Scheduled Maintenance on March 25th, 2017:
Single article purchases and IEEE account management will be unavailable from 4:00 AM until 6:30 PM (ET). We apologize for the inconvenience.
By Topic

International 1989 Joint Conference on Neural Networks

0-0 1989

Filter Results

Displaying Results 1 - 25 of 424
  • Neural network learning time: effects of network and training set size

    Publication Year: 1989, Page(s):395 - 401 vol.2
    Cited by:  Papers (5)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (413 KB)

    The learning time for two-layer backpropagation networks is examined in the context of learning Boolean logic equations from examples. In particular, the relationship between the number of inputs, hidden units, and training set vectors and the learning time is investigated. The networks, the training algorithm, and the tasks are described. The parameter variations and the set of simulations perfor... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Consistent inference of probabilities in layered networks: predictions and generalizations

    Publication Year: 1989, Page(s):403 - 409 vol.2
    Cited by:  Papers (34)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (514 KB)

    The problem of learning a general input-output relation using a layered neural network is discussed in a statistical framework. By imposing the consistency condition that the error minimization be equivalent to a likelihood maximization for training the network, the authors arrive at a Gibbs distribution on a canonical ensemble of networks with the same architecture. This statistical description e... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Nonlinear data analysis and multilayer perceptrons

    Publication Year: 1989, Page(s):411 - 415 vol.2
    Cited by:  Papers (16)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (338 KB)

    The correspondence between multilayer perceptrons (MLPs) with linear processing elements and classical data analysis methods (principal component analysis, discriminant analysis) was shown by P. Galinari et al. (see Proc. IEEE ICNN-88, p.I-391-9, 1988). The authors extend their results to the nonlinear case and show that MLP with nonlinear elements approximates the nonlinear data analysis methods.... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • BPS: a learning algorithm for capturing the dynamic nature of speech

    Publication Year: 1989, Page(s):417 - 423 vol.2
    Cited by:  Papers (20)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (514 KB)

    A novel backpropagation learning algorithm for a particular class of dynamic neural networks in which some units have a local feedback is proposed. Hence these networks can be trained to respond to sequences of input patterns. This algorithm has the same order of space and time requirements as backpropagation applied to feedforward networks. The authors present experimental results and comparisons... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Inversion of multilayer nets

    Publication Year: 1989, Page(s):425 - 430 vol.2
    Cited by:  Papers (80)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (758 KB)

    The method of inversion for arbitrary continuous multilayer nets is developed. The inversion is done by computing iteratively an input vector which minimizes the least-mean-square errors to approximate a given output target. This inversion is not unique for given targets and depends on the starting point in input space. The inversion method turns out to be a valuable tool for the examination of mu... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The mathematical theory of learning algorithms for Boltzmann machines

    Publication Year: 1989, Page(s):431 - 437 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (805 KB)

    The author analyzes a version of a well-known learning algorithm for Boltzmann machines, based on the usual alternation between learning and hallucinating phases. He outlines the rigorous proof that, for suitable choices of the parameters, the evolution of the weights follows very closely, with very high probability, an integral trajectory of the gradient of the likelihood function whose global ma... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A network with multi-partitioning units

    Publication Year: 1989, Page(s):439 - 442 vol.2
    Cited by:  Papers (1)  |  Patents (4)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (298 KB)

    The authors propose a fuzzy partition model (FPM), a multilayer feedforward perceptron-like network. The most important point of FPM is that it has multiple-input/output units which are upper compatible with the threshold units commonly used in the backpropagation (BP) model. The number of outputs is called the degree N of that unit, and an FPM unit can classify input patterns into N categories. B... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Solving the problem of overfitting of the pseudo-inverse solution for classification learning

    Publication Year: 1989, Page(s):443 - 449 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (421 KB)

    The authors investigate the pseudoinverse solution for the learning of a binary classification. They address the problem of overfitting of this solution, i.e. the fact that the generalization rate can be relatively low although the learning rate is very high. They interpret this phenomenon with respect to the standard deviation of the eigenvalues of the covariance matrix of the learned patterns. T... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An additional hidden unit test for neglected nonlinearity in multilayer feedforward networks

