Theory and design of fast reconfigurable transformations and loosely connected neural networks
Please sign in so that we can notify you about a reply
The monograph gives a modern presentation of the theory of designing algorithms for fast transformations and modular neural networks. From the system positions there are rapid transformations and multilayer neural networks with a progeny set of connections and a modular organization. Algorithms for structural design and training of rebuilt rapid transformations and neural networks are proposed. The potentially achievable properties of rapid transformations on plasticity, separating power and generalizing ability are investigated. All considered algorithms are accompanied by examples and software implementations in the MATLAB language. The presentation is focused on practical use in high dimensional signal compression problems, classification and recognition of images, spectral analysis and fractal filtering. In addition to specialists in information telecommunication technologies, the book can be useful for graduate students and students of the master"s level of preparation
Author:
Author:Dorogov Aleksandr Yuryevich
Cover:
Cover:Hard
Category:
- Category:Computer & Technology
- Category:Science & Math
ISBN:
ISBN:9785732510362
No reviews found