By Topic

1998 IEEE International Joint Conference on Neural Networks Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98CH36227)

4-9 May 1998

Go

Filter Results

Displaying Results 1 - 25 of 158
  • The 1998 IEEE International Joint Conference on Neural Networks Proceedings [front matter]

    Publication Year: 1998, Page(s):i - xxxvi
    Request permission for commercial reuse | PDF file iconPDF (2194 KB)
    Freely Available from IEEE
  • Author's index

    Publication Year: 1998, Page(s): A
    Request permission for commercial reuse | PDF file iconPDF (466 KB)
    Freely Available from IEEE
  • Distortion-invariant object representation and discrimination using an FST neural net

    Publication Year: 1998, Page(s):1971 - 1974 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (336 KB)

    A feature space trajectory (FST) neural net is used to represent distorted versions of an object. Its use in determining the class and pose of an object is addressed with attention to: which aspect views and how many are needed to represent an object, which viewpoint gives the best pose animate and the best classification confidence. New eigen features are also advanced to provide improved results View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Pruning a classifier based on a self-organizing map using Boolean function formalization

    Publication Year: 1998, Page(s):1910 - 1915 vol.3
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (476 KB)

    An algorithm is presented to minimize the number of neurons needed for a classifier based on Kohonens self-organizing maps (SOM), or on any other “code-book type” (or “prototype based”) classifier such as Kohonens linear vector quantization (LVQ), K-means or nearest neighbor. The neuron minimization problem is formalized as a problem of simplification of Boolean functions, ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Artificial neural network based planning of generation ready reserve capacity

    Publication Year: 1998, Page(s):1966 - 1970 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (388 KB)

    Describes the use of an artificial neural network (ANN) to plan ready reserve (RR) capacity in a power system. The main role of the ANN is to estimate the needs for RR in the future on the basis of historical data such as maximal and minimum demand not supplied, and the average load that has strained the system. The article deals with the results of ANN capabilities to interpolate from the past da... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Structural learning of recurrent RBF networks with M-apoptosis

    Publication Year: 1998, Page(s):2390 - 2395 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (488 KB)

    The apoptosis is an active form of cell death which plays an important role during embryonic development. We propose a unified approach, called M-apoptosis, to the structural learning of recurrent RBF networks. Minkowski norm of the first order derivatives of RBF networks with respect to input variables is added to the cost function for determining the unknown parameters. The parameters are change... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Detection and classification of insect sounds in a grain silo using a neural network

    Publication Year: 1998, Page(s):1760 - 1765 vol.3
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (404 KB)

    This paper presents the application of a time-delay neural network to the detection and classification of time signatures produced by insect sounds in a stored grain silo. Conventional methods of insect monitoring can only detect some of the adult insects and none of the larvae insects, which are the most destructive to the grain. The acoustic vibrations generated by the adult and larvae when movi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An optimal method for linear threshold neural network synthesis

    Publication Year: 1998, Page(s):1905 - 1909 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (428 KB)

    In the digital design area the minimized function of binary variables may be represented by two levels of AND/OR gates. However, depending upon the application, the design may require a large number of gates. We propose a method that is capable of reducing the required number of gates necessary to realize an N-dim binary function by implementing linear threshold units. Hence, we propose an approac... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural α-feature detector for feature detection and generalization

    Publication Year: 1998, Page(s):1845 - 1850 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (404 KB)

    We propose a neural α-feature detector used to extract a small number of main or essential features in input patterns. Features can be detected by controlling α-entropy for α-feature detectors. The α-entropy is defined by the difference between Renyi entropy and Shannon entropy. The α-entropy controller aims to maximize information contained in a few important α... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A data approach alternative at system identification and modeling using the self-organizing associative memory (SAM) system

    Publication Year: 1998, Page(s):2447 - 2452 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (512 KB)

    We introduce a data-based approach alternative to the rule-based parameter approach toward system identification. Motivated by the design-intensive problem of the parameter approach, the self-organizing associative memory (SAM) system seeks to represent the system using a subset of stored training data. We surmise that knowledge is association between memorized objects, not memorized rules. We pos... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Softmax-network and S-Map-models for density-generating topographic mappings

    Publication Year: 1998, Page(s):2268 - 2272 vol.3
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (436 KB)

    We propose a neural network model for density-generating topographic mappings. The model consists of two parts: the Softmax-network, and the S-Map. The Softmax-network implements the softmax function, so that each neuron's output is a softmax of the weighted sum of the input to that neuron and to its neighbors. The S-Map, based on the Softmax-network, utilises a Hebbian-like learning scheme for th... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Prediction of stochastic fields by RBFNN

    Publication Year: 1998, Page(s):1960 - 1965 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (532 KB)

    A statistical description of stochastic phenomena is utilized to formulate a general modeler of physical laws having the structure of a radial basis function neural network. As a basis for the description of a phenomenon the concept of an auto-regressive field is utilized. Its evolution is represented by a non-linear mapping relation in which the generating function is modeled empirically by a non... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Using an MDL-based cost function with neural networks

    Publication Year: 1998, Page(s):2384 - 2389 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (520 KB)

    The minimum description length (MDL) principle is an information theoretically based method to learn models from data. This paper presents an approach to efficiently use an MDL-based cost function with neural networks. As usual, the cost function can be used to adapt the parameters in the network, but it can also include terms to measure the complexity of the structure of the network and can thus ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recognition of handwritten similar Chinese characters by self-growing probabilistic decision-based neural networks

    Publication Year: 1998, Page(s):1754 - 1759 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (536 KB)

