seconds signal windows. The results show that our method is able to recognize the earthquakes outside the seismic network using only the first second of the seismic records, with a suitable percentage of correct detection. Therefore, this algorithm can be profitably used to make earthquake automatic analyses more robust and reliable. Finally, with appropriate tuning, it can be integrated in multi-parametric systems for monitoring high natural risk areas.

MLP based LOGSIG transfer function for solar generation monitoring

NASA Astrophysics Data System (ADS)

Hashim, Fakroul Ridzuan; Din, Muhammad Faiz Md; Ahmad, Shahril; Arif, Farah Khairunnisa; Rizman, Zairi Ismael

2018-02-01

Solar panel is one of the renewable energy that can reduce the environmental pollution and have a wide potential of application. The exact solar prediction model will give a big impact on the management of solar power plants and the design of solar energy systems. This paper attempts to use Multilayer Perceptron (MLP) neural network based transfer function. The MLP network can be used to calculate the temperature module (TM) in Malaysia. This can be done by simulating the collected data of four weather variables which are the ambient temperature (TA), local wind speed (VW), solar radiation flux (GT) and the relative humidity (RH) as the input into the neural network. The transfer function will be applied to the 14 types of training. Finally, an equation from the best training algorithm will be deduced to calculate the temperature module based on the input of weather variables in Malaysia.

Learning State Space Dynamics in Recurrent Networks

NASA Astrophysics Data System (ADS)

Simard, Patrice Yvon

Fully recurrent (asymmetrical) networks can be used to learn temporal trajectories. The network is unfolded in time, and backpropagation is used to train the weights. The presence of recurrent connections creates internal states in the system which vary as a function of time. The resulting dynamics can provide interesting additional computing power but learning is made more difficult by the existence of internal memories. This study first exhibits the properties of recurrent networks in terms of convergence when the internal states of the system are unknown. A new energy functional is provided to change the weights of the units in order to the control the stability of the fixed points of the network's dynamics. The power of the resultant algorithm is illustrated with the simulation of a content addressable memory. Next, the more general case of time trajectories on a recurrent network is studied. An application is proposed in which trajectories are generated to draw letters as a function of an input. In another application of recurrent systems, a neural network certain temporal properties observed in human callosally sectioned brains. Finally the proposed algorithm for stabilizing dynamics around fixed points is extended to one for stabilizing dynamics around time trajectories. Its effects are illustrated on a network which generates Lisajous curves.

Efficient self-organizing multilayer neural network for nonlinear system modeling.

PubMed

Han, Hong-Gui; Wang, Li-Dan; Qiao, Jun-Fei

2013-07-01

It has been shown extensively that the dynamic behaviors of a neural system are strongly influenced by the network architecture and learning process. To establish an artificial neural network (ANN) with self-organizing architecture and suitable learning algorithm for nonlinear system modeling, an automatic axon-neural network (AANN) is investigated in the following respects. First, the network architecture is constructed automatically to change both the number of hidden neurons and topologies of the neural network during the training process. The approach introduced in adaptive connecting-and-pruning algorithm (ACP) is a type of mixed mode operation, which is equivalent to pruning or adding the connecting of the neurons, as well as inserting some required neurons directly. Secondly, the weights are adjusted, using a feedforward computation (FC) to obtain the information for the gradient during learning computation. Unlike most of the previous studies, AANN is able to self-organize the architecture and weights, and to improve the network performances. Also, the proposed AANN has been tested on a number of benchmark problems, ranging from nonlinear function approximating to nonlinear systems modeling. The experimental results show that AANN can have better performances than that of some existing neural networks. Crown Copyright © 2013. Published by Elsevier Ltd. All rights reserved.

Segmentation of tumor and edema along with healthy tissues of brain using wavelets and neural networks.

PubMed

Demirhan, Ayşe; Toru, Mustafa; Guler, Inan

2015-07-01

Robust brain magnetic resonance (MR) segmentation algorithms are critical to analyze tissues and diagnose tumor and edema in a quantitative way. In this study, we present a new tissue segmentation algorithm that segments brain MR images into tumor, edema, white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). The detection of the healthy tissues is performed simultaneously with the diseased tissues because examining the change caused by the spread of tumor and edema on healthy tissues is very important for treatment planning. We used T1, T2, and FLAIR MR images of 20 subjects suffering from glial tumor. We developed an algorithm for stripping the skull before the segmentation process. The segmentation is performed using self-organizing map (SOM) that is trained with unsupervised learning algorithm and fine-tuned with learning vector quantization (LVQ). Unlike other studies, we developed an algorithm for clustering the SOM instead of using an additional network. Input feature vector is constructed with the features obtained from stationary wavelet transform (SWT) coefficients. The results showed that average dice similarity indexes are 91% for WM, 87% for GM, 96% for CSF, 61% for tumor, and 77% for edema.

Designing Artificial Neural Networks Using Particle Swarm Optimization Algorithms

PubMed Central

Vázquez, Roberto A.

2015-01-01

Artificial Neural Network (ANN) design is a complex task because its performance depends on the architecture, the selected transfer function, and the learning algorithm used to train the set of synaptic weights. In this paper we present a methodology that automatically designs an ANN using particle swarm optimization algorithms such as Basic Particle Swarm Optimization (PSO), Second Generation of Particle Swarm Optimization (SGPSO), and a New Model of PSO called NMPSO. The aim of these algorithms is to evolve, at the same time, the three principal components of an ANN: the set of synaptic weights, the connections or architecture, and the transfer functions for each neuron. Eight different fitness functions were proposed to evaluate the fitness of each solution and find the best design. These functions are based on the mean square error (MSE) and the classification error (CER) and implement a strategy to avoid overtraining and to reduce the number of connections in the ANN. In addition, the ANN designed with the proposed methodology is compared with those designed manually using the well-known Back-Propagation and Levenberg-Marquardt Learning Algorithms. Finally, the accuracy of the method is tested with different nonlinear pattern classification problems. PMID:26221132

Automatic voice recognition using traditional and artificial neural network approaches

NASA Technical Reports Server (NTRS)

Botros, Nazeih M.

1989-01-01

The main objective of this research is to develop an algorithm for isolated-word recognition. This research is focused on digital signal analysis rather than linguistic analysis of speech. Features extraction is carried out by applying a Linear Predictive Coding (LPC) algorithm with order of 10. Continuous-word and speaker independent recognition will be considered in future study after accomplishing this isolated word research. To examine the similarity between the reference and the training sets, two approaches are explored. The first is implementing traditional pattern recognition techniques where a dynamic time warping algorithm is applied to align the two sets and calculate the probability of matching by measuring the Euclidean distance between the two sets. The second is implementing a backpropagation artificial neural net model with three layers as the pattern classifier. The adaptation rule implemented in this network is the generalized least mean square (LMS) rule. The first approach has been accomplished. A vocabulary of 50 words was selected and tested. The accuracy of the algorithm was found to be around 85 percent. The second approach is in progress at the present time.