Multidimensional mutual ordering of patterns using a set of pre-trained artificial neural networks
The article shows that large artificial neural networks can be used for mutual ordering of a set of multi-dimensional patterns of the same nature (handwritten text, voice, smells, taste). Each neural network must be pre-trained to recognize one of the patterns. As a measure of ordering one can use the entropy of patterns "Strangers" that are input to a neural network trained to recognize only examples of the pattern "familiar". The neural network after training reduces the entropy of the examples of the pattern "Familiar" and increases the entropy of examples of pattern "Stranger." It is shown that the entropy measure of the ordering always has two global minima. The first minimum corresponds to the pattern "Familiar", the second to the inversion of the pattern "Familiar". It is also shown that the Hamming distance between the patterns belonging to two different groups (groups of the two global minima) is always as large as possible.
The article discusses development of the segmented characters classifier of the Russian alphabet and of the Arabic numerals on the basis of block neural network structures including the plurality of blocks for each individual character recognition and for the synthesis block decision.
We examine the questions of applying large pyramidal neural (intellectual neuron) networks to solve equipment object control problems. We consider the description of a system for dynamic planning of mobile robot behavior, constructed based on a network of similar elements.
The task of improving the quality of forecasting returns of financial instruments using multivariate mathematical models: regression models and neural networks was analyzed. To construct a multifactor model of returns used the assumption on the influence of market factors that have a different nature. A linear multivariable regression model was constructed using stepwise inclusion algorithm. The multilayer neural network trained using back-propagation algorithm. The quality of the neural prediction models forecast much higher quality, built with the help of a regression model.
Generalized error-locating codes are discussed. An algorithm for calculation of the upper bound of the probability of erroneous decoding for known code parameters and the input error probability is given. Based on this algorithm, an algorithm for selection of the code parameters for a specified design and input and output error probabilities is constructed. The lower bound of the probability of erroneous decoding is given. Examples of the dependence of the probability of erroneous decoding on the input error probability are given and the behavior of the obtained curves is explained.
The dynamics of a two-component Davydov-Scott (DS) soliton with a small mismatch of the initial location or velocity of the high-frequency (HF) component was investigated within the framework of the Zakharov-type system of two coupled equations for the HF and low-frequency (LF) fields. In this system, the HF field is described by the linear Schrödinger equation with the potential generated by the LF component varying in time and space. The LF component in this system is described by the Korteweg-de Vries equation with a term of quadratic influence of the HF field on the LF field. The frequency of the DS soliton`s component oscillation was found analytically using the balance equation. The perturbed DS soliton was shown to be stable. The analytical results were confirmed by numerical simulations.
Radiation conditions are described for various space regions, radiation-induced effects in spacecraft materials and equipment components are considered and information on theoretical, computational, and experimental methods for studying radiation effects are presented. The peculiarities of radiation effects on nanostructures and some problems related to modeling and radiation testing of such structures are considered.