    We introduce a neural network solution that is capable of modeling minor differences among similar characters, and is robust to various personal handwriting styles. The self-growing probabilistic decision-based neural network (SPDNN) is a probabilistic type neural networks, which adopts a hierarchical network structure with nonlinear basis functions and a competitive credit-assignment scheme. Base... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An index-based classification scheme using neural networks for multiclass problems

    Publication Year: 1998, Page(s):1899 - 1904 vol.3
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (476 KB)

    Proposes a novel classification scheme based on a semi-supervised backpropagation (SSBP) learning algorithm for multiclass problems. The proposed approach can derive a fuzzy index as a classification quantifier for each specific class by means of a specially-defined cost function. Misclassifications can be removed through introducing an extra indeterminate class for some complicated non-probabilis... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural network for seismic horizon picking

    Publication Year: 1998, Page(s):1840 - 1844 vol.3
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (232 KB)

    A Hopfield neural network can solve optimization problems. We use a Hopfield net for seismic horizon picking. The peak position of each seismic wavelet corresponds to one neuron. We transform the constraints for detecting local horizon patterns and the constraints for extracting one horizon each time into the system energy function. From the theory of Hopfield nets, changing the values of neurons ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recurrent RBF networks for suspension system modeling and wear diagnosis of a damper

    Publication Year: 1998, Page(s):2441 - 2446 vol.3
    Cited by:  Papers (3)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (452 KB)

    From the beginning of the nineties, artificial neural networks began to be used for the identification and control of nonlinear systems. Thus, new types of nets appeared permitting to model the behaviour of dynamical systems. These recurrent networks extend the range of engineering applications and emerged as new powerful architectures. Besides, local models like RBFN (or neuro-fuzzy structures) p... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Phased backpropagation: a hybrid of BPTT and temporal BP

    Publication Year: 1998, Page(s):2262 - 2267 vol.3
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (420 KB)

    We present a synthesis of backpropagation through time (BPTT) and temporal backpropagation (TB) that permits the efficiency of TB in dealing with delay lines to be combined with the generality of BPTT with respect to arbitrary discrete-time network structures. We express our formulation in terms of an ordered network, subsuming less general network architectures such as time-delay networks and rec... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Face recognition using curvilinear component analysis

    Publication Year: 1998, Page(s):1778 - 1783 vol.3
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (428 KB)

    Automated face recognition can be applied in a number of situations including personal identification, mug shot matching, store security, and crowd surveillance. A large number of techniques based on linear methods of dimensionality reduction, such as principal component analysis (PCA), have recently been proposed. Motivated by the possibility of increased performance, we pursue in this paper a fa... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Model switching by channel fusion for network pruning and efficient feature extraction

    Publication Year: 1998, Page(s):1861 - 1866 vol.3
    Cited by:  Papers (4)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (500 KB)

    Introduces a feature dimension reduction method called channel fusion, and a criterion for redundant channel detection called effective map distance. Channel fusion locally reduces the feature dimension by replacing the redundant channel pair with a single channel, suppressing the map distance between the two models. It is applicable to network model switching such as pruning hidden layer units an... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Totally model-free reinforcement learning by actor-critic Elman networks in non-Markovian domains

    Publication Year: 1998, Page(s):2016 - 2021 vol.3
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (596 KB)

    We describe how an actor-critic reinforcement learning agent in a non-Markovian domain finds an optimal sequence of actions in a totally model-free fashion; that is, the agent neither learns transitional probabilities and associated rewards, nor by how much the state space should be augmented so that the Markov property holds. In particular, we employ an Elman-type recurrent neural network to solv... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A temporal difference method-based prediction scheme applied to fading power signals

    Publication Year: 1998, Page(s):1954 - 1959 vol.3
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (560 KB)

    We first briefly discuss the operating principle of the temporal difference (TD) method. A TD method-based multi-step ahead prediction scheme using the modified Elman neural network (MENN) is then set up. This prediction approach provides for online adaptation and fast convergence rate. Next, it is applied to the prediction of the occurrence of long term deep fading in mobile communication systems... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Ordered fuzzy ARTMAP: a fuzzy ARTMAP algorithm with a fixed order of pattern presentation

    Publication Year: 1998, Page(s):1717 - 1722 vol.3
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (488 KB)

    In this paper we introduce a procedure that identifies a fixed order of training pattern presentation for fuzzy ARTMAP. The resulting algorithm is named ordered fuzzy ARTMAP. Experimental results have demonstrated that ordered fuzzy ARTMAP achieves a network performance that is better than the average fuzzy ARTMAP network performance (averaged over a fixed number of random orders of pattern presen... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The role of excitatory and inhibitory learning in EXIN networks

    Publication Year: 1998, Page(s):2378 - 2383 vol.3
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (460 KB)

    We propose modifications for the learning rules of Marshall's EXIN (excitatory+inhibitory) neural network model in order to decrease its computational complexity and understand the role of the weight updating learning rules in correctly encoding superimposed and ambiguous input patterns. The MEXIN (modified EXIN) models introduce mixtures of competitive and Hebbian updating rules. In this case, on... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Time series prediction by a modular structured neural network

    Publication Year: 1998, Page(s):2501 - 2506 vol.3
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (464 KB)

    This paper proposes a prediction method for nonstationary time series data with time varying parameters. First a modular structured neural network is newly introduced for the purpose of modeling the changing properties of time varying parameters. This neural network is constructed by the hierarchical combination of neural networks NNT for time series data prediction and NNW for weight prediction. ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.