Deep learning enabled inverse design of ultracompact silicon metadevices for mode conversion
Authors:
- Weifeng JiangWJWeifeng Jiang
- Lin ZhangLZLin Zhang
- Zhengguo HanZHZhengguo Han
- Zihan XiangZXZihan Xiang
- Yatai GaoYGYatai Gao
- Zhongmao CaiZCZhongmao Cai
- 12 January 2026
- journal article
- Published by Optica Publishing Group in Optics Express
- Vol. 34 (2) , 1294-1306
Abstract
Deep learning has shown great potential for enabling the rapid inverse-design of high-performance photonic devices. In this work, we develop a forward prediction and inverse design platform for silicon metadevices for mode conversion using deep neural networks (DNNs). The performances of twenty silicon-based mode converters are predicted by using this platform, with a prediction time of just 5 ms and a root mean square error (RMSE) of less than 0.018. Five silicon-based mode converters were inverse-designed using DNN, achieving an ultracompact dimension of only 4.08 µm × 1.68 µm and a design time of only 10 ms. Experimental results verified the performance of the inversely designed devices and compared them with their forward predictions. The measured insertion losses deviated from the desired and predicted values by less than 0.887 and 0.878 dB, respectively. The DNN-based forward prediction and inverse design approach enables the rapid development of highly integrated and high-performance silicon-based mode converters, providing key components for mode-division multiplexing systems.Keywords
Funding Information
- National Natural Science Foundation of China (62275128)
- The Startup Foundation for Introducing Talent of NUIST (2023r108)
- the Postgraduate Research & Practice Innovation Program of Jiangsu Province (SJCX25_0492)
This publication has 37 references indexed in Scilit:
- Inverse design and demonstration of on-chip silicon high-order mode pass filterAPL Photonics, 2024
- Analysis of Deep Neural Network Models for Inverse Design of Silicon Photonic Grating CouplerJournal of Lightwave Technology, 2021
- Deep neural networks for the evaluation and design of photonic devicesNature Reviews Materials, 2020
- Mode-Oriented Permutation Cipher Encryption and Passive Signal Switching Based on Multiobjective Optimized Silicon Subwavelength MetastructuresACS Photonics, 2020
- Ultra‐Compact and Ultra‐Broadband Guided‐Mode Exchangers on SiliconLaser & Photonics Reviews, 2020
- Arbitrarily routed mode-division multiplexed photonic circuits for dense integrationNature Communications, 2019
- Efficient TE-Polarized Mode-Order Converter Based on High-Index-Contrast Polygonal Slot in a Silicon-on-Insulator WaveguideIEEE Photonics Journal, 2019
- Deep Neural Network Inverse Design of Integrated Photonic Power SplittersScientific Reports, 2019
- Compact Silicon Waveguide Mode Converter Employing Dielectric Metasurface StructureAdvanced Optical Materials, 2018
- Design of Compact TE-Polarized Mode-Order Converter in Silicon Waveguide With High Refractive Index MaterialIEEE Photonics Journal, 2018