[Proceedings] 1992 RNNS/IEEE Symposium on Neuroinformatics and Neurocomputers

- 1992

Filter Results

Displaying Results 1 - 25 of 150
  • Statistical research and multilayered neural networks

    Publication Year: 1992, Page(s):1172 - 1178 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (309 KB)

    Discusses two possible trends in research into three-layered neural networks synthesized from the paradigm of variable-structure neural networks and used for pattern recognition. These trends are: (i) minimizing the dimension of the space of attributes, and (ii) evaluating the permissible spread in weighting coefficients in order to determine the class of accuracy of electrical parameters of the c... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Structure optimization of multilayer neural networks with cross connections

    Publication Year: 1992, Page(s):509 - 520 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (400 KB)

    Problems of choosing the structure (the number of layers and the number of neurons in a layer) for multilayer neural networks with cross connections and consisting of neurons with two lattices are considered for the solution of pattern recognition problems. Consideration is given to multilayer neural networks with complete cross connections where the attribute set of each layer consists of initial... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • RNNS/IEEE Symposium on Neuroinformatics and Neurocomputers (Cat. NO.92TH0483-8)

    Publication Year: 1992
    Request permission for commercial reuse | PDF file iconPDF (35 KB)
    Freely Available from IEEE
  • A multilayer neural adaptive network as a model of nonlinear plants

    Publication Year: 1992
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (45 KB)

    Summary form only given. A multilayer neural adaptive network is chosen as a model for complex nonlinear plants. The basic element of the neural network (NN) is the sigmoid perceptron. The tuning of NN weights is based on measurements of the vector input and output of a nonlinear plant, the output of which is affected by an additive random noise. Optimal convergence rate algorithms for tuning the ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On a possible creative neurocomputer system

    Publication Year: 1992
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (54 KB)

    Summary form only given. It is suggested that one way to develop neural net computer systems may be through the use of the computer as a generator but not as a transformer of information. The core of these hypothetical computer systems must consist of two functional modules: a knowledge generator, based on the physical dissipative process, and a transformer of information into a structural form si... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Influence of stimuli on the comportment of an artificial neural population

    Publication Year: 1992, Page(s):846 - 854 vol.2
    Cited by:  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (416 KB)

    The dynamic behavior (comportment) of an artificial neural network depends not only on the synaptic weights, but also on the stimuli input. Without stimuli, the network stable state is permanent or oscillatory. The external stimuli input modifies the neural state, changing it from one stable permanent state to another or from an oscillatory to a permanent state, or changing the frequency of oscill... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Segmentron: a neuronal network simulating the structure and the function of the visual cortex

    Publication Year: 1992, Page(s):414 - 422 vol.1
    Cited by:  Papers (1)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (328 KB)

    The authors present a neuronal network of the visual cortex referred to as the Segmentron. It is based primarily on neuromorphological and neurophysiological data obtained in animal experiments. The response profiles of the neuron-like elements in the model are shown to be very similar to the profiles of the real neurons in the visual cortex. Moreover, by the running of the model, properties of th... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Some problems of mathematical theory of neural nets

    Publication Year: 1992, Page(s):438 - 448 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (332 KB)

    The formal mathematical theory of neural nets is considered. A novel approach to neural net research is proposed, based on the representation of a trained neural net as a set of five mathematical objects: two Boolean rings and three functions, determined on these rings. This approach made it possible to obtain some new results on the capacity of the Hopfield-Hebb net. Using this approach, some res... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Equilibrium states and oscillation of an artificial dynamic neural network

    Publication Year: 1992, Page(s):928 - 939 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (384 KB)

    Some theoretical results related to the behavior of a large collection of neurons composed of two and three families are presented. Each of them is represented by the mean value of the characteristic parameters such as the potential membrane and weights of the connections. Description in the state space has been retained. Attention is given to the determination of the equilibrium states and to the... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A bibliographic data base `Neural Networks and Neurocomputers'

    Publication Year: 1992, Page(s):1125 - 1126 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (88 KB)

    A neural networks and neurocomputers database at Moscow University containing about 7000 references is ready for use. The earliest works date from the 1940s and the latest ones reflect the present status of investigations. The database includes publications on mathematical and computer simulation studies of properties of neural network schemes, biological investigations explicitly using connection... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Representativeness of learning samples for paradigm of variable-structure neural networks

    Publication Year: 1992, Page(s):449 - 456 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (324 KB)

    The authors discuss the problem of the representativeness of a learning sample for the paradigm of the variable-structure neural network which synthesize neural networks for pattern recognition. They describe briefly the paradigm and classify recognition problems by the capability to simulate the learning sample. One problem involving a partially simulated learning sample is given as an example to... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Synchronization in stochastic neural model with excitatory connections

    Publication Year: 1992, Page(s):940 - 947 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (308 KB)

    The dynamics of a stochastic neural model with excitatory connections is investigated. The exponential fast synchronization of firings is proved View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the transfer function of macro-neuron

    Publication Year: 1992, Page(s):1127 - 1132 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (184 KB)

    In artificial intelligence models which are characterized by large volumes of associative memory, it is very important to find a compact description of the memory elements, while preserving the neural logic. The authors describe results of a numerical experiment which shows a way to correctly replace neurons by macro-neurons (ensembles) View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Phase portrait characteristics of random neural networks

