Поиск по всему репозиторию:
Просмотр по издательству "BPI"
Отображаемые элементы 21-40 из 40
-
Neuroelectronic Systems: Binding Neurons to Electric Circuits
(BPI, 1999)This paper is devoted to description of a new type of information system based on interaction of biological neurons and electronic components – neuroelectronic (NE) system. Main features and components, like neural network sensor, of NE are briefly put into consideration followed with an example of ...2021-07-27
-
New Approach of the Recurrent Neural Network Training
(BPI, 1999)In this work the technique o f creation o f adapthre training algorithms for recurrent neural networks (RNN) is cortsidered. These algorithms have high convergence and accuracy on a comparison with traditional backpropagation. The original technique of calculation of an adaptive training step with use ...2021-07-27
-
On the Detecting of Symmetries of Switching Functions for Eflicient Design of Decision Trees and Decision Diagrams
(BPI, 1999)This paper presents an approach to detect symmetries of switching functions for efficient design of Reed-Muller decision trees (DTs) and decision diagrams (DDs). The information theory measures of switching functions are used to determine possible symmetric variabies and therefore to reduce search ...2021-07-27
-
Program Improvement of Quality of the Images at 2D Acoustic Visualization
(BPI, 1999)The various program methods are offered (linear filtration, processing on special functions with an adjustable weights, color-coding) to improve the images quality at 2D acoustic visualization of solidstate products, which are illustrated on an example of researches of friction pair. Overlapping the ...2021-07-27
-
Reactive Control of a Mobile Robot Based on Neural Networks
(BPI, 1999)This paper describes the neural system for reactive control of the mobile robot. The neural system consists of different types of neural networks, which are combined in the intelligent system. The efficient techniques for the training of neural networks are considered. The main problem of the neural ...2021-07-27
-
Research of Optimum Depth of Overlapping in Cellular Automata
(BPI, 1999)The speed of calculations homogeneous cellular automata is determined depending on depth of overlapping with use of a method of overlapped Windows. Optimum depth of overlapping also is investigated depending on parameters cellular automata.2021-07-27
-
Self-Organizing System Forming Strategy of the Global Behavior for Control of an Autonomous Mobile Robot
(BPI, 1999)Autonomous mobile robots typically require a preconceived and very detailed navigational model (map) of their intended operating emironment. It requires the presence of a priori information. The creating world model of eironment is a difficult problem requiring a detailed description of all possible ...2021-07-27
-
Self-training Neural System for Autonomous Control of a Mobile Robot
(BPI, 1999)Unsupervised training is of great importance for adaptation to the emironment. This adaptation is performed by means of continuos training as a result of interaction between robot and emironment. In this case the teacher is the emironment. This paper presents the way for the implementation of unsupervised ...2021-07-27
-
Social Dynamics and Self-Organizing
(BPI, 1999)A view of dynamics and self-organizing of social systems is presented. Both a separate Homo sapiens and human society are viewed as social systems. The dynamics of the development of such systems is also examined.2021-07-27
-
Synthesis of a Test Generator for a Built-ln Self-Test Scheme
(BPI, 1999)This paper presents a new algorithm for the automated synthesis of pseudo-random test patterns generators for Built-ln Self Test schemes with a mixed testmode. The experimental resulls showan opportunity of using the given method on a design stage of circuits producing In this paper it is shown that ...2021-07-27
-
Temporal Processing Neural Networks for Speech Recognition
(BPI, 1999)Application of the temporal processing neural networks (TPNNs) to the speech recognition is justifled by the nature of the task. Indeed ASR is a sequence recognition problem and assumes incorporation of time into decision process. Static models treat elements of sequence as independent patterns, which ...2021-07-27
-
The Application of Self-Organizing Neural Networks in Financial Market Analysis
(BPI, 1999)The problem of comparison of different companies is facing, when analyzing company's performance in stock exchange market. Due to many different fmancial ratios and parameters sometimes it is almost impossible to decide which company is a leader or not. One of the ways to solve this problem is the use ...2021-07-27
-
The Approach of the Shortest Path Planning Problem Solution on the Basis of a Neural Network
(BPI, 1999)This paper describes the neural network solving the shortest path planning problem. It consists of the layer set, where each layer forms a path-candidate for the fixed number of cities. During competition the layer-winner determining the sequence order of cities in the shortest path is selected. In ...2021-07-27
-
The Architecture of the Neural System for Control of a Mobile Robot
(BPI, 1999)Building autonomous mobile robots has been a primary aim of robotics and artificial intelligence. Artificial neural networks are capable of performing the different aspecis of autonomous drmng, such as collision-free motions, avoiding obstacles, mapping and planning of path. This paper describes the ...2021-07-27
-
The Regularities of the Development of the Nonlinear Dynamics of the Control Systems with Pulse-Width Modulation
(BPI, 1999)The paper reports on the dynamics of the control systems with pulse-width modulation of the second kind. The bifurcations of the stationary motions are studied. The regularities of occurrence of chaotic processes are revealed.2021-07-27
-
The Training of Feed-Forward Neural Networks
(BPI, 1999)The training of multilayer perceptron is generally a difficult task. Excessive training times and lack of convergence to an acceptable solution are frequently reported. This paper discribes new training methods of feedforward neural networks. In comparison with standard backpropagation algorithm it ...2021-07-27
-
The Universal Contact Device for Monitoring Printed Circuit Cards
(BPI, 1999)Shoot, V. The Universal Contact Device for Monitoring Printed Circuit Cards / V. Shoot, I. Prozherin // International Conference on Neural Networks and Artificial Intelligence ICNNAI'99 = Международная конференция по нейронным сетям и искусственному интеллекту ICNNAI'99 : Proceedings, Brest, Belarus, ...2021-07-27
-
Unsupervised Training Algorithm for Recirculation Neural Network
(BPI, 1999)Unsupemised learning is the great promise of the future. In such training the network is provided with inputs but not with desired outputs. Unsupemised learning is used for the principal component networks. This paper describes a nerw method for training of the recirculation networks. Such method is ...2021-07-27
-
Walsh Operators of Edge Detection for Gray-Scale Images
(BPI, 1999)Walsh operators are proposed that may be used in image processing for edge detection. These operators based on discrete two-dimensional Walsh functions of appropriate dimensionality. The algorithms of a calculation of operators and the algorithm of edge detection are represented.2021-07-27
-
Є-insensitive Unsupervised Learning
(BPI, 1999)One of the major paradigms for unsupervised learning in Artificial Neural Networks is Hebbian learning. The standard implementations of Hebbian learning are optimal under the assumptions of Gaussian noise in a data set. We derive e-insensitive Hebbian learning based on minimising the least absolute ...2021-07-27
