Despite the significant progress made in deep learning on digital computers, their energy consumption and computational speed still fall short of meeting the st
