| Presentation | 2022-10-20 Structural Design by Deep Learning for Improving Coupling Efficiency between Si Thin Wire and Topological Waveguide Itsuki Sakamoto, Tomohiro Amemiya, Sho Okada, Hibiki Kagami, Nobuhiko Nishiyama, Xiao Hu, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | We propose a structure design method using deep learning to achieve highly efficient coupling between a normal Si waveguide and a topological transmission line. In topological photonic systems, it is not realistic to design the parameters of all unit cell elements individually, because the number of parameters increases in powers of unity. Therefore, we have applied deep learning to the structure designing. In the procedure of structural design using deep learning, a dataset consisting of 6,000 data was obtained using the finite difference time domain method by randomly shifting the distance from the center of the unit cell to the air pore for each unit cell. Next, to use the datasets, we constructed a neural network consisting of five layers, including a convolutional layer. By training the network, we obtained a regression function with a correlation coefficient of 0.943. By exploring the structural parameter space from the regression function, we were able to derive structural parameters that exceed the highest coupling efficiency in the dataset, demonstrating the effectiveness of the method using deep learning. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Topological photonics / Silicon photonics / Deep learning |
| Paper # | OCS2022-25,OPE2022-71,LQE2022-34 |
| Date of Issue | 2022-10-13 (OCS, OPE, LQE) |
| Conference Information | |
| Committee | OPE / OCS / LQE |
|---|---|
| Conference Date | 2022/10/20(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Toshikazu Hashimoto(NTT) / Takeshi Hoshida(Fujitsu) / Junichi Takahara(Osaka Univ.) |
| Vice Chair | Taro Arakawa(Yokohama National Univ.) / / Kosuke Nishimura(KDDI Research) |
| Secretary | Taro Arakawa(Kochi Univ. of Tech) / (Mitsubishi Electric) / Kosuke Nishimura(NTT) |
| Assistant | Yuhei Ishizaka(Kanto Gakuin Univ.) / Takeshi Umeki(NTT) / / Yoshiaki Nishijima(Yokohama National Univ.) / Nobuhiko Nishiyama(Tokyo Inst. of Tech.) |
| Paper Information | |
| Registration To | Technical Committee on OptoElectronics / Technical Committee on Optical Communication Systems / Technical Committee on Lasers and Quantum Electronics |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Structural Design by Deep Learning for Improving Coupling Efficiency between Si Thin Wire and Topological Waveguide |
| Sub Title (in English) | |
| Keyword(1) | Topological photonics |
| Keyword(2) | Silicon photonics |
| Keyword(3) | Deep learning |
| 1st Author's Name | Itsuki Sakamoto |
| 1st Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| 2nd Author's Name | Tomohiro Amemiya |
| 2nd Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| 3rd Author's Name | Sho Okada |
| 3rd Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| 4th Author's Name | Hibiki Kagami |
| 4th Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| 5th Author's Name | Nobuhiko Nishiyama |
| 5th Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| 6th Author's Name | Xiao Hu |
| 6th Author's Affiliation | National Institute for Materials Science(NIMS) |
| Date | 2022-10-20 |
| Paper # | OCS2022-25,OPE2022-71,LQE2022-34 |
| Volume (vol) | vol.122 |
| Number (no) | OCS-216,OPE-217,LQE-218 |
| Page | pp.pp.45-50(OCS), pp.45-50(OPE), pp.45-50(LQE), |
| #Pages | 6 |
| Date of Issue | 2022-10-13 (OCS, OPE, LQE) |