General Neural Network

摘要

In this paper the authors propose a general neural network that has the capacity to adapt to the problem that has to be solved, reducing itself to a particular case by classical neural network, being a "specialized" character. The advantages that result from this character that appertain also to classical neural networks are the following: a high power of calculation, tolerance to deterioration, the capacity to processes more incomplete information etc. On other hand through its capacity to reduce to an unlimited number of particular cases by classical neural networks, the general neural network has also an "universal" character being capable to solve a very big number by type of problems. The neural computers achieved from this general neural network would be self-configuration being solve a big number by type of problems during short time, processing high capacities of information and being capable to process the incomplete input data into noise conditions.