Keywords:spintronics, physical reservoir computing, delayed feedback circuit
Physical reservoir computing based on phyiscal reservoirs, such as soft materials and quantum systems, has attracted much attention for next-generation AI technology. In particular, the physical reservoir computing based on spintronics device, which consists of magnetic multilayers in nanometer scale, is of great interest because of its small size and low power consumption. However, the figure of merit for the physical reservoir computing, such as a short-term memory capacity, still remains low in spintronics devices. In this work, we study the calculation performance of spintronics physical reservoir computing in the presence of delayed feedback, and find a high short-term memroy capacity compared to single device.