Recognition Accuracy Enhancement using Interface Control with Weight Variation-Lowering in Analog Computation-in-Memory

Abstract

As AI technology develops, it is necessary to verify the technical feasibility of Memory-Centric convergence technology. Previously investigated resistive synaptic devices (RSDs) can successfully mimic the function of biological synapses. However, the effect of the system recognition rate reflecting the variation of 16 weight states has not been studied yet. In this article, we perform simulations of various weight variation sets through real resistive synaptic device (RSD) engineering in Analog Computation-in-Memory (ACiM) system. These simulation results can provide guidelines for the continued design and optimization of a resistive synaptic device for realizing ACiM.