A spintronic Huxley-Hodgkin-analogue neuron implemented with a single magnetic tunnel junction
Authors:
Davi R. Rodrigues,
Rayan Moukhader,
Yanxiang Luo,
Bin Fang,
Adrien Pontlevy,
Abbas Hamadeh,
Zhongming Zeng,
Mario Carpentieri,
Giovanni Finocchio
Abstract:
Spiking neural networks aim to emulate the brain's properties to achieve similar parallelism and high-processing power. A caveat of these neural networks is the high computational cost to emulate, while current proposals for analogue implementations are energy inefficient and not scalable. We propose a device based on a single magnetic tunnel junction to perform neuron firing for spiking neural ne…
▽ More
Spiking neural networks aim to emulate the brain's properties to achieve similar parallelism and high-processing power. A caveat of these neural networks is the high computational cost to emulate, while current proposals for analogue implementations are energy inefficient and not scalable. We propose a device based on a single magnetic tunnel junction to perform neuron firing for spiking neural networks without the need of any resetting procedure. We leverage two physics, magnetism and thermal effects, to obtain a bio-realistic spiking behavior analogous to the Huxley-Hodgkin model of the neuron. The device is also able to emulate the simpler Leaky-Integrate and Fire model. Numerical simulations using experimental-based parameters demonstrate firing frequency in the MHz to GHz range under constant input at room temperature. The compactness, scalability, low cost, CMOS-compatibility, and power efficiency of magnetic tunnel junctions advocate for their broad use in hardware implementations of spiking neural networks.
△ Less
Submitted 13 April, 2023;
originally announced April 2023.
