Physics‐Informed Neural Network for Nonlinear Dynamics in Fiber Optics

Abstract

A physics-informed neural network (PINN) that combines deep learning with physics is studied to solve the nonlinear Schrödinger equation for learning nonlinear dynamics in fiber optics. A systematic investigation and comprehensive verification on PINN for multiple physical effects in optical fibers is carried out, including dispersion, self-phase modulation, and higher-order nonlinear effects. Moreover, both the special case (soliton propagation) and general case (multipulse propagation) are investigated and realized with PINN. In existing studies, PINN is mainly effective for a single scenario. To overcome this problem, the physical parameters (i.e., pulse peak power and amplitudes of subpulses) are hereby embedded as additional input parameter controllers, which allow PINN to learn the physical constraints of different scenarios and perform good generalizability. Furthermore, PINN exhibits better performance than the data-driven neural network using much less data, and its computational complexity (in terms of number of multiplications) is much lower than that of the split-step Fourier method. The results show that PINN is not only an effective partial differential equation solver, but also a prospective technique to advance the scientific computing and automatic modeling in fiber optics.

Fast Predicting the Complex Nonlinear Dynamics of Mode‐Locked Fiber Laser by a Recurrent Neural Network with Prior Information Feeding

“…[27,28] The prediction accuracy and the generalization ability of the AI model are bound to suffer a tremendous loss when missing the imperative prior information. The recent rise of physics-informed neural network (PINN) manifests itself as a method with superior performance in modeling nonlinear partial derivative equations [29,30] and seems to be a good candidate for femtosecond mode-locked fiber laser modeling. Nevertheless, suffering from the strong bond between the PINN and the modeled physical equations, the generalization ability of the PINN is considerably confined.…”

A review of machine learning-based failure management in optical networks

“…Furthermore, previous ML algorithms used in failure management were based on data-driven modeling without other prior knowledge. In addition, recently, physics-informed machine learning has been attracting wide attention from various areas, because it combines the benefits of machine learning and physical principles instead of functioning according to a purely data-driven approach [86,87]. In optical communications, lots of physical knowledge have been explored and could provide helpful information and insightful analysis for failure management.…”

Multiscale Physics-Informed Neural Networks for Stiff Chemical Kinetics

“…a good application prospect in the fields of fluid, 9 wave field, 10 and materials. 11 Raissi et al 12 first proposed the PINN to solve forward and inverse problems of partial differential equations. Li and Mei 13 added a data loss item to PINN loss function and pointed out that the accuracy of the neural network can be improved through small sample learning combined with the PINN.…”

“…This method gets rid of the dependence on a large number of data and is able to solve differential equations (i.e., forward problems) and inferring parameters based on observations (i.e., inverse problems). The PINN has a good application prospect in the fields of fluid, wave field, and materials . Raissi et al first proposed the PINN to solve forward and inverse problems of partial differential equations.…”