Use of machine learning to efficiently predict the confinement loss in anti-resonant hollow-core fiber

Abstract

The fundamental mode confinement loss (CL) of anti-resonant hollow-core fiber (ARF) is efficiently predicted by a classification task of machine learning. The structure–parameter vector is utilized to define the sample space of ARFs. The CL of labeled samples at 1550 nm is numerically calculated via the finite element method (FEM). The magnitude of CL is obtained by a classification task via a decision tree and k -nearest neighbors algorithms with the training and test sets generated by 290700 and 32300 labeled samples. The test accuracy, confusion matrices, and the receiver operating characteristic curves have shown that our proposed method is effective for predicting the magnitude of CL with a short computation runtime compared to FEM simulation. The feasibility of predicting other performance parameters by the extension of our method, as well as its ability to generalize outside the tested sample space, is also discussed. It is likely that the proposed sample definition and the use of a classification approach can be adopted for design application beyond efficient prediction of ARF CL and inspire artificial intelligence and data-driven-based research of photonic structures.

“Machine micro/nano optics scientist”: Application and development of artificial intelligence in micro/nano optical design

“…(linear regression, LR) [81] 、支持向量回归(support vector regression, SVR) [82] 、 KNN [83] 、 随 机 森 林 (andom forest regression, RFR) [84] 及梯度提升回 归(gradient boosting regression, GBR) [85] Wang等 [47] 提出一种利用伴随法(adjoint method)…”

“…

于化学 [26−31] 、材料学 [32−35] 、量子力学 [36−40] 、粒子 物理 [41−45] 等领域. 微纳光学设计方面, AI已经应 用于手性材料 [46] 、功率分配器 [47] 、微结构光纤 [48] 、 光子晶体光纤 [48] 、钙钛矿太阳能电池 [49] 、等离子体 波导 [50] 等设计研究. 相较于传统数值仿真方法,

…”

Design of porous-core photonic crystal fiber based on machine learning approach

“…To address this issue, machine learning (ML) techniques have emerged as a promising alternative for optimizing waveguide designs. In fact, ML techniques have been successfully applied to other photonic applications such as sensors, 15,16 the design of optical couplers, 17 microresonators, 18 hollow-core anti-resonant fibers, 19,20 prediction of the chromatic dispersion of PCFs, [21][22][23][24] cross-layer optimization of software-defined networks, 25 quality of transmission estimation, 25 design of nano-photonic structures, 26 and prediction of nonlinear phenomena in optical fibers. [27][28][29] For instance, Rodrigues-Esquerre et al 21 reported a multilayer perceptron (MLP) artificial neuronal network (ANN) to test and predict the chromatic dispersion of PCFs.…”

“…Furthermore, He et al 32 reported an inverse design method based on a neuronal network that allows to study and optimize the weak coupling problem in few mode fibers when they are employed in mode division multiplexing systems. Meng et al 20 predicted the confinement losses in anti-resonant hollow-core fibers by employing an ML technique involving decision trees and k-nearest neighbors (k-NN) and tandem-neuronal networks (T-NN). 19 Finally, ML-based techniques have been used to enhance the efficacy of optical fiber sensors in recent years.…”

Simple structure dual-core photonic crystal fiber polarization beam splitter covering the O + E + S + C + L + U band based on the surface plasmon resonance effect

“…The diameter of other air holes is d2, the hole to hole pitch is Λ, and the substrate material is the silica. The finite element method (FEM) is used to carry out the calculation [26]. In order to absorb the radiation energy, a perfect matching layer (PML) with the thickness of d2 is added to the outermost layer of the proposed SS-DC-PCF, and the refractive index of the PML (nPML) is 0.03 higher than that of the silica material [27].…”