    Publication Year: 1992, Page(s):145 - 156 vol.1
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (356 KB)

    Considers statistical properties of the phase portraits of random binary networks such as the mean number of fixed points, the stability of these points, the convergence time, and the volumes of the attraction basins. The phase portrait characteristics are analyzed in terms of the vector lengths, i.e., the distances between sequential trajectory states in the Hamming metric. The trajectories are c... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural networks processing systems in recognition and control problems

    Publication Year: 1992, Page(s):820 - 828 vol.2
    Cited by:  Papers (1)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (284 KB)

    The construction principles and architecture of neural network processing systems (NPSs) are considered. Attention is given to mechatronic system adaptation using NPS-based control, an NPS-based robot adaptive control architecture, threshold-polynomial training algorithms for recognition, and probabilistic training algorithms of logical NPSs for recognition View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Pseudo-inverse with increasing threshold: an error-recovery pattern recognition algorithm

    Publication Year: 1992, Page(s):888 - 894 vol.2
    Cited by:  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (436 KB)

    A pseudo-inverse algorithm with an increasing threshold value for pattern recognition is presented. This method is particularly useful when the original matrix is not a square matrix. The algorithm and its implementation are studied. Using English capital letters as samples, the error-recovery capability of the proposed algorithm is observed to be better than that of the fixed threshold algorithm.... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A macrodynamical approach to the analysis of neural networks

    Publication Year: 1992, Page(s):948 - 958 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (328 KB)

    General features of an asymptotical method for an analyzing complex system like neural networks are presented. The method is analogous to the mean-field approach and allows treatment not only of steady states but also of dynamical properties of networks. It may also be interpreted as a Galerkin procedure for the master equation. The types of neural networks and related problems to which the method... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Some problems of building and learning of neural networks while creating user's expert systems

    Publication Year: 1992, Page(s):1133 - 1135 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (116 KB)

    An experimental confirmation has been obtained of the possibility of neural network expert systems based on learning that uses examples without using any rules of logical deduction. This approach is used in the environment NEXSY, which provides users with the means for expert system creation by themselves. It has been possible to learn the fields of optimal use of algorithms and regimes of neural ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Principles of optical implementation of time-domain neural networks based on time-delayed four-wave mixing

    Publication Year: 1992, Page(s):157 - 160 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (76 KB)

    Properties of time-delayed four-wave mixing (TDFWM) are theoretically investigated for the purpose of implementing a neural network model in optics. The analysis is carried out on the assumption of low-intensity excitation pulses with spectral widths which significantly exceed the characteristic width of an inhomogeneously broadened line. In particular, the temporal properties of TDFWM can be used... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The hypothesis of inverse image sequence

    Publication Year: 1992, Page(s):684 - 695 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (412 KB)

    Certain considerations are presented that lead to the conclusion that the organization of the real trace sequence of activated neurons requires a special mechanism which would ensure synaptic potentiation in an inverse direction, from the first to the second neuron. The simplest mechanism is probably the rebound effect, whereby neuron activation and the subsequent inhibition period are followed by... View full abstract»

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

    Publication Year: 1992, Page(s):467 - 477 vol.1
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (484 KB)

    The development of the evolutionary computer is based on three sources: the concept of adaptive identification in the specified class of structured models, algorithmic facilities and software of data processing developed for the purpose of classification, recognition, prediction, reconstruction, and design; and a component base developed for realizing the hardware support of evolutionary algorithm... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The retrieval possibilities of autoassociative networks under variation of the interconnection architecture

    Publication Year: 1992, Page(s):1011 - 1016 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (168 KB)

    The dependence of the main retrieval characteristics of autoassociative neural networks-the memory capacity and the basins of attractors-on the architecture of connections is studied in the framework of the macrodynamical approach. The continuous-time version of the macrodynamical system of equations is analyzed. It is shown that in the limit of diluted connection structure the retrieval possibili... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Artificial neural networks in weather forecasting

    Publication Year: 1992, Page(s):829 - 835 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (224 KB)

    A method of statistical weather forecasting using databases of meteorological information is presented. The authors propose a special technique for constructing the learning set of images which represent the actual data. Forecasting of forthcoming weather events is based on the recognition of known data images using artificial neural networks and metric methods. Significant probabilities of true f... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural model of multijoint movement learning and control

    Publication Year: 1992, Page(s):895 - 906 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (372 KB)

    The authors consider motor control neural networks as specific dynamic systems and assume the choice of the particular movement among the different implementations of the motor plan to be a consequence of their organization and operation. They attempt to develop a model of multijoint movement control and learning in the context of the whole neuro-musculo-skeletal system, whose observed movement is... View full abstract»

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

    Publication Year: 1992, Page(s):1056 - 1067 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (364 KB)

    It is necessary to introduce many parameters describing the structure and input signal of a pattern recognition system during the construction of open-loop structures of multilayer neural networks in order to provide maximum probability of correct recognition in practice. The availability of a large number of parameters, i.e., hundreds and thousands, poses some difficulties for learning and for th... View full abstract»

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