Resource-Optimal Programmable Photonics for Neural Networks
We study phase-shifter efficiency in programmable photonics, deriving an information-theoretic lower bound for phase shift in universal circuits, and proposing a “3-MZI” mesh that saturates this bound and requires 10 × less phase than conventional designs.
Authors
- Alexander Sludds (ORCID: https://orcid.org/0000-0002-7443-1378)
- Jasvith Raj Basani (ORCID: https://orcid.org/0000-0001-8453-8104)
- Dirk Englund (ORCID: https://orcid.org/0000-0002-1043-3489)
- Sri Krishna Vadlamani (ORCID: https://orcid.org/0000-0003-4602-0771)
- Ryan Hamerly (ORCID: https://orcid.org/0000-0003-4976-2236)
Institutions
- Vassar College (US)
- Diagnostics for the Real World (United States) (US)
- University of Maryland, College Park (US)
Publication Details
- Published
- 2025-01-01
- DOI
- https://doi.org/10.1364/cleo_si.2025.ss184_1
- Primary Topic
- Neural Networks and Reservoir Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00