Implementation of Memory Elements and Combinational Logic in the Class of Pyramidal Neural Networks of Fast Learning
Alexander Yu. Dorogov · 2023
The paper considers the use of fast learning pyramidal neural networks for the implementation of memory elements and combinational logic on the example of an encoder and decoder of binary codes. Networks of the class under consideration are representable by linear operators, have a self-similar structure, and are a special case of the fast Fourier transform algorithm. Methods of topological construction of pyramidal networks are given. An algorithm for fast learning of the networks with absolute convergence is presented. It is shown that the pyramidal memory network provides storage and accurate recovery of images similar to storing numbers in random access computer memory. It is noted that pyramid networks belong to the category of deep learning networks.