SWITCHED DIFFUSION ANALOG MEMORY FOR NEURAL NETWORKS WITH HEBBIAN LEARNING-FUNCTION AND ITS LINEAR-OPERATION

We have fabricated a new analog memory for integrated artificial neura l networks. Several attempts have been made to develop a linear charac teristics of floating-gate analog memorys with feedback circuits. The learning chip has to have a large number of learning control circuit. In this paper, we propose a new analog memory SDAM with three cascaded TFTs. The new analog memory has a simple design, a small area occupan cy, a fast switching speed and an accurate linearity. To improve accur ate linearity, we propose a new charge transfer process. The device ha s a tunnel junction (poly-Si/poly-Si oxide/poly-Si sandwich structure) , a thin-film transistor, two capacitors, and a floating-gate MOSFET. The diffusion of the charges injected through the tunnel junction are controlled by a source follower operation of a thin film transistor(TF T). The proposed operation is possible that the amounts of transferred charges are constant independent of the charges in storage capacitor.