Skyrmion-based artificial synapses for neuromorphic computing
The electrical current-induced creation, motion, detection and deletion of skyrmions in ferrimagnetic multilayers can be used to mimic the behaviour of biological synapses, providing devices that could be used for neuromorphic computing tasks such as pattern recognition.
Abstract
Since the experimental discovery of magnetic skyrmions achieved one decade\nago, there have been significant efforts to bring the virtual particles into\nall-electrical fully functional devices, inspired by their fascinating physical\nand topological properties suitable for future low-power electronics. Here, we\nexperimentally demonstrate such a device: electrically-operating skyrmion-based\nartificial synaptic device designed for neuromorphic computing. We present that\ncontrolled current-induced creation, motion, detection and deletion of\nskyrmions in ferrimagnetic multilayers can be harnessed in a single device at\nroom temperature to imitate the behaviors of biological synapses. Using\nsimulations, we demonstrate that such skyrmion-based synapses could be used to\nperform neuromorphic pattern-recognition computing using handwritten\nrecognition data set, reaching to the accuracy of ~89 percents, comparable to\nthe software-based training accuracy of ~94 percents. Chip-level simulation\nthen highlights the potential of skyrmion synapse compared to existing\ntechnologies. Our findings experimentally illustrate the basic concepts of\nskyrmion-based fully functional electronic devices while providing a new\nbuilding block in the emerging field of spintronics-based bio-inspired\ncomputing.\n