Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic Devices
Gradient-descent-based digitized adjoint method offers a way to realize the high-efficiency inverse design of digital nanophotonic devices with diverse functions. However, the vanishing gradient problem encountered in the design of high-dimension devices may lead to significant inefficiencies, makin...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Photonics Journal |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10026670/ |
| _version_ | 1811167518358765568 |
|---|---|
| author | Qiaomu Hu Zishan Zeng Zinan Xiang Kaiyuan Wang Minming Zhang |
| author_facet | Qiaomu Hu Zishan Zeng Zinan Xiang Kaiyuan Wang Minming Zhang |
| author_sort | Qiaomu Hu |
| collection | DOAJ |
| description | Gradient-descent-based digitized adjoint method offers a way to realize the high-efficiency inverse design of digital nanophotonic devices with diverse functions. However, the vanishing gradient problem encountered in the design of high-dimension devices may lead to significant inefficiencies, making it difficult to integrate novel functions on a single chip. Here, we propose a highly efficient digitized adjoint method for large-scale inverse design, called adaptive gradient-descent with momentum. It uses the first- and second-order momentum, instead of the gradient, to update the device pattern during adjoint optimization. To demonstrate the efficiency of the proposed method, we design a coarse wavelength division multiplexer and a three-mode power divider with design dimensions of 800 and 1360, respectively, which are approximately 2-4 times that of conventional digital nanophotonic devices. The simulation results show that, compared with the conventional gradient descent method, the momentum-assisted adjoint method has about 4-6 times higher efficiency and obtains better optimization performance, which provides a powerful tool for the inverse design of novel digital nanophotonic devices. |
| first_indexed | 2024-04-10T16:11:48Z |
| format | Article |
| id | doaj.art-ac753207b4cc4cd6924a51023357217b |
| institution | Directory Open Access Journal |
| issn | 1943-0655 |
| language | English |
| last_indexed | 2024-04-10T16:11:48Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Photonics Journal |
| spelling | doaj.art-ac753207b4cc4cd6924a51023357217b2023-02-10T00:00:07ZengIEEEIEEE Photonics Journal1943-06552023-01-0115111010.1109/JPHOT.2023.324018910026670Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic DevicesQiaomu Hu0https://orcid.org/0000-0001-6375-6161Zishan Zeng1Zinan Xiang2Kaiyuan Wang3Minming Zhang4https://orcid.org/0000-0002-3742-1445School of Optical and Electronic Information, Wuhan, ChinaSchool of Optical and Electronic Information, Wuhan, ChinaSchool of Optical and Electronic Information, Wuhan, ChinaSchool of Optical and Electronic Information, Wuhan, ChinaSchool of Optical and Electronic Information, Wuhan, ChinaGradient-descent-based digitized adjoint method offers a way to realize the high-efficiency inverse design of digital nanophotonic devices with diverse functions. However, the vanishing gradient problem encountered in the design of high-dimension devices may lead to significant inefficiencies, making it difficult to integrate novel functions on a single chip. Here, we propose a highly efficient digitized adjoint method for large-scale inverse design, called adaptive gradient-descent with momentum. It uses the first- and second-order momentum, instead of the gradient, to update the device pattern during adjoint optimization. To demonstrate the efficiency of the proposed method, we design a coarse wavelength division multiplexer and a three-mode power divider with design dimensions of 800 and 1360, respectively, which are approximately 2-4 times that of conventional digital nanophotonic devices. The simulation results show that, compared with the conventional gradient descent method, the momentum-assisted adjoint method has about 4-6 times higher efficiency and obtains better optimization performance, which provides a powerful tool for the inverse design of novel digital nanophotonic devices.https://ieeexplore.ieee.org/document/10026670/Inverse designsilicon photonicsadjoint method |
| spellingShingle | Qiaomu Hu Zishan Zeng Zinan Xiang Kaiyuan Wang Minming Zhang Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic Devices IEEE Photonics Journal Inverse design silicon photonics adjoint method |
| title | Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic Devices |
| title_full | Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic Devices |
| title_fullStr | Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic Devices |
| title_full_unstemmed | Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic Devices |
| title_short | Momentum-Assisted Adjoint Method for Highly Efficient Inverse Design of Large-Scale Digital Nanophotonic Devices |
| title_sort | momentum assisted adjoint method for highly efficient inverse design of large scale digital nanophotonic devices |
| topic | Inverse design silicon photonics adjoint method |
| url | https://ieeexplore.ieee.org/document/10026670/ |
| work_keys_str_mv | AT qiaomuhu momentumassistedadjointmethodforhighlyefficientinversedesignoflargescaledigitalnanophotonicdevices AT zishanzeng momentumassistedadjointmethodforhighlyefficientinversedesignoflargescaledigitalnanophotonicdevices AT zinanxiang momentumassistedadjointmethodforhighlyefficientinversedesignoflargescaledigitalnanophotonicdevices AT kaiyuanwang momentumassistedadjointmethodforhighlyefficientinversedesignoflargescaledigitalnanophotonicdevices AT minmingzhang momentumassistedadjointmethodforhighlyefficientinversedesignoflargescaledigitalnanophotonicdevices |