Monte Carlo network machine. Its application in problems of direct numerical modeling

• Autorzy: Kuzyakin D.
• Konferencja: Proceedings of the Third Seminar on Recent Research and Design Progress in Aeronautical Engineering and its Influence on Education/seminar(III; 23-25.11.1998; Warszawa, Polska)
• Języki publikacji: EN

Abstrakty

EN

In this report the speech will go about neural distinguished from a usual such as network with the back propagation method of learning. This network machine has another structure, mode of a construction and design of a element of a network - neuron at last. Describing network machine solves usual task solve usual neural networks. Such as task of approximation, prediction and discernment From macroscopic points of view (velocity of tutoring / work, the volumes of a processed in unit of time, possibility of a in hardware) given neural machine (NM) shows the most acceptable performances, on a comparison with usual. It is necessary to note that NM method has distinctive ability to be on tasks with a large contribution of a stochastic in the entering data flow, that is very important for tutoring on tasks of aerodynamics. We hope that this report will bring the contribution in research connected with artificial intelligence. The given work is based on the up-to-date knowledge about on a structure and functioning of the elements of a higher nervous system (first and second signal systems) and of course Monte-Carlo methods.

Słowa kluczowe

EN

direct numerical modeling; neural network; Monte Carlo network

PL

bezpośrednie modelowanie numeryczne; sieci neuronowe; sieć Monte Carlo

• Wydawca: Institute of Aeronautics and Applied Mechanics, Warsaw University of Technology
• Czasopismo: Research Bulletin / Warsaw University of Technology, Institute of Aeronautics and Applied Mechanics
• Rocznik: 1999
• Tom: nr 9
• Strony: 51--54
• Opis fizyczny: Bibliogr. 3 poz., rys., wykr.

Twórcy

autor

Kuzyakin D.

• Moscow Institute of Physics and Technology, Russia,