CONFERENCE PROCEEDINGS
Advanced Search
Home > Proceedings > Volume PC12746 > Article
Open Access Presentation
4 October 2023 Leveraging highly data-efficient computational intelligence for the engineering of photonic devices: a case study on vortex phase mask coronagraphs
Nicolas Roy, Lorenzo König, Olivier Absil, Charlotte Beauthier, Alexandre Mayer, Michaël Lobet
Author Affiliations +
Proceedings Volume PC12746, SPIE-CLP Conference on Advanced Photonics 2023; PC127460E (2023) https://doi.org/10.1117/12.2688331
Event: SPIE-CLP Conference on Advanced Photonics, 2023, San Diego, California, United States
Abstract
This study employs computational intelligence techniques to optimize complex photonic devices. Traditional surrogate models have limitations in accurately modeling complex devices like vortex phase masks (VPMs). VPMs are essential for observing faint light sources near bright objects, such as exoplanets near stars. To address this, a data-efficient surrogate optimization setup using a deep neural network (U-Net) and particle swarm optimization is proposed. The U-Net achieves high accuracy and efficiency. The resulting surrogate optimization setup outperforms both carefully devised grid-based searches and optimizers.
Conference Presentation
© (2023) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Citation Download Citation
Nicolas Roy, Lorenzo König, Olivier Absil, Charlotte Beauthier, Alexandre Mayer, and Michaël Lobet "Leveraging highly data-efficient computational intelligence for the engineering of photonic devices: a case study on vortex phase mask coronagraphs", Proc. SPIE PC12746, SPIE-CLP Conference on Advanced Photonics 2023, PC127460E (4 October 2023); https://doi.org/10.1117/12.2688331
PROCEEDINGS
 PRESENTATION
WATCH
PRESENTATION SAVE TO MY LIBRARY
GET CITATION
Advertisement
Advertisement
KEYWORDS
Photonic devices
Coronagraphy
Engineering
Particle swarm optimization
Simulations
Instrument modeling
Neural networks
RELATED CONTENT
Subscribe to Digital Library
Receive Erratum Email Alert
Back to Top