TR2021-045
Nano-Optic Broadband Power Splitter Design via Cycle-Consistent Adversarial Deep Learning
-
- , "Nano-Optic Broadband Power Splitter Design via Cycle-Consistent Adversarial Deep Learning", Conference on Lasers and Electro-Optics (CLEO), DOI: 10.1364/CLEO_SI.2021.SW4E.1, May 2021.BibTeX TR2021-045 PDF Presentation
- @inproceedings{Tang2021may3,
- author = {Tang, Yingheng and Kojima, Keisuke and Koike-Akino, Toshiaki and Wang, Ye and Jha, Devesh K. and Parsons, Kieran and Qi, Minghao},
- title = {Nano-Optic Broadband Power Splitter Design via Cycle-Consistent Adversarial Deep Learning},
- booktitle = {Conference on Lasers and Electro-Optics (CLEO)},
- year = 2021,
- month = may,
- doi = {10.1364/CLEO_SI.2021.SW4E.1},
- url = {https://www.merl.com/publications/TR2021-045}
- }
- , "Nano-Optic Broadband Power Splitter Design via Cycle-Consistent Adversarial Deep Learning", Conference on Lasers and Electro-Optics (CLEO), DOI: 10.1364/CLEO_SI.2021.SW4E.1, May 2021.
-
MERL Contacts:
-
Research Areas:
Communications, Electronic and Photonic Devices, Machine Learning, Optimization, Signal Processing
Abstract:
A novel generative deep learning model with a cycle-consistent adversarial network is introduced for optimizing 550 nm broad bandwidth (1250 nm to 1800 nm) power splitters with arbitrary target splitting ratios.