By Topic

Neural Networks, 1991. 1991 IEEE International Joint Conference on

Date 18-21 Nov. 1991

Filter Results

Displaying Results 1 - 25 of 444
  • 1991 IEEE International Joint Conference on Neural Networks (Cat. No.91CH3065-0)

    Publication Year: 1991
    Request permission for commercial reuse | PDF file iconPDF (47 KB)
    Freely Available from IEEE
  • Behaviors of transform domain backpropagation (BP) algorithm

    Publication Year: 1991, Page(s):349 - 354 vol.1
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (272 KB)

    Several discrete orthogonal transforms have been used to study the behaviors of transform-domain backpropagation (BP) algorithms. Two examples of computer simulation show that, on selecting the appropriate parameters and the suitable structures of a neural network, the performance of the transform-domain BP algorithm is somewhat better than that of the original time-domain BP algorithm, regardless... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis of time series by neural networks

    Publication Year: 1991, Page(s):355 - 360 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (228 KB)

    Neural networks have been constructed to analyze artificial time series derived from the Tent and Henon map as well as population data of the Canadian lynx. Simple three-layer forwardfeed networks, trained on a small sample data set, provided reasonably good fit to the data and performed well on short-term predictions. Simple neural network models trained on small data sets can perform quite well ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A general purpose neurocomputer

    Publication Year: 1991, Page(s):361 - 366 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (260 KB)

    Presents a neural network, composed of linear units with threshold, as the CPU of a stored program MIMD architecture. The Caianiello formalism, is introduced as an aid to implement the arithmetic and control algorithms, needed for the smooth running of this general-purpose system. That is, in the neural net both the arithmetic and logic algorithms and the operating system have been implemented. Th... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Classification of hand-written digits by a large scale neural network `CombNET-II'

    Publication Year: 1991, Page(s):1021 - 1026 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (276 KB)

    The authors describe the network structure and the learning procedure of CombNET-II together with experimental results on hand-written digit classification. CombNET-II uses the self-growing neural network learning procedure for training the stem network. After training the stem network, all input data are partitioned into category groups. Then, branch networks are trained for every category group.... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The analysis of the augmented ART1 neural network

    Publication Year: 1991, Page(s):2658 - 2663 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (272 KB)

    The dynamics of the augmented ART1 neural network (AART1-NN) introduced by L. Heileman and M. Georgiopoulos (1991) are described by a set of nonlinear differential equations that facilitate the real-time implementation of the ART1 neural network. The AART1-NN equations are summarized. It is shown that under certain parameter constraints the AART1 model behaves in the same manner as the ART1 model.... View full abstract»

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

    Publication Year: 1991, Page(s):671 - 677 vol.1
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (312 KB)

    The authors show how to synthesize, given a Bayesian network description of a probability distribution, an instance of P. Smolensky's harmony network (1986) that computes maximum-likelihood completions to partial value assignments, according to the given distribution. As an application, they present a scheme for translating a high-level description of a conceptual hierarchy, with default values an... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A face graph method using a fuzzy neural network for expressing conditions of complex systems

    Publication Year: 1991, Page(s):1600 - 1605 vol.2
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (248 KB)

    The face graph method with such varying elements as dyes, eyebrows, mouth, etc. is used for expressing multidimensional data. Since human beings are very sensitive to human faces, one can easily evaluate the multidimensional data expressed by the face graph. The authors present a novel approach of the face graph method using a fuzzy neural network for expressing conditions of complex systems. Expe... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Complexity measures for classes of neural networks with variable weight bounds

    Publication Year: 1991, Page(s):2624 - 2630 vol.3
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (300 KB)

    The author derives complexity measures for classes of single-hidden-layer feedforward networks based on the capacity and metric entropy of a class of functions. Based on these measures, some deficiencies in commonly used complexity-penalty terms implemented to prevent overfitting are indicated View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An enhancement to MLP model to enforce closed decision regions

    Publication Year: 1991, Page(s):729 - 733 vol.1
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (304 KB)

    Describes a modification of the basic MLP (multilayer perceptron) model implemented to improve its capability to enforce closed decision regions. The authors' proposal is to use hyperspheres instead of hyperplanes on the first hidden layer, and in turn combine them through the next layers. After training, the decision regions will be naturally closed because they are built on simple computational ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • ARENA, a rule evaluating neural assistant that performs rule-based logic optimization

    Publication Year: 1991, Page(s):678 - 683 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (300 KB)

    ARENA, a rule evaluating neural assistant, has been integrated with a logic optimization system and used to determine which logic transformation out of many available should be used. Starting with no prior knowledge, the network learned to recognize the proper time to apply logic optimization rules and to select a useful rule accordingly. By selectively repeating previously learned patterns, the n... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Goal directed model inversion

    Publication Year: 1991, Page(s):2422 - 2427 vol.3
    Cited by:  Papers (2)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (292 KB)

    A new neural network technique for model inversion called goal directed model inversion (GDM) is presented. It allows the system to produce an inverse model in a goal directed manner. The major advantage of an inverse model created in this matter is that it can adapt to unexpected changes in the system with which it must interact. As an example of the GDMI technique, a simple kinematic controller ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A recursive neural system for memorizing systems of values arranged in a tree like structure

    Publication Year: 1991, Page(s):1776 - 1781 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (308 KB)

    It is pointed out that, in general, adaptive automata have a cost function for organizing relations between input signals and output signals. But most of these automata have been studied with an a priori fixed cost function. For this reason the authors introduce a self-constructing value system with profits or losses. This ability helps the automaton to adapt to its environment. In the proposed me... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Synapse-X: a general-purpose neurocomputer architecture

