Towards scalable Photonic Neural Networks - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Towards scalable Photonic Neural Networks

Résumé

In recent years Neural Networks or Neuromorphic Computing has significantly shifted the limits of what is computationally possible [1], and by now Neural Networks offer a practical approach to solve problems not addressable with classical algorithmic programming. Inspired by the human brain, Neural Networks utilize a large number of nonlinear transformations carried out in parallel. In architectural terms, Neuromorphic processor therefore go conceptually far beyond the von Neumann computational principle. Current neuromorphic hardware is strongly limited by the number of connections implemented in parallel, and implementations in 2D integrated photonic platforms do not scale with size
Fichier non déposé

Dates et versions

hal-03186622 , version 1 (31-03-2021)

Identifiants

  • HAL Id : hal-03186622 , version 1

Citer

Johnny Moughames, Javier Porte, Michael Thiel, Laurent Larger, Maxime Jacquot, et al.. Towards scalable Photonic Neural Networks. Physics of Quantum Electronics, Jan 2020, Snowbird, United States. ⟨hal-03186622⟩
41 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More