Dr. Xavier Porte Parera

FEMTO-ST, France

Xavier Porte was born in Manacor, Spain, in 1981. He received the Msc. degree in 2011 and the PhD. degree in 2015, both in Physics, from the University of the Balearic Islands (UIB), Spain. His Msc. and PhD. main research topic was nonlinear dynamics of delay-coupled semiconductor lasers. From 2016 to 2018 he was with the Technical University Berlin, Germany, where he was engaged in research work on external optical coupling of quantum-dot microlasers for nanophotonic applications. He is currently a postdoctoral researcher at the FEMTO-ST Institute in Besançon, France, working on photonic spatio-temporal neural networks. His research interests range from nonlinear delay dynamics to nanophotonics and neuromorphic computing.

Website – https://www.femto-st.fr/en/femto-people/javierporte

Talk Title 

Greedy Boolean Learning in Photonic Recurrent Neural Networks

Abstract

We demonstrate the implementation of a large scale recurrent photonic neural network with up to 2025 photonic neurons. All network internal and readout connections are physically implemented with fully parallel photonic technology. Based on a digital micro-mirror array, we can train the Boolean readout weights using a greedy version of reinforcement learning paradigm. We find that the learning excellently converges. Furthermore, it appears to possess a conveniently convex-like cost-function and demonstrates exceptional scalability of the learning effort with system size.