Long-term memory in a silicon microring network for photonics-based neuromorphic computing
Photonic neural networks generally lack energy efficient and resettable long-term memory required to tackle time dependent tasks with relatively slow time scales, e.g. for optical sensing applications. We experimentally demonstrate machine learning-exploitable long-term memory (at least 10 microseconds) in a network of silicon microring resonators.
Authors
- Stefano Biasi (ORCID: https://orcid.org/0000-0002-3361-133X)
- Lorenzo Pavesi (ORCID: https://orcid.org/0000-0001-7316-6034)
- Peter Bienstman (ORCID: https://orcid.org/0000-0001-6259-464X)
- Alessio Lugnan (ORCID: https://orcid.org/0000-0002-6587-2614)
- Alessandro Foradori (ORCID: https://orcid.org/0009-0001-3195-8620)
Publication Details
- Published
- 2025-01-01
- DOI
- https://doi.org/10.1364/cleo_si.2025.ss103_1
- Primary Topic
- Neural Networks and Reservoir Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00