Magnetic Skyrmion Neurons with Homeostasis for Spiking Neural Networks

Area of Science:

• Neuromorphic Engineering
• Materials Science
• Computational Neuroscience

Background:

• Spiking Neural Networks (SNNs) mimic the human brain, driving advancements in artificial intelligence.
• Hardware-based SNNs (H-SNNs) are key for on-chip neuromorphic computing, implementing SNN algorithms like leaky integrate-and-fire (LIF).
• Neuronal homeostasis is vital for SNN learning but lacks single-device hardware implementation.

Purpose of the Study:

• To introduce a novel skyrmion-based spiking neuron device.
• To achieve hardware-level implementation of neuronal homeostasis in a single device.
• To demonstrate the feasibility of skyrmion applications in electronic devices.

Main Methods:

• Fabrication of a skyrmion-based spiking neuron device.
• Demonstration of neuronal homeostasis through gate voltage modulation.
• Characterization of the device's threshold modulation capabilities.

Main Results:

• Successful hardware implementation of neuronal homeostasis in a skyrmion-based neuron.
• Demonstrated modulation of the neuron's threshold value via applied gate voltage.
• Validation of skyrmion-based devices for neuromorphic applications.

Conclusions:

• The novel skyrmion device successfully implements neuronal homeostasis at the hardware level.
• This work paves the way for advanced on-chip SNN devices.
• Skyrmions show significant potential for future electronic device applications.