Artificial Neural Network-Assisted Anti-Resonance Hollow Core Fiber Confinement Loss Prediction

One of the critical performance metrics of anti-resonant fibers(ARFs) is the confinement loss of the fundamental mode, and achieving a lower confinement loss is one of the primary objectives in current design endeavors. Traditionally, the process of obtaining the confinement loss of designed fibers through simulation software like COMSOL is excessively time-intensive. To expedite the modeling process of anti-resonant fibers, this paper explores how to leverage a well-trained artificial neural network for rapid simulation of anti-resonant fibers while achieving high-precision output predictions. In the proposed method, an artificial neural network model is established to effectively map the feature space to the target optical characteristics. The effectiveness of this method is demonstrated using a specific structure of anti-resonant fiber as an example.