Demonstration of Pavlov associative memory by implementation of rate coding using magnetic tunnel junction neurons

A fully functioning Pavlov associative memory system is designed by a basis of rate coding. In order to implement rate coding characteristics to the system, a neuron signal is constructed by utilizing telegraphic switching characteristics of the magnetic tunnel junction (MTJ). Binary neural signal created by the MTJ is used as a weight update method for AgSi memristor synapses. The proposed Pavlov associative memory conducts not only basic associative memory characteristics such as learning and forgetting, but also conducts complex characteristics such as contrived/natural forgetting, consolidation learning, rate variation, etc.