Today, neural networks are everywhere: they are the virtual assistants in our smartphones, they are powering our search engines and they are the key to big data classification. They can even beat humans at image recognition or at strategy games such as Go. However neural networks are still running as software on our current computers, consuming tens of kilowatts and working rather slowly. In comparison, the brain from which they are inspired, operates with 20 W and can realize incredibly complex tasks in fractions of seconds. Building chips more closely inspired from the brain architecture is the path to cognitive computing at low energy cost. Applications of such chips span from embedded automatic pattern recognition for big data management, going through unmanned vehicle control to bio-medical prosthesis. The challenge of fabricating brain-inspired hardware relies in the ultra-high density networks that have to be built, out of complex processing units interlinked by tunable connections providing memory. Magnetic nanodevices can be a key technology in this context thanks to their ability to provide massive access to memory, their multiple and tunable functionalities, their non-linear dynamics and compatibility with CMOS. In this talk I will give an overview of recent advances in the field of hardware brain-inspired computing. I will show how magnetic nanodevices can be used for brain-inspired computing, and discuss the working principle of neuromorphic models that can be implemented in Spintronics. Finally, I will show our first results of cognitive pattern recognition with magnetic oscillators.
Julie Grollier is a researcher director in the CNRS/Thales lab in France. Her Ph.D was dedicated to the study of a new effect in spintronics: the spin transfer torque. After two years of post-doc, first in Groningen University (Netherlands, group of B.J. van Wees), then in Institut d’Electronique Fondamentale (France, group of C. Chappert), she joined CNRS in 2005. Her current research interests include spintronics (dynamics of nanomagnets under spin torque), and new devices for cognitive computing.
Julie has over 75 publications, and is a frequent invited speaker in international conferences. She is also a Fellow of the American Physical Society. In 2010 she was awarded the Jacques Herbrand prize of the French Academy of Science, for her pioneering work on spin transfer. She is the recipient of two prestigious European Research Council grant: “NanoBrain” project (Memristive Artificial Synapses and their integration in Neural Networks, 2010-2015) and “BioSpinSpired” project (Bio-inspired Spin-Torque Computing Architectures, 2016-2021).
Julie is now leading the team “nanodevices for bio-inspired computing” that she initiated in 2009. She is also chair of the interdisciplinary research network GDR BioComp coordinating national French efforts to progress towards the hardware realization of bio-inspired systems.