Abstract
In the letter, a stable magnetic tunnel junction (MTJ) at 300 K that is changed stochastically by spin Hall effect (SHE) and voltage controlled magnetic anisotropy (VCMA) is designed as a switching component of a neuron, which is used to implement restricted Boltzmann machine (RBM) as on-chip feature extractor. With the effect of VCMA, the naturally probability of the device can be regulated to be acted as the activation function of RBM. An online training scheme with VCMA-assisted SHE-MTJ and modified crossbar architecture is presented, which achieves high accuracy as well as neuron and synaptic sparsity. To meet the low power and low cost requirements, the training procedure and crossbar architecture is adjusted to be an offline training scheme, which achieves 72.50% energy consumption reduction of extraction and 2.84% accuracy promotion to 95.48%.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
