Machine learning at the speed of light

Watch the seminar here

Photonics has long been considered an attractive substrate for the next generation of computing and machine learning. Reservoir computing significantly simplified the implementation of artificial neural networks in analogue hardware. In this seminar, we will review the motivation for photonic computing and discuss the basic principles of reservoir computing. We will illustrate the concepts through several opto-electronic implementations of reservoir computing, such as large-scale systems with high computational power and time-delay dynamical systems with high processing speed.

Piotr ANTONIK was born in Minsk, Belarus, in 1989. He graduated in physics from the Université libre de Bruxelles (Brussels, Belgium) in 2013 and defended his PhD in 2017 under the direction of Prof. S. Massar. During his PhD, he studied implementations of machine learning methods in photonic systems. 
His PhD thesis won the Springer Theses Award and was published in the Springer Theses collection.
In October 2018, he obtained a permanent position of Associate Professor at CentraleSupélec, Metz Campus, with the LMOPS EA-4423. His research activities combine machine learning, photonics, and FPGA programming, with applications in telecommunications, computer vision and biomedical imaging. To this day, Dr. Antonik authored and co-authored 2 books, 12 articles in peer-reviewed journals and more than 20 contributions to international conferences.