Evaluation of Memory Capacity and Time Series Prediction Using a Spin Hall Oscillator as Reservoir

Spintronic devices, such as spin Hall oscillators (SHOs) and spin transfer torque oscillators, have become popular candidates for use as the reservoir in the reservoir computing (RC) architecture. The memory capacity of a reservoir quantifies the amount of information retained at any given time, indicating its significance for different RC tasks. In this work, we consider an SHO, consisting of a bilayer structure of platinum/Permalloy, as the reservoir. Using micromagnetic simulations, we evaluate the change in memory capacity of the SHO system as the magnetization dynamics varies from transient state oscillations to limit cycle oscillations. We also perform a three-bit parity task to study the performance of the oscillator in carrying out a nonlinear task. A time series prediction task, namely, the nonlinear auto regressive moving average (NARMA)2 task, is also performed. The results of both tasks confirm a correlation between the evaluated memory capacity of the reservoir and its efficiency in performing temporal tasks. The best performance in RC tasks is observed when the output magnetization dynamics of the oscillator consists of both transient state and limit cycle oscillations.

Automated summary

In this study, a spintronic device called spin Hall oscillator (SHO) is used as the reservoir in reservoir computing (RC) architecture. By evaluating the memory capacity of the SHO system with micromagnetic simulations, it is found that the reservoir's memory capacity is correlated with its efficiency in performing temporal tasks.