Название: Artificial Intelligence Systems Based on Hybrid Neural Networks: Theory and Applications
Автор: Michael Zgurovsky, Victor Sineglazov
Издательство: Springer
Год: 2020 (2021 Edition)
Страниц: 527
Язык: английский
Формат: pdf (true)
Размер: 22.8 MB

This book is intended for specialists as well as students and graduate students in the field of artificial intelligence, robotics and information technology. It is will also appeal to a wide range of readers interested in expanding the functionality of artificial intelligence systems.

One of the pressing problems of modern artificial intelligence systems is the development of integrated hybrid systems based on deep learning. Unfortunately, there is currently no universal methodology for developing topologies of hybrid neural networks (HNN) using deep learning.

The development of such systems calls for the expansion of the use of neural networks (NS) for solving recognition, classification and optimization problems. As such, it is necessary to create a unified methodology for constructing HNN with a selection of models of artificial neurons that make up HNN, gradually increasing the complexity of their structure using hybrid learning algorithms.

The construction of hybrid neural networks (HNN), consisting of various types, each of which is trained according to a certain algorithm in layers, in many cases can significantly increase the efficiency of the ANN. The study of the principles of hybridization of ANNs, fuzzy logic and genetic algorithms allows you to create new types of models that have a higher quality of recognition, prediction, decision support while reducing computational costs for training.








[related-news] [/related-news]