    Publication Year: 1991, Page(s):2168 - 2176 vol.3
    Cited by:  Papers (6)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (400 KB)

    A neurocomputer architecture is described which features the following characteristics: the compute-bound elementary operations are extracted from the set of neural algorithms; the elementary operations are executed by a specific VLSI neural signal processor MA16, and noncompute-bound operations by commercially available digital signal processors (DSPs) or microprocessors; the fine-grain systolic ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A novel approach to electrical load forecasting based on a neural network

    Publication Year: 1991, Page(s):1172 - 1177 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (256 KB)

    The authors demonstrate how an artificial neural network can be used to forecast electrical load demand. This network is based on the nonstatistical neural paradigm, backpropagation, which is found to be effective for accurate forecasting of electrical load. The major advantage of using an artificial neural network as opposed to other techniques for electrical load forecasting is that the network ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An accelerated back propagation training algorithm

    Publication Year: 1991, Page(s):165 - 170 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (192 KB)

    The concept of backpropagation training with a gradual increase in accuracy is presented. The concept may be incorporated with the basic backpropagation algorithm or with the backpropagation algorithm with a momentum weight change to result in significant speed improvements without increasing the size of the network or requiring additional support hardware. Experimental results show that the propo... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Determination of the parameters in a modified Hopfield-Tank model [2] on solving TSP

    Publication Year: 1991, Page(s):2455 - 2460 vol.3
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (244 KB)

    The work of S.V.B. Aiyer et al. (IEEE Trans. Neural Net., vol.1, pp.204-215, June 1990) is extended. Through the analysis of eigenvalues, the eigenvectors, and the energy functions, some relations among the parameters are given theoretically, especially for D and A 1, which play an important role in the model. Theoretically, these relations can be used to solve traveli... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The backpropagation unfold recurrent rule

    Publication Year: 1991, Page(s):734 - 739 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (200 KB)

    A novel learning algorithm for recurrent backpropagation (BP) networks is introduced. Existing learning algorithms of the recurrent models are mostly based on the static generalized delta-rule of the BP algorithm which does not give satisfactory results for a recurrent network. The authors derived the BP unfold recurrent rule (BURR) by unfolding the structure of the recurrent network. Using this l... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Reduction of necessary precision for the learning of pattern recognition

    Publication Year: 1991, Page(s):1795 - 1800 vol.2
    Cited by:  Papers (1)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (272 KB)

    The authors propose a novel learning algorithm with weighted error function (WEF). They have reduced the necessary precision for the learning of multi-font alpha-numeric recognition to 10-bit fixed point precision using the WEF. The WEF raises the recognition accuracy by more than 25% when the precision of all operations (including multiplication and addition) and the precision of all data (includ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neuromorphic sensing and control-impact control of robotic manipulator

    Publication Year: 1991, Page(s):1967 - 1972 vol.3
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (308 KB)

    A new approach for impact control of a robotic manipulator by a neural network (NN)-based controller is presented. Collisions are very quick phenomena and have strong nonlinearity. Therefore, it is difficult to sense collisions and to control a robotic manipulator undergoing collisions. The proposed approach has robustness against the impact force. It also has capabilities for sensing and percepti... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Pattern classification for remote sensing using neural network

    Publication Year: 1991, Page(s):652 - 658 vol.1
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (272 KB)

    The authors propose a pattern classification method for remote sensing data based on neural network theory. From geographical knowledge and Kohonen's self-organization feature maps, training areas for each pattern are selected. Using the backpropagation algorithm, a layered neural network is trained such that the training patterns can be classified within a level. After training the network, some ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Self-learning neural M-ary tree classifier

    Publication Year: 1991, Page(s):1612 - 1617 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (232 KB)

    A novel version of a multilayer neural network, called the self-learning neural model (SLNM), is presented. The different level structures, dynamics, and learning strategies of the SLNM are investigated. This neural model can be used as adaptive nonparametric neural-net classifiers or clusters, which can be trained by unlabeled data. An M-ary decision tree structured classifier with the b... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural network models for rule-based reasoning

    Publication Year: 1991, Page(s):503 - 508 vol.1
    Cited by:  Papers (7)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (392 KB)

    Neural network (connectionist) models of rule-based reasoning are investigated, and it is shown that while such models usually carry out reasoning in exactly the same way as symbolic systems, they have more to offer in terms of commonsense reasoning. A connectionist architecture for commonsense reasoning, CONSYDERR, is proposed to account for commonsense reasoning patterns and to remedy the brittl... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Control of a robotic manipulating arm by a neural network simulation of the human cerebral and cerebellar cortical processes

    Publication Year: 1991, Page(s):1444 - 1447 vol.2
    Cited by:  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (188 KB)

    The authors propose a system commanding a robotic manipulating arm under visual control, based on brain modeling. In this model, the movement command is learned by a network which links two subsystems together: a cerebral subsystem which can learn a goal, and a second subsystem responsible for quantitative adjustments and coordination. Those two subsystems are complementary because each subsystem ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural networks based adaptive predictors for nonlinear dynamical systems

    Publication Year: 1991, Page(s):777 - 782 vol.1
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (268 KB)

    Develops a novel predictive model based on multilayer neural networks for nonlinear dynamical systems. Two isomorphic multilayer neural networks are used together to implement the proposed predictor. One is the learning network which learns the input-output behavior of the system. The other is the prediction network, which gets its weights mapped from the learning network and generates the predict... View full abstract»

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