Bioinspired Computing Leveraging the Physics of Magnetic Nano-Oscillators

Brains display many features typical of non-linear dynamical networks, such as synchronization or chaotic behavior. These observations have inspired a whole class of models that harness the power of complex non-linear dynamical networks for computing. In this framework, neurons are modeled as non-linear oscillators, and synapses as the coupling between oscillators. However, there are few hardware implementations of these systems, because large numbers of interacting non-linear oscillators are necessary. In this talk, we will see why coupled magnetic nano-oscillators are very promising for realizing cognitive computing at the nanometer scale. Then, we will present our experimental and theoretical results. We will show how speech recognition can be performed using the transient dynamics and the synchronization of a few harmonic spin torque oscillators [1]. These results highlight key opportunities and requirements for harnessing spintronic physics for bioinspired computing. We will also show how superparamagnetic oscillators can code and transform information in a robust population-type scheme [2]. These results highlight that some apparently undesirable phenomena like superparamagnetism can become compelling for bioinspired schemes. We will finally discuss how this line of research can take inspiration from both neuroscience and machine learning, and finish by open questions raised by our research.

[1] M. Romera, P. Talatchian, S. Tsunegi, F. A. Araujo, V. Cros, P. Bortolotti, J. Trastoy, K. Yakushiji, A. Fukushima, H. Kubota, S. Yuasa, M. Ernoult, D. Vodenicarevic, T. Hirtzlin, N. Locatelli, D. Querlioz and J. Grollier, Nature, Vol. 563, p. 230, 2018.
[2] A. Mizrahi, T. Hirtzlin, A. Fukushima, H. Kubota, S. Yuasa, J. Grollier and D. Querlioz, , Nature Communications, Vol. 9, Article number: 1533 (2018).