UID:
edocfu_9959186308002883
Umfang:
1 online resource (IV, 157 p.)
Ausgabe:
Online edition Springer Lecture Notes Archive ; 041142-5
ISBN:
3-540-70819-7
Serie:
Lecture Notes in Control and Information Sciences, 210
Inhalt:
Providing an in-depth treatment of the main topics in neural networks this volume concentrates on multilayer networks and completely connected networks, as well as discussing both analog and digital networks. The central themes are dynamical behaviour, attractors and capacity. Boltzmann machines are also discussed. The subject is developed from scratch and does not use statistical physics. Because the volume adopts a pedagogical approach, explaining all steps in full, the style which the book takes it will appeal to engineers, computer scientists and applied mathematicians. The reader will learn the parameters and results that are most important for the design of neural networks and will be able to critically assess the existing commercial products or design a hardware or software implementation for him- or herself.
Anmerkung:
Bibliographic Level Mode of Issuance: Monograph
,
Key concepts in neural networks -- Backpropagation -- Neurons in the brain -- The fundamental system of differential equations -- Synchronous and discrete networks -- Linear capacity -- Capacity from a signal to noise ratio -- Neural networks and Markov chains.
,
English
In:
Springer eBooks
Weitere Ausg.:
ISBN 3-540-19995-0
Sprache:
Englisch
URL:
http://dx.doi.org/10.1007/BFb0034478
