Reinforcement learning for photonic component design
We present a new fab-in-the-loop reinforcement learning algorithm for the design of nano-photonic components that accounts for the imperfections present in nanofabrication processes. As a demonstration of the potential of this technique, we apply it to the design of photonic crystal grating couplers...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
AIP Publishing LLC
2023-10-01
|
Series: | APL Photonics |
Online Access: | http://dx.doi.org/10.1063/5.0159928 |
_version_ | 1797634129399906304 |
---|---|
author | Donald Witt Jeff Young Lukas Chrostowski |
author_facet | Donald Witt Jeff Young Lukas Chrostowski |
author_sort | Donald Witt |
collection | DOAJ |
description | We present a new fab-in-the-loop reinforcement learning algorithm for the design of nano-photonic components that accounts for the imperfections present in nanofabrication processes. As a demonstration of the potential of this technique, we apply it to the design of photonic crystal grating couplers fabricated on an air clad 220 nm silicon on insulator single etch platform. This fab-in-the-loop algorithm improves the insertion loss from 8.8 to 3.24 dB. The widest bandwidth designs produced using our fab-in-the-loop algorithm can cover a 150 nm bandwidth with less than 10.2 dB of loss at their lowest point. |
first_indexed | 2024-03-11T12:04:10Z |
format | Article |
id | doaj.art-ebc19cb40e974ed3aec076d7684ab01b |
institution | Directory Open Access Journal |
issn | 2378-0967 |
language | English |
last_indexed | 2024-03-11T12:04:10Z |
publishDate | 2023-10-01 |
publisher | AIP Publishing LLC |
record_format | Article |
series | APL Photonics |
spelling | doaj.art-ebc19cb40e974ed3aec076d7684ab01b2023-11-07T18:01:03ZengAIP Publishing LLCAPL Photonics2378-09672023-10-01810106101106101-1210.1063/5.0159928Reinforcement learning for photonic component designDonald Witt0Jeff Young1Lukas Chrostowski2Department of Electrical and Computer Engineering, The University of British Columbia, Vancouver V6T-1Z4, British Columbia, CanadaStewart Blusson Quantum Matter Institute, Vancouver V6T-1Z4, British Columbia, CanadaDepartment of Electrical and Computer Engineering, The University of British Columbia, Vancouver V6T-1Z4, British Columbia, CanadaWe present a new fab-in-the-loop reinforcement learning algorithm for the design of nano-photonic components that accounts for the imperfections present in nanofabrication processes. As a demonstration of the potential of this technique, we apply it to the design of photonic crystal grating couplers fabricated on an air clad 220 nm silicon on insulator single etch platform. This fab-in-the-loop algorithm improves the insertion loss from 8.8 to 3.24 dB. The widest bandwidth designs produced using our fab-in-the-loop algorithm can cover a 150 nm bandwidth with less than 10.2 dB of loss at their lowest point.http://dx.doi.org/10.1063/5.0159928 |
spellingShingle | Donald Witt Jeff Young Lukas Chrostowski Reinforcement learning for photonic component design APL Photonics |
title | Reinforcement learning for photonic component design |
title_full | Reinforcement learning for photonic component design |
title_fullStr | Reinforcement learning for photonic component design |
title_full_unstemmed | Reinforcement learning for photonic component design |
title_short | Reinforcement learning for photonic component design |
title_sort | reinforcement learning for photonic component design |
url | http://dx.doi.org/10.1063/5.0159928 |
work_keys_str_mv | AT donaldwitt reinforcementlearningforphotoniccomponentdesign AT jeffyoung reinforcementlearningforphotoniccomponentdesign AT lukaschrostowski reinforcementlearningforphotoniccomponentdesign |