General considerations for neural networks implemented in (optical) hardware

par Daniel Brunner (FEMTO-ST, CNRS & Université Bourgogne Franche-Comté)

lundi 7 oct. 2019, 14:00 → 15:00 Europe/Paris

Salle de Conférence 05-2021 (LLR)

The implementation of neural networks in analogue substrates is gaining substantial momentum in recent years. Besides encouraging progress, one needs to keep in mind and address the essential challenges: the scalability of such a computing architecture. I will highlight some of the relevant challenges and put them into relationships with recently demonstrated experimental systems.

AfficheSeminaireBrunner.pdf