A Associate Memorizing Neural Network With Self-Learning

In this paper,a self-learning,associatively memorizing network is proposed.Basingon biological concepts,the synapse in the net is simulated,each with 4 diodes and 4 batteriesThe learning process of the net is carried out by charging a specific battery in each synapse andthe process is controlled by adjusting the duration of charging and input sample pattern.Thestructure of the synapse is simple.The network composed of such synapses is capable of self-learning and is competent to develop into large scale network It is also capable of simultaneous(parallel) multi-input learning instead of scanning as most computers do Its memory capacity isdiscussed and the function of a preliminary 4-neuron,16-synapse model network has shown inte-resting features