    Publication Year: 1989, Page(s):451 - 455 vol.2
    Cited by:  Papers (34)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (499 KB)

    The author presents a statistical test of the hypothesis that a given multilayer feedforward network exactly represents some unknown mapping subject to inherent noise against the alternative that the network neglects some nonlinear structure in the mapping, leading to potentially avoidable approximation errors. The tests are based on methods that statistically determine whether or not there is som... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An investigation of adaptive learning implemented in an optically controlled neural network

    Publication Year: 1989, Page(s):457 - 463 vol.2
    Cited by:  Papers (4)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (549 KB)

    The authors used a synaptic array of amorphous silicon photoconductors to build a feedforward adaptive neural network. Using backpropagation learning, this network can be taught to perform simple tasks of analog computation. The performance of the network compares well with that of an idealized model, despite significant component variation and externally imposed constraints not accounted for in t... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Experiments of learning in optical perceptron-like and multilayer neural networks

    Publication Year: 1989, Page(s):465 - 471 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (477 KB)

    An optical multilayer network with backpropagation learning capability is constructed and tested. The modifiable connection weights and the unit plane are realized by using a photorefractive crystal and a microchannel spatial light modulator, respectively. First, in a simple optical perceptron-like network, the experiment of learning is conducted. The learning rate is optimally determined by setti... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optical implementation of a high order associative memory

    Publication Year: 1989, Page(s):473 - 475 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (337 KB)

    An experimental implementation of a 10-input quadratic associative memory based on a compact optical 'chip' design is described. Such a design lends itself easily to the implementation of kth-order associative memories. The authors have experimentally investigated the performance of the network, finding that it stably stores up to four vectors and has error correction capability. Experimental resu... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A new approach to a GaAs/AlGaAs optical neurochip with three layered structure

    Publication Year: 1989, Page(s):477 - 482 vol.2
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (253 KB)

    A GaAs/AlGaAs optical synaptic interconnection device for neural networks is reported. It consists of a light-emitting-diode array, an interconnection matrix, and a photodiode array, which are integrated in a hybrid-layered structure on a GaAs substrate. The device structure and characteristics are reported. The fabricated device can simulate a 32-neuron system. The experimental results for a Hopf... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optical orthogonal neural network associative memory with luminescent rebroadcasting devices

    Publication Year: 1989, Page(s):483 - 485 vol.2
    Cited by:  Papers (4)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (272 KB)

    The operation of high-resolution luminescent rebroadcasting devices is described, including analog addition and multiplication. The outer-product formulation reminiscent of Hebb's learning rule, is used to store associated vectors for a 32-bit orthogonal heteroassociative memory. Experimental results show that the luminescent rebroadcasting material is ideal for this type of associative memory. Th... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Highly-parallel optical adaptive neural computers with layered and segmented structures

    Publication Year: 1989, Page(s):487 - 494 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (474 KB)

    In order to make highly parallel large-scale optical adaptive neural computers for recognition, vision, and other man-machine interface applications, a new technological approach to all-optic neurons and adaptive synapses of layered and segmented neural architectures is approached. Basic optical elements will be used in the proposed three-layer segmented neural architecture in which many fully int... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Self-organisation: a derivation from first principles of a class of learning algorithms

    Publication Year: 1989, Page(s):495 - 498 vol.2
    Cited by:  Papers (27)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (309 KB)

    A novel derivation of T. Kohonen's topographic mapping learning algorithm (Self-Organization and Associative Memory, Springer-Verlag, 1984) is presented. Thus the author prescribes a vector quantizer by minimizing an L/sub 2/ reconstruction distortion measure. He includes in this distribution a contribution from the effect of code noise which corrupts the output of the vector quantizer. Such code ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Combining self-organizing maps

    Publication Year: 1989, Page(s):499 - 502 vol.2
    Cited by:  Papers (10)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (357 KB)

