Abstract
We present three different approaches to apply deep learning to inverse design for nanophotonic devices. The forward and inverse regression models use device parameters as inputs and device responses as outputs, and vice versa. The generative model to create a series of improved designs. We demonstrate them to design nanophotonic power splitters with multiple splitting ratios.
© 2020 The Author(s)
PDF ArticleMore Like This
Keisuke Kojima, Mohammad H. Tahersima, Toshiaki Koike-Akino, Devesh Jha, Yingheng Tang, Ye Wang, Kieran Parsons, Fengqiao Sang, and Jonathan Klamkin
Th1A.6 Optical Fiber Communication Conference (OFC) 2020
Keisuke Kojima, Toshiaki Koike-Akino, Yingheng Tang, and Ye Wang
IF3A.6 Integrated Photonics Research, Silicon and Nanophotonics (IPRSN) 2021
Minwoo Jung, Keisuke Kojima, Toshiaki Koike-Akino, Ye Wang, Dayu Zhu, and Matthew Brand
SW5E.6 CLEO: Science and Innovations (CLEO_SI) 2022