TY - JOUR
T1 - A binarized spiking neural network based on auto-reset LIF neurons and large signal synapses using STT-MTJs
AU - Liu, Haoyan
AU - Ohsawa, Takashi
N1 - Funding Information:
This work was supported by the VLSI Design and Education Center (VDEC), the University of Tokyo in collaboration with Cadence Corporation and Synopsys Corporation, and by JSPS KAKENHI Grant No. JP20K04626. It was partly executed under the cooperation of organization between Kioxia Corporation and Waseda University.
Publisher Copyright:
© 2023 The Japan Society of Applied Physics.
PY - 2023/4/1
Y1 - 2023/4/1
N2 - A binarized spiking neural network using auto-reset leaky integrate-and-fire neurons with a two-transistor and three-magnetic tunnel junction core and large signal synapses with two-transistor and two-magnetic tunnel junctions is designed. The network is applied to a classifier of the MNIST handwritten digit dataset with a 784 × 400 synapse crossbar array. The weights are trained offline using the spike-timing-dependent plasticity learning algorithm and deployed to the spin-transfer torque magnetic tunnel junction (STT-MTJ) resistances in the synapses after being binarized. Its performance is evaluated by HSPICE using the STT-MTJ device model, which takes the stochastic change in the angle between the two magnetic moments in the free and pinned layers into consideration. 75% test accuracy is achieved for 1200 patterns with 1 ns read and 1 ns write operations and 0.23 pJ/SOP energy consumption.
AB - A binarized spiking neural network using auto-reset leaky integrate-and-fire neurons with a two-transistor and three-magnetic tunnel junction core and large signal synapses with two-transistor and two-magnetic tunnel junctions is designed. The network is applied to a classifier of the MNIST handwritten digit dataset with a 784 × 400 synapse crossbar array. The weights are trained offline using the spike-timing-dependent plasticity learning algorithm and deployed to the spin-transfer torque magnetic tunnel junction (STT-MTJ) resistances in the synapses after being binarized. Its performance is evaluated by HSPICE using the STT-MTJ device model, which takes the stochastic change in the angle between the two magnetic moments in the free and pinned layers into consideration. 75% test accuracy is achieved for 1200 patterns with 1 ns read and 1 ns write operations and 0.23 pJ/SOP energy consumption.
KW - auto-reset neuron
KW - leaky integrate-and-fire (LIF) neuron
KW - spiking neural network
KW - spin-transfer torque magnetic tunnel junction (STT-MTJ)
UR - http://www.scopus.com/inward/record.url?scp=85154618551&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85154618551&partnerID=8YFLogxK
U2 - 10.35848/1347-4065/acc9f4
DO - 10.35848/1347-4065/acc9f4
M3 - Article
AN - SCOPUS:85154618551
SN - 0021-4922
VL - 62
JO - Japanese journal of applied physics
JF - Japanese journal of applied physics
IS - 4
M1 - 044501
ER -