    The author proposed a learning rule for a single-layer network of modules representing adaptive tables of the type formed by T. Kohonen's vector quantization algorithm (Rep. TKK-F-A601, Helsinki Univ. of Technol., 1986). The learning rule allows combination of several modules to learn more complicated functions on higher dimensional spaces. During learning each module learns a function, which is a... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast self-organization by the probing algorithm

    Publication Year: 1989, Page(s):503 - 507 vol.2
    Cited by:  Papers (12)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (379 KB)

    A new computational algorithm, the probing algorithm, is introduced for the subproblem of finding the best matching unit in Kohonen's self-organization procedure (Self-Organization and Associative Memory, Springer-Verlag, 1988). It is compared to exhaustive search and to four other algorithms and shown to be roughly six to 10 times faster for the case of high-dimensional vectors.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A model for self-organization in WTA networks and its application to map prediction problems

    Publication Year: 1989, Page(s):509 - 516 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (856 KB)

    A mathematical model for long-term memory (LTM) reorganization in self-organizing winner-take-all (WTA) networks is developed. The model describes the temporal evolution of the density of neural LTM states using a diffusive partial differential equation. Solutions to this equation show that, in the long run, LTM states tend to cluster about the modes of the stimulating source's probability density... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Variants of self-organizing maps

    Publication Year: 1989, Page(s):517 - 522 vol.2
    Cited by:  Papers (6)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (668 KB)

    Self-organizing maps have a connection with traditional vector quantization. A characteristic which makes them resemble certain biological brain maps, however, is the spatial order of their responses which is formed in the learning process. Two innovations are discussed: dynamic weighting of the input signals at each input of each cell, which improves the ordering when very different input signals... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Enhancing supervised learning algorithms via self-organization

    Publication Year: 1989, Page(s):523 - 530 vol.2
    Cited by:  Papers (10)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (786 KB)

    A neural network processing scheme is proposed which utilizes a self-organizing Kohonen feature map as the front end to a feedforward classifier network. The results of a series of benchmarking studies based upon artificial statistical pattern recognition tasks indicate that the proposed architecture performs significantly better than do conventional feedforward classifier networks when the decisi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Reducing transmission error effects using a self-organizing network

    Publication Year: 1989, Page(s):531 - 537 vol.2
    Cited by:  Papers (18)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (576 KB)

    The author shows that a Kohonen self-organizing network can be used to reduce the effects of errors in digital transmission of speech, Kohonen learning is used to cause the network nodes to mimic the input data distribution, while self-organization is used to arrange the nodes so as to reduce sensitivity to transmission errors. The number of bit substitutions that can be tolerated without unaccept... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural models of cursive script handwriting

    Publication Year: 1989, Page(s):539 - 542 vol.2
    Cited by:  Papers (4)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (403 KB)

    The main characteristics of cursive script handwriting are reviewed, and a basic coding scheme is described. Two types of neural networks are described: (i) a self-organizing network, similar to that used by T. Kohonen for his phonetic typewriter (see IEEE Comput., p.11-22, 1988), which is able to discover Graphotopic Maps; (ii) a three-layer perceptron called NetWrite in analogy with the NeTalk a... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neuroplanners for hand/eye coordination

    Publication Year: 1989, Page(s):543 - 548 vol.2
    Cited by:  Papers (9)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (489 KB)

    The authors generalize a previously described architecture, which they now call a neuroplanner, and apply it to an extension of the problem it was initially designed to solve-the target-directed control of a robot arm in an obstacle-cluttered workspace. By target directed they mean that the arm can position its end-effector at the point of gaze specified by a pair of stereo targetting cameras. Hen... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Connectionist modelling of a cognitive model to face identification: simulation of context effects

    Publication Year: 1989, Page(s):549 - 556 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (697 KB)

    From a cognitive model of face recognition, a connectionist layered network, FACENET, has been specified and implemented. FACENET system's architecture and functioning are presented, as well as tests on recognition dynamics and experimental results concerning the variability and specificity of encoding contexts.<> View full abstract»

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