Dynamic Memristors for Physical Reservoir Computing

Abstract

Reservoir computing (RC) has garnered considerable attention for its efficient handling of temporal signal and lower training costs. As a nonlinear dynamical system, RC can map low-dimensional inputs into high-dimensional spaces and extract task-relevant features using a simple linear readout layer. Memristor inherently exhibits high-order dynamic characteristics due to their physical processes, which renders them an ideal choice for the implementation of physical reservoir computing (PRC) systems. This review focuses on PRC systems based on memristor, explaining the resistive switching mechanism at the device level and emphasizing the tunability of their dynamic behavior. The development of memristor-based reservoir computing systems is highlighted, along with discussions on the challenges faced by this field and potential future research directions.