Pluralization of Learning in PRC using MEMS Nonlinear Resonator Array

Recently, physical reservoir computing (PRC) has attracted attention as a fast and power-saving implementation of machine learning. In this research, we proposed a method that uses the motion states of multiple resonators for machine learning (pluralized method) and evaluated its performance, aiming to improve the learning performance of PRC using MEMS nonlinear resonator array. The results showed that the pluralized method increased the Memory Capacity for the Short Term Memory (linear) task and Parity Check (nonlinear) task.