| Issue |
EPJ Web Conf.
Volume 287, 2023
EOS Annual Meeting (EOSAM 2023)
|
|
|---|---|---|
| Article Number | 13008 | |
| Number of page(s) | 2 | |
| Section | Focused Sessions (FS) 4- Machine Learning and Photonic Artificial Intelligence / Optical Neural Networks and Neuromorphic Computing | |
| DOI | https://doi.org/10.1051/epjconf/202328713008 | |
| Published online | 18 October 2023 | |
https://doi.org/10.1051/epjconf/202328713008
A scalable and fully tuneable VCSEL-based neural network
1. Institut FEMTO-ST, Université Bourgogne Franche-Comté, CNRS UMR6174, 15B Avenue des Montboucons Besançon, France
2. Instituto de Física Interdisciplinar y Sistemas Complejos, IFISC (UIB-CSIC), Campus UIB, E-07122 Palma de Mallorca, Spain
3. Institute of Photonics, Department of Physics, University of Strathclyde, 99 George str., Glasgow G1 1RD, UK
4. Institut für Festkörperphysik, Technische Universität Berlin, Hardenbergstraße 36, 10623 Berlin, Germany
* Corresponding author: anas.skalli@femto-st.fr
Published online: 18 October 2023
We experimentally demonstrate an autonomous, fully tuneable and scalable neural network of 350+ parallel nodes based on a large area, multimode semiconductor laser. We implement online learning strategies based on reinforcement learning. Our system achieves high performance and a high classification bandwidth of 15KHz for the MNIST dataset. Our approach is highly scalable both in terms of classification bandwidth and neural network size.
© The Authors, published by EDP Sciences, 2023
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.