Artificial intelligence designer for optical Fibers: Inverse design of a Hollow-Core Anti-Resonant fiber based on a tandem neural network

• DOI: 10.1016/j.rinp.2023.106310
• Publication Date: 2023-02-25T00:50:53Z
• Authors: Fanchao Meng; Jinmin Ding; Yiming Zhao; Hongwei Liu; Weiquan Su; Luyun Yang; Guangming Tao; Andrey Pryamikov; Xin Wang; Hongqian Mu; Yingli Niu; Jingwen He; Xinghua Zhang; Shuqin Lou; Xinzhi Sheng; Sheng Liang
• Keywords: Hyperparameter; Tandem

In this work, artificial intelligence (AI) is trained to "study" optical fibers as an AI fiber scientist. The dataset constructed on the structural parameters and confinement loss of hollow-core anti-resonant divided into different datasets. An effective approach overcome non-uniqueness challenge inverse designs tandem neural network (T-NN), which consists a forward prediction model design model, are by impact hyperparameter performance proposed also studied. For mean square error reaches lowest value 0.0007 with 6 hidden layers; nodes each layer 800, corresponding coefficient determination R2 0.9154. Moreover, R2CL T-NN can reach maximum 0.9881 for dataset-1-0. comparison between target, result, calculation result proves that way intelligently. Compared current numerical simulation methods, presented based more "intelligent". This study both accelerate provide guidance development scientists in optics.