P-Bits Using Stochastic Magnetic Tunnel Junctions for Probabilistic Acquisition and Processing of Big Data [Phase I]

Intelligence and energy-efficiency in information processing have made significant leaps during the past decades, allowing more and more data to be processed around the clock..... Nonetheless, the increasing inflation in the amount of generated data by sensors and internet-of-things (IoT) devices is surpassing the capacity of computing and artificial intelligence systems, making it very difficult to entirely process such data due to the enormous energy budget associated with such processing. This triggers a need for more energy-efficient processing of big data, as well greater intelligence – not only in processing of sensory data - but also in the generation and acquisition of such data in the first place. A holistic approach to the big data processing bottleneck is needed, where: (1) selective and intelligent data-acquisition as well as (2) energy-efficient big data processing are both achieved. This project aims at advancing both of these two important frontiers by employing an emerging spintronic technology based on advanced magnetic material heterostructures; the stochastic magnetic tunnel junction (sMTJ).